Micsig CP1003/CP503 100MHz/50MHz Current Probe

[amar87]

i will conntinue tomorow after reading your posts, i bought the probe from Meilhaus electronic last year probably the price is the same now i waited for it a long time and it also came with a china USB- 230V  adapter, and you can't power it from the scope because it needs some more juice than the scope can provide, it also seems that it doesn't want to work with other phone chargers i had arround .
 

[Martin72]

Hi,

Test Setup for Accuracy 6 A range :

- same as above + the following:
- DC electronic load
- Fluke 8846A Multimeter (not calibrated)

If you compare this with a multimeter, I would also measure the output of the current clamp with a multimeter.
There are BNC to banana plug adapters.
Apart from that, I'm no longer surprised that the current clamp is so "off" at very low mA values after I've seen the specs.
There the measuring range is specified from 20mA or 50mA (6A/30A).

A bit sobering as far as the 6A range is concerned.
Seen earlier, Batronix now also has them on offer, both variants in stock.
https://www.batronix.com/versand/messtechnik/tastkoepfe/Micsig-CP503B.html

[amar87]

Yes i expected it to be off at currents lower than they give in the datasheet, you are absolutely right. i just put those measurements in there because i wanted to see how bad it is or to better put it if it could be used. But what i observed is that the error (below 20mA in the 6A range and below 50 mA in the 30A range) is mainly caused by the offset that it has after zeroing. What i also observed that this offset  has a tendency to drift after zeroing. maybe some interesting experiments would be observe when it stops drifitng. Strictly speculating with the info i have at the moment this could be the the reason why they give the +/- 10Ma after the percentage.

I can also plug it in the multimeter witha BNC Banana adapter, but then how to compare it ? The DC load i used (Rigol DL3021A) needs to be stable. (which i have no idea how stable it is ), i could take a reading DC load on X mA in the shunt of the multimeter, then same with current probe plugged in the multimeter on voltage? Then we don't know how accurate the multimeter is on voltage vs on current. Then the sampling rate and bandwidth from the multimeter is lower than the scope.......... This makes me go crazy....

How i did it before i am actually showing the error from the scope + the error from the probe actually(which is also not very scientific ).... this is not easy to do...

What confuses me here is the spec of +/1%+/-10mA(6A Range) and it measures above 20mA whatever is under that you can't trust. Let's assume my fluke is perfect. i read on the fluke 22.9 mA on the scope 25 mA. Is the probe in spec or not ?     

Also the reason for not testing over 6 A is that i have no power supply that provides more than 6A.   

[Martin72]

Also the reason for not testing over 6 A is that i have no power supply that provides more than 6A.   

You can turn windings over the core of the clamp.
Example:
Turn a wire (Measure lead) 5 times around the core, feed in 1A, the currentclamp will measure 5A...
Like I did in the pic below.

What confuses me here is the spec of +/1%+/-10mA(6A Range)

For me it reads as +/-1%, but at least +/-10mA.

[_Wim_]

For me it reads as +/-1%, but at least +/-10mA.

I would expect this to be the sum of both: so +-1% of the measured value plus an additional +-10mA. This means when putting 6A through, you end up with +-70mA.

[_Wim_]

But what i observed is that the error (below 20mA in the 6A range and below 50 mA in the 30A range) is mainly caused by the offset that it has after zeroing. What i also observed that this offset  has a tendency to drift after zeroing. maybe some interesting experiments would be observe when it stops drifitng. Strictly speculating with the info i have at the moment this could be the the reason why they give the +/- 10Ma after the percentage.

This is typical for a probe like this. My fluke I50 also drifts after zeroing. I think you are correct that the +-10mA comes at least partly from this drift, and hence the accuracy is specified as the sum of the drift and the measurement error.

I can also plug it in the multimeter witha BNC Banana adapter, but then how to compare it ? The DC load i used (Rigol DL3021A) needs to be stable. (which i have no idea how stable it is ), i could take a reading DC load on X mA in the shunt of the multimeter, then same with current probe plugged in the multimeter on voltage? Then we don't know how accurate the multimeter is on voltage vs on current. Then the sampling rate and bandwidth from the multimeter is lower than the scope.......... This makes me go crazy....

