Wide Band RF Current Probe Empire/Singer/Gertsch CP-315

[_Wim_]

Does anybody has some more info on this one:

https://www.ebay.com/itm/Empire-Singer-Gertsch-CP-315-Wide-Band-RF-Current-Probe-Clamp-On-Transformer/283121969474?hash=item41eb626942:g:-hAAAOSwBydbfhje

Especially max frequency spec would be interesting, and also current rating (probably high enough for typical electronics use)

I have looked and looked on google, but could not find any info at all. Found a few Ebay listings of a CP-105, but no datasheets or any info on the web. Is my google fu failing me?

[tautech]

The seller lists 70KHz as a tested -3dB point and that's a but low for much advanced electronic work. 1mV/mA sensitivity is good though. However it looks like it could manage a fairly big conductor however the clamp provides for zero insulation which would be a big minus for lots of work.
$150 is OTT for such a current clamp IMHO.

[_Wim_]

Thanks for the info. For DC to 100kHz I use one of these: https://www.lem.com/sites/default/files/products_datasheets/lts_6-np.pdf They are quite cheap (12€) and work quite nicely with a sensitivity of 100mV/A. High frequency is however much more difficult. I have repaired a fluke I-50 current probe (goes up to 50Mhz), but for small currents (1mA) is it to noisy. This one however with a sensitivity of 1V/A could yield a lot more info.

I have been looking for quite a while on Ebay for a clamp on probe like this, but most go for much higher prices.  $150 being OTT is a bit late for me, as I already bought one...  (he had 4 available, 3 were already sold). When they were selling fast I gave up in looking for the datasheet, and made the gamble. I was now hoping somebody got some additional info.

[_Wim_]

$150 is OTT for such a current clamp IMHO.

I think you are indeed correct, one sold for only 51$ a couple of weeks ago. Did not show up in my search on Ebay the previous time. I think I may have jumped the gun a bit because the other were selling quite fast... Och wel, my TEA will be satisfied anyway (for a short while...)

[tautech]

Sure but study the datasheet a bit harder for the frequency spec. At a glance I didn't see one for the industry std -3dB and only -1db to 200KHz.....plus it's a AC-DC current sense and for 12 Euro that's a good deal IMHO.

Just be damn careful what you measure with the Empire/Singer probe as there's zero physical insulation.

I've never cared too much about fine accuracy for a scope current probe as that's not what I'm normally looking for, instead switching timing, inductor saturation and back EMF indications. 

[_Wim_]

Sure but study the datasheet a bit harder for the frequency spec. At a glance I didn't see one for the industry std -3dB and only -1db to 200KHz.....

I am not sure what you mean here. Bandwith is at least 200kHz then no?

Just be damn careful what you measure with the Empire/Singer probe as there's zero physical insulation.

Thanks for the warning, I will...

[_Wim_]

It seems to be only farnell who is selling them for 12€ (+tax). Other sites sell the LEM-units for about 25€ 

https://be.farnell.com/lem/lts-6-np/current-transducer-6a-pcb/dp/2146860

[tautech]

Sure but study the datasheet a bit harder for the frequency spec. At a glance I didn't see one for the industry std -3dB and only -1db to 200KHz.....

I am not sure what you mean here. Bandwidth is at least 200kHz then no?

Correct.
But as this is a precision device and not so much test equipment, it's BW spec is rated to tighter limits than -3dB.

[mmagin]

I have a Ion Physics Current Monitor that's constructed like this, but not split core.  They made (make?) a lot of different models with different ratios and different bandwidth.  Back when I bought mine used on ebay, they were going for much more reasonable prices, but it seems they're all $150+ now.

[_Wim_]

I have a Ion Physics Current Monitor that's constructed like this, but not split core.  They made (make?) a lot of different models with different ratios and different bandwidth.  Back when I bought mine used on ebay, they were going for much more reasonable prices, but it seems they're all $150+ now.

The spilt core I do find a hande feature, as it not always possible to fit the cable through.

I discovered the previous seller selling the Singer CP-315 at $51 had 5 identical pieces, and they were bought by this seller (buyer and seller both have an Ebay feedback rate of 918, what is the chance...), and he is now reselling them easily at$150, so he's is making a nice profit, and keeping 1 for himself.

[_Wim_]

I have a Ion Physics Current Monitor that's constructed like this, but not split core. 

Another alternative is Fischer Custom Communications. These are also spilt core and high bandwidth, but $150 seems whichfull thinking for one of these...

