Current Probe Recommendation

[IDEngineer]

That one looks functionally a lot like the Hantek CC-65:


http://www.hantek.com/en/ProductDetail_15_77.html

...which I already own. The CC-65 is actually a reasonable current probe for DC-20KHz for currents above 10-20mA or so. Goes to at least 65A, so way above what we're talking about here, but isn't as useful in the mA ranges and utterly useless below ~10mA. Noise floor is just too high no matter how aggressively you limit your scope's input bandwidth.

Looks like the K2 has about 100x implied better sensitivity based on the front panel switch (10mV/1mA). The CC-65 offers 1mV/10mA. The big question... how is the noise floor on the K2? All the gain in the world isn't useful if the signal is buried in noise.

[HighVoltage]

Looks like the K2 has about 100x implied better sensitivity based on the front panel switch (10mV/1mA). The CC-65 offers 1mV/10mA. The big question... how is the noise floor on the K2? All the gain in the world isn't useful if the signal is buried in noise.

I will hook it up tomorrow to my scope and measure the noise floor.

[IDEngineer]

I will hook it up tomorrow to my scope and measure the noise floor.

Excellent, thank you!!! 

[korlatos]

Applied Physics 428D can measure down to 20uA from DC to 200 Hz. It is a modernized version of the old HP 428B.

[IDEngineer]

Applied Physics 428D can measure down to 20uA from DC to 200 Hz. It is a modernized version of the old HP 428B.

And the price...? Their website doesn't say. Maybe if you have to ask.... 

It does say the bandwidth goes to 600Hz, though!

EDIT: These appear to be flux gate based devices, similar to the AIM TTI I-Prober 520. I suspect they'll be higher performance, though.

EDIT: Just found a non-authoritative comment that suggests the price is ~$4000.

[korlatos]

Per 428D data sheet, the -3dB BW is 200 Hz on the lowest range (10 mA FS). If you are patient, you can buy a used 428D without the probe for a couple hundred bucks on fleabay. Used probes go for about $50.

[HighVoltage]

I will hook it up tomorrow to my scope and measure the noise floor.

Excellent, thank you!!! 

OK, I hooked the K2 current probe up to two scopes and one DMM.
And here are the results for signal and noise floor.

As long as you are using "high resolution" or "average mode" for the acquisition, the slow signal looks perfectly alright and the noise is suppressed.

In normal acquisition mode, the Agilent MSO7034B has a peak to peak noise of 5.8 mA.
The 6000X scope shows about 5 mA Pk/Pk noise

The real signal has a 100 ns rise time.
The scope shows 185 us rise time because of the probes bandwidth limitation.

For slow signals, this probe is working well.
In this case, the 15 mA signal has a duration of 1 ms and it shows OK.

The price is reasonable and as I said before, the built quality is excellent.

[HighVoltage]

And here are some pictures of the probe

CHAUVIN ARNOUX K2

[IDEngineer]

Thank you very much for taking and posting this data, complete with screen shots!

Looks like this probe would be useful down into the low 10's of mA, perhaps even a bit lower. And I am a fan of the clamp style over the flux ring style, since getting repeatable results with the latter is so orientation sensitive (and I don't mean to the Earth's magnetic field, I mean even to the trace/wire you're measuring).

[krisk]

Which one do You think will be better: AM503B with A6302 or TCP202
I briefly read the spec the A6302 and TCP202. They seems to be very similar in the therms of performance ...
But TCP202 seems to be much newer design
What would you chose ?

[capt bullshot]

Don't know the TCP202, never had one to try out myself. Sometimes I use a Hioki (available as Yokogawa, Agilent, ... rebranded), so I can give a bit of comparison:

A modern Hioki one (e.g. the 3267 model) is clearly better in terms of drift and noise than the A6302 / AM503 combo, so I'd expect the TCP202 being better too, but I don't know if Hioki or other technology is inside the TCP202, so don't nail me down on the TCP202.

Noise and drift results most from the Hall sensor element, so the choice of a particular AM503 model doesn't make a big difference with the A6302.

[Someone]

I will hook it up tomorrow to my scope and measure the noise floor.

Excellent, thank you!!! 

OK, I hooked the K2 current probe up to two scopes and one DMM.
And here are the results for signal and noise floor.