I am not completely sure why you want to do this, but if it is to rule out errors from the scope, I would just check the scope DC voltage accuracy vs your multimeter. But as you are using a 10bit scope, I do not expect here anything significant vs the 1% error you are trying to verify. For me it looks like the probe is meeting is spec for DC accuracy everywhere apart from the 2A measurement on the 6A range where it just barely exceeds it, but this could be due to tolerances in the test setup. Noise is also quite ok from the looks of it.

For the AC noise, can you put the bandwidth limit to 20MHz on your scope and report the ACrms and ACpp? This should according to the datasheet be below 1.4mArms when limited to 20MHz.

[amar87]

The requested screenshots with 20 Mhz bandwidth:

6A range :


30A range 100 ms/div  :




I am not sure which figure to take looking at the STDEV measurement on the scope which i think is the RMS without DC component but not sure and playing with the timebase the noise can be lower


Having  20 ms/div i get this :





Then 20 ms/div and change aquisition to average i get this:




So i don't know how to measure to verify the ≤ 1.4 mA RMS (Bandwidth at 20 MHz, Range 30A, 10X) spec

[amar87]

I tried doing some bode plot using the internal signal generator from the scope that goes up to 25 Mhz but i think i am doing something wrong and don't know what but i would have expected the bode plot to be flat. I am attaching a few pictures  so don't laugh.....

i wanted to see if the magnitude line is flat and does not change up to 25Mhz




Here is the test setup and the scope setup for the measurement:

In Ch1(1 Meg )  i pluuged the output from the current probe in CH2(50 Ohm ) the input from the internal generator and the red wire is also looped through the probe as much as it fits.








I have to try with more windings i thing that is why i am getting crap.

[Martin72]

Assuming the generator really delivers a 3vpp level, you would have to measure approx. 21mA with just ONE winding (i.e. simply connect the clamp to the cable).
Because 3vpp is 1.06Vrms.
This voltage divided by the 50 ohms is 21mA.
The output of the clamp delivers 500mV/A, i.e. 10.6mVrms at 21mA.
That would be a factor of -40dB for one winding, you have several windings and approx. -67dB.

[_Wim_]

The requested screenshots with 20 Mhz bandwidth:

6A range :
(Attachment Link)

30A range 100 ms/div  :

(Attachment Link)


I am not sure which figure to take looking at the STDEV measurement on the scope which i think is the RMS without DC component but not sure and playing with the timebase the noise can be lower


Having  20 ms/div i get this :

(Attachment Link)



Then 20 ms/div and change aquisition to average i get this:

(Attachment Link)


So i don't know how to measure to verify the ≤ 1.4 mA RMS (Bandwidth at 20 MHz, Range 30A, 10X) spec

I would just put the scope channel into AC-coupling to avoid DC offset errors. The standard deviation is not the correct measure. Subtracting the "mean" value from the "RMS" will be closer, but I am not sure is this is mathematically correct also. 

[amar87]

For the bandwidth i have an 45W RF amplifier that can do 150kHz -250 Mhz , i need a high power terminator which i have to get my hands on i already asked arround to borow one ,  i will try to do it like this. I guess it could work if i remove the shielding from a coax and them solder a wire somehhow over the shielding so i can put the probe over the inner conductor + teflon  would work.
I can use my siglent scope on the FFT function (hope it has MAXHOLD) with the current probe plugged into it and then sweep the output of the RF Amp into a terminator and monitor the coax with the current probbe.

The above procedure should give a nice line and rolloff so we could see the bandwith i guess. Please correct me if you see something wrong in what i described above. 

[Martin72]

Hi,

I can use my siglent scope on the FFT function (hope it has MAXHOLD)

Has it:

https://www.eevblog.com/forum/testgear/siglent-sds2000x-hd-12bit-(published-for-chinese-domestic-market-only)/msg5213874/#msg5213874

[Martin72]

Hi again,

I guess it could work if i remove the shielding from a coax and them solder a wire somehhow over the shielding so i can put the probe over the inner conductor + teflon  would work.

This will also work, you may not even have to solder anything.
I have done this before and am doing it again.
At the moment the Bodeplot is running and the current clamp is the tektronix A6302/A501A combination.
I have set a 1Vrms and the load is simply the input resistance of the scope, as previously discussed and as you have already done.
The Bodeplot should have no problem with these low levels because the scope itself has no problem with them.
I had tested up to 100Mhz and am currently running it up to 50Mhz and maximum number of points, which takes a while....
Final then still with 20Mhz bandwidth limitation.
Pictures will follow soon.

