XJW01 LCR-Bridge, need help for repair

[Gio@EEVBlog]

Hello,
I came across this very informative forum while looking for a cheap LCR tester. I decided on the xjw01 and bought it. At first it seemed to work. However, after a relatively short time it stopped showing stable values, i.e. the values ​​displayed kept changing even though the same component was being measured. Originally, short and open calibrations could be carried out. Now short calibration no longer works. Regardless of whether the probes are connected to each other or not, an open calibration is always carried out. Sometimes it shows "open calibration.." while probes are connected and at the end "short calibration...OK", which is absurd. To rule out errors outside the circuit board, I connected Lc/Lp and Hp/Hc directly on the circuit board. Sometimes, when the device was switched off for a long time, short calibration seems to work. However, only a few times. After that, the error appears as described. Even if this short calibration worked once, the measured values ​​are not really stable or reliable. Damage caused by electrolytic capacitors that have not been discharged can be ruled out. Now I don't know whether I should send the device back, which I don't really want to do because I waited a long time for the delivery from China and would have to repeat the problem when I buy a new one, or whether it can be repaired relatively easily. Perhaps someone has already had such an error and repaired it. I could imagine that a TL082 (unlike in the circuit https://www.eevblog.com/forum/testgear/xjw01-auto-lcr-meter-review-($120-bench-top-lcr-meter)/ my board has TL081 installed), TL084 or CD4052 has a (thermal) defect.
Thank you in advance for your help

[coromonadalix]

thread plus schematics of it
https://www.eevblog.com/forum/testgear/xjw01-auto-lcr-meter-review-($120-bench-top-lcr-meter)/

[Gio@EEVBlog]

Thanks, I know the thread (I also linked it in my post above). Very interesting, but unfortunately I didn't find any answers to my questions there.

[coromonadalix]

well you have to check many input parts,  even the diode  etc ...  bad solder, bad contacts, op amps   up to the relay,  if you are sure  nothing charged blew up things    return it 


for the pricing   an DE-5000  would have been way better than this one     yeah i know  you have may or will receive the same comments

the open and short calibration is important to null out things, probes resistance  etc ..  even some kelvin ones  can be problematic,  quality is not there on cheap stuff,  try other probes

i don't know if the fw  received some updates,  worth to check,  or any patch on the pcb ?


for other answers    you'll have to be patient  ... if some can help or want to help

[floobydust]

There are a few different builds of the XJW-01. Which one do you have? Any pics would help.

What does it read in resistance mode?

You just bought it? I would check for bad soldering or solder balls inside etc.
Next check the PSU voltage +/-5V is OK. Some wall adapters are quite noisy and common-mode noise can be a problem depending on your version of the board.
The input clamp diodes, input op-amp and sometimes the CD4052 mux can get damaged from charged caps - but you say that has never happened.
There is a relay for the output excitation, it might have poor contacts.

[BillyO]

I don't think the "real" XJW01 is available anymore.  There seem to to be many cheap imitations available on AliExpress.  Not that all stuff on AliExpress is totally useless, but some certainly is.  Who knows what schematic they are following with your model.

As others have suggested, do a careful visual inspection for bad soldering and good power.  Failing that, the input op-amps, genuine ones, are cheap.  I'd buy some and replace the existing ones.

If that doesn't get it working your in for some serious trouble shooting.  That might not be possible without the correct schematic and some experience with how the utilized circuitry works.  If you can get teh right schematic and you understand how the design works, you'll need a decent DMM and scope to find the fault.

[Gio@EEVBlog]

Hello,

first of all, thank you very much to all of you for giving me  support.

It is the following version:

PCB REV. 2.0
Circuit REV. 3.1
DATE 2020/05

See attached photo.

I also found the corresponding circuit diagram here in the forum. It is with TL081 and the component position numbers match.

I did a visual inspection of the soldering points with a magnifying glass but could not find anything unusual.

I had desoldered the buzzer so that I could hear the relay clicking. It obviously switches, but I did not check whether the contacts close properly.

A problem with the sockets or probes can be ruled out because I made the short directly on the circuit board.

A problem with the power supply can also be ruled out. With 12V via a laboratory power supply (Toellner 8742) I measured with Fluke 89IV: +5V: 5.0583V and -5V: 5.0564V on the board.
The input clamp diodes are OK. I measured a breakdown voltage of 0.599 V (with diode tester in Fluke 89IV).
The TL081 actually look "cheap".
I think I'll replace the TL081 first as recommended. I would like to read the firmware first. Can I use the adapter according to the attached circuit diagram and would the AIAPP-ISP tool be suitable for this?
Sorry for the possibly bad translation from German and so many questions.

