Peaktech 4250 AC/DC current clamp teardown and repair.

[The Soulman]

I've owned this clamp now for a while but never was satisfied with the low current readings there was to much noise and drift to be useful.
After reading here:  https://www.eevblog.com/forum/testgear/hantek-cc-65-acdc-current-probe-teardown-and-testing/
about insufficient decoupling in the Hantek CC-65 I decided to open up my (similair but not exactly the same) Peaktech.
And discovered there weren't any electrolytic capacitors fitted at all! 
Fitted some 33uf 35V caps because that is what I had and it functions much better now. 
Anyone has an idea about who could be the OEM of these clamps and what each of those trim pots does?

[The Soulman]

More pics..

[The Soulman]

Caps fitted..

[schmitt trigger]

I have that exact same probe!
Between today and tomorrow I will open it up and see if it also has the same capacitors missing.

To your question, I see a pair of probable answers:
-cost savings. When a company sells these so inexpensively, every cent counts!
-manufacturing error. The board is a mixed mode SMT/TH component assembly. Perhaps the TH components are hand fitted, and they forgot installing the electrolytics.

[The Soulman]

I have that exact same probe!
Between today and tomorrow I will open it up and see if it also has the same capacitors missing.

To your question, I see a pair of probable answers:
-cost savings. When a company sells these so inexpensively, every cent counts!
-manufacturing error. The board is a mixed mode SMT/TH component assembly. Perhaps the TH components are hand fitted, and they forgot installing the electrolytics.

Cool, yeah my idea is that the guy or gall responsible for fitting them came in late that day or had a long lunch break...

[mcinque]

Maybe Hantek? I found that current clamp under various brands (Pico for example) and two different versions (oscilloscope and dmm).

In my experience, Peaktech only rebrands instruments and sell them at higher prices, without ANY added value.

[schmitt trigger]

I believe this is a trend for many of the lower cost test instruments.

I've a differential probe which I found that Dave had made a teardown analysis video.

However, Dave's probe brand and model are different from mine, although inside they are essentially identical.

[technogeeky]

I have two questions:

* is there any reasoning for the exact value of the electrolytics?
* is there any reason to use some other type of capacitor for noise purposes? MLCC? tantalum? etc?

[The Soulman]

I have two questions:

* is there any reasoning for the exact value of the electrolytics?
* is there any reason to use some other type of capacitor for noise purposes? MLCC? tantalum? etc?

I used what I had that would fit (2,5mm spacing, 5mm diameter, 12mm height) for me that happened to be 22uf 35V Yageo "low-esr".
Just to see what would happen, I liked what happened so left them in. 
If the design is reverse engineered than each capacitor could be optimized, perhaps off-set and bandwidth performance could also be improved.

[schmitt trigger]

Mystery solved!
Finally on this weekend I had some spare time to open my probe.

I found that those empty capacitor holes are in parallel with the SMT tantalums nearby.

As a fact, one can read C7 and C7A, C1 and C1A.............meaning that you could fit either a SMT or TH component on the board depending on availability.

[technogeeky]

Mystery solved!
Finally on this weekend I had some spare time to open my probe.

I found that those empty capacitor holes are in parallel with the SMT tantalums nearby.

As a fact, one can read C7 and C7A, C1 and C1A.............meaning that you could fit either a SMT or TH component on the board depending on availability.

To me, that really begs the question of why they don't include more capacitance. So the through hole electrolytic capacitor spots are just alternates for the tantalums they included. But surely if they included larger tantalums, the noise would be lower as this modification suggests.

I really don't understand.

[SeanB]

Larger value tantalums would be expensive, and possibly not usable in the PNP process they have due to size or mass, so they used either ones on the machine already or the largest value they could reliably place instead. Going for the TH part and hand soldering would be more expensive as well, and using SM electrolytics would probably not work either as they would have the same mass and height limits that were there for the other caps already. In any case 22uF added to the existing 10uf and 22uf ones would be a good improvement.

[The Soulman]

Mystery solved!
Finally on this weekend I had some spare time to open my probe.

I found that those empty capacitor holes are in parallel with the SMT tantalums nearby.

As a fact, one can read C7 and C7A, C1 and C1A.............meaning that you could fit either a SMT or TH component on the board depending on availability.

