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

[jrf]

Try reading the posts above or looking at the PDF schematic attached above.
ie TLC272 & TL062 as marked on drawing.

JRF

[jrf]

Further testing of CC-65:
AC: 50Hz
Dodgy is the best way to describe the performance of the clamp.  At 65A it worked well. But at low currents ie <6A it was poor. The Zero offset was high & took a long time to stabilise. ie Zero of 400mA common.
DC:
At low currents accuracy +/- 10%. ie <2Amps.
Above this accuracy improved. At 60Amps it was good.
BUT 60A DC magnetisd the head. So when removed it still read 2A.
Reversing the current direction through the meter re-zeroed it.
I also put it back on 60A AC to make sure. (I used a Microwave transformer converted to a spot welder pluged into a variac to generate the current. I could do 1000A if I had a big enough variac & was fast enough.) Search for info on DIY spot welders.
I used a Fluke clamp meter & multimeter to compare. The Fluke uses more modern systems & is not affected by magnetic fields etc. It indicates down to 10ma but drifts up to 50ma over 10min. So the CC-65 is more sensitive & stable at low readings so long as not phisically moved.

Verdict: It does the job it was designed to do. ie Automotive testing.
It is not a accurate measuring instrument. It is an indicator. It shows what is going on.
As such I am happy. For $72AU it does the job I need.
Looking at what else is available I would need to pay 10-20X that to get something better enough to make a difference.

Cheers,
John

[mqsaharan]

Hi jrf,
Thanks for the schematic. I think the 3.3k resistor attached to Zero button and 470uF capacitor marked as R34 is actually R24 and I also have doubts about the R269, its assigned number as well as its value. Please reconfirm it before making any changes.

I don't own this clamp. I have Peaktech 4250, instead. And the schematic is almost the same with difference in some component values.
The power supply ripple was horrible at the output of its voltage regulator U5. I did replace C1A, C7A and C9A with double values but the ripple remained there. Then I installed a 6.8uF capacitor at the input of U5 and all the ripple was gone.
I didn't install 100nF capacitors that you suggested in your schematic. The little testing that I have done with this probe, I think these are not needed.

Regards,
Qasim.

[jrf]

Agreed R34 is actually R24, Typo. Thanks! Good to see others interested enough to check & let me know.
R269 is on the bottom of the board & clearly marked on the photos I have.
Colours Brown-black-black-green---brown(I think).  Looks like 1-2W.  2 other schematics here show it but one says 100M.
(will not dismantle again, now, as the power switch can be a pain if the wipers fall off!)
Not sure why it is there, under the board as well as in the circuit, or the size!??
Often large capacitors have dischage resistors to make them safe to work on some time after the power is removed or for ballancing but neither apply in this circuit.

I agree adding 100nF after the reg is actually in parrallel with C9, so is unlikely to do much. Missed that!. C1 has nothing to do with the regulator ripple & C7 is large already.
By putting a capacitor on the input to the reg it implies a resistance in the battery supply circuit. ie dirty switch contacts? or poor battery, but that is often why capacitors are put there! ie my second new 100nF. May be undersized!. Did you put it to battery -ve or circuit gnd? Not impressed with the power/selector switch at all! Designed to give problems!.

Cheers,
John.

Cheers,
John.

[Martin72]

Here is the schematic I´ve drawn from my owon cp-05+ (200Khz, 4/40/400A, appx 170€):

https://www.eevblog.com/forum/testgear/pintek-pa622-current-probe-crap-or-just-cheap/msg2220903/#msg2220903

It seems, it´s nearly the same design regardless of the price.

[mqsaharan]

Agreed R34 is actually R24, Typo. Thanks! Good to see others interested enough to check & let me know.
R269 is on the bottom of the board & clearly marked on the photos I have.
Colours Brown-black-black-green---brown(I think).  Looks like 1-2W.  2 other schematics here show it but one says 100M.
(will not dismantle again, now, as the power switch can be a pain if the wipers fall off!)
Not sure why it is there, under the board as well as in the circuit, or the size!??
Often large capacitors have dischage resistors to make them safe to work on some time after the power is removed or for ballancing but neither apply in this circuit.

I agree adding 100nF after the reg is actually in parrallel with C9, so is unlikely to do much. Missed that!. C1 has nothing to do with the regulator ripple & C7 is large already.
By putting a capacitor on the input to the reg it implies a resistance in the battery supply circuit. ie dirty switch contacts? or poor battery, but that is often why capacitors are put there! ie my second new 100nF. May be undersized!. Did you put it to battery -ve or circuit gnd? Not impressed with the power/selector switch at all! Designed to give problems!.

Cheers,
John.