[_Wim_]

Would it be realistic to expect simular performance for the CP-315 as for the FCC  F-33-2?

https://www.arbenelux.com/wp-content/uploads/2015/01/DS-F-33-2-Rev-RLSE_1d22.pdf

Both have a -3db at 70kHz, a transfer impedance of 1 ohm, and simular dimensions (can't check exactly, as I did not yet receive mine).

For the Singer there is really no info on the net... 

[Hydron]

What sort of work are you planning to use it for? 70kHz lower limit is a bit high for general purpose use, but if you're looking at EMC debugging etc and it performs at all similar to the FCC clamp then $150 is a steal (though $51 even more so).

For lower frequency work you do occasionally see Pearson current monitors in clamp form, but they don't come up often and can be pricey. I was lucky enough to get a model 3525 (0.1V/A) and 7800 (0.01) for 100 and 70 GBP respectively, but at that price it was a race from receiving the eBay email to hitting "buy now" - 50% higher would have been a good deal still.

[_Wim_]

EMC is certainly of interest. The other goal was to measureme the actual current of modulated light sources.

Just did a few tests with the LEM module, it quite noisy (much more then my fluke i50 probe). See comparison below:

red: reference voltage into 50ohm load
green: fluke i50
blue: lem module

Both AC coupled because of 2.5V dc offset of the lem module. Lem module is wired for max sensitivity.

[nctnico]

The LEM module is using a HAL sensor and these are quite noisy in general. Besides that the Fluke probe is probably made by LEM as well.

[_Wim_]

Bandwidth of the LEM module. So the +1db at 200kHz was not a typo in the document… It really has a rising responce at the high end

[_Wim_]

Just for comparison, this is the fluke i50 (same 1 Vpp signal into 50 ohm, so +-10mA)

Remark: FRA is limited to 20Mhz with my scope, so it was not possible to test higher up (-3db should be at 50Mhz according to the datasheet)

[_Wim_]

Besides that the Fluke probe is probably made by LEM as well.

Yes, it is the LEM PR50 with a different label...

[_Wim_]

For lower frequency work you do occasionally see Pearson current monitors in clamp form, but they don't come up often and can be pricey.

I had a detailed look to these, these are indeed ideal for general purpose work. When using a pre-amp for the scope, what is the lowest current signal you can measure?

[Hydron]

My clamp-on ones aren't that sensitive at 0.1 and 0.01 V/A, but I also have some smaller fixed core 1V/A units. I've never tried to measure very low currents with these (or used a pre-amp), but I imagine that given the 50R source impedance that the scope's own noise would be the limiting factor.

On the topic of pre-amps for scopes, has anyone got any suggestions for a good DIY design? Thinking of something like 10x gain, 50MHz bandwidth with 50R input Z.

[Hydron]

You may also want to see what the Fluke i50 does when the power is off. Depending on the design it may be passively terminated at high-frequency when the power is off, so you'd be able to see what part of the noise is from the hall sensor/amplifier vs the CT section. It could unfortunately default to 0.1V/A with the power off (rather than 1V/A), but maybe worth a quick test? If it works at 1V/A then it may give a ballpark idea of how noisy a Pearson or similar current monitor would be.

[_Wim_]

You may also want to see what the Fluke i50 does when the power is off. Depending on the design it may be passively terminated at high-frequency when the power is off, so you'd be able to see what part of the noise is from the hall sensor/amplifier vs the CT section. It could unfortunately default to 0.1V/A with the power off (rather than 1V/A), but maybe worth a quick test? If it works at 1V/A then it may give a ballpark idea of how noisy a Pearson or similar current monitor would be.

No signal att all when powered off.

I think (hope) noise wise the Singer probe I bought will be comparable or better at 100kHz to the pearson at lower frequencies.

[Hydron]

Ah, that's a shame. The following tech note may be of interest: https://www.gmw.com/electric_current/Bergoz/documents/CT-TECH-NOTE-02-08.pdf
(From https://www.gmw.com/electric_current/Bergoz/ct_Support.html)

I'm hoping to find time at some point to build a DIY amplifier for my A6302&A6303 probes (to replace an AM503), and I'll be measuring the output noise as part of this to determine if a pre-amp would be useful for a 10V/A gain option (with A6302 probe), or whether the noise from the hall sensor and bucking amplifier is too high to make it worthwhile. Would also include 1V/A (straight 50R term on scope) and 0.1V/A (20dB T attenuator, partly bypassed at DC to allow for higher bucking current) options. This would be similar to the TCPA300 amplifier operation, with the scope doing most of the offset/attenuation work rather than the amplifier as in the AM503.