[Martin72]

Here are the results of the good old Tekprobe...
It has a bandwidth of 40Mhz (-3dB), the pictures show the bodeplot once with and without faded-in phase, then again only this time with active bandwidth filter of 20Mhz (switched on at AM501), plus finer resolution of 2dB/div.
That was the setup with external generator, "tomorrow" I will reproduce your first attempt 1:1 with the internal generator.

[Martin72]

I have now also tried it with the internal AWG of our 2000X HD, without any problems.
Once with your settings, i.e. 10pts per decade, once with 50.
The finer you resolve the amplitude, the more "terrible" the transmission looks, but if the plot is between +/- 0.5dB, I think it's good.
If you take a resolution of 6dB, you will of course have a dash.
Well, what does that tell us now?
That it works as you had previously thought, even without several windings through the core.
So something must have gone wrong with your attempt, check again.
Oh well:
And the course of the transmission as shown in the data sheet of the CP1003B can safely be regarded as an idealized plot, it will not be the truth.

Martin

[Martin72]

Last pic.

[cameronb]

I picked up the CP503 from SAELIG.  It looks to require only a 5V/3A supply.  Noise is not critical as it drives a buck converter and linear regulator.

Here's what's inside
  GDE232E (Arm Cortex). 
  TI54331 buck converter + AMS1117 provide a clean 3.3V rail for the analog end of things.
  OPA4171 probably related to range selection from the DAC on the GDE232E
 
As far as I can tell, there are no electrical connections to the 3.3V and -5V rails. 
I've confirmed all the digital bits are running directly off the 5V rail.

It's probably as easy as passing a USB cable into the housing.  I'll try it tonight and report back.

[cameronb]

There's definitely a second rail on an unknown-to-me pin.  I did, however, get the device and relay up and running off the 3.3V rail.  I'm probing out a schematic right now.

I'll have more updates later this week.

[amar87]

I tried some more experiment:



First would be with the internal AWG from the scope.







Second with the signal generator where i manually swept up to 100 MHZ, recoded the measurements into a spredsheet, converted to DB and made an ugly bode plot.






if i compare the bode plot from the scope and the one from the spredsheet they don't really look alike. And that dip at 30 Mhz i don't get why it is. Maybe the probe is actually like this ... 

[idolclub]

i will conntinue tomorow after reading your posts, i bought the probe from Meilhaus electronic last year probably the price is the same now i waited for it a long time and it also came with a china USB- 230V  adapter, and you can't power it from the scope because it needs some more juice than the scope can provide, it also seems that it doesn't want to work with other phone chargers i had arround .

The maximum power consumption of CP1003B is about 5V / 800mA, typical 400mA.

CP1003B also can work with Apple or Samsung 5V / 2A charger.

[amar87]

I picked up the CP503 from SAELIG.  It looks to require only a 5V/3A supply.  Noise is not critical as it drives a buck converter and linear regulator.

Here's what's inside
  GDE232E (Arm Cortex). 
  TI54331 buck converter + AMS1117 provide a clean 3.3V rail for the analog end of things.
  OPA4171 probably related to range selection from the DAC on the GDE232E
 
As far as I can tell, there are no electrical connections to the 3.3V and -5V rails. 
I've confirmed all the digital bits are running directly off the 5V rail.

It's probably as easy as passing a USB cable into the housing.  I'll try it tonight and report back.

If you have the time could you also try to do same experiment as Martin and me and report back the result?

[Martin72]

I tried some more experiment:

I can't see any difference between your setup and mine, should the current clamp really be that "bad"...



But I can't believe that this is normal for a model costing over €1000.
Perhaps the people at Meilhaus can check this on another Micisg.

[_Wim_]

if i compare the bode plot from the scope and the one from the spredsheet they don't really look alike. And that dip at 30 Mhz i don't get why it is. Maybe the probe is actually like this ...

I would first try to figure out why you get different results. First thing I would do is make a bode plot of the signal generator only (so feed the signal both to the 50ohm input AND the 1M input. This should give a flat response, if not, investigate further.

I also would try a figure out why with the external the signal goes down when the frequency goes up, and it is inverse with the internal signal generator.

If you repeat the plot multiple times (with for example the internal signal generator), do you get everytime the same result, or is the result more random in nature?

[_Wim_]

One more question, are you sure the slider is fully closed (should be in the "lock" position to measure)

[Martin72]

Woah, Eagle-Eye! 



It could be, no, it will be.