[floobydust]

I can't see the eight clamp diodes on the PCB, are they on the backside? Older PCB revs (1.1) they were there on the top. If these are missing then that is a big problem.

I think you need a scope to dig in and see where the signal problem is. A multimeter on ACV can work up to a point.

I would check for about 0.18V sine-wave coming out of U106, also can be seen at the two (+) output BNC's.
You could put a test resistor in, and then look at what the input op-amp is doing U104 and U105 outputs.
Pretty much signal tracing is what I would do.

edit: check soldering on R152, it looks way too crooked.

[Gio@EEVBlog]

I replaced the two TL081C and the short calibration works again. The measured values ​​are also quite accurate. Nevertheless, I wanted to calibrate the device according to the "XJW01 LCR Bridge Manual". There are 19 steps. In the steps R3x, G1x and G2x, the corresponding resistors should be used to adjust to 0 pF.

Unfortunately, that doesn't work because the display is not stable. It changes about 3 times per second between small (also negative) pF values ​​up to values ​​of over 100 Henry. No idea if that's supposed to be the case. I don't have any other device to compare it to. The phase shift of -90.04° for the measured capacitor with 185pF (+-1pF) also seems strange to me. The frequencies also deviate from the target values ​​and the peak-to-peak voltage is 400 mV (with 10:1 probe, the scope has not been calibrated for a long time, I also checked the frequencies with a frequency counter). Here, too, I don't know if this is normal.
The clamp diodes are on the other side. The resistor R152 is a bit crooked but the soldering is OK.
Anyway. For my purposes, it is sufficient as it is working now. But I would still be grateful for further information. In particular on calibration and how the software can be read from the STC, as I asked before, so that it can be replaced if necessary.

[floobydust]

I have my old notes from when I was doing calibration. R3x, R4x did not seem to do anything.

If your readings are noisy, I would suspect the power supply (PE ground) might be needed or there is a problem with perhaps another IC.
You can follow the signal with the scope and see where it goes noisy.

I have not worked with this MCU, STC 12C5A60S2. Seems to be an enhanced 8051, for the people.
I thought originally the XJW01 was open-source, until its creator (professor) got pissed off people were manufacturing them and making money, so he stopped.
I have .hex files and a bunch of others that are old.  Altium files, source code but it has been years since I was in it.
PCB, silkscreen "XJW01, Putian  Program REV: 5.5, LCR Meter Rev:5.5, DATE: 2016/11/27

[daisizhou]

As far as I know, this is an electronic LCR production project from a middle school teacher.
I remember it had a special frequency of 7.8kHZ，And if you want to get 100khz you need to email the author,And provide MCU code

It is an open source project, and there are schematics, PCB production diagrams, and source code on the Internet.

Cheap knockoffs are sold on the Internet,Especially some capacitors and op amps are low-quality components.The original author hopes that everyone can make it by themselves, but it may cost more money.
The fun and sense of accomplishment of making things by yourself cannot be bought with money.

[Gio@EEVBlog]

The capacitors look like (fake) Sprague "orange drops capacitors". Since the display is not stable, I would change all the capacitors. Which types should be used (KS, KP, MKP, FKP...) and what maximum tolerance should they have?

[floobydust]

Those orange film caps - you can take one apart if you think it's fake. Only a few are critical. I think they are a copy of old Siemens parts and I've had no issues with them.

Did you try to just PE ground the case/BNC and see if readings are affected?
I cannot find details on how the Toellner PSU output is, as far as common-mode noise, if they have Y-caps on the output etc.
I think you'd have to poke around the XJW01 with a scope and look for the noise.

The XJW01 power supply section was revised a few times. It was originally an 8-0-8VAC mains transformer then changed to single 12VDC wall-wart power.
To make the split +/-5V rails, an artificial (0V) rail-splitter is used.  Possibly your U302 and is that working decent in the presence of CM noise?

[Gio@EEVBlog]

Sorry for the late reply. I don't think it's noise from the power supply. The display changes regularly 3 times in 1 second. When the XJW01 is operated with the 12VDC wall-wart power adapter, interference is superimposed on the signal at the terminals, so that the sine signal can hardly be triggered (1st image). In contrast, it is clean when operated with the Toellner laboratory power supply (The Toellner is a high-quality linearly regulated power supply) (2nd image). Incidentally, both the 12VDC wall-wart power adapter and the Toellner PSU are connected to the mains via an isolation transformer. I have attached two videos (as zip-file) here. One on the 12VDC wall-wart power adapter (1st Video) and one on the Toellner (2nd Video). The same behavior can be observed in both cases.