EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: KungFuJosh on October 04, 2024, 08:06:41 pm
-
What do you recommend for a decent accuracy low(ish) current probe? I want something that has good accuracy in mA range and up to 5A (though I probably won't probe anything over 2A, so 2A or 3A is fine, whatever). Needs to cover AC and DC.
Also, hopefully not insanely expensive. Ideally, I'd like something under $500. Used is fine. If somebody has something awesome they want to trade for my 3532-50, PM me. 😉
Most common use would be tube guitar amp stuff, which mostly ranges in the low mA, but heater current might be significantly higher in the 1.5A level. Also my FUtracer tube curve tracer, but that's mostly the same levels. So 5A is overkill.
I understand there are other ways to measure this, that's not the point. I want to see things on the scope screen for fun, not due to any necessity.
Thanks,
Josh
-
Any bandwidth requirement?
https://www.eevblog.com/forum/beginners/current-probes/ (https://www.eevblog.com/forum/beginners/current-probes/)
https://www.eevblog.com/forum/testgear/micsig-cp2100b-current-probe/ (https://www.eevblog.com/forum/testgear/micsig-cp2100b-current-probe/)
https://www.eevblog.com/forum/testgear/which-dso-current-clamp-to-buy/ (https://www.eevblog.com/forum/testgear/which-dso-current-clamp-to-buy/)
-
Not really any BW requirements. Some stuff is DC, and most of what I would be testing are audio circuits.
The power tubes I use are probably going to run between 20mA and 100mA, and those are more important than the preamp tubes that might only see .5mA to ~1.5mA on the plates.
I've already read a lot of threads and opinions on here, but mostly it points to the common recommendations being poor subjects for low current readings. The CP2100B was one I was curious about, but that apparently sucks in the low mA range too. 🤷 The Pico TA189 is also interesting, but it's pricey. I don't need a current probe, I want one to screw around with in the range I need.
Thanks,
Josh
-
One option for lower currents is looping the wire through the aperture multiple times. This works if you can afford the extra wire length, extra inductance, and likely impact on bandwidth.
You might find the Aim-TTI I-prober 520 interesting: https://www.aimtti.com/product-category/current-probes/aim-i-prober-520 (https://www.aimtti.com/product-category/current-probes/aim-i-prober-520). The main selling point is the ability to measure PCB traces directly, but it does come with a ferrite clip thing to measure current through wires. It quotes ±5% accuracy from ±10mA to ±10A in wire mode and has 5MHz bandwidth. I have one. The DC offset is manually adjusted with a potentiometer, and it takes some time to stabilize (thermal, I guess).
The Pintek PA-699 is a little more traditional. It claims ±3% ±10mA on its 8A range and ±4% ±100mA on its 80A range. 1.5MHz bandwidth.
https://www.pintek.com.tw/productDetail/land-ctop-2/index/pscsn/33672/psn/185730 (https://www.pintek.com.tw/productDetail/land-ctop-2/index/pscsn/33672/psn/185730)
https://www.globalmediapro.com/dp/A2H116/Pintek-PA-699-Current-Probe-DC-15MHz/ (https://www.globalmediapro.com/dp/A2H116/Pintek-PA-699-Current-Probe-DC-15MHz/)
Looks like the same thing as an Elditest CP6990 at a better price.
https://www.caltestelectronics.com/product/CP6990-NA (https://www.caltestelectronics.com/product/CP6990-NA)
The Pico TA189 might be better than either of those in terms of price and accuracy if you're ok with the bandwidth.
Another "out there" option is to build your own measurement device based on an isolated hall sensor like a TMCS1126A5A https://www.ti.com/product/TMCS1126 (https://www.ti.com/product/TMCS1126). It's less convenient than a clamp because you'd have to wire it into the circuit under test. It also outputs 2.5V at zero current input, so opamps (likely with a potentiometer for offset trim) would be needed to convert to proper zero-referenced bipolar levels for a scope input.
If you're keen on the homebrew route, the AMC3301 is another option https://www.ti.com/product/AMC3301 (https://www.ti.com/product/AMC3301). It's an isolated amplifier meant to interface with a shunt and has an internal isolated DC/DC converter that powers the circuitry on the sensing side. Again, you'd have to wire it into the circuit. This one has a differential output so it likewise needs some opamps to interface with a scope.
-
I have used Prova Model 11 extensively and always found it to be reliable.
https://www.prova.com.tw/product_detail.asp?seq=235 (https://www.prova.com.tw/product_detail.asp?seq=235)
-
I´m quite happy with the Peaktech 4250 as a very reasonable priced that you can use together with a DMM and scope:
https://www.eevblog.com/forum/testgear/peaktech-4250-calibration/msg5575529/#msg5575529 (https://www.eevblog.com/forum/testgear/peaktech-4250-calibration/msg5575529/#msg5575529)
If you need more bandwidth I have heard good things about the MicSig-clamps available for around $500,-.
I´d always be sceptical if someone promises a large aperture for large conductors together with a high DC-sensitivity. A large aperture principally collects more stray fields than a small one and with resolutions around 1mA the stray fields get very significant.
For higher resolutions you practically need to use a shunt or a CT if you want to measure AC.
-
Thanks everybody. So far, it seems the Pico TA189 is going to be the best option for my needs. I'll put it on my wishlist. 😉
-
An isolation or differential amplifier with current shunt might be a better option. Especially if you have a low noise oscilloscope with 12 bit resolution and filtering. A current clamp is typically HAL sensor based and these have a lot of noise and offset (due to the earth's magnetic field).
-
An isolation or differential amplifier with current shunt might be a better option. Especially if you have a low noise oscilloscope with 12 bit resolution and filtering. A current clamp is typically HAL sensor based and these have a lot of noise and offset (due to the earth's magnetic field).
Are we talking any specific products, or a project? Any examples of what I should look at?
Thanks,
Josh
-
An isolation or differential amplifier with current shunt might be a better option. Especially if you have a low noise oscilloscope with 12 bit resolution and filtering. A current clamp is typically HAL sensor based and these have a lot of noise and offset (due to the earth's magnetic field).
Are we talking any specific products, or a project? Any examples of what I should look at?
There are some DIY designs out there for low noise differential amplifiers. Typically around an instrumentation amplifier. But you'd be limited to tens of volt of common mode range. Another option is to built a device around a current sensing amplifier. You should get to much higher voltages that way but it will be limited to a minimum supply voltage and there will be an extra current drawn from the source.
-
For DC and really slow stuff you might like an HP 428B if you can find one with a good probe. Boat anchor but they work really well. One caution, and I know this from sad experience, don't let the probe slide down an insulated high voltage wire and contact the bare terminal at the end. Flash, bang and time for a new probe.
-
An isolation or differential amplifier with current shunt might be a better option. Especially if you have a low noise oscilloscope with 12 bit resolution and filtering. A current clamp is typically HAL sensor based and these have a lot of noise and offset (due to the earth's magnetic field).
Are we talking any specific products, or a project? Any examples of what I should look at?
There are some DIY designs out there for low noise differential amplifiers. Typically around an instrumentation amplifier. But you'd be limited to tens of volt of common mode range. Another option is to built a device around a current sensing amplifier. You should get to much higher voltages that way but it will be limited to a minimum supply voltage and there will be an extra current drawn from the source.
That sounds like a lot of effort that might be fun in the future when I have the time and ambition. I guess now it's race between waiting for that or budgeting and/or excusing the cost of the TA189.
Thanks,
Josh
-
An integrated current sensor IC could be useful. Like this one, with a 5A range:
https://www.allegromicro.com/en/products/sense/current-sensor-ics/zero-to-fifty-amp-integrated-conductor-sensor-ics/acs723 (https://www.allegromicro.com/en/products/sense/current-sensor-ics/zero-to-fifty-amp-integrated-conductor-sensor-ics/acs723)
That would mean no ready-to-use probe, but less effort than an amplifer design project.
Of course, tht means the IC has to be connected to the circuit, but the point is that these sensors have very little inductance and burden voltage.
-
An integrated current sensor IC could be useful. Like this one, with a 5A range:
https://www.allegromicro.com/en/products/sense/current-sensor-ics/zero-to-fifty-amp-integrated-conductor-sensor-ics/acs723 (https://www.allegromicro.com/en/products/sense/current-sensor-ics/zero-to-fifty-amp-integrated-conductor-sensor-ics/acs723)
You can buy these as a product off-the shelve (I have stock 8)) but the noise and offset prevent any accuracy below 10mA.
-
An isolation or differential amplifier with current shunt might be a better option. Especially if you have a low noise oscilloscope with 12 bit resolution and filtering. A current clamp is typically HAL sensor based and these have a lot of noise and offset (due to the earth's magnetic field).
Are we talking any specific products, or a project? Any examples of what I should look at?
I just checked: A couple of years ago I created a (dual channel) design around an AD8421 which has a common mode range of +30 / -30V but I never turned it into a product.
-
I have no experience at all with oscilloscope current probes, but I do know that many of the voltages will be in the 200 to 400 VDC range. VAC will generally be much lower, probably sub 12VAC, but at most would likely be US line voltage.
-
The cheap $60 Hantek CC-65 might be useful, there's some post on here about improving the performance. We've had one for years, it works OK, and we did the BW mod.
Best
-
I have no experience at all with oscilloscope current probes, but I do know that many of the voltages will be in the 200 to 400 VDC range. VAC will generally be much lower, probably sub 12VAC, but at most would likely be US line voltage.
In that case something contactless is to be preferred. Another option is to use a fast DMM. Some can sample at thousands of measurements per second. That should be enough to get up to few kHz of bandwidth but you'd have to read it remotely and plot a graph to get a decent output.
-
Over the years, I had used many inexpensive current clamps for a scope.
Hantek, Micsig, Owon, Instrance...
If you want to measure with low noise and accurately in the lower mA range, none of them was good enough.
There are current clamps that can do the job, but they also cost accordingly; you can't get anything for less than €2300.
Unless...
You don't mind buying something used and don't mind taking up some space on the table.
Then the Tektronix combo would be the right choice, consisting of A6302, TM501, AM501...
It has an accuracy of 1%, a bandwidth of up to 50Mhz and a maximum current of 30A.
You can use it to measure currents from 1mA/div.
With a bit of luck, you can get the Tek for well under 1000, usually around 500...800€.
-
Hi,
A custom project might be an option (although this would take longer than buying a ready-made product of course : (
Anyway, if you went the DIY project route, then it's possible to amplify the voltage from a sense resistor, using (say) INA 21x series (an example circuit is attached, to monitor the current flowing where the red arrows are indicated; you could simplify a lot by removing U4 and U9, since they just optionally remove a small offset).
The output at J2 (green arrow) could be monitored with a multimeter, and the circuit could be powered from a battery, since you want everything isolated.
However, for measuring at a higher speed, then the output at the blue arrow could be connected to an ADC board (there are some boards on AliExpress, although it's a bit pricey, but it saves the effort to enter the schematic/create a PCB for that part). And then, something like a Pi Pico Wireless (Pico-W) could be used to send the data to a PC, i.e. PC is isolated, and it would log or chart. The code would be quite simple. If that's a problem, I don't mind helping to write a bit of code, (and I'm sure others would help if there was a major problem, since it's easy for anyone to explore the code if it's on say GitHub).
Also, there's no need to assemble a circuit board for the INA 21x board, since the eval module is cheap (https://www.ti.com/tool/INA210-215EVM).
A custom PCB would be neater of course, but for a one-off maybe it's all feasible using the ready-made modules approach. I've not done the calculation, but I believe a single sense resistor value should get you several amps of measurement, with mA resolution or better with the above combination of bits and pieces. Others may have better ideas.
EDIT: Just read above that @slugrustle has a DIY suggestion too, using an isolated AMC3301. Neat idea. It looks better than an 'analog linear optocoupler' that I tried in the past, which was disappointing.. I couldn't get decent performance out of it (had noise issues that made low-level measurements not good.).
-
I appreciate the thought of that kind of project, but I'm really just looking to connect to my oscilloscope to entertain and educate myself a little.
-
Should I buy a lottery ticket and win so I can get a Hioki CT6701 clamp and the required 3269 PSU? $5600 seems like a bargain to accurately measure down to 1mA. 🤣
-
https://siglentna.com/product/cp6030-current-probe/
-
https://siglentna.com/product/cp6030-current-probe/
Got any lying around? 😉
-
https://siglentna.com/product/cp6030-current-probe/
Got any lying around? 😉
Sorry no.
But have the SAPBus variant SCP5030 coming in a few weeks to play with....much cheaper than a new Tek P6021 which I have a few ePay finds of, but P6021 is an AC only current probe.
https://siglentna.com/product/scp5030-current-probe/
You'll need a Siglent DSO with SAPBus active probe support to use these.
-
Owon CP-07+
https://files.owon.com.cn/specifications/OWON%20AC_DC%20Clamp%20Current%20Probe%20technical%20spec.s.pdf (https://files.owon.com.cn/specifications/OWON%20AC_DC%20Clamp%20Current%20Probe%20technical%20spec.s.pdf)
https://eleshop.eu/cp-07-current-probe.html (https://eleshop.eu/cp-07-current-probe.html)
https://www.digikey.com/en/products/detail/owon-technology-lilliput-electronics-usa-inc/CP-07/15911696 (https://www.digikey.com/en/products/detail/owon-technology-lilliput-electronics-usa-inc/CP-07/15911696)
https://www.youtube.com/watch?v=3SwKDBjTBX8 (https://www.youtube.com/watch?v=3SwKDBjTBX8)
-
Owon CP-07+
It looks like 400mA is the low end of what it can reliably measure?
I'm looking to get down to ~1mA or at least down to 10mA.
But have the SAPBus variant SCP5030 coming
Fiiiine. I'll take that and a 3000X HD. 😉
-
https://siglentna.com/product/cp6030-current-probe/ (https://siglentna.com/product/cp6030-current-probe/)
Short review:
https://www.eevblog.com/forum/testgear/siglent-cp6030-current-probe-a-very-very-short-review/msg5352083/#msg5352083 (https://www.eevblog.com/forum/testgear/siglent-cp6030-current-probe-a-very-very-short-review/msg5352083/#msg5352083)
It's on my list, but it's rather low on the list because of the price. Until then, I'm happy with my Tek combination.
-
Owon CP-07+
It looks like 400mA is the low end of what it can reliably measure?
According to the specs it has 2 ranges: 0 to 400mA and 0 to 4A. So in one range you can measure up to 400mA and in the other 4A. No specification for noise but it looks like the Owon CP-07+ could very well fit the bill. Unfortunately the video from above is pretty much useless to get some real numbers. Still, IMHO the CP-07+ is worth trying if 1MHz bandwidth is good enough.
-
There's a special current probe from Little Bee that's discontinued now, you might be able to find them on eBay. These actually work well at lower current levels (why we got them), don't know why they stop making them tho.
Best,
-
The design of that littlebee probe looks interesting. Clever use of readily available ring cores with a gap.
Random google find:
https://www.split-corecurrenttransformer.com/sale-27008294-hall-effect-current-sensor-core-split-gapped-ferrite-core.html (https://www.split-corecurrenttransformer.com/sale-27008294-hall-effect-current-sensor-core-split-gapped-ferrite-core.html)
-
According to the specs it has 2 ranges: 0 to 400mA and 0 to 4A.