As long as you are using "high resolution" or "average mode" for the acquisition, the slow signal looks perfectly alright and the noise is suppressed.

In normal acquisition mode, the Agilent MSO7034B has a peak to peak noise of 5.8 mA.
The 6000X scope shows about 5 mA Pk/Pk noise

The real signal has a 100 ns rise time.
The scope shows 185 us rise time because of the probes bandwidth limitation.

For slow signals, this probe is working well.
In this case, the 15 mA signal has a duration of 1 ms and it shows OK.

The price is reasonable and as I said before, the built quality is excellent.

Thanks for the details as it looks like it has more noise than the contemporary Tek or Hioki, we'll assume that first image with the rise time is in averaging mode? Tek have come down a long way with their current probe pricing and still have the best noise performance, what pricing are you seeing for the Chauvin-Arnoux K2?

[Neomys Sapiens]

Which one do You think will be better: AM503B with A6302 or TCP202
I briefly read the spec the A6302 and TCP202. They seems to be very similar in the therms of performance ...
But TCP202 seems to be much newer design
What would you chose ?

I think that they are very close regarding parameters. If you want flexibility, get an A6302 with AM503x, as you can operate with any scope without a pesky 1103 Power Supply. If you use a scope with a TekProbeII interface, and maybe even travel with it, try to get a TCP202. In this application I like it much, along with the Tek HV differential isolator, which also works directly from the interface.

[krisk]

I did probe comparison and hmm I don't understand System noise parameter... it seems like A6302 has only 250uA (up to 200MHz) where  TCP202 has 2,5mA (for only 20MHz bandwidth)? Or I'm missing something???
For the rest of parameters It looks like TCP202 has much better performance in high frequencies (above 1MHz) it also has better "Ampero-second product" so it theoretically better for high current spike measurement even if A6302 has higher current rating (20A).

Does anybody know, what is minimal current which can be measured by A6302, TCP202 probes?

I'm also wondering if A6302 and TCP202 are the some product class. Or at the time when they were sold there was some thing better ? (for A6302 there is 100MHz version, I think A6312)

[krisk]

I'm also wondering how TCP202 will work with Rigol scope DS4024?
Rigol T2R1000 Active Probe Adaptor and TCP202 seems to be nice compact solution which can save space on my desk:))

[alm]

Never tried it, but keep in mind that the TCP202 can draw more current than most voltage probes, and sometimes even induce ripple in the voltage rails feeding the probe power. The way these probes work is that an opposing current is fed through the transformer to keep the current measured by the Hall sensor zeroed. So if you are measuring 10 A, it will have to source 10 A / N (turns ratio). This is where the TM-500/5000 series amps are superior, since they have a dedicated power supply. No idea if this will be a problem with the Rigol adapter.

[MarkL]

I have both a TCP202 (w/1103 power supply) and a A6302 (w/AM503 amp).  Below is a comparison of the noise levels for a 0 to 2.0mA square wave @ 1MHz using a Keysight MSOX3104A.

The first photo shows the two probes on their most sensitive scale (Ch1 = TCP202, Ch2=A6302).  Since the TCP202 has no scale adjustment (fixed 0.1V/A into 50ohms), all the amplification happens inside the scope.  As you can see, the TCP202 is pretty much useless for signals that low.  Admittedly, much of that noise is from the scope itself since it's on maximum sensitivity (1mV/div), so if you had a low noise front end you could do better.  BW limiting to 20MHz just about halves it, so that's another approach.

If your signal permits, you can use averaging for a much cleaner result.  The second photo shows a 256 average.  However, the TCP202 signal is still too low to be very useful.  This particular scope has a "magnify" feature, which can be used to increase the amplitude of the displayed waveform, third photo, magnify in purple.  It can be handy if your scope has it too, but it's still pretty noisy.

If you're wanting to measure small signals, I would definitely go for a A6302 before a TCP202.  The A503x amplifiers can also be used with the larger A6303 probe for even higher current measuring capability.

I also have the adapter that allows me to plug the TCP202 directly into the Keysight scope, so truth be told, I actually end up using the TCP202 most often because of the convenience factor.  The Rigol T2R1000 adapter lists the TCP202 as supported, so one would presume they've done the design and testing to meet the power requirements for the list of supported probes.

