Reward 150 USD input that help to repair my Advantest R6243 (ADCMT 6243)

[aronake]

Hi,

I have an Advantest R6234 with some errors. I pay a reward of 150 USD to whoever can give me the information to repair it. If many people provide valuable input  that least to a working unit I will split the reward as I see split of amount of contribution.

When I turn on the machine it fail with self test with error +336 VSVM 3.2 V +FS VSVM 3.2 V +FS.

Info here on EEVBLOG on this machine is mainly these two threads:

https://www.eevblog.com/forum/repair/adcmt-r6243-dc-voltage-current-sourcemonitor-error-337-mini-teardown/
https://www.eevblog.com/forum/testgear/advantest-r6243-teardown-and-repare/

According to these seems broken resistors on the board on top of the heatsink (lets call it High Voltage Board) is a common problem. I have checked all resistors hear and all are fine. I have also mapped out some parts of the machine in quite some detail, an a general overview of the whole machine.

What I have found out if unplug the cable to the High Voltage Board the machine make it until error +340 VSVM 32 V +FS VSVM 32 V +FS test error in the self test..

Attached photos of how generated voltages develop during self test. It seem that with the cable plugged in, the machine fail to reach 3.2V and that the board somehow disturb this. Without the cable, the machine can not generate the 32 volt needed.

Clearly the High Voltage Board is supsect.

Some questions to people who have some knowledge about Advantest R6234 that would help me in continued search for the error:

- What i call the High Voltage Board seems to have the purpose of generating higher voltages. But the machine anyway have voltage rails with much higher voltage than output. And I don't really see any of the components usually needed for generating higher voltages (transformer, SMPS controller, capacitors, etc). All voltage rails are lead to this board, but I can not find any control line (or are there). Only two output lines. Any though what this board is doing and how it is doing it?
- For the test point voltages on this board they are direct power rails from the main PSU and are unregulated, and real voltage much higher than marked (same as what people found in the other treads). But what voltage do the two output lines (VMT and VPT) supposed to be? Anyone know of could measure in their unit? I my case I see VPT 43.4v and VMT 43.5v. While doing self test they keep the same value. If this indeed is the high voltage board it is not much of high voltage with the machine able to output 100V. And I can not find any control lines to the card that would increase these voltages. So this is clearly further a suspected error source.
- I see a voltage drop in my machine as explained here in this thread: https://www.eevblog.com/forum/repair/adcmt-r6243-dc-voltage-current-sourcemonitor-error-337-mini-teardown/ In my case slightly lower. Rails are supposed to be 15V and -15V and mine are 12.5V and -12.5V after the voltage drop. Is that a fault with some short further down that rail or how a working Advantest R6234 do?

In meantime I will continue my search. But any help here would be appreciated.

[Swake]

Is there any (service) documentation available about this SMU? PDF's, schematics, manuals, pictures of the boards, measurement points, anything?

ps: It seems the model number is wrong in the title: R6243, not R6234

[HighVoltage]

Rails are supposed to be 15V and -15V and mine are 12.5V and -12.5V after the voltage drop. Is that a fault with some short further down that rail or how a working Advantest R6234 do?

In meantime I will continue my search. But any help here would be appreciated.

Advantest instruments are sometimes very difficult to repair, because there are no schematics or service manuals available. But if you invest the time and follow the circuit step by step, you might get success. I have repaired two Advantest instruments, but it took a long time.

However, it seems you have found the problem already.
The 15 V rail has to be 15 V maybe +/- 0.1 V but definitely not 12.5 V

Either your rail has a supply problem or something takes too much power from the rail.

Start with measuring the ripple on the rail and maybe you only have bad electrolytics or regulators.

Follow the rail and and disconnect different circuits and see how the rail voltages improve.

When this much power is drained from the rail, you can use a thermal camera and see what component is hot and has a shortage.

Good luck!

[gamalot]

I measured the voltages of test point VPT, VMT, and GNDF, all relative to GNDC. The output sweeps from -110V to 110V in steps of 1V, with no load or 100mA load. The results are placed in two spreadsheet files (actually there is almost no difference between them).

[gamalot]

I also measured the voltage of all test points relative to GNDC whit the output is set to 0V and enabled.

On my unit the +15VP is 14.9 and -15VP is -15.2. The mains power here is about 230V.