Where does it say that? The datasheet says "Test Range 400mA - 4A" and some of the retail sites say minimum is 10mA, but I don't see that anywhere else.
There's a special current probe from Little Bee that's discontinued now, you might be able to find them on eBay. These actually work well at lower current levels (why we got them), don't know why they stop making them tho.
Now that's interesting! Especially since they made it as an open source hardware project.
https://www.eevblog.com/forum/crowd-funded-projects/little-bee-current-probe/ (https://www.eevblog.com/forum/crowd-funded-projects/little-bee-current-probe/)
https://github.com/westonb/little-bee-B1 (https://github.com/westonb/little-bee-B1)
Thanks,
Josh
-
According to the specs it has 2 ranges: 0 to 400mA and 0 to 4A.
Where does it say that? The datasheet says "Test Range 400mA - 4A" and some of the retail sites say minimum is 10mA, but I don't see that anywhere else.
You should interpret that as having two test ranges: 0 to 400mA and 0 to 4A.
Eleshop has this spec sheet: https://static.eleshop.nl/mage/media/downloads/OWON_AC_DC_Clamp_Current_Probe_technical_specs.pdf (https://static.eleshop.nl/mage/media/downloads/OWON_AC_DC_Clamp_Current_Probe_technical_specs.pdf). I'd go from there and ignore the specs quoted on websites as these can be wrong. Manufacture provided data is the authorative source.
-
I'm cured of Owon when it comes to current clamps; you might want to look into CP05+...
Of course, it's possible that these pliers are a lucky find, unfortunately it's always a kind of lottery with these manufacturers.
-
https://siglentna.com/product/cp6030-current-probe/ (https://siglentna.com/product/cp6030-current-probe/)
Short review:
https://www.eevblog.com/forum/testgear/siglent-cp6030-current-probe-a-very-very-short-review/msg5352083/#msg5352083 (https://www.eevblog.com/forum/testgear/siglent-cp6030-current-probe-a-very-very-short-review/msg5352083/#msg5352083)
It's on my list, but it's rather low on the list because of the price. Until then, I'm happy with my Tek combination.
I did consider CP6030 but instead opted for the more PlugnPlay SCP5030 version with SAPBus connectivity and automatic input attenuation and not need a power source for the probe.
When spending this sort of $ a little bit more to provide simplicity of use gives value IMO.
-
That's right, this model is still available and yes, I do have a scope where it fits. ;)
But even there, low priority applies because of the price...
You don't get them on every corner either, so you could hope for a discount.
-
That's right, this model is still available and yes, I do have a scope where it fits. ;)
But even there, low priority applies because of the price...
You don't get them on every corner either, so you could hope for a discount.
Cost is just one consideration.
Let's look vs a personal fav Tek P6021 of which I have a few.....AC only, 60 MHz 15A
$ 2700 :o
https://www.tequipment.net/Tektronix/P6021A/Passive-Oscilloscope-Probes/ (https://www.tequipment.net/Tektronix/P6021A/Passive-Oscilloscope-Probes/)
Makes SCP5030 seem cheap. :P
-
According to the specs it has 2 ranges: 0 to 400mA and 0 to 4A.
Where does it say that? The datasheet says "Test Range 400mA - 4A" and some of the retail sites say minimum is 10mA, but I don't see that anywhere else.
You should interpret that as having two test ranges: 0 to 400mA and 0 to 4A.
That's the datasheet I quoted. I don't like interpreting things, especially not with brands like Owon. Other brands explicitly state the ranges as 1mA to whatever or 10mA to whatever, and Owon's datasheet says "Test Range 400mA - 4A". Maybe you're right, and I should assume that means 0 to 400mA, but I think that's a leap of faith in a brand I have no faith in. Has anybody personally used the Owon and found the results to be any good? Especially in the 1mA to 50mA range?
Thanks,
Josh
-
In that case, I would also assume two areas.
Up to 400mA and up to 4A.
Up to 400mA, the output is 1mV/1mA, up to 4A, 1mV/10mA.
The +/-1.5% tolerance... Almost uninteresting.
More important would be an indication of the intrinsic noise, which is "of course" missing.
-
Then again, if you buy it from a place with hassle free returns (like Amazon) you can just try it and return it if it doesn't work as required. That is better than endless rounds of guessing.
-
I viewed the entire video before making my other post and it has the answers to all of these questions. No one even bothered to actually watch it, I guess. Yes, mute the audio and you'll need to pay attention to the subtitles and what he's doing to keep track of what is going on. But he's got a $5k+ probe and a $25k+ scope on his bench, so despite some issues it's still worth a watch if you're thinking about buying either probe.
-
I viewed the entire video before making my other post and it has the answers to all of these questions. No one even bothered to actually watch it, I guess. Yes, mute the audio and you'll need to pay attention to the subtitles and what he's doing to keep track of what is going on. But he's got a $5k+ probe and a $25k+ scope on his bench, so despite some issues it's still worth a watch if you're thinking about buying either probe.
I skimmed through the video, and didn't see anything below 500mA. Do you have a timestamp for when he went to lower mA range?
-
13:06 you can see a 15mA 70kHz measurement: youtu.be/3SwKDBjTBX8?t=786 (http://youtu.be/3SwKDBjTBX8?t=786)
Looks like the 500kHz version is only $100 on aliexpress: https://www.aliexpress.com/item/737723863.html (https://www.aliexpress.com/item/737723863.html)
Maybe it can be upgraded.
edit: it does seem to pick up a lot of noise. Though not really worse than the littlebee (22mA p-p, 2.5mA RMS), but that is 10MHz.
-
13:06 you can see a 15mA 70kHz measurement: youtu.be/3SwKDBjTBX8?t=786 (http://youtu.be/3SwKDBjTBX8?t=786)
Looks like the 500kHz version is only $100 on aliexpress: https://www.aliexpress.com/item/737723863.html (https://www.aliexpress.com/item/737723863.html)
Maybe it can be upgraded.
Turning on high-res mode (or using input filtering) will very likely reduce the noise greatly and make the Owon probe more useable. Ofcourse a $1000 current probe will be better and a $10000 current probe will be fantastic but if the budget isn't there then a better option is to invest a little bit of time to try a cheap model. If it doesn't work only some time is lost and likely some useful experience gained to make a further selection.
-
Oh guys...
Normally, I'm also a big fan of cheap but still usable – but there are limits and I see them especially with current clamps for the mA range.
Please forgive me. ;)
-
As mentioned in the video comments, there is speculation that the yellow CP-07+ might be of lower quality than the official red product, but even then I have to say the video still shows some impressive performance from the Owon given the price point.
There is also a 100mA/1A version called the CP-07A where the coax is directly connected: https://www.ebay.com/itm/125291709845 (https://www.ebay.com/itm/125291709845)
https://www.youtube.com/watch?v=x3F7-bd9apQ (https://www.youtube.com/watch?v=x3F7-bd9apQ)
-
Oh guys...
Normally, I'm also a big fan of cheap but still usable – but there are limits and I see them especially with current clamps for the mA range.
Please forgive me. ;)
Can you give some details?
I always thought the low current accuracy was more or less defined by the clamp area and the strey fields that are received.
How much better are expensive current clamps? Can they measure e.g. 1mA with 5% accuracy?
I've worked for some time with an expensive Tektronix and Lecroy clamp and the only things I remember that these things have performed better are a functional degaussing and better frequency response >10kHz. And of course, for a professional lab these functions are worth the money, but e.g. for repairing audio equipment they may be compromisable...
-
There are several reasons why these current clamps are so expensive.
First of all, there is the core material, which is needed to achieve bandwidths of 10, 20, 50 or 100 MHz.
Then there is the measuring principle with current compensation.
A counter-current is applied via a winding on the core until the difference is zero. This corresponds to the measured current and is processed accordingly.
This has the advantage that the core is not or hardly magnetized, the intrinsically good Micsig CP2100, for example, does not have this, its accuracy decreases with increasing current (tolerance up to 15%).
If a current clamp has a degaussing function, you also know that it follows this principle.
Then the amplification must also have a high bandwidth with the lowest possible noise and high linearity.
These are µV that have to be boosted so that a few mA of current do not get lost in the intrinsic noise.
Finally, the supply of the circuits must also be very low-noise, and all of this cannot be sold for €200 afterwards.
Even the Chinese can't do magic in this regard... ;)
As you might expect, the issue of degaussing/current compensation plays a role in DC current measurements.
The core (its material composition) and the amplifier circuit are responsible for the bandwidth and accuracy.
The amplifier circuit and the supply of this stage are responsible for the intrinsic noise.
-
That is only half of the story.
High bandwidth current clamps typically consist of a HAL sensor to get from DC to a tens or hundreds of kHz and a current transformer for the higher frequencies. But, the core in the current transformer will have a different inductance depending on DC current which causes a shift in the crossover frequency between HAL sensor and current transformer. So a counter winding is used to make the DC magnetic flux through the core zero (so the core sees AC magnetic flux only) and thus keep the crossover frequency the same.
The cheap current clamps working up to 1MHz tops likely don't have have a current transformer and rely on a HAL sensor only.
Either way, HAL sensors are quite noisy and susceptible to the earth's magnetic field. The earth's magnetic field (and other fields) are nulled through degaussing / nulling before taking a measurement.
-
That is only half of the story.
Let's agree on 75%... ;)
-
*Hall effect sensor
-
That is only half of the story.
Let's agree on 75%... ;)
Well.....as this thread doesn't explicitly mention AC/DC current probes, consider the performance of yesteryears entirely passive AC current probes of which P6021/6022 were the industry leaders albeit for current levels that today are anemic.
60 & 100 MHz performance from these great old probes was outstanding for designs now approaching 1/2 a century old.
-
This AliExpress listing shows the two ranges on the CP-07 and CP-07+, 0.1-400mA (1mV/1mA out), and 0.1-4A (1mV/10mA out).
https://www.aliexpress.com/item/32797769120.html (https://www.aliexpress.com/item/32797769120.html)
I noticed there is also a CP-071 which is 0-100mA (10mV/1mA out) and has auto calibration (nulling), but is limited to 1kHz.
https://www.aliexpress.com/item/1005004147673762.html (https://www.aliexpress.com/item/1005004147673762.html)
What do others think about these specifying accuracy in terms of % of reading plus/minus 5 digits?
These are analog out, and there is no ADC or display. So how can there be a digit component to the accuracy spec?
-
Who needs a datasheet when you have an aliexpress ad? 😉😉
ETA: It says 0.1mA to 4A range, but it also says 2mA noise. Not a big issue at 4A, but...
-
Who needs a datasheet when you have an aliexpress ad? 😉😉
Indeed. Also I should note those AE listings do not mention Owon at all, nor did the ones others posted earlier. And none of the yellow units in the photos or videos I've seen have Owon branding on them.
-
It is possible to crudely measure fractions of a mA on the cheap with a DIY probe (photo shows a sawed-off ferrite that was attached to a TI magnetometer eval board, which only provides a bandwidth from DC to just a few kHz, but is very sensitive).
The 'scope trace shows the output (yellow) for 1 mA RMS current (there is sometimes the possibility to switch on averaging etc if an even cleaner output is required).
However it might be more practical to use a current sense resistor if possible if such low current measurement is required and that's a very budget-friendly option.
There are some cheap current transducers (e.g. LEM offers them) which might be an option too, typically with a BW of a few hundred kHz.
EDIT: Second 'scope trace shows a 100uA RMS current, this time at 10 kHz which is pushing it. This time averaging was needed to see it as cleanly as shown. Strapping an amplifier on the output, it's possible to easily measure 10 uA RMS.
-
ETCR007AD - Is the one that all the cool kids are going for. https://www.aliexpress.com/item/1005004233101404.html? (https://www.aliexpress.com/item/1005004233101404.html?)
If you are into dubstep, you gonna love it... the noise is pretty intense with that clamp-meter.
but its so noisy that maybe its defect. https://i.imgur.com/Cj2B5AJ.mp4 (https://i.imgur.com/Cj2B5AJ.mp4)
https://www.eevblog.com/forum/testgear/etcr-007ad-acdc-current-clamp-(ac-45hz-400hz-dc-100khz) (https://www.eevblog.com/forum/testgear/etcr-007ad-acdc-current-clamp-(ac-45hz-400hz-dc-100khz))
But likely why nobody went for it, they got the memo - as pricewise ETCR007AD is between 40 to 50$ and with better paper-specs than fx CC-65.
Though CC-65 struggles with tangible readings down in the single digit mA.
10mA DC
https://i.imgur.com/ILHG1YP (https://i.imgur.com/ILHG1YP)
-
Thanks Martin72 for all the explanations, this quite well aligns with the experiences I had with all the probes.
Then there is the measuring principle with current compensation.
A counter-current is applied via a winding on the core until the difference is zero. This corresponds to the measured current and is processed accordingly.
This has the advantage that the core is not or hardly magnetized, the intrinsically good Micsig CP2100, for example, does not have this, its accuracy decreases with increasing current (tolerance up to 15%).
If a current clamp has a degaussing function, you also know that it follows this principle.
I have two cheap current clamps in my private equipment, one of them also once showed a strong non-linearity (loosing sensitivity for currents >5A). I used an electrical degausser on it and since then it is in its 5%-window again.
From my point of view these cheap probes also are useful if you know about their limitations and have a healthy general scepticism about displayed values. In the low DC-ranges they are affected by stray fields as much as the expensive ones. Of course you need to null them out at zero current, but this has only limited use in changing magnetic fields (as in many metallic electrical structures...) or if the clamp moves during measurement.
And 95% of the measurements I do with this devices are low current DC. I sometimes use the Peaktech-clamp for inrush current measurements but these don't have to be super precise and frequencies >5kHz don't matter much for this. In many cases it´s much better to have the cheap clamp available than to have no clamp ::)
-
ETCR007AD - Is the one that all the cool kids are going for. https://www.aliexpress.com/item/1005004233101404.html? (https://www.aliexpress.com/item/1005004233101404.html?)
Well, the price is cheap enough that it's worth a shot. But wtf does this mean:
Frequency
AC: 45Hz-400Hz
DC: DC-100kHz
DC is DC. Does it mean there's 2 frequency ranges? Why did they display it like that in the manual?
Thanks,
Josh
-
The device itself only says "DC-100kHz".
And from the comment here its really 8kHz: https://www.eevblog.com/forum/testgear/etcr007ad-clamp-proble/ (https://www.eevblog.com/forum/testgear/etcr007ad-clamp-proble/)
Maybe it was cut down to reduce price, or it was just always a bad design.
-
If there isn't even a decent data sheet, I would give it a wide berth.
I was briefly tempted to order one, it is also available from Amazon and is delivered within two weeks.
I was put off by the strange information regarding AC and DC.
Then tolerance...3%...over the entire range?
Speaking of ranges, 100mV/A at one point, then 10mV/A.
Great.
100mV/A from...to...? 10mV/A from...to?
As a rule, the lower the voltage/A, the greater the range.
But let's take the range with 100mV/A.
That's 100µV for 1mA... Have fun measuring with low currents.
-
The device itself only says "DC-100kHz".