Looking at the pics it is actually quite obvious,    when time comes I will try to draw a schematic of this thing, gotta know what each of those pots do..

[mqsaharan]

I replaced the caps C1A, C7A and C9A with double values as suggested in this thread without checking or even thinking what I was doing. The power supply ripple remained there. I didn't bother to actually measure it but it was substantial. Then I soldered a 6.8uF capacitor at the input of voltage regulator U5 and all the ripple was gone. This was the only small value capacitor I had at hand that did the trick. The inputs of U5 in this circuit are the negative pad of C7A or pin 1 of U5 and the tab or middle pin of the U5 is the positive input.
I would suggest to install around 10uF cap before replacing or adding to the suggested (C1A, C7A and C9A) capacitors in order to get rid of the ripple in the power supply. I didn't replace the caps again and left these double value caps installed.
I haven't used this probe too much. But the little testing that I have done, it is good enough for my hobby work.

Edit: I forgot to mention that the schematic, forum member jrf, posted in that Hantek CC-65 thread (also attached here) is almost the same for this probe with difference in some components values.

Edit2: Added the picture and schematic.

[grumpee]

Old thread, but thank you.  I added a 10uF electrolytic capacitor in parallel with a 12nF ceramic on the input of regulator U5 as described above and it completely got rid of the low frequency ripple on the output.  The attached images show the output ripple before and after the mod.  I'm sorry I didn't record it, but when I measured the voltage on the input of that regulator, the ripple was quite high (several hundred mV I think) and was definitely causing the output to ripple at the same frequency (~590Hz).

Cheers!

[MikePie]

Thank you to all for your most helpful inputs.
First thing I did on my just-received PeakTech 4250 was to check its output ripple.
Sure enough, there was this 550Hz ripple. Since the only cap I had on hand which would fit into the available space was a 1000uF/6V3, I did not want to add this across U5 input in order to not expose D4 to the resulting switch-on inrush current peak. Instead, I soldered it in parallel to D4, since I thought that this negative supply could well benefit from additional bypassing anyway. I soldered its (+) lead into the thru-hole of C7 (-) mounting position, and its (-) lead to D4 Vb- pad.
After that, the ripple was gone.
I also replaced the coiled output cable with its non-retracting banana plugs with a coaxial BNC cable of the same diameter (which I took from a defective pH probe). This made the broadband output noise amplitude a little less sensitive to handling and also reduced stray signal pickup.
I added some red-marking to the schematic diagram which was posted by mqsaharan (Jul 07, 2020) to show differences I found on my probe, and the added capacitor. The PCB in my probe is Rev.E; there has been some relayout around the ZERO switch SW1, apparently to accommodate a switch with a different footprint, relative to the Rev.D PCB which is shown in above posts.
I attach the scanned schematic drawing. It includes the measured current consumption (9.6mA) and the measured Vb- voltage.
I also checked the squarewave response using a 50mA peak, 10kHz signal, and the zero stability of about +/-3mA during 10 minutes (Plots attached).
Overall, I am quite happy with this current probe.

[schmitt trigger]

I have modified my own probe, adding a 10 uF cap, and can confirm that IT WORKS!!
The ripple disappeared.

Thanks everyone for their contributions. Special thanks to Mike Pie for taking the time to draw a schematic.

[MikePie]

I updated the schematic with some more differences I found. Thought I might as well share it.

On the less sensitive range, I noticed large spikes at fast transitions as if the load was capacitive (which it was not, just simple carbon resistors).
I decided to reduce the bandwidth by adding a 2n2 ceramic capacitor between the wipers of VR5 and VR6. This I did not show in the schematic.
BW is now about 150kHz on the 100mA/mV range, closer to the about 50kHz of the 10mA/mV range. Before, it was about 500kHz, but apparently with artefacts.
The four plots show this (with/without 2n2, 100mA and 10mA range).

[pgodefroy]

There was indeed this 550Hz ripple.
I soldered a 1000uF/6V3 capacitor parallel to D4: I soldered its (+) lead into the thru-hole of C7 (-) mounting position, and its (-) lead to D4 Vb- pad.
After that, the ripple was gone.