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

[Fraser]

I just ordered a Hantek CC65 from AliExpress delivered to the UK for £40. That was the best price I could find as we now pay VAT

Hantek can produce some very competitively priced test equipment and I own a few of their TME products (USB DSO, USB  ARB FG, USB LA) but sadly it is the software side of things that often lets them down. Thankfully the CC65 does not need firmware or software to do its job 

Thank you to all those forum members who have helped to reverse engineer, analyze and improve this current clamp. Your efforts are much appreciated. I intend to use this current clamp with my new MICSIG ATO1102 DSO for both automotive and lab measurements. I repair thermal imaging cameras and you can learn a lot about a thermal camera by monitoring its current draw during start-up. This CC65 probe appears to be the best value for money solution on the market and, thanks to this thread, may be improved if desired.

My one disappointment with the design is the poor implementation of the Zeroing function. My higher current Beckman AC/DC clamp uses a conventional potentiometer adjustment which I prefer. This charging an electrolytic capacitor lark seems somewhat over complicated and flawed. I will see how my probe behaves when it arrives and may fit a manual zeroing control.

Regarding magnetisation of the clamp head, I have seen at least one automotive video where the tech using a similar looking clamp declared it in need of replacement due to a permanent current offset that was not zero’d out by the Zero button. Sadly I suspect many perfectly functional clamps go in the bin when all they need is a good degaussing. I would be interested to hear what others use to degauss their clamp heads. I think I have a tape head degausser somewhere. I note that in one review I saw of the CC65, the presenter thought the Zeroing button activated an internal degaussing circuit but we know the truth about that function now.

Before I ‘pulled the trigger’ on the CC-65 purchase, I bought a HEME LEM LTA 50P / SP1 AC/DC current transducer. It cost me £10 and offers 50A nominal operating AC current, 160A Max AC current and +/-50A DC current monitoring and measurement at DC to 100KHz. The unit offers both current and voltage output. Current is 1000:1 ratio and the voltage output is 100mV/1A. It seems like a neat little unit to have mounted on my lab power supply positive output cable to monitor DUT current draw behaviour on my DSO. LEM are known to make some decent Current sensors. I just need to set the sensor up with a +/-15V power supply. It will be interesting to compare the LTA 50P performance with that of the CC65 probe..... very different formats of sensor but each with its own applications.

Fraser

[rhb]

Hmm,

I just got one today and came wandering over here looking for more information.  From an initial inspection it appears to be eminently hackable. And in need of it ;-)

So far I have just skimmed this thread, but I'll more closely read backwards from here.  I observed a lot of flux on the unit, so I recommend a scrub with isopropyl.  There was no residue on mine on the other side and you *really* don't want to tangle with the switch contacts!

I don't understand the charge on a capacitor bit.  If it's just a reference voltage, the cheapest voltage reference and voltage divider would do  that.  It seems to me far better.

But I don't yet understand how it works in sufficient detail to build one from scratch.  So I'll stop until I know more.

Have Fun!
Reg

[dcac]

I replaced C4 with a 1000uF/10V high quality cap

Did you use a bipolar cap or still a polarized cap?

I see the original capacitor is polarized, which means the negative voltage range is very limited if we want leakage to remain low. However, if we go to a bipolar cap, we can live with an increased voltage over this cap,
which means we can reduce its weight in the circuit (increase value of R23). This will make it less sensitive to leakage of this capacitor. There's still the effort of finding a cap with low leakage though.
Edit: its possible to do the same trick with a polarized cap obviously, but then there a risk of the reverse voltage being large enough to cause increased current.
If we go too far we might need to increase R269 too, and then there's the question of how the leakage of the cap depends on the voltage over it when the applied voltage is so much lower than rated voltage

Sorry for the late reply, I used a regular polarized cap, I didn't actually think about possible reverse polarity when testing the replacement cap - only that it seemed to have much less leakage - or perhaps mainly less dielectric absorption like when a cap is acting a bit like a battery. The replacement cap seemed to settle much faster on the 'sample' voltage when the zero button was pressed. Even though it was twice the capacitance, 1000uF compared to 470uF for the original.

[toli]

Been while since I was able to find the time to play with this probe. I've had a free hour to do so tonight, so I looked into a few of the things that were of interest to me. First, the gain seems to be off by >10% on my unit (DC current measurement), so well outside the spec. Its easy enough to adjust, but somewhat disappointing non the less. I see there is also ~2% error between polarities, so I think it would also make sense to adjust the balance of the two sensors when I get to adjusting the gain.