And from the comment here its really 8kHz: https://www.eevblog.com/forum/testgear/etcr007ad-clamp-proble/ (https://www.eevblog.com/forum/testgear/etcr007ad-clamp-proble/)
Maybe it was cut down to reduce price, or it was just always a bad design.
Could be, it looks like there is other ETCR-clamp meters using this design, not least one with an on-board-display (stand alone clamp meter)
http://file.yizimg.com/492225/2016112-16458399.pdf (http://file.yizimg.com/492225/2016112-16458399.pdf) (Old ETCR007AD specs*)
ETCR is making a shutload of leakage clamp products, so one would reckon they know how to do it. https://www.etcr-instruments.com/index.php (https://www.etcr-instruments.com/index.php) but their AC+DC ETCR007AD seems like a bad product and also worrying how the details changed, on the "newer" ETCR007AD while the older units sold on Aliexpress' shows values that don't hold water in practice, both datasheet and the units own body-markings.. looks like amp-specs & frequency been dialed down while the 100kHz that previously were highlighted, now been removed. https://tinyurl.com/yckcnjv4 (https://tinyurl.com/yckcnjv4)
Datasheet https://tinyurl.com/fz26h6p9 (https://tinyurl.com/fz26h6p9)
- My ETCR007AD is barebone..
-
Has anybody tried this this generic "C10B" probe?: https://www.aliexpress.us/item/3256805410377735.html (https://www.aliexpress.us/item/3256805410377735.html)
I'm tempted to try that and the ETCR007AD. I assume both will be disappointing, but who knows.
-
Has anybody tried this this generic "C10B" probe?: https://www.aliexpress.us/item/3256805410377735.html (https://www.aliexpress.us/item/3256805410377735.html)
I'm tempted to try that and the ETCR007AD. I assume both will be disappointing, but who knows.
With an output of 400mV/A it seems to be quite sensitive but not in the single digit mA range. Just make sure your scope can be set to 400mV/A scaling otherwise getting a proper reading will be confusing. I think you'll still need high-res mode to see currents in the mA range.
-
Has anybody tried this this generic "C10B" probe?: https://www.aliexpress.us/item/3256805410377735.html (https://www.aliexpress.us/item/3256805410377735.html)
I'm tempted to try that and the ETCR007AD. I assume both will be disappointing, but who knows.
With an output of 400mV/A it seems to be quite sensitive but not in the single digit mA range. Just make sure your scope can be set to 400mV/A scaling otherwise getting a proper reading will be confusing. I think you'll still need high-res mode to see currents in the mA range.
I have the SDS2504X HD. Noise from the probes is what will likely be the issue.
For example, the ETCR007AD says it can measure 1mA, but it also says there's 2mA noise, so... GLWT? 😉
The LOTO C10B seems more interesting, but it's cheap, so my expectations are low. I'll mostly be probing DC in the 1mA to 40mA range. Possibly some AC stuff, and possibly a little higher current.
-
For very low DC-only currents, the Fluke 771 or similar is unbeatable for the price in my opinion (eBay). You are stuck reading the display, but otherwise it's very accurate. At the lowest range, with care you will get +/- 0.05mA or better, quick and dirty you will always be within 1mA...
CP-07+ is typically off by up to a few mA, but otherwise still seems pretty reasonable for the price. "Is this 0mA or 20mA or 40mA" will be trivial. Also, with a clamp you have the luxury of being able to characterize it out of circuit. That info can be applied when testing in-circuit.
-
As low as my expectations were for the LOTO C10B, I don't think they were low enough. Before I assume that this is absolute trash, let's make sure it's not a PEBKAC error.
The ad says it's full current at 4V, and it's a 10A model. So 400mA/V, correct?
It claims that it can measure 1mA, but the noise says that's bullshit. The ad also claims 1% accuracy, which seems totally believable. 🤣
I tried zeroing as best I could, but the pot is too sensitive, so it wasn't easy.
I tried looping the wire through the probe to reduce the noise, but I didn't see anything to compensate for the loops on the scope. Do I have to manually offset that? So 4 loops would be 1.6V/A? Scope is Siglent SDS2504X HD.
Thanks,
Josh
-
Your trigger is “out” on every image....
You get more signal if you put more lines through the current clamp, that's right.
e.g. 10mA actual current times 4 turns through the clamp makes 40mA measured current.
Noise reduction on the scope:
bandwidth limit on, use Eres or averaging.
How come I knew in advance that the clamp was a piece of junk...
-
How come I knew in advance that the clamp was a piece of junk...
I think we all knew it was a piece of junk in advance. 😉
It was cheap enough to mess with...before I return it anyway. 🤣
I would have kept it if it sucked any less than expected, but it disappointed my already low expectations.
-
Is that the same LOTO that makes some USB scopes?
If so, we had someone from the company as a member here - but it looks like he hasn't been online for quite a while.
https://www.eevblog.com/forum/profile/?u=419857 (https://www.eevblog.com/forum/profile/?u=419857)
-
How come I knew in advance that the clamp was a piece of junk...
I think we all knew it was a piece of junk in advance. 😉
It was cheap enough to mess with...before I return it anyway. 🤣
I would have kept it if it sucked any less than expected, but it disappointed my already low expectations.
Have you tried adding an RC filter to the output of the current probe? With a cut-off say around 1kHz.
-
Is that the same LOTO that makes some USB scopes?
Yes, which doesn't help my impression of their scopes at all. 😉
Have you tried adding an RC filter to the output of the current probe? With a cut-off say around 1kHz.
Might be a fun experiment, but I've satisfied my curiosity with this lame probe. Now I'll wait until I find a good deal on one that doesn't suck...as much.
Thanks,
Josh
-
Test setup is SPD3303X to ET5420A+, 18V and 200mA. Wire wrapped through probe clamp 4 times.
-
What is that, is that supposed to be a DC current?
Or does the ET load have so much ripple...
Do you possibly have a normal load resistor?
-
The ET load very probably causes the ripple. It is 100Hz so it must be mains related.
-
Load resistor is a good idea, I'll try that next.
It could be ripple from the ET, it is supposed to be DC. It did get better at higher current.
I had to reduce wraps after 3A for probably obvious reasons.
-
It is 100Hz so it must be mains related.
It would have to be 120Hz then.
This reminds me to finally test the Siglent DC load properly. ;)
-
It's useless for my needs with lower currents, but it seems for somebody looking for something super cheap for use above 2 or 3A, it's not terrible. I'm tempted to keep it for that range, but then it will probably just be another probe I eventually forget I own. 🤷
-
As I said, I had already tested a few current clamps for an oscilloscope.
Owon, Hantek, twice Micsig, an Instrance (whatever that name is supposed to be), Siglent, Tektronix.
The last two are suitable for low currents – but also the most expensive.
However, I would place the Micsig current clamps just behind them.
-
As I said, I had already tested a few current clamps for an oscilloscope.
Owon, Hantek, twice Micsig, an Instrance (whatever that name is supposed to be), Siglent, Tektronix.
The last two are suitable for low currents – but also the most expensive.
However, I would place the Micsig current clamps just behind them.
I wish you also tried out the Pico TA189. That one looks promising. I think your company needs you to try one out. 😉
-
In fact, I am looking for a current clamp in exactly this design for our test field, to replace the outstanding, but no longer manufactured AP011 from LeCroy.
However, 30A is not enough for us, the AP011 could handle 120A.
This Pico has trustworthy data.
-
In fact, I am looking for a current clamp in exactly this design for our test field
Perfect! You should test out a combo of the TA189 (30A) and the TA167 (200A/2000A) probes. 😉
-
Hehehe....
If I am looking for a current clamp that can measure small currents as accurately as possible, then it will rarely be a current clamp that can also measure 100A or more.
Also, such current clamps are physically large because the lines that can transport more current are large.
So one thing bites the other: if I want to measure mA, the clamp should be able to properly detect physically small wires.
That's why we have different clamps at work.
Regarding the picos:
The TA189 has the measuring principle I prefer, so its data will be “correct”.
However, it is basically “huge”.
The TA167 is only suitable for “power”, where it doesn't matter if you have an ampere more or less.
-
Load resistor is a good idea, I'll try that next.
It could be ripple from the ET, it is supposed to be DC. It did get better at higher current.
Try a resistor for sure. From previous threads about DC loads and my own experience with a Korad DC load it looks like many of the lower end DC loads have trouble with the magnetic field from the mains transformer influencing the analog control circuitry and causing a mains ripple at the output. In my Korad KEL2010 I have moved the transformer about 20cm away AND packed it in mu-metal which took care of most of the mains ripple but not all. The Korad DC load casing is huge, if you have a more compact DC load, chances of improving it are slim. Another problem the KEL2010 had was the magnetic field from the fan also causing ripple in the current regulation so I have replaced the fan as well (which solved the ripple due to the fan completely). And it is also possible the PSU doesn't like a DC load (after all a current sink / source has a high impedance by definition) and the PSU starts to oscillate a bit. When testing the KEL2010, I used a Li-ion battery pack to have a super clean, low impedance DC power source to rule out any possible influence that may be caused by the control loop of a PSU.
@Martin72: you are right about 120Hz.
Edit: on second inspection it looks like there is more going on in KungFuJosh's screendumps and I suspect there is an interaction between the PSU and DC load as well.
-
From previous threads about DC loads and my own experience with a Korad DC load it looks like many of the lower end DC loads have trouble with the magnetic field from the mains transformer influencing the analog control circuitry and causing a mains ripple at the output.
In my opinion, this is a serious design flaw, so it will be all the more interesting to find out whether my siglent last behaves in the same way.
-
In my opinion, this is a serious design flaw, so it will be all the more interesting to find out whether my siglent last behaves in the same way.
I'm curious about that too. I dunno if I'd get one anyway, this load was free. Hard to beat that. 😉
Hmm, it might be an interesting project to see what can be done in the ET to get rid of the ripple.
-
Perfect! You should test out a combo of the TA189 (30A) and the TA167 (200A/2000A) probes. 😉
I have the TA167, it was a cheap 2nd hand by-catch. Since I rarely need to measure current, it was laying around unused, until tonight. Why not finally trying it out...
Setup:
R&S HMP4040 set to 1.998A, K2001 displays 1.99982A DC + 0.15mA AC RMS (these are two measurements, with the AC measurement forced to the 2A range).
The wire makes 4 turns through the TA167, 200A range, a roll of cloth makes sure that the wire is centered in the clamp.
I thought a current of 2A with 4 turns fits well, because this is 4% of the range, approx. the same as the 200mA, 4 turns for the TA189.
Results, K2001 (for accuracy), after zeroing the TA167, average values:
0A: 0.09mV DC => 9mA => 2.25mA per turn (offset)
2A: 79.92mV DC => 7.992A => 1.998A per turn
-2A: -79.90mV DC => 7.990A => 1.9975A per turn
gain error: -0.12%
Results, Rigol MSO5074, for noise, 20MHz BW limit, 16 avarages, parameter Vavg2 is averaged over trace with stats:
2A: Pico TA167 2A 4turns.png => flat line, 158,57µV std => 15.9mA => 4.0mA per turn
0A: Pico TA167 0A.png => flat line, 57,943µV std => 5.8mA => 1.44mA per turn
Of course, these Vavg2 values are averaged over the full data set of 1000, acquired in 160s... The trace is from one measurement and looks good, though.
-
I have the TA167, it was a cheap 2nd hand by-catch. Since I rarely need to measure current, it was laying around unused, until tonight. Why not finally trying it out...
Thanks for sharing that! That supports my good impression of the Pico stuff.
I also rarely ever need to measure current, so the price tag for something I'll use a lot for a week then store in a case forever isn't inspiring. 😉 I'll try and be patient now and see if a good deal pops up on eBay or something.
Thanks,
Josh
-
interesting to find out whether my siglent last behaves in the same way.
Looks like you might have a load improvement project too: https://www.eevblog.com/forum/testgear/micsig-cp1003cp503-100mhz50mhz-current-probe/msg5689885/#msg5689885 (https://www.eevblog.com/forum/testgear/micsig-cp1003cp503-100mhz50mhz-current-probe/msg5689885/#msg5689885)
-
Just done...
Conventional linear power supply, then the Siglent load.
The current clamp is the Tektronix, the clamp itself is small, but the supply for it takes up space.
The current is 1A, the resolution is 200mV/div (on the AM501) or 10mV/ on the Magnova.
As you can see, there is a small ripple that disappears completely with averaging.
If you increase the resolution (AC coupling) and set the cursors, it is exactly 100Hz - classic mains ripple.
But for me, this is almost too “small” to seriously tinker with the load.
A “final” proof would be to use a DC load via an isolating transformer and a battery as the source to see if the ripple then becomes smaller.
-
If I'm not mistaken the ripple is around 15mA peak-peak (assuming a scale of 50mV/A).
-
It is 200mA/div, with an output of 10mV.
A bit stupid to calculate, that's the disadvantage if you want to document something
The ripple height always remains the same, regardless of whether 1A or 100mA is flowing, whether CC or CR mode.
I think it's an interference.
But I'll add a few more pictures.
-
The ripple height always remains the same, regardless of whether 1A or 100mA is flowing,
Not quite, but not dramatically more either.
Settings for the ripple pics: 2mA/div on the AM501.
Trigger source set to AC.
-
I'm tempted to keep it for that range, but then it will probably just be another probe I eventually forget I own. 🤷
How dare you! :-DD
I just discovered another clamp meter in a drawer that I had forgotten all about.
-
How dare you! :-DD
I just discovered another clamp meter in a drawer that I had forgotten all about.
:-DD :-DD
-
For reference, the one I forgot I had is a Fluke i30s AC/DC Current Probe with BNC connector.
On review the specs are pretty similar (almost identical) to the Pico TA189:
- 30A with 100mV/A output
- 1ma resolution (surely that depends on the scope)
- 1% of reading +/- 2mA accuracy
- DC to 100kHz (0.5dB)
The only differences I note are different colours, control positions, and maximum conductor/opening size (Fluke is 19mm, Pico is 25mm).
There is a supplement for the Fluke i30s which changes:
- "specified" measuring range from "20A AC RMS or DC" to "30mA to 30A DC, 30mA to 20A AC RMS"
- "usable" measuring range from "+/-30A" to "5mA to 30A DC, 30mA to 20A AC RMS"
The reason I forgot about The Fluke i30s is because I typically use my Micsig CP2100B, but I had misplaced that and during the search for it I also found the Fluke.
-
That's hilarious.
Unfortunately, most of what I want to measure will be between 1mA and 30mA. I don't know if the Pico can really do that either. 🤷
Thanks,
Josh
-
The problem there is that stray magnetic fields very often are in the range of 1-30mA, even with the best current clamp with small clamp area.
The good scope clamps are magnetically better shielded than the big DMM-type ones. But anyhow, single mA are not the scale where current clamps feel at home.
-
That's all true, but the Tektronix is a current clamp that can handle this best of all.
That's why I still have it... ;)
At 1mA/div, you also have to use averaging, but hey... 1mA... per division... 8)
You could do without it from 2mA/div, and anyway above that.
Here are a few pictures, split because of the maximum number.
-
The second batch....
Now with “current”, simple DC, simple with resistance - a DC load makes no sense at such currents (lower mA range), you have to emphasize that across brands again, because I'm afraid many believe that.
