"Cheap" Chinese Current Oscilloscope Probe

[ivaylo]

Well, only as cheap as current probes go, paid $235 delivered...

Felt a bit adventurous I guess and got myself one of these http://bit.ly/Wgvm6A. Did not believe the spec the seller is posting (unbelievable from what I've seen) but thought even if it is like twice exaggerated it may be still usable. Anyways, completely impulsive buy...

Long story short, the item arrived, it kinda worked but the noise level at DC was unacceptable I thought (around 10mV, a couple of MHz nasty noise, it was supposed to work from 1mA, 1mV/mA, so the first 10mA were absolutely lost for any practical use). Complained to the seller, he confirmed what I see, blamed it on 1/f noise since I was measuring DC (whatever...), then asked don't I want to try the "++" model (this one is "CP-07+") where this is taken care of. Knowing that I can't screw up more than this I said "sure, bring it on...". Yesterday the "++" came and it may be actually usable (horrendous PCB execution though, see bellow). Attaching a short description, a few pictures, and an oscilloscope snapshot of the thing measuring the current through a RC circuit driven by a square wave (will try to do better measurements later on, main FG broken at the moment).

(Advertised) Spec:

• AC/DC current Measurement
• Range: 1mA to 4A
• Bandwidth: DC~1MHz(±3db).

(Advertised) Accuracy:

Range	Accuracy	Sensitivity
400mA	±1.5%	1mV/1mA
4A	±1.5%	1mV/10mA

Exterior
Kinda OK. Not extremely solid or anything but perfectly usable. Wouldn't use it anywhere outdoors, but for the bench should be fine. Not sure how accurate they are but they did list reasonable IMO safety warnings and such. The battery compartment is kinda flimsy behind a self tapper screw.

Interior
Really bad, hand soldered, mixed through hole and surface mount PCB. More of a prototype level than a finished product. They seem to be constantly tweaking it, I think I know what they changed to produce the "++" model (I believe the two white wires coming from the "passive" side of the clamp weren't there on the "+" model, no pictures unfortunately from that one).

And a few questions if the analog gurus in this great forum can clarify:

• Can someone point me to the schematic of a similar device (google did not return much)? Especially interested in the sensor part. Looks like multiple coils around the half toroid there but can't see much behind all the glue. I think I've seen a similar ribbon cable with like 8 wires going into the clamp head on a Fluke clamp meter but that was a while ago.
• Again, sensor related - what are the two tiny rectangle things they've epoxied to the "active" side of the clamp between the two halves of the toroid core (placed question marks against those on the clamp head pucture)? If it measures DC it should have those Hall sensors, is this what this is?
• Are these guys onto something (forget the execution for a moment)? Is this spec even plausible? Haven't seen anything on the market even remotely close (the LeCroys and Agilents measure from 20/50mA? or so and have much narrower bandwidths, not to mention the prices). That same seller advertises another model as well http://bit.ly/XZhYB4, a bit different spec. Not an expert in any of this, this is the first current probe I use, need some education here I guess.

Thanks for reading...

[PA4TIM]

O my god, what a piece of junk. The soldering is done by a blind kid ?
The trace on your scope is real bad too. It is not much money for a current probe but much to much for this.
I made one myself once, using a 2,50 euro currentsensing element, it was only DC upto 500 KHz or so but it gave better response then I see here.

I have a Tek P6042 that starts at 1 mA/ div and does DC to 50 MHz. Also have some Tek AC probes. The secret there is the construction of the ferite/hallsensor. The flats are like mirrors and close seemless to each other. The schematics are on the internet. Jim Williams used one and designed an even beter frontend for the probe amp. It is in some application note. Can not recall itch one. You could build that and use the sensor of the one you have now. http://www.pa4tim.nl/?p=3135  schematic parts