[aronake]

I managed to fix the main problem. After drawing schematics for quite some part of the meter and doing lots of measurements, showed out resistors R68 R175 was broken in open state.

It still seems to be a bit oversensitive toward overload conditions quite easily throwing an overload error, so will investigate this. But time to sleep now.

[aronake]

Rails are supposed to be 15V and -15V and mine are 12.5V and -12.5V after the voltage drop. Is that a fault with some short further down that rail or how a working Advantest R6234 do?

In meantime I will continue my search. But any help here would be appreciated.

Advantest instruments are sometimes very difficult to repair, because there are no schematics or service manuals available. But if you invest the time and follow the circuit step by step, you might get success. I have repaired two Advantest instruments, but it took a long time.

However, it seems you have found the problem already.
The 15 V rail has to be 15 V maybe +/- 0.1 V but definitely not 12.5 V

Either your rail has a supply problem or something takes too much power from the rail.

Start with measuring the ripple on the rail and maybe you only have bad electrolytics or regulators.

Follow the rail and and disconnect different circuits and see how the rail voltages improve.

When this much power is drained from the rail, you can use a thermal camera and see what component is hot and has a shortage.

Good luck!

The whole voltage drop happens over two resistors and no ripple, so power delivery seems ok. In very unjapanese fashion Advantest seems to have been pretty unacruate with voltage markings on some other test points. These voltages come from unregulated rails though. But for instance one 91V test have 120V. Gamalot did some test measurements posted here too so his is the same. 

[aronake]

Is there any (service) documentation available about this SMU? PDF's, schematics, manuals, pictures of the boards, measurement points, anything?

ps: It seems the model number is wrong in the title: R6243, not R6234

No service manual at all.

ah! thanks! title fixed!

[aronake]

I measured the voltages of test point VPT, VMT, and GNDF, all relative to GNDC. The output sweeps from -110V to 110V in steps of 1V, with no load or 100mA load. The results are placed in two spreadsheet files (actually there is almost no difference between them).

thats some serious measurements! Thanks!

Could you also check voltage between TP223 and +15VF, TP223 and -15VF. These are on the below board.

This "High Voltage" board is a rather odd things.

[aronake]

I managed to fix the main problem. After drawing schematics for quite some part of the meter and doing lots of measurements, showed out resistors R68 R175 was broken in open state.

It still seems to be a bit oversensitive toward overload conditions quite easily throwing an overload error, so will investigate this. But time to sleep now.

Quick test show that it then to go into overload above around 400 mA current both when sinking sources and when driving loads.

Solving this may be trickier than the fault that made self test not pass. But always good with a challenge!

[aronake]

Some more investigation of the overload problem. It happens at around 360 mA both when sourcing and sinking and voltage does not matter. The machine also refuses to be calibrated at highest current range, and when trying to calibrate full scale in highest range it is sending out 360 mA. Good thing that I then have some way of putting it close to failure when investigating by turning on calibration and error seems to be related to highest current range.

[gamalot]

I measured the voltages of test point VPT, VMT, and GNDF, all relative to GNDC. The output sweeps from -110V to 110V in steps of 1V, with no load or 100mA load. The results are placed in two spreadsheet files (actually there is almost no difference between them).

thats some serious measurements! Thanks!

Could you also check voltage between TP223 and +15VF, TP223 and -15VF. These are on the below board.

This "High Voltage" board is a rather odd things.

I can't find the +15VF and -15VF rails on the board, did you mean +15VP and -15VP, they are next to the TP223 and can easily be misread.

If so, their voltages are 10.4V and -10.7V relative to the TP223.

[aronake]

I measured the voltages of test point VPT, VMT, and GNDF, all relative to GNDC. The output sweeps from -110V to 110V in steps of 1V, with no load or 100mA load. The results are placed in two spreadsheet files (actually there is almost no difference between them).

thats some serious measurements! Thanks!

Could you also check voltage between TP223 and +15VF, TP223 and -15VF. These are on the below board.

This "High Voltage" board is a rather odd things.

I can't find the +15VF and -15VF rails on the board, did you mean +15VP and -15VP, they are next to the TP223 and can easily be misread.

If so, their voltages are 10.4V and -10.7V relative to the TP223.

Hi! Thanks! Yes, should be VP and not VF. Similar voltages to what i see. Pretty odd with such inaccurate marking of test points.