-
Finally, a few pictures with AC signals, current is 2mA/div., various frequencies/waveforms.
With/without averaging.
Devices involved:
DC linear supply, Tektronix A6302 with AM501A unit in TM501 housing, various real resistors, SDG2122X, SDS3104X HD, “special adapter” for the A6302 for the AC measurements.
Pictures of the setup will follow.
-
Setup from the afternoon(SDG2122X not shown)....
-
One only hopes screenshots get taken with A/div settings....... :P
-
That's why I marked the pictures accordingly.
-
Just to show how bad it can be with cheap equipment:
[attachimg=1]
Ch1 is the output of a signal generator with a squarewave with 10V_pp. It is connected to a 330 \$\Omega\$ resistor, so ca. 33mA_pp are flowing. Ch2 is the output of the Peaktech 4250. The screenshot speaks for itself - if really needed you may estimate the amplitude of the waveform by filtering & cursors - but it´s just not suited for such a job.
As a DMM attachment it is a little more useful for small currents. But definitely not comparable to the very decent results of Martin72s gear.
Maybe I can try the same test with the Micsig CP2100B end of the week, but last time I compared them against each other there was not a huge difference.
-
Just to show how bad it can be with cheap equipment:
FYI, there are some potential improvements you can try
https://www.eevblog.com/forum/testgear/peaktech-4250-acdc-current-clamp-teardown-and-repair/ (https://www.eevblog.com/forum/testgear/peaktech-4250-acdc-current-clamp-teardown-and-repair/)
Not sure if they will help with low current measurements.
-
Just to show how bad it can be with cheap equipment:
FYI, there are some potential improvements you can try
https://www.eevblog.com/forum/testgear/peaktech-4250-acdc-current-clamp-teardown-and-repair/ (https://www.eevblog.com/forum/testgear/peaktech-4250-acdc-current-clamp-teardown-and-repair/)
Not sure if they will help with low current measurements.
I don't think so. With a range of 60A, the noise is just too much to measure in single digit mA range. However, I think the lower range clamp which KungFuJosh has tested a few posts up, has the ability to measure currents in the mA range. Only nulling it precisely will be tricky and noise will need to be filtered / averaged.
-
Maybe I can try the same test with the Micsig CP2100B end of the week, but last time I compared them against each other there was not a huge difference.
I had once opened a thread about a Micsig CP2100, with various tests:
https://www.eevblog.com/forum/testgear/micsig-current-probe-cp2100b-tests-and-comparing/msg3070834/#msg3070834 (https://www.eevblog.com/forum/testgear/micsig-current-probe-cp2100b-tests-and-comparing/msg3070834/#msg3070834)
(and the following pages)
-
Tried with the Hantek CC65 (also 60A rated clamp) to see if it could muster minute current clamp values.
The setup is as crude as it gets..
Top of a washing machine, no shielding..
Sig gen DG800PRO 50Hz sine, sq w. with a cheap BNC cable to alligator fitted with 10R metal film resistor.
1mA pr div. (as low as the scope will go) with lpfiltering & avg.
The noise from the clamp can be cleaned up quite a bit on these Micsig scopes with both variable bandwidth filtering and averaging, so even minute mA, triggered..
[attachimg=4]
[attachimg=2]
[attachimg=1]
[attachimg=3]
Setup.
https://www.youtube.com/watch?v=ylOdk6SJ6U8 (https://www.youtube.com/watch?v=ylOdk6SJ6U8)
-
Tried with the Hantek CC65 (also 60A rated clamp) to see if it could muster minute current clamp values.
Could you please explain your setup in more detail (the video is a bit blurred)? I'm a bit puzzled trying to get the numbers together: The Hantek has, in its more sensitive range, 1 mV / 10mA, which is equivalent to 100µV / 1mA. I checked the Micsig range at Batronix and could not find a model with a sensitivity higher than 1mV/DIV. The 2nd screenshot tells 1.52mA PK-PK and the trace covers about 1.5DIVs. Doesn't that require a 100µV/DIV range? OTOH, the AWG seems to read 700mV, and, if I understand correctly, you used a 10Ω resistor as load? With 50Ω output impedance, this would lead to 700mV/60Ω = 11.67mA. Should this be the actual current, the resolution of the measurement would still be impressive, especially because of the stable triggering, since the cleanup through averaging cannot happen without stable triggering on the single acquisition to average. So please tell something on the variable filtering your instrument has!
-
The pictures aint consistent, it cycles between 1x and 10x, the vpp-values on the sig-gen, also vary.
It was more an attempt to see how little noise it could muster at relatively low values, the LP-bandwidth filters do make quite an impact...
Yes 1mV/10mA on CC65, scope with 30KHz LP, with averaging and 10 Ohm resistor as load...
The values in itself, take it with a pinch of salt, its 9v battery was down in the high 6v region, so far from ideal but the only 9v I had on hand.
Will try to take a look at both ETCR007A and CC65 when I get some 9v batteries..
CC65 (60A 20KHz) 2 settings. [1mV/10mA & 1mV/100mA]
ETCR-007A (50A 100KHz) 2 settings [10mV/A & 100mV/A]
-
I have a CC65 coming today to play with. I'll see how it does compared to the crappy one I'm returning today. 😉
-
My CC-65 gives me VERY inconsistent results. I test it with known small currents
of under 10 mA and some times it's right on the money, and sometimes it is off by 20 mA. I find that I can't trust it's results. I've tried it with high end DMMs and oscilloscopes. :scared:
-
I'll mention the HP428 again. It has a full scale range of 1 mA. They're not uncommon and can often be had for under a hundred bucks. A DC instrument. Some detailed info:
https://www.prc68.com/I/HP428.html (https://www.prc68.com/I/HP428.html)
https://www.diyaudio.com/community/threads/modernized-hp428-clone.270510/ (https://www.diyaudio.com/community/threads/modernized-hp428-clone.270510/)
-
The CC65 arrived today, and my expectations are that it will be better than the other crappy probe, and besides that, probably still pretty crappy.
I ordered the Tek TM502A / AM503B / A6302 Probe on eBay. I expect that to be a little better.
Thanks,
Josh
-
I expect that to be a little better.
YMMD... :-DD
-
:-DD :-DD
I'm actually surprised; the CC65 is lightyears better than the other piece of crap. I dunno about repeatability or accuracy or anything that matters, but I couldn't get the other probe to do even this...
I have the SDG2122X connected with some mini-grabbers to a piece of copper wire going through the current probe. Settings are 1kHz, 1Vpp. Probe set to 100mV/A.
ETA: ignore the filename, the last screenshot has averaging turned off.
-
The low mA range on the CC-65 is mediocre, but better than the previous piece of junk.
Switching to the 10mA/V setting, the CC-65 was pretty consistent with the DCA stepping through to 6A.
-
One more thing:
The termination impedance is not 50 ohms, as far as I can remember regarding the CC-65.....
-
Their amazing manual says typical load is 10kΩ.
Here's a capture with DC1M on CH1.
-
Holy shit, Batman! Martin, when you said this thing was big, I had no idea how ridiculous it is. 🤣🤣🤣
(https://www.eevblog.com/forum/testgear/accurate-low-amp-current-probe-advice/?action=dlattach;attach=2423121;image)
Only problem is that the probe that came with the set has a small issue with the switch. |O
(https://www.eevblog.com/forum/testgear/accurate-low-amp-current-probe-advice/?action=dlattach;attach=2423125;image)
-
Only problem is that the probe that came with the set has a small issue with the switch. |O
Which forms part of the jaw locking device.......
-
I still have a few parts... 8)
-
I set the SDG2122X output to 369mV at 1kHz, and that measured 5.00mA on my SDM3065X.
Then I setup the AM503B at 1mA/DIV with 20M limit.
This looks pretty good, right? Did I do anything stupid in my setup? 🤔
...because after reconnecting to the SDM, I'm getting 1.713mA. I assume there's an equation I should know here. 😉
-
As I mentioned before, you set 10mV/div, 1:1 and 50Ohm on the scope itself and leave it as it is.
Everything else (bandwidth limitation, scaling, AC, DC) is set on the 503B.
I set the SDG2122X output to 369mV at 1kHz, and that measured 5.00mA on my SDM3065X.
So you used a 73.8 ohm resistor as a load?
-
As I mentioned before, you set 10mV/div, 1:1 and 50Ohm on the scope itself and leave it as it is.
Everything else (bandwidth limitation, scaling, AC, DC) is set on the 503B.
I set the SDG2122X output to 369mV at 1kHz, and that measured 5.00mA on my SDM3065X.
So you used a 73.8 ohm resistor as a load?
A piece of copper wire with some insulation on it.
-
I see, so you use the scope's internal 50 ohms as a load and connect the generator to the scope with this copper wire?
-
I see, so you use the scope's internal 50 ohms as a load and connect the generator to the scope with this copper wire?
No, I connected the copper wire between the positive/negative grabbers on the cable from the AWG. Then ran that through the current clamp. I also tried the same way with a 100Ω resistor, but I guess I'm doing it wrong. 😉
-
No, I connected the copper wire between the positive/negative grabbers on the cable from the AWG.
That would be a direct short circuit and you don't know how the SDG handles something like that (current limiting).
I would continue trying with the right resistors.
Or you have your ET load, use that first to see if the clamp is working properly.
Since the button was off the clamp:
Did you make sure that no “Probe open” indicator was lit on the AM503B?
-
No, I connected the copper wire between the positive/negative grabbers on the cable from the AWG.
That would be a direct short circuit and you don't know how the SDG handles something like that (current limiting).
I would continue trying with the right resistors.
Or you have your ET load, use that first to see if the clamp is working properly.
Since the button was off the clamp:
Did you make sure that no “Probe open” indicator was lit on the AM503B?
I switched to DC with the ET load already, but I don't know what I should set things as. So how would you do it?
It's the AM503B, so I can set specific mA/DIV values (1,2,10,20,50,100 etc.)
I want to see what the low end is that this can measure, but I'm having trouble getting it to do anything in that range without lots of guessing.
How would you set this up? Tell me everything from the current amplifier to the scope and trigger settings. I know the scope settings are supposed to be 10mV/DIV, but why can't we change the display to A on the modern scope? There must be settings we can use related to the settings on the current amplifier?
Right now I have the PSU supplying 18V with no current limit. The ET load is set to draw 30mA. The current probe is on the negative lead with the arrow pointing towards the PSU negative, as the manual shows.
Thanks,
Josh
-
Hi,
Right now I have the PSU supplying 18V with no current limit. The ET load is set to draw 30mA. The current probe is on the negative lead with the arrow pointing towards the PSU negative, as the manual shows.
I would then proceed as follows:
- As I said, set the scope to 10mV/div and 1:1 coupling, 50 Ohm DC
- Set the following on the AM503:
5mA/div, make sure that none of the Probe Open lights up, make sure that “DC” is lit - Because after switching on, the AM503 is always set to “REF” and you can't see anything on the scope....
And before you put the current clamp in the line, do a auto-degaussing.
At 30mA and 5ma/div on the AM503, you should set the vertical line on the scope to the last division (5mA times 6 boxes).
Try it out first before we worry about the conversion on the scope.
-
No, I connected the copper wire between the positive/negative grabbers on the cable from the AWG.
That would be a direct short circuit and you don't know how the SDG handles something like that (current limiting).
I would continue trying with the right resistors.
Or you have your ET load, use that first to see if the clamp is working properly.
Since the button was off the clamp:
Did you make sure that no “Probe open” indicator was lit on the AM503B?
I switched to DC with the ET load already, but I don't know what I should set things as. So how would you do it?
It's the AM503B, so I can set specific mA/DIV values (1,2,10,20,50,100 etc.)
I want to see what the low end is that this can measure, but I'm having trouble getting it to do anything in that range without lots of guessing.
How would you set this up? Tell me everything from the current amplifier to the scope and trigger settings. I know the scope settings are supposed to be 10mV/DIV, but why can't we change the display to A on the modern scope? There must be settings we can use related to the settings on the current amplifier?
Right now I have the PSU supplying 18V with no current limit. The ET load is set to draw 30mA. The current probe is on the negative lead with the arrow pointing towards the PSU negative, as the manual shows.
Thanks,
Josh
Been thinking about too this as my Siglent active current probe is soon to arrive....
SDG models can supply 200mA......
If we set one for DC supply and some V output into an appropriate load we can measure/check the current drawn from the SDG using a meter we have on hand and trust.
Then check our current probes are displaying similar values......
Then check again with waveforms activated......
-
I would then proceed as follows:
- As I said, set the scope to 10mV/div and 1:1 coupling, 50 Ohm DC
- Set the following on the AM503:
5mA/div, make sure that none of the Probe Open lights up, make sure that “DC” is lit - Because after switching on, the AM503 is always set to “REF” and you can't see anything on the scope....
And before you put the current clamp in the line, do a auto-degaussing.
At 30mA and 5ma/div on the AM503, you should set the vertical line on the scope to the last division (5mA times 6 boxes).
Try it out first before we worry about the conversion on the scope.
I (think) I did all of that, and this is what I'm getting. I set to -60mV on the position so the wave would be visible. It's triggering off AC instead of the DC signal.
30mA current draw of the load is confirmed with the SDM in circuit showing 29.9mA.
Thanks,
Josh
-
WTF.... :o
Off the top of my head, I would say that either the current clamp or the AM503B module is defective – it should be a straight line, just DC current.
But let me recreate this at my place tomorrow and make a short clip of how I set it up.
If you can't do it exactly the same way, something is defective.
-
WTF.... :o
Off the top of my head, I would say that either the current clamp or the AM503B module is defective – it should be a straight line, just DC current.
But let me recreate this at my place tomorrow and make a short clip of how I set it up.
If you can't do it exactly the same way, something is defective.
I assume there's AC interfering with the DC current, but the frequencies are goofy (to me, who knows nothing 😉). The measure counter shows frequencies around 85Hz, then it jumps to 667Hz with that test.
Now I'm trying 1A DC to lose whatever AC is below that, but it seems like maybe the amplifier is an issue. I have it set to 100mA/div and the scope is showing about 85mV, and the frequencies are jumping around in the low kHz range.
Maybe the filter caps in the current amp need to be replaced? Or maybe I'm doing something dumb. I dunno. I'll try AC again with a 100Ω resistor.
-
Stick with the 1A DC, set the 503 to 0.2A/div and close all the measurement orgies on the scope so that you only see the screen.
-
Stick with the 1A DC, set the 503 to 0.2A/div and close all the measurement orgies on the scope so that you only see the screen.
I assume this is about what should be expected at 200mA/DIV?
-
If you could now set the scope to 10mV/div...
-
If you could now set the scope to 10mV/div...
I did and changed the image faster than you replied. 😉😉
-
OK,
Except for the ripple, it looks good.
Now leave the settings on the scope as they are (except timebase, set it to say 1ms/div)and reduce the current, while simultaneously reducing the sensitivity on the 503.
e.g. 500ma current, 100mA/div on the 503.
Then 250mA, 50mA/div and so on.
-
WTF.... :o
Off the top of my head, I would say that either the current clamp or the AM503B module is defective – it should be a straight line, just DC current.