[_Wim_]

Ah, that's a shame. The following tech note may be of interest: https://www.gmw.com/electric_current/Bergoz/documents/CT-TECH-NOTE-02-08.pdf
(From https://www.gmw.com/electric_current/Bergoz/ct_Support.html)

I'm hoping to find time at some point to build a DIY amplifier for my A6302&A6303 probes (to replace an AM503), and I'll be measuring the output noise as part of this to determine if a pre-amp would be useful for a 10V/A gain option (with A6302 probe), or whether the noise from the hall sensor and bucking amplifier is too high to make it worthwhile. Would also include 1V/A (straight 50R term on scope) and 0.1V/A (20dB T attenuator, partly bypassed at DC to allow for higher bucking current) options. This would be similar to the TCPA300 amplifier operation, with the scope doing most of the offset/attenuation work rather than the amplifier as in the AM503.

Thanks for the doc. Reading this doc, a 20µA should be possible with an 1V/A probe and a preamp with an NF of 6db. Nice 

Making your own version of an AM503 seems like quite an endeavor. Keeps us updated if it happens...

[hwalker]

Kenneth Wyatt wrote an interesting article on high frequency current probes.  A link to the article can be found at http://www.interferencetechnology.com/wp-content/uploads/2012/04/Wyatt_NA_DDG12.pdf.  He gives a DIY procedure for characterizing a home made probe which can be applied to the CP-315.


-Herb

[_Wim_]

Kenneth Wyatt wrote an interesting article on high frequency current probes.  A link to the article can be found at http://www.interferencetechnology.com/wp-content/uploads/2012/04/Wyatt_NA_DDG12.pdf.  He gives a DIY procedure for characterizing a home made probe which can be applied to the CP-315.


-Herb

Thanks. When the CP-315 arrives, I will give this a try...

[_Wim_]

I received the probe, and it appears to have a frequency responce from 70 kHz to 15MHz. Attached is the frequency responce measured with my Picoscope (goes only to 20MHz for the FRA bode plot) and also with my CMU200 (from 10MHz up to 100MHz). I used a coax cable with 2 100 Ohm parallell resistors to generate the current (see calibration fixture). The CMU200 was set to 0dbm and had a 10 dB attenuator at the output to protect it from reflections. The picoscope had a 50 ohm adaptor at its input. The probe outputs approx 1V/A (-32db with 50 ohm load).

[_Wim_]

So I think it is more equivalent to an FCC F-40 probe:

https://www.arbenelux.com/wp-content/uploads/2015/01/DS-F-40-Rev-RLSE_57ad.pdf

Any other tests I could do?

[Hydron]

Maybe worth having a play with a differently constructed calibration jig? Possibly with a larger diameter centre conductor - current one will be far from 50R in that portion. If you get significantly different results then maybe it can go higher, otherwise it adds confidence in your original measurements.

[hwalker]

Your test fixture is going to have some radiation loss at the high end so I would guess that the actual range is up to 20 MHz as shown for the FCC F-40.  The probe manufacturers use a shielded test fixture with a vswr of less than 1.2 when calibrating their probes for transfer impedance.  I have built a fixture using a large Bud box terminated with a 50 ohm load and was able to generate transfer impedance curves that were within a dB of those I received from the manufacturer.  The fixture you constructed should be fine for verification purposes.  I would run another test and generate actual transfer impedance curves if want to compare the sensitivity of your probe against some commercial equivalents.  If you can export your specan readings to Excel you can semi-automate the measurements.

First column:  Frequency
Second column: Specan measurement in dBuV with generator connected to the input of the fixture, current probe is installed in fixture and terminated with a 50 ohm load, output of fixture is connected to 50 input of spectrum analyzer.
Third column: Fixture current in dBuA (equal to second column minus 34)
Fourth column: Probe output in dBuV with generator connected to the input of the fixture, current probe is installed in fixture and connected to 50 input of spectrum analyzer, output of fixture is terminated with a 50 ohm load.
Fifth column: Transfer Impedance  in dB ohms (equal to fourth column - third column)

Sample measurement (use same generator level throughout measurements):

Freq_MHz   Fixture_dBuV    Fixture_dBuA     Probe_dBuV     Probe_dBZt
  1                      -11              -45                      -46.5          -1.5


- Herb

[_Wim_]

Hi, did the test as suggested by Herb above. Fixture current in dBuA was calculated using second column - 32 (divide by 40, 1V/A into 50 ohm) (incorrect, must be -34 as correctly stated by Herb below)

[hwalker]

Nice job on the measurements and photo documentation.  They will definitely help others who need to characterize/verify their own rf current probes.  I'm not familiar with any probe I have used in the past which has a similar transfer impedance curve to the CP-315, so I don't know which test standard it was designed for.  Price wise, I think you got a good deal on the probe and if it doesn't suit your needs you should easily be able to make a profit on it with the addition of the verification data you can now provide.

