Micsig CP2100B current probe

[gregb]

Had been working on a project using Trinamics stepper drivers (specifically the https://www.trinamic.com/support/eval-kits/details/tmc5160-bob/), and hit a point where I couldn't tune things further. Following the app notes as recommended required being able to view current waveforms. Current sense resistors weren't easily accessible in this case, and measuring voltage across an additional shunt wasn't yielding anything useful.

I started keeping an eye out for decent current probes on EBay. Seems there's a lot of vintage Tektronix gear (eg TM502A with A6302 probe), but offered prices seem very large compared to sold prices o_O

The big problem with most new probes is low bandwidth. The chopper frequencies of the steppers started at 20kHz, so I thought I might be mislead by 50-100kHz probes.

But late last year I found MicSig have some current probes on offer. There are two models - CP2100A (800kHz) and CP2100B (2.5MHz).

The 800kHz version is about €180 and the 2.5MHz version (which I purchased) is about €310; this is substantially cheaper than the vintage probes I've been following. I got them from https://eleshop.eu/ who were good, but there seem to be many other suppliers.

The probe is powered via usb, and connects directly to the scope. The controls are pretty self explanatory - has 10A (0.1 V/A) and 100A (0.01V/A) ranges, and a current 'zero'

I've only done minimal testing so far, AC waveforms and on a 50 load at low voltages / currents, nevertheless things seemed correctly scaled on these low ranges. Waveforms are noisy because of the low range. At higher frequencies there was a very slight delay (order of a few %) but that could be due to my testing procedure, and isn't really important to my intended usage.

I don't have ready access to the original project that got me interested in this equipment, but I'll try it on that when I get the chance. Bought the probe mainly because it seemed generally useful, but also to feed that growing Test Equipment addiction I seem to have acquired.

Advice on usage / questions welcome!

Greg

[bob91343]

I recently upgraded my HP628A to 628B, which has oscilloscope output and ranges up to 10A.  It reports DC on the meter but the 'scope output is good to a few hundred Hz.

I used to have a Tektronix current probe but it didn't work down to dc.  But it was good to rather high frequency.

In any case, while I would love to have a high frequency current probe, the cost is prohibitive.

[Ivan7enych]

Can you tell how much noise it has in 10A range?
I see in specification "Measuring range : 50mA~10Apk(10A)", but nothing about noise...

[tautech]

At higher frequencies there was a very slight delay (order of a few %) but that could be due to my testing procedure, and isn't really important to my intended usage.

It is perfectly normal for current probes to show a propagation delay WRT a passive probe.
It is addressed/corrected using the scopes probe deskew feature so voltage and current peaks align.

It matters for scopes with Power Analysis capabilities.

[gregb]

@bob91343 - I hadn't seen those, will look them up. Haven't tested near DC yet.

@Ivan7enych - On the 10A range I was only testing currents around 100mA - so the noise I was seeing was around 10-20% on 1-2mV - very consistent and about the floor of my scope and not a fair test really. However in average mode the waveforms followed *really* tightly and followed the applied voltage waveform very closely. It's on my list of things to do to test a current amplifier with 1-2A.

@tautech - that's really interesting, I hadn't twigged that they were useful for that. Maybe handy for switch mode supplies, or even RF?

[tautech]

@tautech - that's really interesting, I hadn't twigged that they were useful for that. Maybe handy for switch mode supplies, or even RF?

Affordable current probes don't have many RF applications as frequencies are too high however for 'things' switching they are a necessary tool. Good ones will allow you to see when you're approaching saturation of your magnetics if needing to push a design to the limits.

In your OP you talked about the high prices of them but if you know what you want and are patient you can pick up good Tek current probes for just a few hundred $ on eBay. I've got a few from there over the years and never paid more than $300 for 50 and 100 MHz AC current probes. They are specialist bits of gear and expensive to replace so care need be taken not to abuse them in any way.

[nctnico]

$300 for an AC current probe is way too much. You can make such a probe yourself using a small current transformer which costs a few dollars at most. Getting 10MHz of bandwidth is easy.

[gregb]

@nctnico - re: current transformers - thanks for the tip, I haven't used those before; would be interesting to compare. For the stepper control case I'm certainly mainly interested in AC.

[tautech]

@nctnico - re: current transformers - thanks for the tip, I haven't used those before; would be interesting to compare. For the stepper control case I'm certainly mainly interested in AC.

Don't get your hopes up too high for a simple solution.....been there done that !

My first current probe was a CT rescued from an old PC PSU where it presumably did duty at a few 10's of KHz.
Connecting a coax and BNC for the old CRO I had at the time terminates it into the scopes 1M inputs then where you need something to calibrate it against, in that case a borrowed Tek P6021 and shunted the CT output with a resistor so that the 2 current probes readings matched to some reasonable degree.
Enevidably sensitivity wasn't sufficient for my needs and the clumsiness of use needing to feed a wire through the core sealed its fate of hanging on a nail in the shed where it still is and to this day my focus still remains on the wonderful decades old Tek designs like P6021 and P6022.