Remember about the interference from the transformer affecting performance of DC loads at these current levels... IOW: I don't think the clamp is broken, it is the test setup.
@KungFuJosh: Test the DC level with a battery and resistor. This way you eliminate all the unknowns that may be introduced by DC load and/or PSU. Secondly, measure the current using a shunt resistor. At low currents in the tens of mA range, a 1 Ohm resistor is a good choice and will give a solid foundation to base an conclusions on.
-
OK,
Except for the ripple, it looks good.
Now leave the settings on the scope as they are (except timebase, set it to say 1ms/div)and reduce the current, while simultaneously reducing the sensitivity on the 503.
e.g. 500ma current, 100mA/div on the 503.
Then 250mA, 50mA/div and so on.
-
I see that you keep going “out” with the vertical zero line, is the ripple that big?
Now you would have to test whether it is the load, as Nico had already posted.
I will do the same with my Siglent load and a power supply “tomorrow”, I “only” have the AM503A, but that doesn't matter in this case.
-
I don't have a THT 1Ω resistor handy, so I used a 100Ω resistor connected to a 9V battery. In series with the SDM3055X-E it measured 82mA.
10mA/DIV on the current amp.
-
In series with the SDM3055X-E it measured 82mA.
Then please set 503 to 20mA/div....
-
In series with the SDM3055X-E it measured 82mA.
Then please set 503 to 20mA/div....
I did that too, and it looked the same except the mean was about 39mV. I could do it again and screen capture it if you want.
-
That would be nice, because I want to be able to see the zero line on the scope.
-
That would be nice, because I want to be able to see the zero line on the scope.
I switched to a 1.21k resistor to lower the current. SDM in circuit shows 7.19mA.
Here the current amp is set to 5mA/div. What should I actually set it to? 😉
Thanks,
Josh
-
As I said, I will make a short clip later in which I measure various currents with my Tek System.
If you get it into your head that nothing more needs to be set on the scope except for the basic condition, the rest only on the 503, then it will be easier to work with.
At the moment, it's just a matter of determining whether your Tek System is working properly.
And we will find that out later.
-
The inside of the amp doesn't look too bad. The stickers think it was last calibrated in 2014, and the "cal void if cut" stickers were already cut.
-
https://www.youtube.com/watch?v=Q9vVEaNeqZQ (https://www.youtube.com/watch?v=Q9vVEaNeqZQ)
Short clip about initial conditions and degaussing the probe.
-
https://www.youtube.com/watch?v=lGVDFRFfKEE (https://www.youtube.com/watch?v=lGVDFRFfKEE)
Various loading situations with the electrical DC load.
You can clearly see the influence of the load's ripple, which naturally decreases as the current increases.
-
I had already done current measurements with normal resistances:
https://www.eevblog.com/forum/testgear/accurate-low-amp-current-probe-advice/msg5692771/#msg5692771 (https://www.eevblog.com/forum/testgear/accurate-low-amp-current-probe-advice/msg5692771/#msg5692771)
I'll have to revive my SDL1000X thread soon... ;)
-
These are the transmission ratios that you have to set in the channel setup under user defined probe to get correct measurement results.
The only disadvantage is that you have to set them for each current range, but you don't change that often and you have two user probes that can be set.
(https://www.eevblog.com/forum/testgear/accurate-low-amp-current-probe-advice/?action=dlattach;attach=2424301;image)
-
These are the transmission ratios that you have to set in the channel setup under user defined probe to get correct measurement results.
The only disadvantage is that you have to set them for each current range, but you don't change that often and you have two user probes that can be set.
(https://www.eevblog.com/forum/testgear/accurate-low-amp-current-probe-advice/?action=dlattach;attach=2424301;image)
Yep, active current probe terminations are a PITA whereas a passive termination is set and forget.
Once you go down the DC capable current probe route things become somewhat more complex and harder to use compared to an current probe with plug and play scope support.
-
I would have preferred an SCP5030, but it's not going to fit into my budget this year.
Probably not next year either. ;)
-
I took apart the amp again and took better photos. I like the chassis design; it makes repairs easier.
Here's the electrolytic caps in the 503B:
35V 100uF LS:5mm x7
16V 220uF LS:5mm x3
50V 1uF LS:2mm? x1
50V 2000uF LS:7.5mm D:16mm H:35mm x2
50V 1000uF LS:7.5mm D:14mm H:25mm x2
Marked 85C
It looks like 4 of the caps were already replaced at some point.
Thanks,
Josh
-
Here are my test results using the ET load.
I also did screenshots with the probe not connected to anything.
Thanks,
Josh
-
Hi,
Now do the same tests with "real" resistors..
-
Hi,
Now do the same tests with "real" resistors..
What resistor values? What power source? I assume batteries, but specify which type please.
Thanks,
Josh
-
Battery does not have to be.
For small currents I used a 1k resistor, for larger currents 100 Ohm, and then adjusted the voltage and current with the power supply unit to achieve the desired values.
See link in reply #152
-
Speaking of batteries, the battery in my AM503B is weak (.6V, should be ~1.5V), so I removed it to see what it is (390 or SR1130SW). Removing it cleared the NVRAM and reset the amp to factory default (355 error after power on).
You had mentioned before that the AM503 normally starts up with ref coupling, but mine had always defaulted to DC, never ref on start. After reinserting the dying battery, now it starts up as you said in ref coupling. I wonder what other weird things were set. 🤷 Anyway, I ordered a new battery.
Thanks,
Josh
-
This probe has some serious drift. Note the timestamps and scale offset. This was after over 30 minutes warmup time. I'll let it warm up longer and see if it ever stabilizes at all.
Thanks,
Josh
-
Here are those tests. They're not terrible, but none are dead on. I degaussed multiple times throughout the process to try and get it closer.
I also had one of my SDMs in circuit to verify current level was accurate within at least 1 or 2 decimal places.
Thanks for the tests, now I think I understand how the current amplifier works with regards to the scope scale. 10mA/div is 1:1 with 10mV/div on scope, and each variation changes the ratio.
This needed over an hour to get to the mediocre stability it had. I'm assuming that's more the probe's fault than the amplifiers. Am I also correct to assume the amplifier seems good, but the probe is not great?
Thanks,
Josh
-
If you lost your NVRAM, you also lost your gain calibration constants.
Take a look here:
https://w140.com/tekwiki/wiki/AM503B
And specifically the documents:
AM503B/5030, A6300 067-0271-00 Verification and Adjustment Kit Manual 070-9352-01
AM503B/AM5030 Calibration Adapters
You might want to go through the verification procedure to at least determine if the inaccuracy is due to the amp or the probe.
For the AM503B, calibration requires temporary installation of the GPIB option card (as found in the AM5030), a TM5000 mainframe, and a computer with GPIB. Fortunately, someone has reverse-engineered the contents of the NVRAM, and it's possible to just write the new cal values directly to the NVRAM:
https://groups.io/g/TekScopes/files/AM503B%20Calibration%20without%20GPIB/AM503B_AM5030%20NVRAM%20Replacement_V01.pdf
-
Now I recall why I never bought that Tektronix current clamp... 8)
-
Are you kidding? The custom FRAM chip update project looks fun! I'm getting my Mouser order ready now. 😉
It's already pretty close, but if a fun project and some tweaking gets it dead on, that's pretty cool. I'm not going to spend $5000 on a probe any time soon, so this is the best option.
I think the best reason to avoid this probe is the physical size of the current amp it connects to.
-
Fortunately, someone has reverse-engineered the contents of the NVRAM, and it's possible to just write the new cal values directly to the NVRAM:
Thanks! Do you know if they shared the spreadsheet shown in the PDF? I can't find it anywhere.
Thanks,
Josh
-
Now I recall why I never bought that Tektronix current clamp... 8)
Because you have a functioning crystal ball. ;)
The battery of my 503A was dead too, but changing it was no problem because I had followed these instructions.
-
Fortunately, someone has reverse-engineered the contents of the NVRAM, and it's possible to just write the new cal values directly to the NVRAM:
Thanks! Do you know if they shared the spreadsheet shown in the PDF? I can't find it anywhere.
Thanks,
Josh
I haven't seen it anywhere. However, the data in NVRAM is sparse and it looks to be only about 25 bytes. It wouldn't take long to type it in if you're looking to copy their NVRAM. Some of the values may already have been initialized to defaults if the battery was detected as dead.
-
Now I recall why I never bought that Tektronix current clamp... 8)
Because you have a functioning crystal ball. ;)
The battery of my 503A was dead too, but changing it was no problem because I had followed these instructions.
It's not the same for the 503B. My test mode maxes out at 53. 52 is to degauss the amp itself. Nothing else is really listed in the manual. The battery replacement section in the manual also doesn't say anything about tests after replacing the battery.
My tests were pretty close anyway, I don't know if it really lost any data. It's been a decade since it was calibrated according to the stickers, so I would guess nothing was lost based on that alone considering how close it was.
However, if I want to get it to be more accurate, I'm guessing the easiest way is with the NVRAM replacement game.
-
I haven't seen it anywhere. However, the data in NVRAM is sparse and it looks to be only about 25 bytes. It wouldn't take long to type it in if you're looking to copy their NVRAM. Some of the values may already have been initialized to defaults if the battery was detected as dead.
Well, I was hoping it autogenerated the cal constant hexadecimal codes so I could be lazy about that part. 😉
-
Now I recall why I never bought that Tektronix current clamp... 8)
Because you have a functioning crystal ball. ;)
The battery of my 503A was dead too, but changing it was no problem because I had followed these instructions.
The AM503A has a PCF8582A EEPROM, which is probably used to store the cal data. The AM503B only has a PCF8570 low-power CMOS memory.
The AM503A also has a PCF8570, and I guess the procedure is copying the data between the two memory chips, but I'm not clear which one it means when it says "NVRAM".
I guess my crystal ball is not so good. I have 2x AM503, 2x AM503B, 2x A6302, 1x A6303, and 2x TCP202. They all have various drift, noise, and fidelity issues. I use the AM503B the most because I like the auto balance, but it has processor clock noise in the output which can be annoying. In those cases, I put up with the fiddly drift in the AM503.
-
They all have various drift, noise, and fidelity issues. I use the AM503B the most because I like the auto balance, but it has processor clock noise in the output which can be annoying.
Is there anything we can do to mitigate any of those issues? Currently (pun intended), the drift is the most annoying issue.
-
This is kinda funny. I think I scared the 503B. I removed the PCF8570P and installed a socket there, then inserted the chip. Now it's behaving a little better. 🤣
The drift is still annoying, but it seems to improve depending on warmup time and room temp. I degaussed between every value change, but not between the two 100mA tests.
-
They all have various drift, noise, and fidelity issues. I use the AM503B the most because I like the auto balance, but it has processor clock noise in the output which can be annoying.
Is there anything we can do to mitigate any of those issues? Currently (pun intended), the drift is the most annoying issue.
I don't know; I haven't spent time to figure out the source of drift. And it could be multiple things. It's why Tek added the auto-balance feature. And DC measurements can magnetize the core and add an offset, so depending how accurate you want to be, it may be important to de-energize the circuit and check for a zero reading.
I will set up a quick test and plot the drift after a 30 minute warmup so you can compare and determine if your drift is out of line.
On the clock noise, I was considering a circuit that would pause the clock when making sensitive measurements, but it was easier to swap in the AM503 in those cases.
These probes aren't great for long term current measurements or in the range of a few ones or tens of mA. I think it was mentioned earlier, but in those cases a differential preamp and shunt resistor would probably be better provided the preamp can sufficiently reject the CM range at the shunt.
-
These probes aren't great for long term current measurements or in the range of a few ones or tens of mA. I think it was mentioned earlier, but in those cases a differential preamp and shunt resistor would probably be better provided the preamp can sufficiently reject the CM range at the shunt.
I got this primarily for entertainment/education. My most common range will likely be between 20mA to 40mA DCI. I don't need long term measurements, but I do want accuracy within a few minutes of zero/degauss.
Stability was better after a minimum of 60 minutes. However, my probe isn't great. The seller is allegedly sending a replacement.
I'll probably spend more time messing with the probe/amp than actually using it. 😉
-
I should have the replacement probe and the parts from Mouser tomorrow, and I will be doing the chip change and calibration.
Rico, who created that NVRAM replacement mod/documentation shared his spreadsheet with me, and said I can share it. Another member of the group also shared his version, and I've combined things from both for this new version (attached).
Rico's .BIN file for NVRAM comparison is also attached, along with the PDF instructions.
Thanks,
Josh
-
The replacement probe arrived today, and it's much better. Without any warmup time, I tested it and it was right on for 100mA test. Drift after 10 minutes of warmup time is still within 4mA (2mA on screen). I'll test more later after proper warmup.
The previous probe was never this accurate or stable, even after an hour warmup. The replacement is in much better shape physically too.
Thanks,
Josh
-
Here it is after 25 and 30 min warmup time.
-
Some more tests...
5mA and below was less stable, but not bad. I think that's mostly because what I'm seeing is +/-1mA if I wait long enough.
After 12 minutes, 10mA was showing 11mA. That's significantly better. The previous probe would drift down down down and fall off the screen in a few minutes. 🙄
-
Here's a quick stability test. Not terrible! 😉
-
Upgrading the NVRAM chip was easy using a $10 CH341A programmer.
I used Rico's BIN file and cleared the data and wrote that to the chip. I dunno if that was really necessary, but I figured it was worth it to test the programmer before inserting the new chip.
I don't have the adapters needed to do an actual calibration, but thankfully, it's pretty accurate already. I'll eventually see about finding or making the needed adapters so I can calibrate it for fun.
I installed the new chip and powered up the 503B, and it worked as normal with the factory default warning. Then I removed the chip, and wrote the serial number to it and reinstalled it in the 503B. No more default settings warning.
I attached the 341A programmer software since I had trouble finding a virus free version. This one was clean.
Thanks,
Josh
-
The Tektronix current clamp is still one of the best you can get, especially if it's used and therefore cheaper, because new it costs a fortune.
It's all the more annoying that you made such a bad purchase, I was luckier.
-
The Tektronix current clamp is still one of the best you can get, especially if it's used and therefore cheaper, because new it costs a fortune.
It's all the more annoying that you made such a bad purchase, I was luckier.
Were you though? 😉
My crap probe got replaced with a good probe for no extra fee. Now I'm messing with the amplifier to see how new lower ESR filter caps affect it.
The new probe measurements looked good, right?
-
They do.
Were you though?
My mixed system (A6302 from the US, AM503A from the Czech Republic, TM502 from Germany) works without any problems. 8)
-
My mixed system (A6302 from the US, AM503A from the Czech Republic, TM502 from Germany) works without any problems. 8)
Have you tried a stability test over time?
All my quick tests looked good, and the stability test was decent with the newer probe. After I replace the larger filter caps I'm going to try again after a nice warmup. As I've said before, this is for entertainment and education. I could have left it alone and it would be fine for measuring. 😉
-
Ya know, I thought replacing some caps might help, but I wasn't expecting this result. 😎
So far, I only replaced 4 caps (2 each):
50V 2000uF Old: 60mΩ ESR New: 13mΩ ESR
50V 1000uF Old: 120mΩ ESR New: 11mΩ ESR
That looks slightly better than the test in reply #181.