What I was interested in for the moment was the LDO output voltage to check for any instability. I've observed it with the scope, and there was something there (at ~3.5KHz, although nothing too extreme), so I've connected to my audio measurement setup which is better suited for such low frequencies.
See 2 first attached images, first is time domain, second is the spectrum of that same signal. Notice this is with a 40dB gain so scale Y axis accordingly. Placing some decoupling at the input has improved it somewhat as can be seen in the 2 last images attached. The peak was reduced by ~7dB, and total RMS value at <100KHz is ~6dB lower (<200uVrms integrated noise up to 100KHz which is fairly typical figure for a low power LDO). Modifying the amount of capacitance further had negligible effect, as did modifying decoupling at the output and to the negative rail. This is a "free" improvement over what was there before the change, and nothing too extreme, but still something I might revisit when I have a chance. Perhaps replace with another LDO (one I actually know the part# for ) such as the LP2980/1. This tone at 3.5KHz can standout over the noise, and therefore I'd prefer to get rid of it completely if its cheap and quick.

I hope to get to the more interesting stuff of extending BW and reducing noise in the next few weeks. The shortage of parts is affecting this too, as some of the opamps I wanted to use cannot be obtained for a while now, with no solution in the near future. I will most likely go for plan B with other parts of lesser spec, or else this will probably never get done. Will try to keep posting when I have additional data.

[toli]

In case it will be of interest to anyone, I've posted on my blog about the measurements and modifications I did to my probe.
In the blog post(s) I go into much more detail as there's no one there to stop me , but I'll try to summarize this briefly here as well.

The changes included:
- Extension of -3dB BW to 1MHz. This can be stretched to a few MHz if desired, but I chose not to do so as for my needs it will result in increased noise with no practical use.
- Reduction on noise density. At 10KHz it is now ~12dB (X4) lower than it previously was.
- Removal of the tone observed at the supply (and output) of the probe. This was caused partially by instability of the LDO (which capacitor at its input fixed), and more dominantly by instability of the bias circuit which was also apparent at the LDO output due to limited load regulation. This was fixed by adding local negative feedback in the bias circuit by an additional capacitor, that has also reduced bias loop BW and noise.
- Reduction of minimum operating voltage to 6V by modifying the negative rail to have a 2.5V shunt regulator instead of the zener diode.
as well as a few other tweaks that are mostly nice to have.

I'm attaching a few images:
1 - Schematic I've implemented (with the exception of the part in green which I've selected not to implement, at least not for the moment). The green part is there to stabilize the offset drift over temperature, which I didn't see as a major drawback for the moment.
2 - Measured frequency response after the modifications for both ranges
3 - Transient response (with 4A pulse so it has some reasonable swing at the output, and not only small signal measurement) before mods
4 - Same as (3) after mods
5 - 1KHz 3mArms sine wave before mods (192KSPS so BW limited to ~90KHz, although probe BW limits this much earlier with 26KHz -3dB point before mods)
6 - Same as (5) with the probe after mods (far higher noise BW since now its limited to ~90KHz by the measurement setup)
7 - Same as (6), but this time with 48KSPS so BW is limited to <24KHz which is somewhat closer to the stock BW, just for comparison.

[Maciej]

I would like to better understand supply / bias of Hall sensor.
At first, opamp U1-2 seems to have positive feedback thru R1, R2, or just circuit drawing mistake?
As I understand the Hall is current driven. Hall input resistance is in this case the current sensing resistance, diode D1 provides reference.
Input resistance of Hall changes exponentialy with falling temperature, if we look at here:
" alt="" class="bbc_img" />
Would it be better to stabilise supply current with just another resistor?
I thing that bandwidth of U1-2 regulator should be nearly as low as temperature changes. It would help to cut off lot of noise. Just an idea.

[toli]

R1/R2 provide negative feedback, the PNP device adds its gain which is inverting.

Making the loop slower for reduced noise is indeed a valid point (but the question is where the noise is coming from, at the input to the amplifier, or due to the components in the current path at which case the BW should actually be higher to clean it). In this case you are indeed right, its better to make it slower, which is exactly what I did in my probe via the new 1uF capacitor which reduced noise significantly. It provides local negative feedback, stabilizing the loop, but also making the BW far lower in the process which improved noise further.

