LeCroy LC684DL displays random signals [success]

[Cat]

Hello,

I got a LeCroy LC684DL which is only showing random signals at all four channels  (see attached images and video at https://drive.google.com/open?id=0B1j-dj5uKJrYS19VcDBYUGdEbUk ).
Channel 2 and 4 have some "flickering" bits, Channel 1 and 3 remain mostly static. It does not change when the coupling is changed and even showing when "grounded" or fed with the Cal.-signal. The Amplitude does not change when gain is changed, however the pattern can be moved horizontally when the delay time is changed.
The Pattern of the random signals did not change the last two years (got this scope two years ago, but had no ESD protected area to disassemble it).

Do you have any hint what could be wrong or even had a similar behaviour?

The previous owner told me it had gone bad from one day to another. Since it was already old, they just purchased a new one and dumped this one. Inside it was full of dust, which maybe affected cooling or was slightly conductive. I cleaned everything and reseated the RAM modules (in case something corroded) but it did not change anything.

KO4BB has the schematics and service manual, unfortunately this scope seems not to be designed for real repair. Measurements at the main board can not be done in working condition because the cover blocks most access. Anyway, J11 at the controller board has a few supply voltages.
From the manual:                               
Vcc : + 5.00V ( Min = + 4.9 V, Max = + 5.10 V )        Measured: 4,97 V       
Vee : - 5.1 V ( Min = - 5.05 V, Max = - 5.15 V )    Measured: -5,19 V
+15 : +14.95 V ( Min = +14.90 V, Max = +15.05V )    Measured: 15,03 V
- 15 : - 14.95 V ( Min = -14.90 V, Max = -15.05 V )    Measured: -14,97 V
Since only Vee seems slightly out of range and I don't know how accurate my multimeter is, I didn't touch the pots at the power supply. No AC ripple could be measured.
Any advice how to measure the voltages at the main board while it's in the scope?

Since the pattern can be moved I would suspect the error near the DAC or RAM. However it is strange that all four channels failed like this.
What would you say?

My next approach would be cleaning every RAM connector with IPA and also removing and cleaning the DACs.

Regards
Cat

[Cat]

Some teardown images...
They seem to have reworked a GAL on the processor board. Also there was some sticky liquid inside the LCD connector (maybe from reworking?).

[Cat]

Mainboard in original condition...

[Cat]

The mainboard cleaned with a vacuum cleaner and compressed air.

[nctnico]

I'd start by looking for something which is common for all channels (sample clock for example). It can be a poor contact or a cracked component. If you have another scope you can check the power supply rails for execessive ripple. A DMM in AC mode could hint towards that as well if it isn't limited to 50Hz. My first wild guess would be no data goes into the ram and you are seeing the random patterns in the RAM after power up.

[Cat]

The hunt continues...
Yesterday I opened the powersupply to check the capacitors. They seem fine (altough I'll change them once I get the sope working again).
After my initial doubt how to measure the voltages at the mainboard I tried to assemble it without lower cover... and succeed.
Then I could sleep peacefully and think about how to find the fault.

[Cat]

A DMM in AC mode could hint towards that as well if it isn't limited to 50Hz.

I checked my DMM, but it seems only good up to 400 Hz.

I'd start by looking for something which is common for all channels (sample clock for example). It can be a poor contact or a cracked component.

Today I prepared everything to get some measurements. Installed a fan for cooling and got my old HAMEG 312 (10 MHz Scope) and a thermal camera ready. Then I turned the scope on...

After a few seconds the maximum temperature was above the limit (150°C). I turned it off to find the component. It was the voltage regulator A117 (LM79L12) which was getting hot.
Also a transistor at the regulator build around A6 (LM324) was getting warm. The schematic says it's powering the frontend with 1,35A from 5,8V to 5V. That's ~1W so it should be OK.

The glowing red hot VR A117 is powering the ADC with -12V. Between the Output and GND I measured 0,5 Ohms, so there is a short somewhere. I suspect a tantalum capacitor but first need to study the schematic further.

regards
Cat

[nctnico]

If you supply 1V at 1 amp or so you should be able to find the culprit with a DMM. The lower the voltage drop the closer you are to the short. But a thermal camera should show the problem maker instantly.

[Cat]

The culprit is found!
I checked the schematics and A117 is only powering four OP-Amps, 1  Ta. cap and 6 MLCCs.
I first removed the tantalum, but nothing changed. Then i lifted the output pin from A117 and removed the 0,1µ MLCC . Still nothing.
Then I removed the MLCC at A112 (AD705)...and got 37kOhm between -12VADC and GND! With the voltage regulator connected it dropped to 7kOhm.
Now I need to get some replacement capacitors...

If you supply 1V at 1 amp or so you should be able to find the culprit with a DMM. The lower the voltage drop the closer you are to the short. But a thermal camera should show the problem maker instantly.

Unfortunately I don't have a laboratory supply here right now. But with this techniques I got some faults earlier. Before I got the termal Imager I also measured voltage drop to find the fault. Unfortunately once there was a 100µ-trace which happily burned away (I think it was at 2 Amps).  Now with the termal imager much lower current is needed to find faults.

[Ice-Tea]

Nice thread and some good fault hunting! 

[Cat]

Thank you very much for your support.
After desoldering the caps I replaced them, but omitted the tantalum one. I only got THT Ta.-Caps which are probably 30 years old and of unknown origin.
Then I plugged it in and it worked!
The regulators stay cool, the scope is showing the Calibration trace.

Voltages Measured @ TP100:
+2V:   1,998 V
+5V:   5,00 V
+12V:  11,94 V
-4,5V: -4,46 V
-5V:   -4,95 V
-12V:  -11,94 V(the broken one)

Now I'll build the scope together again, I'm afraid I'll fry something if it is not cooled properly.
The calibration signal looks a bit noisy, but maybe it gets better with the shielding.
The glowing 78L12 might be replaced in near future (read: never )

I'll be right back...
Cat

[Cat]

The Scope is now completed and still working 
Only the internal printer is not shown in the settings...
Let's see how long it lasts until the next failure.

Regards
Cat

[scopeman]

Hi Cat,

Sam here moderator of the LeCroy Yahoo! Group. Sorry I missed your post but been only on line once in a while lately. However great catch on the fix!

On the internal printer there is a large electrolytic (usually a blue Philips one about 4000uF) tied to +5V on the printer board with the printer mechanics. Replace it even if it tests okay. These are know to fail. Common across all LeCroy series scopes that use this printer.

Sam
W3OHM