Thanks,
Josh
-
You can see the quantization in the blue trace, so the red trace surely has different averaging applied on the scope.
-
You can see the quantization in the blue trace, so the red trace surely has different averaging applied on the scope.
The red trace is an SDM3055X-E. The blue trace is the SDS2404X HD. The settings are identical between both stability tests, except I think the 2nd test sampled more frequently (log rate in TC).
I attached both tests to this reply to make it clearer.
-
Sorry for the delay in getting this test done, but here's my AM503B drift with an A6302. I duplicated your test with a 30mA test current over a 30 minute measurement time and after a 40 minute warmup. I pressed the "AUTOBLANACE" with the power supply disconnected just before the recording started. The AM503B was set to 10mA/div. By the time the test was done, the AUTOBALANCE button was blinking, indicating it was time for a re-balance. but I don't know exactly when the blinking started.
This particular AM503B came to me with a dead battery, which I have since replaced, but I have not gone through the re-calibration procedure. So, the absolute current reading is expected to be off.
I think it does not exhibit as much short-term variation as shown in your test.
Edit: Additional details: Constant 30mA current supplied by a HP 3631A and measured by an HP 3478A DMM. AM503B output measured by an HP 3456 DMM with a 50R pass-through terminator.
-
Here's another drift run using the exact same setup and procedure as outlined previously. The only difference is that everything has been powered for two or three hours, and the monitoring time is a little longer.
The absolute current output from the AM503B is now closer to the actual (as read by the 3478A) and drift is slightly less.
-
Here's another drift run
It looks like your PP is ~0.5mA? After switching the 4 larger caps, mine went from 4.5mA to ~0.2mA PP. I'll run the test again later today with my DMMs instead of using the scope.
Thanks,
Josah
-
Here's another drift run
It looks like your PP is ~0.5mA? After switching the 4 larger caps, mine went from 4.5mA to ~0.2mA PP. I'll run the test again later today with my DMMs instead of using the scope.
...
Do you mean Pk-Pk drift? Yes, it's in that range for both plots. If you mean Pk-Pk noise, with a scope I'm measuring 5mA, and the AC RMS noise as 0.5mA, both with the 20MHz BW limiter on and no current going through the probe.
I'm surprised that replacing the filter caps improved your drift, if that's what you mean. It's completely within reason that it improved the noise, especially if there was a 120Hz component. An FFT would show that clearly.
Given the random variation in your plot, perhaps the scope was measuring the DC level AND the noise. A DMM is a better approach for measuring slowly changing DC levels, since it's going to reject all the other garbage. I'm interested to see your results.
-
I know we're talking about the AM503B in this thread, but since I have it set up I thought I would share what the drift looks like on the older AM503. Same test scenario, and using the same A6302 probe.
The AM503 is an all analog implementation, with front panel buttons and knobs for manual degaussing, balance, and DC level. The DC level requires constant attention, even in short time spans, which is why the AM503B is such a pleasure to use with its AUTOBALANCE button.
The amount of drift over a 30 minute span is much improved in the AM503B by around a factor of 3 or 4 in these tests, as is the output stability for a minute or less.
-
I know we're talking about the AM503B in this thread, but since I have it set up I thought I would share what the drift looks like on the older AM503. Same test scenario, and using the same A6302 probe.
The main focus of the thread was for the probe, which I settled on the A6302. It's good to see data with different amps with it...and I'm glad I got the 503B, even if it did need some work. 😉
-
Stability Test 3 Setup:
SPD3303X at 3.05V with a 100Ω CMF resistor.
SDM3055X-E in-circuit for current reference.
A6302 on negative lead of PSU.
AM503B connected to SDM3065X via 50Ω termination.
PP (min and max values): 235.5826µV over ~50 minutes.
Thanks,
Josh
-
The red trace is an SDM3055X-E. The blue trace is the SDS2404X HD. The settings are identical between both stability tests, except I think the 2nd test sampled more frequently (log rate in TC).
Ah, sorry. I recognized the screenshot as being TestController, but it didn't register that the measurements were from two different devices. I was thinking you were showing before and after.
-
Stability Test 3 Setup:
SPD3303X at 3.05V with a 100Ω CMF resistor.
SDM3055X-E in-circuit for current reference.
A6302 on negative lead of PSU.
AM503B connected to SDM3065X via 50Ω termination.
PP (min and max values): 235.5826µV over ~50 minutes.
Thanks,
Josh
So... If you're still using 10mA/div to be comparable with previous tests, that's a drift of 0.235mA. Half of what I was getting. Not bad!
Edit: Ah, right. The numbers say it was 10mA/div. Great!
-
A good result that I am tempted to reproduce.
-
A good result that I am tempted to reproduce.
I'd love to see how your AM503A compares!
For anybody with an AM503B that wants to recap or convert the NVRAM, this is my Mouser BOM (https://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=b82a04814a) of what I ordered.
So far I've upgraded the NVRAM and changed only the 4 large caps. I will probably do the rest of the caps on that list too.
Thanks,
Josh
-
So... If you're still using 10mA/div to be comparable with previous tests, that's a drift of 0.235mA. Half of what I was getting. Not bad!
Thanks! I'm happy with the result.
I'm curious now if replacing the rest of the electrolytic caps with lower ESR caps will change anything, and if so, for better or worse. 🤔
ETA: Mark, can you please confirm the values of your electrolytic caps in your AM503B? 4 caps were already replaced before I got mine.
Thanks,
Josh
-
Then, to be consistent, you would have to monitor the voltages that the electrolytic capacitors filter.
If these are/were unstable due to aging of the electrolytic capacitors, it will affect everything that is supplied by them.
-
Then, to be consistent, you would have to monitor the voltages that the electrolytic capacitors filter.
If these are/were unstable due to aging of the electrolytic capacitors, it will affect everything that is supplied by them.
What will monitoring them help? I don't have a schematic or reference of any kind for the 503B. If I swap out a set of caps and test the performance of the probe, won't that tell me what I need to know?
I'll also be connecting all caps to the IM3570 to test ESR at 100kHz and µF at the proper range for the value. I tested some of them in circuit with the ST42, and it did pretty well too.
-
If I swap out a set of caps and test the performance of the probe, won't that tell me what I need to know?
Of course you can do it that way...
I'm always interested in the background of such things, because even years ago I always had a problem in the test field.
Don't ask, just replace...
That was always not enough for me.
Forget what I said. ;)
-
I'm always interested in the background of such things
Sure, but you're also more likely to try and trace out the circuit and make your own schematic for it. I won't. 😉
-
Please summarize your setup again.
Power supply, resistor, SDM3065X with the Tester software, then the Tektronix.
You connected the output of the AM503 to the SDM via 50 ohms, right?
-
Please summarize your setup again.
Power supply, resistor, SDM3065X with the Tester software, then the Tektronix.
You connected the output of the AM503 to the SDM via 50 ohms, right?
There's a mediocre photo in reply 196, but here's my setup:
PSU red terminal to DMM-1 current terminal, DMM-1 black wire to 100Ω resistor to PSU black terminal.
A6302 connected on negative wire near PSU (obviously pointed at PSU).
AM503B connected to DMM-2 via 50Ω terminator and BNC to banana adapter.
TestController set to monitor reference current on DMM-1 and AM503 output on DMM-2.
My DMMs had 4 hours of warmup. AM503B and PSU had over an hour warmup. Maybe 2, I don't remember. I wouldn't do this test with less than an hour warmup on anything.
Thanks,
Josh
-
Ah, that's right, you had used two DMMs.
I'll have to solve that differently then, I only have one that can connect to the test controller software.
I would warm up the DMM and the resistor, and I would record the AM503A including the cold start.
I need to see what the trend function on the SDS3104X HD can do before I have to buy a second DMM.
-
Ah, that's right, you had used two DMMs.
I'll have to solve that differently then, I only have one that can connect to the test controller software.
I would warm up the DMM and the resistor, and I would record the AM503A including the cold start.
I need to see what the trend function on the SDS3104X HD can do before I have to buy a second DMM.
In reply 181 I used my SDS2504X HD instead of the second DMM with TC. You can try that too. Replace DMM-2 with your scope to measure the current from the AM503.
I added the SDS3000X HD series to the attached device file for TestController. Let me know if it works for you. It should.
Thanks,
Josh
-
Ah, that's right, you had used two DMMs.
I'll have to solve that differently then, I only have one that can connect to the test controller software.
I would warm up the DMM and the resistor, and I would record the AM503A including the cold start.
I need to see what the trend function on the SDS3104X HD can do before I have to buy a second DMM.
What DMM do you have? If you have one that has a jack in common with a volts input and a current input, you can wire it up and use your computer to read current through the 100R resistor, then switch DMM functions to read the voltage output from the AM503B. Keep flipping back and forth reading current and volts.
-
...
Mark, could you please check your AM503B and count the electrolytics? Mine has three 220uF caps, but I want to confirm that's the correct value before I replace them (they were the only new caps present when I received the 503B).
Thanks,
Josh
-
...
Mark, could you please check your AM503B and count the electrolytics? Mine has three 220uF caps, but I want to confirm that's the correct value before I replace them (they were the only new caps present when I received the 503B).
Thanks,
Josh
Sure! I have:
(3) 220uF @ 16V, Nichicon VX(M) series
(6) 100uF @ 25V, Nichicon VX(M) series
(2) 1000uF @ 50V, Nippon Chemi-Con SME series
(2) 2200uF @ 50V, Nippon Chemi-Con SME series
-
Sure! I have:
(3) 220uF @ 16V, Nichicon VX(M) series
(6) 100uF @ 25V, Nichicon VX(M) series
(2) 1000uF @ 50V, Nippon Chemi-Con SME series
(2) 2200uF @ 50V, Nippon Chemi-Con SME series
Thank you! There's (7) 100uFs in mine, I think one is hiding from you. 😉
There's also a cute little 1uF near the 220uFs.
Thanks for checking that, now I can replace those last 4 caps.
Thanks,
Josh
-
Sure! I have:
(3) 220uF @ 16V, Nichicon VX(M) series
(6) 100uF @ 25V, Nichicon VX(M) series
(2) 1000uF @ 50V, Nippon Chemi-Con SME series
(2) 2200uF @ 50V, Nippon Chemi-Con SME series
Thank you! There's (7) 100uFs in mine, I think one is hiding from you. 😉
There's also a cute little 1uF near the 220uFs.
Thanks for checking that, now I can replace those last 4 caps.
Thanks,
Josh
You're right I missed the 1uF @ 50V near the output amplifier.
Double checking, the 503B I've been using for the above tests indeed only has (6) 100uF. However, the other 503B I have has (7) like yours, and I now see the empty pads on the first unit. The pads look factory fresh, so I guess Tek decided it was unnecessary. The board layouts look identical and also have the same part number (A9F-1631-00, at the edge near the front panel).
For the record, the 100uF that's missing is the one that's in the center, flanked by two diodes on the top and two TO-92 transistors on the bottom.
-
Double checking, the 503B I've been using for the above tests indeed only has (6) 100uF. However, the other 503B I have has (7) like yours, and I now see the empty pads on the first unit. The pads look factory fresh, so I guess Tek decided it was unnecessary. The board layouts look identical and also have the same part number (A9F-1631-00, at the edge near the front panel).
For the record, the 100uF that's missing is the one that's in the center, flanked by two diodes on the top and two TO-92 transistors on the bottom.
Interesting! Now the question is, does one of them appear to have better stability? Then again, there might be something else different component wise that eliminated the need for the 7th cap.
Thanks,
Josh
-
Double checking, the 503B I've been using for the above tests indeed only has (6) 100uF. However, the other 503B I have has (7) like yours, and I now see the empty pads on the first unit. The pads look factory fresh, so I guess Tek decided it was unnecessary. The board layouts look identical and also have the same part number (A9F-1631-00, at the edge near the front panel).
For the record, the 100uF that's missing is the one that's in the center, flanked by two diodes on the top and two TO-92 transistors on the bottom.
Interesting! Now the question is, does one of them appear to have better stability? Then again, there might be something else different component wise that eliminated the need for the 7th cap.
Thanks,
Josh
I'll run the same stability test on the second AM503B in the next day or two for curiosity's sake. You're right that there might be other differences between the two. One test could be adding the cap to the first, or removing it from the second. But if they're both equally stable, I'm probably not going to mess with either.
-
I'll run the same stability test on the second AM503B in the next day or two for curiosity's sake. You're right that there might be other differences between the two. One test could be adding the cap to the first, or removing it from the second. But if they're both equally stable, I'm probably not going to mess with either.
If one is more stable than the other, I'd start with the big 4 caps with low ESR versions like I used. Those had a bigger impact than anything else so far.
I have it all warming up now, and we'll see if the 220uFs did anything useful. The old caps were ~450mΩ ESR, and the new ones are ~45mΩ.
Is your 1uF cap bipolar?
Thanks,
Josh
-
Did you measure the old ones you had already swapped?
-
Did you measure the old ones you had already swapped?
Yup:
| 35V 100uF | Old: 1.3Ω ESR | New: 55mΩ ESR |
| 16V 220uF | Old: 450mΩ ESR | New: 45mΩ ESR |
| 50V 1uF | Old: 1.3Ω ESR | New: 3.4Ω ESR |
| 50V 2000uF | Old: 60mΩ ESR | New: 13mΩ ESR |
| 50V 1000uF | Old: 120mΩ ESR | New: 11mΩ ESR |
All values measured on IM3570 after warmup and corrections.
-
Well, this is also interesting. I replaced the final 4 caps (220uF x4 and 1uF).
Before replacing those, after degauss/zero button would show around 0.25mV on the DMM.
Now it's slightly better.
So far:
Big caps (1000uF and 2200uF) affected the stability/drift drastically. Went from ~5mV PP to 0.2mV PP.
100uF caps...nothing noticeable.
220uF and 1uF: much closer to zero after degaussing.
Now to check the stability again.
Thanks,
Josh
-
The TM501 unit houses the power supply for the plug-in units.
Ultimately, the capacitors would also have to be replaced there.
-
The TM501 unit houses the power supply for the plug-in units.
Ultimately, the capacitors would also have to be replaced there.
I agree, those should be replaced too. I'll get there eventually. 😉
-
The TM501 unit houses the power supply for the plug-in units.
Ultimately, the capacitors would also have to be replaced there.
I'd do a frequency analysis (FFT) of the noise (it it isn't mains related) and try to correlate that with noise in the same frequency band(s) from power supply lines or other places in a circuit before attempting to make seemingly random changes to equipment. It is very possible there are external sources which cause the noise. For example: in my own lab I can see interference around 30MHz coming and going at random times.
I'd also be wary about changing capacitors for ones with a much different ESR without checking the specification for the original capacitor. Tektronix may have designed the circuit for a specific ESR range. For example: in the TDS500/600/700 series scopes there are a few electrolytic capacitors which can't be changed by any other type like ceramic or tantalum because the ESR is too different.
-
When the capacitor was replaced, it was more a matter of drift than noise, at least that's how I understood the posts, and that's how an improvement was achieved.
So if he's already replacing electrolytic capacitors, then he might as well replace the electrolytic capacitors of the actual supply.
If I remember correctly, all supplies in the system are linear and not switched, so the ESR question hardly arises in this respect.