Double check your measurements by generating a known current in your 50 ohm loop and verifying that you read the expected value on your spectrum analyzer.

For example:

With -40 dBm into your 50 ohm loop you should have, -40 dBm + 107 (dBm -> dBuV) - 34 dB_ohms = 33 dBuA.

Therefore at 50 MHz, per your data, you should be reading approximately 33 dBuA + -5 dBzt = 28 dBuV with your current probe on the spectrum analyzer.

- Herb

[_Wim_]

With -40 dBm into your 50 ohm loop you should have, -40 dBm + 107 (dBm -> dBuV) - 34 dB_ohms = 33 dBuA.
- Herb

Thanks for challenging me on my "calculations", I indeed made some incorrect assumtions above. I am now not in the lab, but convinced myself the value read on the spectrum analyzer is correct for a 1V/A probe. I used the following table for this: http://wera.cen.uni-hamburg.de/DBM.shtml

Input into load:
-10 dBm (0 dBm output on the siggen with 10 dBm attenuator)
=> 70.7mVrms =>  1.41mArms (for a 50 ohm load)

1.41mArms in a probe with an expected output of 1V/A
=> 1.41mVrms output => -44dBrms on spectrum analyzer

And this is exactly the value I am seeing in the passband (@10MHz)

What still strikes me as quite strange, is that the probe continues outputting quite a strong signal up untill almost 100Mhz. After I posted the above I also ran a simular test up until 500Mhz, but I unsure how to interprete this result.

[_Wim_]

But indeed at 50Mhz something does not add up. Let me investigate...

[_Wim_]

Ok, still had a -2dB error in my Excel from using an incorrect conversion factor. Attached is the corrected graph.

I also added a plot for V/A as I personally find this more instructive than dBOhms. Probe outputs a little more than 1V/A

I also attached my Excel, maybe it is of use for somebody else... (warning: I use a Dutch version of Excel, so some formulas might be different when opened in an English version).

[hwalker]

My QA personnel requires me to perform 3 repeatable tests runs to verify any calibrations I perform, so it is second nature for me to suggest a sanity check be performed when someone else is doing a calibration.  Your attached spreadsheet data is correct by my calculations and anyone else who uses a CP-315 should be able to spec out their probe using your data.


Happy Measurements,


Herb 

[_Wim_]

My QA personnel requires me to perform 3 repeatable tests runs to verify any calibrations I perform, so it is second nature for me to suggest a sanity check be performed when someone else is doing a calibration. 

Absolutely true. Will do in the next couple of days when I have some time to "play"...

[_Wim_]

Ok, here is another test.

Test setup:

Signal generator (CMU200) => Test fixture => channel 3 (pink) Rigol DS1054Z (with 50 ohm adaptor on input)
Current probe output => E&G 5185 pre-amp => channel 1 (yellow) Rigol DS1054Z (with 50 adaptor on input)

Goals: compare signal level between both to estimate frequency response
Limits: preamp only to 150MHz, Scope only to +-100Mhz

Results, see screenshots.

Conclusion: simular behavior is seen (altough frequencies are a little shifted, probably to the use of the much longer cables to test this)

[tautech]

Set signal source and scope to display and measure 1V p-p at low frequency.
Set cursors to 0.707V p-p and increase test frequency until waveform diminishes to 0.707 cursor levels.
Result = frequency @ -3dB point.

Can also be done with a swept frequency that better shows linearity over the sweep range.

[_Wim_]

Set signal source and scope to display and measure 1V p-p at low frequency.
Set cursors to 0.707V p-p and increase test frequency until waveform diminishes to 0.707 cursor levels.
Result = frequency @ -3dB point.

Can also be done with a swept frequency that better shows linearity over the sweep range.

Thanks for your reply. For a well behaved frequency response (one that can be characterized by a single -3dB point) I totally agree with the method you stated.

In this case however the goal was to check the measured impedance plot in post https://www.eevblog.com/forum/testgear/wide-band-rf-current-probe-empiresingergertsch-cp-315/msg1811753/#msg1811753

As you can see from that plot, is not so well behaved (between 0.1 and 20MHz is it reasonable ok, above that, it still outputs quite some signal, but it is not flat any longer).