Cheers,
John.

Hi John,
I am sorry I missed that through hole resistor. I do have seen those pictures but I forgot it is on the other side of the board.
As I mentioned before my current probe is Peaktech 4250. It does not have any component on the other side of the board except for two LEDs, the physical Zero switch and on/off switch contacts. I have attached pictures here borrowed from the other thread (https://www.eevblog.com/forum/testgear/peaktech-4250-acdc-current-clamp-teardown-and-repair/).

What I am calling ripple is actually oscillations due to lack of a capacitor at the regulator input terminals. The one that is in Peaktech seems to be unstable without the input capacitor. I tried 100n first, then 1uF but these didn't make much difference. The next up with me was 6.8uF and it did the trick.
I added the capacitor at the input of U5 and circuit ground. And it removed all the oscillations and regulator started giving clean output.

Regards,
Qasim.

[mqsaharan]

Here is the schematic I´ve drawn from my owon cp-05+ (200Khz, 4/40/400A, appx 170€):

https://www.eevblog.com/forum/testgear/pintek-pa622-current-probe-crap-or-just-cheap/msg2220903/#msg2220903

It seems, it´s nearly the same design regardless of the price.

Hi Martin,
Thanks for the review of CP05. Some pictures from your forum thread are not showing up.
Did you happen to test the probe with increased supply voltage? Did it perform any better?
You are right. MC33078 does need 5 volts minimum dual supply. I don't know why the designer couldn't find any other part that can work at 3v like other op amps used in this clamp. Or why didn't he made sure of the proper supply voltage for this particular part.

Regards,
Qasim.

[Martin72]

Hi Qasim,

No, increasing the supply didn´t take any effect  - Except that I´d burned out the hall sensors...
Fortunately they´ve send me a second one without any questions...

Some pictures from your forum thread are not showing up.

Hm, I still can se all of them (Browser: Edge)....

Martin

[mqsaharan]

Hi Quasim,

No, increasing the supply didn´t take any effect  - Except that I´d burned out the hall sensors...
Fortunately they´ve send me a second one without any questions...

Some pictures from your forum thread are not showing up.

Hm, I still can se all of them (Browser: Edge)....

Martin

My bad. Never mind. I forgot to check the rest of the pages of your thread.
Pictures on page 1 where you announced that you've received the CP-05 are missing as well as in the post after the next post. But its all right.
I am using Firefox on Ubuntu.

I just read that you have already checked with higher voltages with no difference in performance. This begs the question, have the designer made some arrangement to make up for the supply of MC33078 or what trick he used to make this op amp work at lower voltages. Or perhaps this op amp can work at lower voltages with minor decrease in performance.

Anyway, good luck with your current probe project.

Qasim.

[jrf]

Martin72,
If you want to test at 5v then you need to modify the voltage regulation circuit feeding the halls. This is normally temperature compensated at about 0.8v. ie maintain -3v at R11-10k or change the 10k to maintain the same voltage at the halls. Can't see any temp comp in that circuit. Any noise on that 3V will be reflected to a degree in the readings. Although your halls supply is +/- the same voltage. A bit of common mode rejection? C4 may cause a low frequency ripple, depending on size, if any noise on -3v there. Cap is required as all circuits have it.
Yes most of these circuits using 4 pin halls only will be very similar. Looks like yours also has a auto zero at power on?
I also would like to have a spec on the halls used!!!
How is yours affected by large DC currents? ie magnetized? Zero shift?

Qasim,
Thanks. Will look at that cap as low dropout regs are known for needing low impedance supplies & oscillation is common. Better to have & not need than have random issues. More testing rqd...

Cheers,
John

[jrf]

Well did some testing today. Very interesting. Much better results on the scope than on the multimeter.
My CC-65 does not seem to suffer the issues others have experienced. ie Zero error is minimal & as such drifts slowly.
It is important not to move the clamp afer zeroing!
Also there is no need for extra capacitors on mine. Noise level changes little when power is applied to meter.

I logged some results using Sign waves & square from 1KHz to 50KHz using about +/- 100mA.

10mA was the smallest current waveform that showed up on the scope. As such I am very happy.

Cheers,
John

[Martin72]

This is what I would expect from a clamp, only showing it´s limited bandwith and not such a crap like the cp-05 does:

https://www.eevblog.com/forum/testgear/siglent-sds2000x-plus-coming/msg3125996/#msg3125996

There must be something going in resonance or whatever.
Nevertheless, the hantek beats it for more than half of the price.