However, I agree with you about other capacitors in other circuits of the system.
-
Personally I rather leave components which work OK alone. Whatever you take apart is not getting better; only thing you can do is try to minimise damage. ;) In Dutch we have a saying which kind of translates to 'repair it until broken' (which is not a good thing).
-
Personally I rather leave components which work OK alone.
I am with you, never touch a running system - normally.
We (the company) consistently replace electrolytic capacitors after 10 years in the device to ensure continued operational reliability.
However, these are also stressed accordingly.
Electrolytic capacitors dry out over time, faster or slower depending on the load/ambient temperature.
That's why I suggested taking the appropriate measurements.
-
Personally I rather leave components which work OK alone. Whatever you take apart is not getting better; only thing you can do is try to minimise damage. ;) In Dutch we have a saying which kind of translates to 'repair it until broken' (which is not a good thing).
If it worked ok, I would have left it alone. 😉
It went from 4.5mV PP to 0.2mV PP. I'd call that an improvement. I like stability.
Electrolytic capacitors dry out over time, faster or slower depending on the load/ambient temperature.
That's why I suggested taking the appropriate measurements.
This is why I went after those caps in ~37 year old device.
-
PSU:
Nippon Chemi-Con
81D
18mF 16V (1) Maybe 81D183M016KD5D ?
4.7mF 50V (2)
LS: 10mm
D: 30mm
H: 40mm
A whopping 3 caps.
-
The date codes on various seem to suggest your unit was manufactered after 1995. What date does it say on the 'tested' sticker? Interestingly, I'm using the same stickers BTW.
-
The date codes on various seem to suggest your unit was manufactered after 1995. What date does it say on the 'tested' sticker? Interestingly, I'm using the same stickers BTW.
That's almost 30 years. I said 37 based on the copyright on the PSU board. The PSU is older than my AM503B. There's a Copyright 1993 marked on the AM503B. Neither are young. 😉
-
I'll run the same stability test on the second AM503B in the next day or two for curiosity's sake. You're right that there might be other differences between the two. One test could be adding the cap to the first, or removing it from the second. But if they're both equally stable, I'm probably not going to mess with either.
If one is more stable than the other, I'd start with the big 4 caps with low ESR versions like I used. Those had a bigger impact than anything else so far.
Was your measurement technique consistent in your comparison? You had swapped in the DMM in place of the scope at some point.
I would expect the large filter caps to affect noise and not so much the slowly changing DC drift.
Is your 1uF cap bipolar?
Yes, on both units.
-
I'll run the same stability test on the second AM503B in the next day or two for curiosity's sake. You're right that there might be other differences between the two. One test could be adding the cap to the first, or removing it from the second. But if they're both equally stable, I'm probably not going to mess with either.
If one is more stable than the other, I'd start with the big 4 caps with low ESR versions like I used. Those had a bigger impact than anything else so far.
Was your measurement technique consistent in your comparison? You had swapped in the DMM in place of the scope at some point.
I would expect the large filter caps to affect noise and not so much the slowly changing DC drift.
I probably missed this but are there any schematics available? The only cause for slow drift I can think of where it comes to electrolytics would be leakage current. But this would mean the large capacitors would be part of some kind of (slow) servo mechanism. Maybe to cancel DC magnetic field in the probe head? The schematics should tell what is what.
-
Was your measurement technique consistent in your comparison? You had swapped in the DMM in place of the scope at some point.
I would expect the large filter caps to affect noise and not so much the slowly changing DC drift.
Yeah, I left everything setup exactly the same. Nothing was moved.
The initial change from 4.5mV to 0.2mV was done using the scope. Test 3 with the DMMs was 0.236mV over a longer period (see attached, and note the names displayed for each device). The only change was the 4 larger caps in the 503B at the time.
Is your 1uF cap bipolar?
Yes, on both units.
Thanks!
-
I probably missed this but are there any schematics available? The only cause for slow drift I can think of where it comes to electrolytics would be leakage current. But this would mean the large capacitors would be part of some kind of (slow) servo mechanism. Maybe to cancel DC magnetic field in the probe head? The schematics should tell what is what.
There is no known publicly released schematic for the AM503B. The AM503A might be similar enough for those parts though.
-
Here's the manual for the TM502A & its PSU: https://w140.com/tekwiki/images/e/ee/070-6502-00.pdf
-
For the PSU caps, I'm leaning towards these:
MAL205658472E3
LGU1E183MELB
Thanks,
Josh
-
Was your measurement technique consistent in your comparison? You had swapped in the DMM in place of the scope at some point.
I would expect the large filter caps to affect noise and not so much the slowly changing DC drift.
Yeah, I left everything setup exactly the same. Nothing was moved.
The initial change from 4.5mV to 0.2mV was done using the scope. Test 3 with the DMMs was 0.236mV over a longer period (see attached, and note the names displayed for each device). The only change was the 4 larger caps in the 503B at the time.
As I mentioned previously, the 4.5mV Pk-Pk measurement looks more like random noise to me, which is completely within reason as being corrected by the new caps. I think it's difficult to conclude anything about slower moving, and possibly much smaller, DC drift because of the randomness of the measurement.
Perhaps the measurement selected was capturing higher frequency noise, or line-related noise not on a cycle boundary (like measuring RMS on the whole screen with only a cycle or two visible). And all of these things depend on the sample rate and possibly capture size (don't know - I don't have a Siglent scope).
I would pick a slow sweep rate (like 100ms/div), hi-res (aka box car averaging), and no trigger (free-run). Then ask the scope to read the RMS of the entire screen or capture buffer (whatever is possible on the Siglents). You could also ask the scope for the statistical average of a number of RMS measurements as a different way to do averaging.
But still a DMM is a better way to measure this drift. The DMM "before" cap replacement measurement is not available for a direct comparison.
-
...
I probably missed this but are there any schematics available? The only cause for slow drift I can think of where it comes to electrolytics would be leakage current. But this would mean the large capacitors would be part of some kind of (slow) servo mechanism. Maybe to cancel DC magnetic field in the probe head? The schematics should tell what is what.
I'm not aware of any schematics either. The board layout is definitely different between the A and B, but a quick glance says that some of the components are the same (processor, attenuator, output amplifier, etc.).
The four big filter caps in the A schematic are involved in supply filtering, and are located in the rear near the backplane connector as would be expected. In the B, they are also in the same area, although the layout differs a little. And the B uses 2200uF instead of 1000uF. They are almost certainly performing power filtering.
My guess is that there was a lot of power line hum showing up in the output which was resolved by the new caps.
-
As I mentioned previously, the 4.5mV Pk-Pk measurement looks more like random noise to me, which is completely within reason as being corrected by the new caps. I think it's difficult to conclude anything about slower moving, and possibly much smaller, DC drift because of the randomness of the measurement.
Perhaps the measurement selected was capturing higher frequency noise, or line-related noise not on a cycle boundary (like measuring RMS on the whole screen with only a cycle or two visible). And all of these things depend on the sample rate and possibly capture size (don't know - I don't have a Siglent scope).
I would pick a slow sweep rate (like 100ms/div), hi-res (aka box car averaging), and no trigger (free-run). Then ask the scope to read the RMS of the entire screen or capture buffer (whatever is possible on the Siglents). You could also ask the scope for the statistical average of a number of RMS measurements as a different way to do averaging.
But still a DMM is a better way to measure this drift. The DMM "before" cap replacement measurement is not available for a direct comparison.
Regarding the scope version of the tests, nothing was changed except the caps. It could very well have been AC garbage causing the higher PP, which is why I tried that. Whether or not that's affecting the drift directly, I dunno. 4.5mV PP noise certainly could mask a much smaller drift that was hiding in there.
The only thing that makes me wonder about that though is that before changing the caps, if I waited long enough, the signal would drift right off the scope screen. 🤷
-
I removed and tested the 3 caps from the PSU, and...it seems pointless to replace them. They look new, I don't know if they're original or not. I assume not original, because there was some flux residue on the PCB around them.
Also, the ESR is super low. I can't get anything lower, so what's the point if they're not bad.
Does anybody else have a TM502A they can check the PSU caps in? I'm curious what's normally in there.
Thanks,
Josh
-
KunfFu: In the lab since 1990s, we have many TM50x MF , AM503A, AM503B, AM503 and probes, manuals.
Can you say what AC/DC measurements you make and what accuracy/resolution you expect?
Happy to have contact via PM.
Have an absolutely fantastic day,
Jon
-
KunfFu: In the lab since 1990s, we have many TM50x MF , AM503A, AM503B, AM503 and probes, manuals.
Can you say what AC/DC measurements you make and what accuracy/resolution you expect?
Hey Jon!
Accuracy of around 1% is fine. Resolution isn't that important; 3 or 4 digits is plenty.
Messing with this right now is mostly education/entertainment. I was going after different caps to see about improving noise and stability. Stability is pretty decent now.
I took a look at the PSU electrolytics to see if any further improvements could be made with those, but as I posted above, it doesn't look like those caps are a problem. They actually seem quite nice.
Thanks,
Josh
-
I noticed on the AM503A schematic there was a manufacturing test serial port. On the AM503B, this serial port is connected to the GPIB option connector. Using the infinite monkeys method, I did some poking at this port and was able to figure out most of the internal commands to control the AM503B, as would the GPIB controller card.
My goal was to find a way to read and write the cal constants without having to modify the board or start pulling chips. If GPIB could do it, then it must be happening through this serial port. (Hint: It's the "CC" command). Along the way I also found most of the internal commands used by the GPIB option card to control the AM503B and it's twin, the AM5030.
There's probably other things I haven't discovered, but I found what I needed so I'm not going to invest much more time in this. At most I will probably write a Linux bash script to make it easier to extract the cal constants for backup, and also allow a user to update the cal constants based on the procedure found in the service manual.
Below are my notes. Comments/corrections are welcome from anyone who wants to give this a try. I'll also post this to the TekScopes group as it may be of interest there too.
AM503B Internal Serial Port Notes
---------------------------------
11/11/2024 - MarkL @ eevblog
The serial port on the AM503B's internal processor (a P87C528),
communicates with the optional GPIB controller card, and this is where
we can connect and issue text commands and receive text responses.
It's not GPIB commands at this interface, but is closely related in
functionality, parameters, and naming.
The interface could be the same as "MFG_DATA_IN" and "MFG_DATA_OUT" on
the AM503A schematic, but I don't have a AM503A to test if the same
command set is in use there.
Some of the remaining mysteries below could probably be figured out by
eavesdropping on the internal GPIB controller header on an AM5030. An
AM503B with the GPIB option card is the same as an AM5030.
References (all are on https://w140.com/tekwiki):
- AM 503B & AM 5030 AC/DC Current Probe Amplifiers Instruction Manual,
publication #070-8766-05
- AM 503B, AM 5030 & A6300 Series, 067-0271-00 Verification and
Adjustment Kit, publication #070-9352-01
- AM 503A Current Probe Amplifier Service Manual, publication
#070-8174-01
- AM503A_Schematics.pdf, unknown
Internal GPIB Controller Connector
----------------------------------
- 2x10 header directly to the right of the processor
- Serial port is 5V TTL level (not 3.3V!)
- Header pins to serial adapter:
16: Rx (yellow, data coming from AM503B)
14: Tx (orange, data going to AM503B)
12: GND (black, there are other GND pins: 2, 3, 11, 20)
Notes: Header square pad is pin 1 (lower right)
Colors refer to FDTI TTL-232R adapter cable
- 4800 baud, 8 bits, no parity, idle high (the usual)
Commands and Queries
--------------------
- A command is two characters followed by optional arguments.
- A query is two characters followed by a "?" and then optional arguments.
- Not all commands have a query format and visa-versa, but if they do
the query form generally shows what was last set via the command
form.
- All numbers are ALWAYS HEXADECIMAL, including error responses.
- Everything is ALWAYS UPPERCASE.
- Commands/queries are terminated with a single carriage return (0x0D).
- Responses are terminated with a single carriage return (0x0D).
- Query forms (xx?) below show an example response, current settings
will vary the response.
AD?: AD (unknown meaning, appears to be constant)
AMx: Set amps per division x=[A-L] (range steps 1mA/div to 5A/div)
AM?: AMB
BWx: Set BW limiter x=[0,1], toggle if x omitted
BW?: BW1
CCrr+xxxx: Set gain cal constant rr=[00,01,02] to xxxx.
CC02?: CC02+61A8 (0x61A8 == 25000, which is the default)
COx: Set coupling x=[A,D,R] (AC, DC, Ref), toggle if x omitted
CO?: COR
DAxxx: OK (unknown effect, xxx is optional and any length)
DC+xxxxx: Set DC level to xxxxx
DC?: DC-00172 (was set to -2.3 from front panel)
DC?: DC-00162 (was set to -2.2 from front panel)
DC?: DC-00012 (was set to -0.1 from front panel)
DC?: DC+00002 (was set to 0.0 from front panel)
DC?: DC+00012 (was set to 0.1 from front panel)
DC?: DC+00142 (was set to 2.0 from front panel)
DC?: DC+00000 (after initialize)
DEx: Perform degauss operation x=(0, 1, 2):
0 or omitted: Just like pressing front panel button
1: Forces gain calibration (see manual)
2: Gain calibration without balancing Hall device (see manual)
EX: Exit error mode (e.g., 266 being displayed on front panel)
FPx: Front panel lock x=[0,1], lock if x omitted
GP?: GP050 (unknown, appears to be constant)
ID?: ID3.0 (returns firmware version)
IN: Initialize
KYxx: Front panel button/knob simulation(?) xx=key to press, 00=release
(not too sure how this is used)
KY?: KY00
OV?: OV0 (status of front panel OVERLOAD light)
PO?: PO0 (status of front panel PROBE OPEN light, returns "p" when
open, huh?)
PT?: PTA6302 (returns probe type)
SNxxxxxxx: Set serial number to xxxxxxx, set to null if xxxxxxx omitted
SN?: SNB012345
TE?: TE000 (first error from self test, 000=no error)
TE?: TE165 (0x165 == 355 dead battery, NVRAM reset)
TMxx: Run test xx, see AM503A service manual, examples:
00 = blink all LEDs
02 = display firmware version on front panel
TRxxx: Set probe trim to xxx
TR?: TR42D (was set to 069 from front panel, 0x42D == 1069)
UN?: UNA (returns output units, A=amps, V=volts, see manual)
Some Responses Seen
-------------------
- Many of the error message appear to be analogous to the ones listed
in the GPIB section in the manual (not surprising).
OK: Command executed successfully
KP: A front panel button was pushed or knob turned
"KP" and other key press events appear to have the 8th bit set
PW: Power up (8th bit set)
ER065: 101: Unknown command
ER0FC: 252: AM argument out of range
ER105: 261: DE argument out of range
ER10A: 266: Auto-balance not 50 ohm terminated
-
Best amp is TEK AM503B.
Look at epay, ham fleas, auctions.
TM50x MF not critical, of the dozens we have just a few were bad, mostly blown pass trans, the lytics hold up.
The AM503 are not stable and low res.
AM503A is not grat but better than AM503.
Lots of bad probes ....dropping a TEK P6302 a few cm on concrete/metal may hariline crack the ferrite jaw.
Symptom is poor LF/DC or unstable...Beware.
j
-
Best amp is TEK AM503B.