I'm not sure I understood your question about the change of bias circuit with just a resistor, do you mean remove the diode completely for zero tempco? Generally speaking, depending on the range of temperature of use and the sensor used, the temp-co of the biasing circuit will need to be adjusted. In this case we are only guessing the sensor (and therefore tempco), but typically VH starts falling as you increase temperature of the sensor above room temp, so increasing the bias voltage to offset for this is a reasonable way of compensating for this.
This one has a reasonably readable figure for this: https://www.nicera.co.jp/wordpress/wp-content/uploads/2019/06/NHE520Fe.pdf so if we assume this data is more or less representative, it actually makes sense to use a positive tempco for the bias voltage to offset the reduced signal. That is assuming we are at or above room temp.

[Maciej]

The biasing circuit looks like this:
" alt="" class="bbc_img" />
with R3 and R4 as Hall sensors. Voltage input is here stabilised at 2*Vref
Spatially and thermally both sensors are preety distanced.
Looking at ambient temperature curve for voltage driving "Vc Const" I see no chance to compensate it with linear Vref tempco.
" alt="" class="bbc_img" />
" alt="" class="bbc_img" />
Both sensor candidates NHE520 and HW101A look very similar is voltage mode. I cannot compare current mode "Ic const" because graphs are not to scale.
BTW. When using P-N junction tempco it is recommended to use BJT junction rather than standard true diode.

[toli]

Using the diode is a simple way to improve temperature stability over a limited temperature range where the signal drops with a given slope (above room temp if we trust the figures from the sensors we think are of interest). It needn't be perfect, simply better than nothing. Is using a constant voltage instead of one with this tempco better for the temperature range in question? No idea, to know for sure we need to know the exact spec of the sensor used, or to measure tempco ourselves.
Seems like some sort of such correction as in this one is included in other similar probes too, so without measuring this to know better, I think it makes sense to leave this tempco there.

[HardyKefes]

Hello, does anyone know whether the ESI 695 current probe is more accurate than the Hantek cc65?
Also, How is the TECPEL CA-60 AC/DC compared to the Hantek?

[Glowas]

Hi Guys,

I've decided to find all components and upgrade my CC-65 current probe. I have problem to get OPA1652AIDGK. Is anybody have proposal for direct substitute for this part ?
I thought about OPA1642AIDGKT but Im not sure  Many thanks for any advice.

Best regards,
Pawel

[apirltag]

OPA1678 is a fully compatible replacement.

[dcac]

I just noticed member MikePie had drawn in the differences between cc65 and the Peaktech 4250 clamp, found here: https://www.eevblog.com/forum/testgear/peaktech-4250-acdc-current-clamp-teardown-and-repair/msg4565725/#msg4565725

And as I always thought the circuit for the zero button on cc65 was a bit odd and not really configured as a regular “sample and hold" like the Peaktech 4250 seems to be. I tried to implement the same circuit on cc65 by removing the 10M resistor and but in a 220n filmcapacitor instead of the bulky 470uF electrolyte.

But this just resulted in a huge drift and the clamp could not hold zero for more than a couple of seconds. I’d already checked the U1-1 272C op amp datasheets from a few different manufactures and they suggested input bias to be something like 2 to 6 pico-amp. So 220n capacitor should be able to withstand this for quite some time without much change to the “sampled” voltage level. But when I measured the output on U1-1 the voltage was dropping very quickly. So ruling out possible faulty capacitor and measuring the resistance on U1-1 input to ‘infinity’. The only thing left was the 272C being faulty - but with the original circuit configuration with the 470uF cap it worked as it should.

So removing the cap again and measured the input bias, and yeah.. about 1800 pico-amp. So I’m thinking these 272C are fake or very bad quality. Anyhow... this can explain why the cc65 is prone to drift - more than it should - and why they used this huge 470u electrolytic capacitor, that has problems like leakage and dielectric absorption. The Peaktech zero circuit is probably much more stable using a filmcapacitor and probably genuine 272C op amps.

 
           

[Fraser]

An interesting discovery. Thank you 

Are you intending to fit genuine/correct quality 272C op-amps in your unit now, or just leave it as manufactured, with the large electrolytic capacitor ?

Fraser

[dcac]

Yes, I think I have a couple of tlc2272 laying around, somewhere. They should not only be a direct replacement but also a noticeable improvement to the 272. Both lower noise and lower input bias current, only 1pA, so they would be worth a try when I find them.