[bob91343]

I would like some guidance on how to make a current probe.  One problem is that the current to be measured would normally be opened for a shunt but the probe needs to clip on so as not to disturb continuity.

So would I have to cut a ferrite to do this?  I don't think that's easy.  Maybe one of those clip on ferrites for interference control?

[tautech]

Not at all simple or everyone would be doing it !
Insertion losses are to be considered at low currents and high frequencies.

Have a study of this:
http://w140.com/tekwiki/images/6/62/062-1146-00.pdf

[nctnico]

I would like some guidance on how to make a current probe.  One problem is that the current to be measured would normally be opened for a shunt but the probe needs to clip on so as not to disturb continuity.

In today's designs with SMD components you have to open the circuit anyway. Using litze wire reduces skin effect.

@Tautech: you have to get a proper current sensing transformer for use in HF switching dc-dc converters and load it with a resistor at the oscilloscope side.

[tautech]

@Tautech: you have to get a proper current sensing transformer for use in HF switching dc-dc converters and load it with a resistor at the oscilloscope side.

Nope, I have 2 perfectly good Tek current probes rated to 50 and 100 MHz and highly respect the signal fidelity and sensitivity these fine old designs preserve.

[nctnico]

@Tautech: you have to get a proper current sensing transformer for use in HF switching dc-dc converters and load it with a resistor at the oscilloscope side.

Nope, I have 2 perfectly good Tek current probes rated to 50 and 100 MHz and highly respect the signal fidelity and sensitivity these fine old designs preserve.

If you have such probes then fine but my point is that you can't take a random current transformer (probably one for 50Hz) and then go on and claiming old Tektronix probes are better. That makes no sense. If you buy a current transformer intended for switching power supplies (these for example: https://www.we-online.com/catalog/en/WE-CST/ ) you'll get very good results for a few dollars. Sure you won't have clip-on probes but in the circuits I design I have to add a wire anyway if I want to measure the current (after lifting a component).

[gregb]

@tautech - Keeping an eye out for a Tektronix setup, one of your earlier posts that got me interested in them in the first place. That Tektronix note's good reading.

@nctnico - Thanks for clarifying your suggestion, I wasn't sure you meant opening the circuit, or that the mini 1MHz transformers were the ones you were talking about. I couldn't find >1MHz, so it's useful to know they can still give a decent picture for higher frequencies. Went looking further and found these discussions:
    https://interferencetechnology.com/the-hf-current-probe-theory-and-application/
   
I did find some split core sensors with claimed bandwidths in the MHz range - but their intended application was detecting current spikes from discharge.

All of this is really useful.

[tautech]

@tautech - Keeping an eye out for a Tektronix setup, one of your earlier posts that got me interested in them in the first place. That Tektronix note's good reading.

If the one you have despite its low BW is doing its job and meets your needs just keep using it. 
Knowing what's 'out there' can be more costly than working with what you have already. 

Looking back, I wouldn't change a thing as I know these old Tek probes are hard to beat.

[Calvin]

Hi,

Vacuumscmelze offers a range of current sensors (Hall-element, 100/200kHz BW) that allow for hig precision current monitoring.
I used T60404-N4646X662, T60404-N4646X400, T60404-N4646X410 and T60404-N4646X412 .... find them at Mouser or al.
They all can be configured to three transformation factors.
The X662 puts out a singleended offset voltage, while the other three put out a bipolar current, thereby making signal conditioning especially easy.
A current loop has to be opened for testing with these devices, but if that is possible, these sensors work like a charm.

regards
Calvin

ps. made single-sided mixed SMD/TH layouts for these sensors including the conditioning amplifiers and a dc-dc converter supply.
If anybody is interested I could publish.
pps. the X662 is available at Pollin.de at the moment for under 2€/pcs.

[mnsergey]

Dear gregb, I wanted to ask: after purchasing CP2100B, did you test it for compliance with the Rice Time parameter less than 175 nS ?
I have a negative experience of purchasing this model (CP2100B pf $350 with delivery). When I received this product and conducted tests, I received a delay of about 400-450 nS which corresponds to the Junior CP2100A model. the Seller refused to compensate for the price difference, I opened a dispute and sent the device to China to the seller back.
I will be grateful for any information on this device, as I want to buy it, but with honest, not with Chinese characteristics




P.S. I wrote to Micsig support and You know what they answered me through the seller ?  "It is impossible to accurately estimate the Rise Time of the probe if the delay of the measured signal is greater than the Rise Time parameter of the probe" I laughed to tears, and replied that all over the world use the delay subtraction method, and I used the ready-made DELAY (1-2) Function in my RIGOL. I sent them a screen, of course.

[mnsergey]

At higher frequencies there was a very slight delay (order of a few %) but that could be due to my testing procedure, and isn't really important to my intended usage.