Look at epay, ham fleas, auctions.
Thanks, I already got it. I got the A6302, TM502A, and AM503B. I also have a Toolbox 016-0362-00 on the way for the other slot to store the A6302 in.
-
Kung, perfect, cant do better.
We degauss after warmup 30 min and before each use,
just be sure to term AM503B in 50 ohm on scope input.
To CAL we use TEK PG506, (or similar,) and wind a 10, 100T magnet wire square loop with a 50 Ohm term R on the loop in series with 50 Ohm on the gen.
Enjoy,
Jon
-
Quick first look at the factory fresh SCP5030A 100 MHz 30A active current probe measuring the current loop output of the Power Analysis deskew fixture.
So easy to use as it's scope self sensing and results are quite satisfactory from a cold start.
Used HiRes mode to reduce trace noise......
(https://www.eevblog.com/forum/testgear/accurate-low-amp-current-probe-advice/?action=dlattach;attach=2433401)
-
Another possibility. Not as good as the Tektronix system, but way more compact and good enough performance for what I need: a Lecroy CP015 (alternatives are CP030, AP015 and AP030)
I run this probe without a Probus interface, just with a +-12V power supply connected directly to the probe (and a BNC 90°-adaptor to connect it to the scope). There probes can be had for 300-500€ on Ebay, less if you are lucky. When power is turned on, the probe runs a degaussing cycle. So for I have not looked into trying to send I2C commands as these protocol is not well documented (and it works good enough without).
Test was run with the Yokogawa GS200 supplying a 30mA current (to mimic test conditions above). I did not log the current of the Yokogawa, but I am sure this is completely flat (max a couple of µA variation). Probe was connected to a BM869s and logged via testcontroller software. As the output of the probe is only 100mV/A, the signal was only 2.6mV. Temperature in the room was not tightly controlled, so some drift could be temperature related of the >4 hours this test has run.
-
Here some more detail how this is connected to the scope. The damage to the sticker was how a bought it, and it also had already the cut-out where the 90°-BNC adaptor fits in. I did solder the power supply cable (and UTP cable, that fits directly to the Fluke I50 power supply I already had).
-
I removed and tested the 3 caps from the PSU, and...it seems pointless to replace them. They look new, I don't know if they're original or not. I assume not original, because there was some flux residue on the PCB around them.
Also, the ESR is super low. I can't get anything lower, so what's the point if they're not bad.
Does anybody else have a TM502A they can check the PSU caps in? I'm curious what's normally in there.
Thanks,
Josh
The electrolytic capacitors in your power supply have definitely been replaced; I can see that these are (good) hand solderings on the circuit board.
I'll check my 502s to see if they still have the originals installed.
We still have three Tek systems in use at work that are calibrated annually.
But they are quite old, so I contacted Tektronix today to ask if there were any documents for the AM503B module and if we could have them.
They responded promptly and sent me links where I could download the original AM503 circuit diagram, among other things.
There is no circuit diagram for the AM503 B, I could just look at the linked diagram to see if there are similarities between the two versions, was Tek's answer. ;)
-
The electrolytic capacitors in your power supply have definitely been replaced; I can see that these are (good) hand solderings on the circuit board.
Even better now that I reinstalled them after testing. 😉😉
I'll check my 502s to see if they still have the originals installed.
We still have three Tek systems in use at work that are calibrated annually.
But they are quite old, so I contacted Tektronix today to ask if there were any documents for the AM503B module and if we could have them.
They responded promptly and sent me links where I could download the original AM503 circuit diagram, among other things.
There is no circuit diagram for the AM503 B, I could just look at the linked diagram to see if there are similarities between the two versions, was Tek's answer. ;)
They're full of shit. There's no way they never had a schematic for the B version.
-
Well, not for public release. Ofcourse Tektronix has a schematic. But for internal use only.
-
Of course, but it's still annoying. They supposedly declassified the whole system, but never released the B documents.
-
To CAL we use TEK PG506, (or similar,) and wind a 10, 100T magnet wire square loop with a 50 Ohm term R on the loop in series with 50 Ohm on the gen.
Hi Jon,
Do you have any of the fixtures they say are needed for cal? I have the connectors to make them myself, but I can't find any info on the feed-thru fixture (015-0670-00 (https://w140.com/tekwiki/wiki/015-0670-00)) for pre cal degaussing. I'm hoping that it might be the same coding resistor as below, with no other connections, but I have no reference to go with.
The other info was on the tek wiki:
015-0598-00 Voltage adapter (https://w140.com/tekwiki/wiki/015-0598-00):
It is constructed from a female BNC terminated to pin 'M' on an Amphenol 165-9 or equivalent (Tek Part# 131-1854-00) connector.
The shield of the coax is connected directly to GND pin 'J' or 'K' in the Amphenol connector. There is no connection to the shell of the Amphenol connector.
A 1.13 kΩ coding resistor is connected between pins 'B' and GND (pins 'J' or 'K').
Could you please explain your calibration process in detail (with photos)? It sounds like what you do is different than the adjustment kit instructions.
Thanks,
Josh
-
I have now opened my TM502 after all.
The capacitors are still the original Sprague ones, date code 1987...
2 x 4700µF and one 18000µF. They obviously wanted to be on the safe side and didn't want to install too little. ;)
And while I was at it, I also opened the AM503 A, where all the capacitors are still original too.
-
I bet those Sprague caps are worth some money! 😉
I think my 503B might have been all original on the caps too? I assumed the blue Nichicon caps were newer than the other caps (and also because a lot of manufacturers will use only the same brand for consistency), but maybe not.
-
I think my 503B might have been all original on the caps too?
You can actually always see that at the solder joints.
You can solder as well as possible, hand soldering always differs from those “professional” from the soldering machine.
-
I think my 503B might have been all original on the caps too?
You can actually always see that at the solder joints.
You can solder as well as possible, hand soldering always differs from those “professional” from the soldering machine.
I know, but I assumed when I saw the different brand of caps, so I didn't pay any attention when I replaced them. Oh well. 🤷
-
The electrolytic capacitors in my 503 module are all from Nichicon.
-
The electrolytic capacitors in my 503 module are all from Nichicon.
Mine only had Nichicon for the 220uF and BP 1uF. The rest were a different brand I can't remember right now. 3 letters, CFE or something like that. I think Mark's was the same with 2 brands.
My 503 is all Nichicon and Panasonic now (I went with whatever had better specs).
-
Thanks to the Tek group, I found the correct values to make the pass through and voltage adapters for calibration (doc attached).
What I did was:
Pass through adapter:
Short M + N, also a 15kΩ coding resistor is connected between pins 'B' and 'K'.
Voltage adapter:
BNC terminated to pin 'M' on an Amphenol 165-9 or equivalent (I used 165-9-1000) connector. The shield of the coax is connected directly to GND pin 'J'. A 1.13kΩ coding resistor is connected between pins 'B' and 'K'.
In my photo there's a 50Ω terminator at the end of the BNC cable, which is all it needs to avoid the 266 error code.
Degaussing was successful, I'll try doing the lower mV calibration later.
Thanks,
Josh
-
We do have some Tektronix stuff that was bought in the 90s.
What I didn't know, but discovered today, was that
we also have all the original documentation that was available at the time.
And the manuals look like they were filed yesterday. 8)
AM503B:
Tektronix Germany has no documentation other than the one already known.
I'll try again at Tek U.S.
-
I was able to do the calibration. The lower levels both improved, but I made the 1A range slightly worse. I'm going to do a stability test now that I've done this and replaced the 1uF cap with a higher quality BP cap. When I do redo the calibration, I'll probably reverse the DMM setup so I have more digits for the spreadsheet.
I will probably toy with calibration more in the future, but the 1A+ range isn't that interesting to me, so there's no rush.
Thanks,
Josh
-
I've been screwing around with calibration more, and have noted the following things:
- It's best to work on one range at a time.
- Getting the source current to a far higher accuracy than suggested is helpful.
- You have to reset back to 61A8 before readjusting, which makes the process slightly more tedious.
I've found, at least with my setup, getting the negative values more accurate to be more challenging. I'm going to try the sequence again and see if I can get my starting numbers better.
For the record, this is entirely silly. The goal is to be within 1.5% margin of error... I started with the 1mA range at 0.326%, and got it to 0.305%, but want it better than that.
I was able to get the 10mA range to 0.006%. Supposedly. 😉
Thanks,
Josh
-
After being distracted by the AM503B internal serial port hacking, I got around to checking the stability of my second AM503B. This is the one with the additional 100uF capacitor discussed in Reply #214.
It's slightly better than the first one, having a drift of about 0.4mA over 30 minutes as compared to 0.5mA. Short-term variability and noise look about the same.
This is close enough for me to not start modifying the first one by adding the additional capacitor. This capacitor is out for a reason, and without a schematic to understand its exact role there may unknown side effects. And not to mention my long-standing mantra: If it ain't broke...
I've ordered a pair of the Amphenol connectors to build the cal adapters. I'll post calibration results for both units when available.
-
This is close enough for me to not start modifying the first one by adding the additional capacitor. This capacitor is out for a reason, and without a schematic to understand its exact role there may unknown side effects. And not to mention my long-standing mantra: If it ain't broke...
I've ordered a pair of the Amphenol connectors to build the cal adapters. I'll post calibration results for both units when available.
After everything I've swapped with mine, I would say the most value caps to replace are the 1000 and 2200uF caps. Maybe the BP 1uF also. Nothing else made much of a difference, though thankfully I have no regrets. 😉 My power supply caps had already been replaced with very high quality low ESR caps, so I didn't change those.
Where did you find the connectors? I found a decent deal on eBay for the 165-9-1000 variant.
Thanks,
Josh
-
...
Where did you find the connectors? I found a decent deal on eBay for the 165-9-1000 variant.
Thanks,
Josh
On ebay. They look like the ones you bought, maybe even from the same seller, but I don't have them in hand yet. I offered $24 for a pair. My backup plan was to find the right size male pins and (carefully) push the few connections that were needed into the AM503B connector for the calibration. That would be ok given how often I'm likely to do this procedure.
Unfortunately I should have ordered 3 since I discovered the AM503 has a different pinout for its cal adapter. Maybe I can make one of them do double duty.
I also have an A6302 with a broken transformer that I was thinking about scrapping for the connector, but I haven't given up looking for a new transformer for it yet.
I did replace my TM503 caps some time ago, and that took care of a significant amount of line noise on an FG504 and a AM502. It also helped both the AM503's, but made little difference on the AM503B's. I think my AM503B's may be in healthy condition already. When I get a chance I will poke at those caps with the LCR meter and compare to your measurements.
-
On ebay. They look like the ones you bought, maybe even from the same seller, but I don't have them in hand yet. I offered $24 for a pair.
Sounds the same as mine, I got 4 for $48 from here: https://www.ebay.com/itm/313033033242 (https://www.ebay.com/itm/313033033242)
I bought this cable and cut off the SMA end for the voltage adapter: https://www.amazon.com/dp/B0BBYTZM23 (https://www.amazon.com/dp/B0BBYTZM23)
I posted the details previously, but in case you didn't see it, coding resistor is 1.13k for voltage adapter, and 15k for the pass through adapter.
Thanks,
Josh
-
I took a look at the caps you found needed replacing in your unit, namely the 2200uF, 1000uF, and 1uF BP. In my older unit they are clearly showing their age, but in the "new" unit they are still doing well. All are still in spec using the EOL information from each series' datasheet. Since both units are working fine, I'm going to leave the caps alone.
I'm still waiting for connectors for the calibration, and will post the results when done. (And that was the same seller for the connectors.)
Edit: Oops, fixed cut & paste error in spreadsheet image.
-
I assume the lower ESR & higher ripple handling of my newer caps helped filtering. Here's what mine were, all tested out of circuit on IM3570:
50V 2200uF Old: 60mΩ ESR New: 13mΩ ESR
50V 1000uF Old: 120mΩ ESR New: 11mΩ ESR
Thanks,
Josh
-
I assume the lower ESR & higher ripple handling of my newer caps helped filtering.
This would make sense for a switched voltage, but I cannot imagine that an auxiliary voltage is generated in a switched manner on the AM503B board.
-
I assume the lower ESR & higher ripple handling of my newer caps helped filtering. Here's what mine were, all tested out of circuit on IM3570:
50V 2200uF Old: 60mΩ ESR New: 13mΩ ESR
50V 1000uF Old: 120mΩ ESR New: 11mΩ ESR
Thanks,
Josh
Yes, no doubt replacing those caps improved the ripple.
However, if the designers did their job, the worst case characteristics of the capacitors (including initial -20% of C, temperature effects, and end-of-life DF) would be taken into account. The next stage of linear regulation would continue to reject enough ripple to meet a unit's specifications, or if no warranted noise specification exists as is the case here, to at least meet nominal characteristics. Sure, that's putting faith in the designers, but I think it's a good bet from this era of Tektronix.
What was the capacitance of your old capacitors (@120Hz)?
-
I assume the lower ESR & higher ripple handling of my newer caps helped filtering.
This would make sense for a switched voltage, but I cannot imagine that an auxiliary voltage is generated in a switched manner on the AM503B board.
Yes. The 1000uf / 2200uf capacitors are all used for mains ripple filtering which means a frequency of 120Hz at most. At this frequency regulators and opamps have a very high attenuation. I don't think you can even measure the difference in ripple with new versus old capacitors. The series resistance of wiring, PCB traces, rectifier diodes and transformer windings is likely higher compared to the ESR of the original capacitors. Maybe the 1uf bipolar capacitor can make a difference; I have not checked where this is at in the schematics.
-
Yes, no doubt replacing those caps improved the ripple.
However, if the designers did their job, the worst case characteristics of the capacitors (including initial -20% of C, temperature effects, and end-of-life DF) would be taken into account. The next stage of linear regulation would continue to reject enough ripple to meet a unit's specifications, or if no warranted noise specification exists as is the case here, to at least meet nominal characteristics. Sure, that's putting faith in the designers, but I think it's a good bet from this era of Tektronix.
What was the capacitance of your old capacitors (@120Hz)?
I don't recall the cap values being out of spec enough to take note of it, however, I was mostly focused on ESR.
I don't doubt the skills of the designers being far greater than mine; but I do know the tech we have now, along with what caps available, are much better than what was around 30 or 40 years ago.
-
...
Maybe the 1uf bipolar capacitor can make a difference; I have not checked where this is at in the schematics.
The 1uF is connected to pin 8 of U171, which is the output amplifier on the AM503A schematic, and confirmed to be the same on the AM503B PCB. The other side is GND. Pin 8 is the variable gain input, and the 1uF is performing a filtering function for a buffered output coming from the DAC.
It's a sensitive input as it affects whatever the step gain is set to by a factor 0.5x per V (range = +/-1V). So, a noisy input here due to an ineffective 1uF could certainly make a mess of the output.
-
Here is an excerpt from the plan.
(https://www.eevblog.com/forum/testgear/accurate-low-amp-current-probe-advice/?action=dlattach;attach=2442789;image)
If the capacitor were defective or “worn out”, it would indeed be a problem.
But the ESR has nothing to do with it; it can serve as an indicator of a (beginning) deterioration if you know the correct value (data sheet).