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

[jrf]

R24 is just a surge current limit to protect the op-amp U4-2. Good work on calculating why we need a 10sec press to get a good zero.
The 10M R269 may be another 'commissioning'  part. It may be a 'bypass' leakage current resistor for the op-amp & without it the capacitor may actually change. Size & location do not make sense.
Regardless if the zero offset is only +/-5mv then the drift is slow & workable, once the capacitor has been charged for a time. Unfortunately it discharges exponentially, so will drift faster initially, & not good for larger offsets. ie when magnetised.

A better way to Zero is as mentioned on my Schematic & as used on all original & many new clamps. ie a potentiometer.
VR7 was probably the original zero adjuster before the PB capacitor was added. ie simply add a wire from the wiper of VR7, replaced with a potentiometer with shaft,  to the U1-1 side of the Push Button, & remove/disable the PB. It may be possible with resistors to range adjust the new VR7 to be single turn.

john.

 

[CDaniel]

In theory with the cap discharged if you can tweek the pots for zero output and measuring right positive and negative currents  then you have the basic adjustments  right .
So far not so easy in my case because the 2 hall sensors are not perfectly equals

[dcac]

It might be an idea to experiment with the value of R24 (3.3K). I’m pretty sure it can be lowered to perhaps 1K or so for a faster setting time.

But it seems this circuit was mostly intended short time measurements. If you want stability over a longer period you probably, as already mentioned, need to remove/replace C4 and rely on the trim pots. And when measuring with a DMM you can then also use REL to clear the offset.

[CDaniel]

Not very sure how the pot could work without some modifications , since the circuit is made to zero the output with no input , that's why is momentarily connected . With the button pressed you can't measure , it is just showing zero , when released and the wire removed from clamp it will show that last current value as the new zero level .
If you don't want that function just remove R23 and use REL all the time .

[jrf]

CDaniel,
When balancing the halls, VR5-6, the idea is not to zero the output but to nullify the difference as the clamp is rotated about a post.
Any change within 2-5mv is probably reasonable.
Once adjusted then Zero, Calibrate the clamp.

Replace VR7 with a pot mounted for shaft access externally. Connect wiper to output of PB going to Zero Circuit Cap. Done!
As I mentioned you may use balancing resistors to change the voltage to/from the pot to achieve a good central zero position on average using a single turn pot. (actually one could just solder wires to either side of VR7 & a third, wiper, to the 470uF side of the PB & connect to an external 10K pot to test. Do not use the PB as will not work! Range +/-0.25V I think giving +/-25mv zero adjust.)
From memory the voltage across the Cap is 10x the zero offset voltage required. ie 10mv zero requires 100mv on cap.
VR7 ranges from +/-0.5V, by my calcs.

John.

[RodgerTheBadger]

John,

Many thanks, that makes sense now.

I suspect two variable resistors are used to try and keep the temp co. the same on each side?  If fixed resistors were used on one side and a trim on the other they would probably have different temperature coefficients, leading to inbalance with temperature change.  Or maybe it is just to keep the parts list down to make assembly cheaper.

Anyway, thank you for your reply


Rodger

[dcac]

Not very sure how the pot could work without some modifications , since the circuit is made to zero the output with no input , that's why is momentarily connected . With the button pressed you can't measure , it is just showing zero , when released and the wire removed from clamp it will show that last current value as the new zero level .
If you don't want that function just remove R23 and use REL all the time .

Keep in mind you can only use REL as long as the offset is not too high. i.e. if you have a 50000 count meter and want to measure in the 50mV range - but you have 20mV offset canceled out with REL - you can then only measure up to 30mV before you get out of range.

[Elasia]

Hmmm I've had one of these in a drawer for some years now... maybe i'll mod it... thing drifts so much its useless really less i want an instant measurement and only that one instant measurement..

[Elasia]

I slapped a 10uF chip cap right on the output to make it semi usable and cut the fuzz down

It still drifts to no end making it utterly useless for time based tests less that charge cap is modded out

[jrf]

Most of the time when I switch mine on at normal room temperature it is with in +/-1mV of Zero volts. (especially after tweaking VR3).
Hence no point using zero button. The button should never be used for timed tests.

This unit is not designed for 'measurements' below 50mA. It gives an indication below 50mA but not something I would trust, given the noise threshold of +/-10-15mA.

Drift is caused by temperature change, using the Zero PB, or physically moving the clamp.
So eliminating those issues will reduce drift to within +/-1mV.
Given that most long term measurements involve measuring the change then initial zero is not critical.
Adding an output filter for slow measurements to reduce the +/-10-15ma noise is also fine.

Remember this unit is not a precision instrument. It is a cheep automotive clamp mostly used to measure currents above 100mA or to show the relative timing of switching events. ie injectors ect...

Precision instrument amplifiers, circuitry & design layout COST & are hard to duplicate when the clamp heads core is included.

This serves the purpose for which is was designed very well & is the cheapest reasonable option I have found. Pico sell a unit, with what I believe is basically the same circuit, for 2-3 times the cost & one can spend 5x the cost on other units for no gain.

At about 10x the cost some units start to out perform it, but one needs to pay at least $700US to get a unit that may be designed to minimise the noise & achieve good accuracy, sensitivity & frequency response.  These units will usually have two separate but complimentary measuring systems for the DC & AC components.

One gets what one pays for has never been truer!

Cheers,
John.

[Elasia]

Part true minus the expensive alternative / cost bit

I actually do have the pico model I bought years ago before knowing they were on aliexpress, it is the same board, just a rebadge

The best current clamp I've seen for the money is micsig's CP2100A for about $200 and goes up to 800 KHz or CP2100B for about $400 and goes up to 2.5 MHz

These blow this design out of the water although it is still good for a dmm and single use / point testing and now is only $60 bucks or so

The micsig's only came out in the last 2 years or so while this design has been around for 5 years or so?