If you're looking to buy a used TCP202 or A6302, make sure you get one where the housing is not cracked around the core closure (or pieces missing, I have seen).  It can prevent proper seating between the core halves and ruin your measurements.

I've also used the i-Prober 520 and I agree with other posters that it's far too fidgety and, I would add, limited in bandwidth for use as a general purpose current probe.  I wouldn't go there.

[krisk]

MarkL thank for your input.

I have found forum topic about AM503B + A6302  and its noise: https://www.eevblog.com/forum/testgear/tektronix-am503b-current-probe-amp-noise-levels/msg1027432/#msg1027432
From this conversation I'm assuming that AM503B + A6302  has noise flor around 5-6mVp-p and it's same in all amplifier ranges.
Does it mean that probe has always noise of the 0.5 division regardless amplifier settings?

For TCP202 i believe it's different, and it has noise level which is not scaling with amplifier settings because TCP202 has only one, fix gain settings 0,1V/A, ma I right?

Admittedly, much of that noise is from the scope itself since it's on maximum sensitivity (1mV/div), so if you had a low noise front end you could do better.  BW limiting to 20MHz just about halves it, so that's another approach.

As you wrote, on the lowest scope range 1mV/div reading is noisy but its also due to scope noise. But what about higher currents? I would expect that TC202 will have much better performance regarding noise.

In the TCP202 manual there is chapter "Increasing Measurement Sensitivity". If we count 4 loops we should get measurement multiplied by the factor of 5. 
Question is what impedance it will introduce .... ? For the air coil with 4 loops inductance is as low as 0,3uH but here we have probe core ....

BTW I'm bit afraid of A6302 & AM503 due to its age and possible fails ... capacitors e.g. and I'm pretty sure that it won't be easy to repair it

[MarkL]

MarkL thank for your input.

I have found forum topic about AM503B + A6302  and its noise: https://www.eevblog.com/forum/testgear/tektronix-am503b-current-probe-amp-noise-levels/msg1027432/#msg1027432
From this conversation I'm assuming that AM503B + A6302  has noise flor around 5-6mVp-p and it's same in all amplifier ranges.
Does it mean that probe has always noise of the 0.5 division regardless amplifier settings?

There is some noise introduced by the AM503 on more sensitive ranges.  To put some numbers to it, using the scope's measurement function and using the previously shown settings (10mV/div @ 200ns/div), the scope itself is 2.8mVpp for a baseline.  With the AM503 and A6302 and no signal input to the probe for different AM503 settings, here is the measured noise:

  1mA/div   8.8mVpp
  2mA/div   7.6mVpp
  5mA/div   3.6mVpp
  10mA/div  3.4mVpp
  20mA/div  3.2mVpp
  50mA/div  3.1mVpp

Clearly there's some additional amplification element that's switched in for the two most sensitive ranges since the noise floor more than doubles.

For TCP202 i believe it's different, and it has noise level which is not scaling with amplifier settings because TCP202 has only one, fix gain settings 0,1V/A, ma I right?

Right.

Admittedly, much of that noise is from the scope itself since it's on maximum sensitivity (1mV/div), so if you had a low noise front end you could do better.  BW limiting to 20MHz just about halves it, so that's another approach.

As you wrote, on the lowest scope range 1mV/div reading is noisy but its also due to scope noise. But what about higher currents? I would expect that TC202 will have much better performance regarding noise.

At higher currents the signal will be much larger than the noise floor, so I'm not exactly sure what you're asking.  What I was trying to point out in my previous post was that if you're trying to measure small signals, the front-end amplifier in your scope will factor largely in your success with the TCP202.  My front end isn't the best, but it's not the worst either.  The AM503 removes this dependency from the scope.

In the TCP202 manual there is chapter "Increasing Measurement Sensitivity". If we count 4 loops we should get measurement multiplied by the factor of 5. 
Question is what impedance it will introduce .... ? For the air coil with 4 loops inductance is as low as 0,3uH but here we have probe core ....

If you take a look further in the manual, there are charts Fig. 15 and Fig. 17 which show the insertion impedance for a single wire (1 turn) and 10 turns.  5 turns would roughly a factor of 5 times the single wire chart, but a more accurate estimate would be to use the 10-turn chart and divide by 2.