It is perfectly normal for current probes to show a propagation delay WRT a passive probe.
It is addressed/corrected using the scopes probe deskew feature so voltage and current peaks align.

It matters for scopes with Power Analysis capabilities.

Hello Ivan,
I had a negative experience of buying CP2100B, after the measurements, the delay showed that I have at least a Junior model CP2100A although the labels and the certificate claimed that I was sent CP2100B. Opened a dispute, the Chinese refused to return the difference in price, I had to send Pribluda back to China. While on the road, it is not yet known what will end. 22+ thousand rubles in a limp state, so buying on ALI, You can easily get pricked.
And this is not an isolated case, just yesterday came the differential probe DP20003 from the same Micsig/ And what do you think ?   10 NSEC instead of 3.5 and 25 MHz frequency band instead of the promised 100.
Here they are Chinese...
I remember Chinese watts on tape recorders; D
So by worrying about noise, you run the risk of getting Poo

[nctnico]

Perhaps your measurements setup is incorrect. To measure the bandwidth of a probe you need to use a 25 Ohm source (50 Ohm output of an RF generator terminated by 50Ohm). For a differential probe you need to twist the wires as well. The wires are definitely part of the probe'd design. For the lower voltage model of MicSig I measured around 70MHz to 80MHz and I think the higher voltage version should give the same results.

[luma]

Perhaps your measurements setup is incorrect. To measure the bandwidth of a probe you need to use a 25 Ohm source (50 Ohm output of an RF generator terminated by 50Ohm). For a differential probe you need to twist the wires as well. The wires are definitely part of the probe'd design. For the lower voltage model of MicSig I measured around 70MHz to 80MHz and I think the higher voltage version should give the same results.

You might be thinking of the differential voltage probes from MicSig, not the CP2100A/CP2100B current probes shown here.  I know you have a good deal of experience with these sorts of things (including your own designs) so while I have you here - What would be a good bench-top setup to measure rise time on a clamp-on current probe?  I have a CP2100A on the way from China and would be interested in confirming the specs before trusting the readings based on mnsergey's experience above.

[nctnico]

I'd connect short across the output of a 50 Ohm function generator and sweep the frequency. Rise time is just a rule of thumb; better measure the frequency.

[2N3055]

First thing, OP used probe with 50 Ohm termination. That is wrong. It should be used with 1 MOhm termination. Manuals are there for reason.
Secondly, @mnsergey seems to confuse probe delay (probe deskew is there for that) with signal rise time. It might be a language barrier... Signal delay is not even specified in manual.

Output of siggen with 50 ohm termination that has enough loop to fit the probe will be ok, but low current.

Deskew calibration source (fixture) is realy best way to do it. If you have current probes you might want to get one anyways, it is counterpart of built in probe calibration source in scopes.
One can actually be even built without much problems.. They are quite simple and it is a shame they are quite expensive, they could be sold at 50 € at a lot of profit, and everybody would buy one...

[gregb]

@mnsergey - I haven't tried to do a characterisation of the bandwidth, or accurate measurement of delay yet. I wanted to ensure I was measuring things correctly before I put information out vs claimed specs. Also the frequency range I was initially interested in was at the lower end (100kHz) of this probe's spec.

Had it setup with a 50ohm source via coax into a 50ohm load - not exactly as @nctnico was recommending, but maybe ok for 1-2MHz.

Also at the moment I can only investigate low current (order of 200mA) from my siggen - would be nice to try it on higher currents, at least at lower frequencies.

@2N3055 - re: "First thing, OP used probe with 50 Ohm termination. That is wrong. It should be used with 1 MOhm termination. Manuals are there for reason." - thanks for the tip; maybe I miscommunicated - my scope is a 1054z which is always 1MOhm, and I wasn't using a 50 ohm terminator on it. The load is 50 ohm, being driven from the 50 ohm output of a siggen.

[mnsergey]

Perhaps your measurements setup is incorrect. To measure the bandwidth of a probe you need to use a 25 Ohm source (50 Ohm output of an RF generator terminated by 50Ohm). For a differential probe you need to twist the wires as well. The wires are definitely part of the probe'd design. For the lower voltage model of MicSig I measured around 70MHz to 80MHz and I think the higher voltage version should give the same results.

You are completely right, according to the method of correct measurement. However, when I did not load the cable output of the oscilloscope-50 Ohms, although the generator output is standard 50 Ohms, it was a very insignificant additional delay-no more than 1nsec, since the cable is short-less than 0.5 m. But even this one nanosecond does not play a role, since I use the difference calculation method, where: Delay = Rise Time 2 - Rise Time 1. And this is confirmed by the screen where a special DELAY 2-1 function is used in my RIGOL.

And how do you like the second screen, where I use sine and stop when the blockage becomes -3dB from the level at the frequency of 1MHz ?
 Admire.