Hantek CC-65 AC/DC Current Probe Teardown and Testing

[Mechatrommer]

not on the pcb, i marked earlier only to indicate the connection to it. hall effect sensor is to sense magnetic field so it must be near that metal core loop, so thats what you see, there are 2 of them as in the picture, nothing more.

[maxspb69]

Maybe useful for someone...
https://yadi.sk/i/qGA6hIVvvBYOPg

[macboy]

The other HP clamp that I have is colsed loop with inductor and hall. I don´t know how this clamp works. There´s more hall sensors in PCB?

Your HP clamp is probably a much better instrument than this. Hall sensors have limited bandwidth; they work from DC to some 10's or maybe 100's of kHz. In contrast, a traditional current transformer can be designed to have 10's MHz bandwidth, but can't give output at DC.  A high bandwidth AC/DC current clamp will combine both techniques, and use other interesting methods to give good measurements. For one thing, the core materials necessary for high bandwidth will saturate at a relatively low DC value, and will distort the waveform even with a DC bias lower than saturation. For that reason, those probes will use the Hall sensor output to drive a DC current into the sense coil (or another coil), to offset the magnetic flux created by the DC current. It will always try to drive the (low frequency component of) Hall sensor output to zero. Then, the magnitude of this "correction" current (not the Hall sensor output) is measured to give DC/LF current measurement, and is combined with the output from the coil for the AC/HF components, giving an overall AC+DC measurement. This gives a good DC current measurement linearity, even if the Hall sensor is non-linear, it prevents saturation of the core, it helps prevents DC current from magnetizing the core thus causing hysteresis errors, and it prevents DC bias from distorting the AC portion of the waveform.  It's also why a "real" AC/DC current probe is so expensive.

[KlausF]

Огромное вам Спасибо. я очень рад.

[KlausF]

Maybe useful for someone...
https://yadi.sk/i/qGA6hIVvvBYOPg

Огромное вам Спасибо. я очень рад.

[mrprecision]

Hantek Current Probe Frequency Response:

[jrf]

Thanks for the Schematic.
I have been looking at options for a clamp & all the cheap ones look to use almost the same circuit, with the same issues, so I wanted one with a circuit I could modify &/or repair.
To this end I hunted for info & from partial schematics, photos, & descriptions I have prepared the attached schematic with notes.

This is not 100% accurate.

I have ordered a clamp but it is still 4-6 weeks away.
If anyone has any suggestions or corrections please help!

Cheers,
John.

[jrf]

Did not see my attachment in last post so will try it this way.

 cc-65-CLAMP.pdf (21.08 kB - downloaded 664 times.)

Cheers,
John.

[belzrebuth]

Isn't this able to perform better with some circuit change or something or it's at the limit of the topology used?
I'm slightly surprised this hasn't be done yet, given how many people have it and its target price.

I mean even exchanging some parts for more expensive ones should still render this tool a great value for hobbyists.
Since it's analog there shouldn't be much in terms of tedious reversing and they must be some known good designs that the hantek could be adapted to.
I know what I say is vague and I admit I don't know much if anything about current probes but I'd expect this would be a nice hacking candidate for knowledgeable people.

[jrf]

If you look at CC-65-clamp.pdf, available above, you could try replicating part of the circuit U3-1&2 & U4-1 on a breadboard. Feed it power & signal from the clamp. By changing U3 & U4 to amps with 10x the bandwidth you may be able to determine the upper limit of the Hall sensors. I have seen others reach 100KHz with similar sensors. BUT to go further best use an AC clamp. With modern cro's just adding the output from AC & Hantek would give the complete picture. This is the approach used in early model AC/DC MHz range probes. Not sure what is used now.  Remember the material used for the core of the clamp will also affect its response. AC ones are likely to start at 20KHz but go to a GigHz if you spend enough! & thats the point. If you spend a lot of time & energy you may be able to push this to 100KHz but it will use more battery power & it all depends on how much more usefull that is! Also  other clamps like this from PICO etc are using basically the same circuit. It is probably an obsolete design from the 80's just using more modern  components built on price/performance compromise.