BTW I'm bit afraid of A6302 & AM503 due to its age and possible fails ... capacitors e.g. and I'm pretty sure that it won't be easy to repair it

I hear you on that and it's a valid concern.  Old equipment can be a purchase risk and a maintenance chore.  I can say I've never had to replace filter caps on mine (had it 15, maybe 20 years).  The only service it's needed was a broken mechanical link on the DC Level control and cleaning of the range contacts.  But no matter how much I cleaned the contacts, it was always a little crunchy after not being used for a while.  I've gotten used to it.

The TCP202 does have a problem picking up power supply noise when used with a TDS3xxx scope.  That's not your situation, I know, but be aware it does need a clean supply if you get one of those adapters.

And both of these probes drift quite a bit when trying to make low-level measurements.  It can be annoying to constantly zero out everything before taking a measurement snapshot.

[nctnico]

The question however is how much accuracy the OP needs. The cheaper current probes from Ebay (say in the $200 to $300 range) do a decent job for SMPS work because you are typically more interested in the shape of the waveform than the value.

[HighVoltage]

It looks like Keysight has a new current probe with a 1mA resolution and much better noise floor.
The Keysight N7026A

This one looks really promising for a clamp-on probe for low currents.
Just the price tag is EUR 4.522,00 + VAT !

[MarkL]

The N7026A looks great in comparison to their N2893A, but I wouldn't say it has a much better noise floor than the old AM503/A6302 combination, even given the higher BW.  Here is a screen shot of the same settings on a AM503/A6302 and roughly the same signal as in the N7026A brochure.  (The AM503 could obviously use some tweaking on the compensation, but I'm just pointing out the noise level here.)

The N7026A BW of 150MHz and 30A continuous is a nice bonus, compared to the A6302 50MHz and 20A.  But US$4800 is pretty steep.  Ouch.

Plus there's also the "lock-in" factor: The N7026A has a proprietary AutoProbe interface, so it can't be used with non-Keysight scopes and other equipment.  That would kill it for me right there, better specs or not.  I don't like my probing investment glued to my scope manufacturer, if I can possibly avoid it, since it creates a larger financial barrier in considering other manufacturers at scope refresh time (gotta re-buy all those probes).  The TekProbeII interface, as old as it is, has become somewhat of a solution to that problem since a bunch of manufacturers support adapters to it.

The OP said they have a Rigol DS4024 so I guess the N7026A is a no-go for them anyway.

[Fludo]

My first recommendation would be a Yokogawa 701933, but those are a little too pricey.
After that would be a fluke i1310s, they have a standard BNC connection with internal 9V battery and will work on fluke multimeters.

[woodchips]

Is it possible to come up with a simple test setup for current probes to measure noise etc?

I would assume this would show two traces, one of the voltage waveform across a 1 ohm resistor and one of what the current probe sees. These can be reduced or increased to highlight the noise. Or anything better?

I have to admit to a certain level of private interest here. I have three each of the 6302/AM503 and 134/6021 probes that are now no longer needed. I have done simple 'does it show a low and high frequency waveform' to ensure both the Hall and transformer parts are working, but perhaps something more is needed? I am reluctant to sell something that the buyer finds is not as described, rather just not sell them to save the hassle.

[MarkL]

Is it possible to come up with a simple test setup for current probes to measure noise etc?
...

All the equipment you mention have performance verification procedures in their respective manuals.  The AM503 procedure includes how to measure noise the Tek way.

If you're trying to minimize returns from picky buyers, doing the verification is probably the best route, but it's a pain and much of it requires special equipment.

Checking operation by sweeping a signal from DC (or whatever the lower limit is) to max BW through a shunt resistor and comparing the traces is a good idea.  Wrapping multiple turns through the probe to simulate a signal with max current is also a good check.  Make sure a square wave looks square at different frequencies.  Cycle through all the range and switch settings and make sure they all work.  Take a look at the noise levels and make sure they're not out of hand (or do the Tek noise measurement; that one is not so complicated).

In your description say how it was tested, offer a 14 day return, and put tamper seals on key areas to deter unscrupulous buyers.  But expect that no matter how much you do, some buyers are always going to claim an issue and ebay will almost always side with the buyer.

You could always try to sell your stuff here in this forum, where it doesn't have to be so impersonal and contentious.  This very thread is from someone looking to buy.  krisk, please meet woodchips...


EDIT: typo...