[jrf]

Martin72,
I have seen similar, just can't remember when/why. Will ask around. It rings bells! That certainly looks to come from a filter. Quite a high Q in your frequency annalysis. May give good single frequency sine wave responce but hopeless for square waves!!.
Possibly an active filter & would certainly not help the noise responce of the clamp. ie ampifying it! In reality, when you look at the phase lag of CP05 & CC-65 with the CC-65 90' at about 40-50KHz compared to 25KHz for CP-05.
I believe my CC-65 uses the output coax as part of the filter circuit. Filtering out much of the HF noise present when looking at the meter signals directly with a 10x probe.
Also I failed to mention the Scope I used for the pictures was a LOTO-OSC482M. This USB unit works with a PC & Android. $120AU.
And the RMS current readings from the scope agreed with my 4 digit RMS meter within unit accuracy. Impressed. So my earlier comments about AC readings with a meter do not apply when using a osc.

So Conclusion:
Pros:
My CC-65 works up to 10KHz with a bandwidth of about 45KHz.
Its zero is stable & gives accurate readings from 30-50ma & up.
Noise is within acceptable limits. Only slightly higher than background.
Cheapest low current OSC clamp I could find at $72AU.
Does the job it was designed for as advertised.

Cons:
High DC currents will magnetise the head, resulting in a large zero offset. (but still read that current correctly!)
This can be reversed by reversing the current through the clamp or using a large AC current. ie I was testing at 60A AC & DC & zero moved ~20mv on 1mv/100mA scale.
Leaving the clamp for a few hours should give a zero offset with discharged capacitor. If > 5mv or so then the head is possibly still magnetised,picking up stray fields in close metal or the meter was poorly zeroed inititally.
The meter picks up magnetic fields & should not be moved once zeroed prior to taking measurements.
Need to hold the zero button for several seconds, 5-10, to stabilise the capacitor charge.

Conclusion: Worth the money if it meets your spec.

Happy,
John.

[Martin72]

On the first (defective) one I began to desolder all the components for having a free look to the printed circuit, therefore I´ve measured most of the caps.

It makes me wonder if there is not a fault with your clamp.

Yes, this is my opinion too - When I got time to investigate, I´ll do it in the "right" thread then.

Martin

[Martin72]

Transport it to the CP-05 thread:

https://www.eevblog.com/forum/testgear/pintek-pa622-current-probe-crap-or-just-cheap/msg3128474/#msg3128474

[jrf]

This is my 'final' schematic with corrections & updated voltage readings.
Schematic: -  CC-65 CLAMP2.pdf (21.33 kB - downloaded 379 times.)
Also the capacitors marked 100nF are 'unknown' small brown ones! 0.5-100nF range.
I had always assumed they where all the same as the same colour but may be not!
Reading indicates they could be anything!!! What a mess! One way of making life difficult for repairers!

Cheers,
John.

[jrf]

Here is the Schematic with extra notes after some playing with it & experiments.
 CC-65 CLAMP3.pdf (22.24 kB - downloaded 405 times.)

NOTE: I have not measured any of the small ceramic capacitors. Brown or white.
My meter works fine so I will not be desoldering them to find out!

From photos & comments on the net I am doubtful there is any reliable better cheap option.
Cheap < 300US. ie you can pay a lot more than this one ($50US) for little to no gain.

Some use 3V batteries. This implies they use switch mode power supplies. Most likely NOISY!

I have also noted physical copies of name brand units (ie fluke/tektronix), often with different specs!
Regardless I doubt they come close to the originals.  They may even have the same electronics but are unlikely to have the same core & sensor head.
Some clamps are copies of copies & don't even use correctly rated components. So not only do they not meet the specs they have poor performance overall & give wildly inaccurate readings at different frequencies. BEWARE!

So, Best of Luck,
John.

[RodgerTheBadger]

Hello.

Firstly thank you for the work on the CC-65 clamp schematic, it has proven to be both interesting and informative.

I was just wondering if you could explain the variable resitors VR5 and VR6.  You mention that they are used to adjust fr reading changes when rotating the clamp.  Is this due to the earth's magnetic field?  If so why is there a change in reading if the magnetic field is constant?  I was wondering if it is because the two hall effect sensors ahve different sentitivty?  But even then wouldn't they just everage out?

Any help would be appreciated as I can work out the rest of the circuit now

Thanks,

Rodger

[CDaniel]

I received my probe ... so I hooked up the signal generator at the sensors output . Indeed , the circuit is limiting the max frequency , it is perfectly linear up to 10KHz . Of course if the amplitude is low the bandwidth could be even worse .