Also note the capacitors are just a guess. I actually suspect the small ones to be closer to 0.01uF rather than 0.1uF but as I am still waiting on my clamp from AliEXPRESS or Ali-snailmail (7 weeks & yet to leave China!- & can't complain till 100 days!) I can not measure them or some other parts of the circuit to confirm accuracy.

Good luck,
john

[CDaniel]

I ordered one too ... AC bandwidth is anyway much better than an usual clampmeter ~400Hz - 1KHz
It should be easy to test what is limiting the bandwidth by replacing the hall outputs with the signal generator ... But the quality of the core is very important so I don't think the circuit is the limiting factor .

[belzrebuth]

I was also going to order one; that's how I found this thread but I'm kind of reluctant since I was going to use it for taking screenshots of current plots directly from my scope.
Since I want the waveforms to be as accurate/nice as possible this is just not going to cut it unfortunately.

Maybe I would be better off with plotting software than this.

See here:
https://youtu.be/uZTicR_eC9A?t=521

I'm not sure if the triangular waveform would actually look like a triangle if frequency was lower..
If signal is okay for say a frequency of 10k that's fine enough for me at least.
Perhaps someone who has this probe could post some screenshots in different frequency ranges.

[CDaniel]

This is usefull only when you can't interrupt the circuit , otherwise a proper shunt will give you all the bandwidth ...
As I see it , it is suitable for field measurements with a portable oscilloscope , not in the lab .

[Noy]

Which Opamps are used within the clamp?

[belzrebuth]

Which Opamps are used within the clamp?

I believe it's mentioned in this video
TLC272 & TL062

[Martin72]

Hi,

Had the CC-65 too, it´s behaviour in general was worser than the owon cp-05, which I got too (still have it).
The noisefloor kills it, probably this should be only used by multimeter.
Also the offset correction wasn´t stable.
The owon costs about 170€ incl. tax and is usable from 1A/div upwards, because of the noisefloor, the hantek in my memory too.
I wouldn´t recommending both, but the hantek is so cheap, it could be worth trying to improve it´s behaviour.

But the quality of the core is very important so I don't think the circuit is the limiting factor .

Remember the tektronix current probes.
With their external, large supply/amplifier, the circuit does matter too.
It´s got current compensating measuring method, so in theory, the core would not get into saturation.
(Current compensation: A winding is on the core, the hall sensor detect the current flowing through the core and feed it into a regular circuit which controls the compensation winding on the core.
A compensating current will be generated which rises up until the current through the core is zero, which means that the core couldn´t get in saturation - this compensating current will be forwarded to the measure output.
If 1A is flowing through the core, then compensating current will be 1A - And this is, what you measure.
Also this method allows you to do easily degaussing.)
But the saturation also depends from the frequency of the current, so the material of the core is mostly not a single one, it´s a mix, an compound of several ones to get as high as possible.

Current Clamps like the hantek or owon or many others, even when they costs 1000, 2000 bucks, don´t have it at all, what current compensating/ degaussing concerns.
What you can do is to reduce the noisefloor, maybe shifting a little bit the bandwith.

[belzrebuth]

Had the CC-65 too, it´s behaviour in general was worser than the owon cp-05, which I got too (still have it).
The noisefloor kills it, probably this should be only used by multimeter.

Have you tried increasing/adding capacitors as described in this thread?

[Martin72]

To be honest, no. Got it, test it, give it away...
Must read the thread completely, what the capacitors concerns.