I've been using several CP2100B for multi phase power and they all work fantastic

[Martin72]

Plus the unbelivable building quality of the micsigs.
I don´t know how they managed it, but you wouldn´t believe they´re so cheap when you don´t know their prices.
800khz and 2.5Mhz....Look around, at which price you normally do get this.
Unbelievable.

[jrf]

Looks very good if you need the frequency. That seems to be the only significant positive. (can work around zero offsets)
Noise level is about the same, 4x cost & needs external 5V.
I have seen reviews indicating some of the better clamps having 1/10 the noise levels of this?

How does it handle magnetic fields???
Does it self magnetise on high DC currents, if NOT then a BIG plus!

Oh & the Hantek design is ancient! Looks like based on clamps from the 80-90's just using more modern components.

Cheers,
John.

[Elasia]

Martin could probly answer that, you can see the noise levels here

https://www.eevblog.com/forum/testgear/micsig-current-probe-cp2100b-tests-and-comparing/

[jrf]

Noise no better than CC-65 from that $400 unit. Much better Freq! though, but for a bit $ more I would hope the noise would be a bit less!

John

[Martin72]

If it´s so, congrats to hantek.

Here a comparison between much more expensive models like lecroy and tektronix, they´re "of course" better but not far away:

[Elasia]

Case of you get what you pay for.. mine is the noisy cricket lol.. have to use a large averager but that improved greatly after i slapped some chip caps on the output, still picks up on a lot of hfq noise but not as bad filtering it out

Mine also loves to drift off into space constantly, i'll slap a constant voltage via a trim pot voltage divider + fixed resistor next time i tinker with it

[jrf]

Thanks for those comparisons.  The best shot showing relative noise I could come up was from my Tektronix with Ch2 shorted & CH1 CC-65.

I have scaled both inputs to 5mV/div=50mA/div as per yours.
It puts into perspective my relative view. ie noise not an issue given my equipment. My USB scope shows less!
I have not 'zeroed' the clamp. Just switched it on, after powered off 24hr, & the recorded result.
No complaints.

Anyone with feedback on external fields/magnetic etc on other clamps? or magnetisation when using DC?

This seems to be the worst feature of a clamp to be used on cars, DC, or switched DC currents only!

Cheers,
John.

[Circlotron]

I bought mine to look at the primary current in car ignition coils. Typically (some exceptions) not much more than 10 amps, often lower. In that application the DC current goes to zero every switching cycle so you only have to adjust the scope so the zero line corresponds to a horizontal graticule line and you're done. Accuracy, drift, noise and frequency response is quite good enough for this, so in this instance it is fit for purpose and I am quite happy with it. I can understand why others may not be completely satisfied though.

[dcac]

I tried a spice simulation on the frequency response of the 272 op-amps. I guessed C6 to 1nF as this was the only value that gave a somewhat reasonable response in high gain mode:

EDIT: please disregard these traces - see next post.





But in low gain mode the response perhaps starts to drop of too quickly:

[dcac]

(lol) Well when I actually started looking at the caps values and the surrounding resistors it slowly hit me "100n" likely was just a 'place holder' for a sooner to be determined more accurate value.

So after some searching I found my CC-65 and decided to measure the caps 'in circuit' - using a technique with a signal gen, a variable series  resistor, a 10x amplifier and a DMM. The purpose of the amplifier was so I only needed to apply 25mV P-P from the signal gen and still be able to get an accurate reading on the DMM, the series resistor was set to match the resistance in parallel with the caps. This so I could measure the signal drop when probing the cap on the PCB. I could use 20Khz sine wave to get good readings on the DMM and i.e. swapping the probes over the caps only changed the reading about 10%.

So here are the measured/calculated caps values:
C5 = 2n2
C6 = 1n  (hey! I guessed it)
C8 = 2n2
C11 = 22n
C14 = 27n

I cannot swear by these values but they're likely better than ball park figures.

Spice Simulation to follow...

[dcac]

Here's the spice simulation - in high gain mode (6.5A) BW is about 23KHz - and in low gain (65A) about 34KHz.

High gain



Low gain


I have no spice model for the hall sensors so these are just simulated with sine generators with 500 ohms output impedance.

Do we know anything more specific about these sensors other than they appear to be in sot-143 size?

[dcac]

While I had my 10x amplifier handy I tried to measure the signal directly from the hall sensors - between pin 1-3 on HALL 1.

My amp can also be set to 100x and is based on OP37 so it has quite low noise and BW is about 250KHz.

The hall sensors seems to have allot of noise however - about 100uV P-P - but it wasn’t an easy measurement as CC-65 has no shielding to speak of - so I really had to fight with the 50Hz mains pickup.

It’s interesting there seem to be frequency peaking already at the hall input formed by R11//C13. When measuring 50mA RMS with the clamp I got:

500Hz = 120uV
1KHz = 123uV
5Khz  = 132uV
10Khz = 153uV
20Khz = 214uV

RMS from the hall sensors. This is then amplified about 50x by the 272s in high gain mode and about 5x in low gain.

[Elasia]

Yeah it has no shielding at all practically... picks up damn near anything, even the earth lol

[dcac]

I wonder why they chosen to frequency limit the 272s - well, they're limited with C5-C6 - but then peaking is applied with C8 and C14. I will try a simulation later without the caps.

Also why does U3-1-2 not have symmetrical feedback resistors R22 and R15 - one is 11K the other 7.5K.