[jrf]

Rodger,
Good Question!
RE: VR5-6 function.
Circuit:
Hall sensors used: 4 wire - ~1V-grd power on 2 & +/- signal on the other 2.
Signal +ve of one & -ve of other go to either side of VR5, & the others to VR6.
The halls are mounted at the ends of the C cores forming the clamp such that a field running say clockwise around the core creates a positive output in both. ie If both mounted at each end of one core then one will be upside down to the other.
They are trying to only measure the field created in the core by the current flowing through it.
Common mode fields, ie fields in the environment passing through both sensors are undesirable & need to be cancelled out.
Ideally both sensors have similar characteristics & just summing their outputs equally in opposite directions will do the trick, hence most photos of meters will show the pots VR5-6 centrally located. ie everything was within specs... As are mine.
If rotating the meter 180' about ANY axis in free air, away from all metal, creates any large reading differences then the 2 halls are not balanced, or there is a circuit fault.  Mine changes by about 2mV. Not enough to worry about given the fluctuations due to temperature & close magnetic fields when taking measurements are similar. ie large >10mv say.
I do not know how they are factory calibrated & can see that it would be difficult to get a 'perfect' balance here.
Other clamps & clamp meters with similar Halls use similar circuits & include the 'balancing' VR's. Some only one. I suspect 2 may be overkill.
Another observation is why the C11 capacitor on one of the 4 signal wire resistors & not the others? Suspect it made an improvement during commissioning. Possibly not 'designed' as such.

Anyhow I hope this gives you a better understanding of the circuit function & remember to make sure your clamp is not magnetised before starting ANY adjustments.  This is my one significant complaint with all these 'cheap' clamps. A strong DC current will magnetise the core! My zero moved 20mV. Reversing the current, in a similar way to its initial application, will mostly remove it, but appling a large AC current & then winding it slowly to zero will work better. A 'degauss' circuit from an old CRT screen does a good job as well, from what I have seen on google. My Fluke clampmeter is unaffected by magnets or fields but does not have a CRO output... Similar sensitivity though.

Good Luck,
John.

[CDaniel]

I don't have a Fluke DC clampmeter , but from videos I see that the behaviour could be better , but in essence the same , drift , magnetising effect if you pass a large current . That's why there is a zero offset button .
Yea , the idea of using 2 sensors should cancel out common magnetic fields like earth magnetic field ... for some reason not all core magnetising can be canceled this way
If the sensors are not balanced right than + and - currents are not the same ( putting the wire in the opposite way through the clamp ) , so if you adjust the pots for this than should be good for any common mode fields .

[jrf]

Any magnetising of the core, within the core, will pass a common mode field through both sensors! ie a field circulating around within the core.
Hence the zero offset.
A large external field passes through the core & sensors with minimal circulating effect, passing through the sensors in differential mode, & hopefully being cancelled out. ie like parallel lines through the whole area of the clamp.
If the Hall sensors are not flat in the same plane then this field will give different signals from the sensors & be more difficult if not impossible to 'balance' out.
Small fields like a magnet close to the clamp will read more in one sensor than the other, hence giving a reading.

The description of the orientation of the sensors in my previous essay above may not be 100% but the logic & function is correct.

I avoid using the zero Push Button as it will result in reading drift with time. My USB oscilloscope has zero offset ability on its inputs so I use that.
If the offset >5mv I can just trim VR3 to zero it. A magnetised head will result in large zero offsets & significant drift if the Push button is used to 'correct' it. The push button is a poor replacement for the potentiometer, especially as it hides a magnetised head unless you short the 470uF cap or leave the unit for many hours to discharge completely. When I switch mine on it is usually within 2mV of zero. It changes 2-3mV depending on orientation & proximity to metal.

Cheers,
john.

[CDaniel]

Doesn't seems logic at all to replace a pot with a push button for some bias voltage and achive good results ... more likely the voltage/current small pulse when you press the button and the output of the op amp is "high" is somehow reseting the sensors  ( not very logic either way )
The cap is discharged anyway by the 10M resistor and internal leakage relatively fast , so the output of the amp should go slowly back to 0V . Why would you want this effect if this a bias

[dcac]

The "sample and hold" circuit used for zeroing has rather strange time constants and a I don't think I've seen an electrolytic timing capacitor before, usually you'd use a low leakage film capacitor for that.

When Zero button is pressed C4 is charged through R24, so 3.3K x 470u = 1.55s - so that could explain needing to hold the button a few seconds to achieve zero - remember you need 5x that time to charge the capacitor to >99%.

And the discharge time through R269, 10M x 470u = 4700s or about 1.5h, so a (slow) offset drift is unavoidable - unless you were lucky and C4 was charged to 0 volts. But then leakage and dielectric absorption in C4 can still cause offset drift.

[CDaniel]

OK , that OUT on the inverting input is like ground and fooled me 
Should be a better way of compensating the drift of this amplifier