[jrf]

CC-65-CLAMP.pdf above shows one capacitor solution as described in this blog, labled NEW caps in the power supply.
These caps are specified by the voltage regulator manufacturer but where not fitted, from what I can understand.
As stated on CC-65 CLAMP.pdf it should be possible to put in a multi-turn trim pot to replace the Zero push button. Then so long as the unit is at 'temperature' & under steady state conditions the zero should be reasonably stable, unlike a charged capacitor zero circuit! Pots where used on all the name-brand clamps & I suspect the push button was used to simplify the unit for rapid use in non-tech fields, copying its use in meters. Unfortunately meters are able to use a computer to do Zero, & do not drift compared to this solution!

I have read here about the coils used in some clamps. On reading a repair blog I found a 10MHz unit said the coil actually measured AC & the Hall the DC component. This makes sense. A coil could also be used to 'zero' the hall reading to reduce DC-saturation, but does nothing for AC saturation. The photos on this blog clearly show steel laminations. The frequency response of these would probably affect the linearity of the clamp at higher frequencies.

The op-amps used & the current circuit will only do as specified. Op-amps with 10x the frequency response are available but with typically higher current draw or more $$.

John.

[Noy]

Do somebody know pincompatible better ones (Opamp)?

[jrf]

Most dual opamps are pin compatible & available in that case size.
Important things you need to look at:
Operation Voltage range. ie same or better than existing.
Frequency response. 10x existing.
most of the other specs equivalent or better.
Availability.

Changing just the op-amps may not change the frequency response of the overall circuit as some of the capacitors in the circuit may act as filters as well.  Breadboard a circuit & experiment with what needs to be changed or just look up a 'standard' instrument amp circuit with the desired frequency response & modify to suit this applictaion.  Capacitors are often added for compensation & can depend on the opamp used.

John.

[belzrebuth]

@jrf
Could you do a basic test when you receive the unit and post a couple of screenshots here?
Maybe post something like a square into a load and the current rising and falling at half its stated BW and close to 20k?
I'd very much like to know when things start to break up.
With other words I wonder what could be the max frequency at which you can get a pretty looking waveform before discontinuities or other artifacts appear.

[jrf]

Look at the posts on this blog for that! Lots of photos of outputs.
10KHz bandwidth means that a 10KHz sign wave is at 3db down (1/2 expected voltage) at 10KHz.
So a 1KHz square wave, which is made up of 1KHz, 3KHz, 5KHz, 7KHz, 9khz etc sign waves,  will be mostly square.
But by 10KHz all you get is a mostly 1/2 voltage 10KHz sign wave.

John.

[jrf]

Well Aliexpress have delivered my Clamp & ,after a few quick measurements, the circuit is as I drew.
Scematic -  CC-65 CLAMP1.pdf (20.94 kB - downloaded 411 times.)
I have updated the voltage regulator values.

I shorted the 470uF zero cap for 30sec & then powered on & the output was 2mv at 1mV/10mA.
I then zeroed the unit & the cap chaged to ~10mv to offset this to zero.
2mV zero error is what I would expect. The temperature of my unit was ~20'c.
I would expect up to +/- 5mV based on temperature fluctuations & existing mag fields.
I turned off the unit for 2 hours & then back on (the capacitor charge should not be affected by the power switch pos).
It then read 1mV. So 1mV drift. (equiv to 10mA.) Very good.
Noise was ~1mV p-p without any mods. Have yet to test adding caps. I think 1mVp-p is reasonable!

I also put a magnet near the head. Big reading.
So if you short the Zero Cap & get a large value (zero offset) then the head may need demagnetising.
My Fluke is unafected by external mag fields!
If demagnetised & still high zero offset then may need to adjust Zero - step 6. VR3-4

Checking the calibration, mine looks to read 10-20% high!
Only tried DC to 2A. Will try AC & if the same error will calibrate to full scale on DC on both ranges.
(as described in the SChematic attached. Last section only, ie no need to adjust hall sensors, so just Steps 7 & 8.)

Further testing/calibration tomorrow.

Cheers,
John.

[Martin72]

Do you know the name of the op-amps ?