[HighVoltage]

I can't find the +15VF and -15VF rails on the board, did you mean +15VP and -15VP, they are next to the TP223 and can easily be misread.

If so, their voltages are 10.4V and -10.7V relative to the TP223.

I never had such huge difference in any of my many instrument repairs.
In one instrument the rail was only off by 600 mV and caused a problem.

Advantest instruments are sometimes weird.

[gamalot]

I can't find the +15VF and -15VF rails on the board, did you mean +15VP and -15VP, they are next to the TP223 and can easily be misread.

If so, their voltages are 10.4V and -10.7V relative to the TP223.

I never had such huge difference in any of my many instrument repairs.
In one instrument the rail was only off by 600 mV and caused a problem.

Advantest instruments are sometimes weird.

I believe these two power rails are connected to power rails with same names on another board via resistors in series (they are 280 ohms by my measurements).

[aronake]

I can't find the +15VF and -15VF rails on the board, did you mean +15VP and -15VP, they are next to the TP223 and can easily be misread.

If so, their voltages are 10.4V and -10.7V relative to the TP223.

I never had such huge difference in any of my many instrument repairs.
In one instrument the rail was only off by 600 mV and caused a problem.

Advantest instruments are sometimes weird.

I believe these two power rails are connected to power rails with same names on another board via resistors in series (they are 280 ohms by my measurements).

yes, I can confirm this. VERY odd design to put series resistor on a power rail and mark it with wrong voltage after the resistors....

[coromonadalix]

be sure you are on the good ground and not a floating one  ...

sometimes you can get wrong values if the real and float are not evenly matched ....

must be some current sensing circuit / sense resistor ...

[aronake]

Some more investigations on the second problem.

Overload error occurs quite exactly at 368 mA output.

A suspected area was the current range switching but that seems to be working well.

The "High Voltage" board have quite exactly same behaviour as the voltages gamalot measured.

I am gradually mapping out more and more of the schematics of the machine. Just took a deep dive in the guard voltage generation circuitry. Not that i thought that was a cause of error, but good to have it checked out.

I am now trying to figure out how the analog board control the amplifier part, to be able to run the amplifier without the analog board and control logic. That may give some hint on the problem.

[aronake]

I figured out that pin 7A and 7B between the analogue board and the amplifier board is the control signal from the analogue board to the amplifier board.

The signal increases to a new stable state for increased voltage and current. At 2mA from where overload happens, voltage is around 0.19V.  As current is 1mA from where overload happens it starts to be unregulated and float up and down. Once it reach around 9V Overload happens. So very likely the problem is that the amplifier can not generate the current it is supposed to generate so the control signal is increased to compensate for this, until the control signal is so high that the machine declare overload.

[fmashockie]

I had a very similar problem with a linear power supply of mine recently (an HP 3612A).  It would only limit to 60mA regardless of what the voltage or current limit was set to (it can normally allow max load 0.5A).  In my case, the problem was a shorted transistor in the series pass transistor circuit.  Not sure what kind of foldback current limiting circuit this PSU employs, but I would definitely take a look there.

[gamalot]

My another R6243 (yes I have two) has a problem that occasionally reports error "+344 VSVM 110V +FS". Because most of the time if I turn off the power and turn it on again, there is a very high chance that it will pass the self-test, so I have never tried to solve it.

I knew that this fault was caused by a certain resistor on the high voltage board, tonight I checked the chip resistors one by one from the positive voltage side of the board, and soon I found that R61 was dead (open circuit), and the rest was replace it with a new one.

[aronake]

My another R6243 (yes I have two) has a problem that occasionally reports error "+344 VSVM 110V +FS". Because most of the time if I turn off the power and turn it on again, there is a very high chance that it will pass the self-test, so I have never tried to solve it.

I knew that this fault was caused by a certain resistor on the high voltage board, tonight I checked the chip resistors one by one from the positive voltage side of the board, and soon I found that R61 was dead (open circuit), and the rest was replace it with a new one.

My R60, also on the picture was open circuit.

[aronake]

I was on vacation and now back to try to get this machine work.

I am quite sure now that the error is on the amplifier board. I am trying to trace the control signal from the analogue board (red line). This then go into some transistors, but strangely enough, I can not find any way where the control signal continue out of these transistors.

The tick traces on the bottom side of the board are power delivery.

Attached pictures of top of amplifier board and bottom of same board but mirrored image for easier tracking of traces.