The problem with DC/AC probes is the cross over point. The Hall sensor operates to about 50 KHz, it starts to preform less around there, the AC transformer has starts to operate a bit decent there.
The amplifier as to merge those. So you adjust it so the cross over is invisible.
I had the P6042 without the probe and found a probe from a TM5000 plugin. It took me a lot of work to merge those and to repair the P6042 but it was worth every minute. It is realy a joy to use.
The P6042 at work, the yellow trace is current. Here 100 mA/div the uppertrace is the gate of a load, the lower trace the current through the FET

P6016 currentprobe measuring 5 mA current, not a DC probe but this is 1KHz so almost DC ;-)

[chickenHeadKnob]

To Ivaylo: have a look at your .jpg 8 at the parts along the left edge of the board, Q8 an C8 and another cap, they look like they have no solder whatsoever. And those white leads through the coil spring would appear to be at risk of wearing through the insulation with use.
 Sorry I can't help with your questions, but thank you for bringing the teardown foto's they are interesting.

[ivaylo]

@pa4tim Yes, blind children is along the lines of what I thought when I first saw this. Thanks for the info and pictures!

@chickenHeadKnob Thanks for pointing those out! I wish they were the only trouble with that board...

[Smokey]

Interesting.  I had actually talked with that same seller before about this item.
http://www.ebay.com/itm/CP-05A-AC-DC-Clamp-Current-Probe-for-Oscilloscope-/110931036252?pt=LH_DefaultDomain_0&hash=item19d401485c

We use this LEM version at work, and it's pretty obvious that's what they are trying to copy.
http://www.tequipment.net/LEMPR30.html

Fluke sells that same probe now for more than LEM used to, which sucks, so I was looking for an alternative for home.  That seller told me it really did have 200kHz bandwidth (vs the LEM's 100kHz), which I didn't believe.  I got in contact with a bunch of people that had bought that same item from him and a couple got back to me with test results.  They all said the same thing you said, which was a lot of noise.  They also said the bandwidth was more like 2kHz not 200kHz.

As a side note, we also use some of these:
http://www.tequipment.net/LEMPR1030.html

Those are 100A/1000A vs the 30A of the PR30.  The 100A one outputs 10mV/A, and if you try to use that thing at lower currents around power switching PWM, the noise band is so bad you can't see anything without REALLY heavy averaging on the scope.  It picks up EMI like an antenna, probably because it's acting like, well, an antenna.  I took them both apart and the PR30 has a full metal shield inside the case while the 1030 is just the plastic case.  I've been meaning to stick some foil in there and see if it reduces the noise any.

If you don't know how these closed loop hall sensors work, you should really look it up.  It's pretty cool.  Check out page 9.
http://www.lem.com/images/stories/files/Products/1-3_applications/CH24101.pdf

[ivaylo]

Interesting. I had actually talked with that same seller before about this item.

How did it go, sounds like he answered your questions and everything? What I find amazing is that this wasn't a typical rip off (take your cash, send you garbage, and then try to convince you otherwise or just never reply). This guy went out of his way to see what my complaint was, measure and confirm my observations, send me confirmation screenshots (attached, bottom signal - old, noisier model, top signal - the 'improved' one), arrange replacement, etc. Then after he spent so much time over this I expected he'd remove that unit from eBay but no, it's still there. May be expecting it would be of use to someone? 

Check out page 9. http://www.lem.com/images/stories/files/Products/1-3_applications/CH24101.pdf

Thank you for that! Thanks to pa4tim's informative response above I also found this http://materias.fi.uba.ar/6644/info/puntas/manuales/p6042%20current%20probe.pdf. Very cool. That thing is copyrighted in 1967! You'd think by now they would be smaller, cheaper, better sensitivity/bandwidth, etc. Niche product I guess (althought with all the switching going on nowadays you'd think otherwise).

[amyk]

These do look handmade, probably they aren't produced in quantity enough to use automated soldering processes. The manufacturer might be making them just to gauge interest, hence the price.

[tesla500]

I did some debugging on the modern Tek version of this, or the amplifier at least, the TCPA300. Repair/operation video below. Also a look at an older version of these probes, also from Tek.