Fluke 5520A - Err 104

[Analogdevice]

Hello everyone,

I have problem with my calibrator 5520A. Someone just connect something to 20AMPS banana which causes some fire at A7 board (burned PPTC's, surge arrester, some OPAmps etc.). I will repair it later, still waiting for mouser package with spare parts.
Meanwhile I checked other functionalities with this result:
Resistance - OK
Capacitance - OK
Temp (RTD, TC) - OK
Voltage DC - NOK (description below)
Voltage AC - NOK (description below).


Voltage have 2 different problems:
at voltage DC, works perfectly up to 33V range. After 35V or more is dialed, I have Error 104 - hardware relay trip occurred;
at voltage AC (f 50Hz) at every range I have Error 1503 - output current lim exceeded.

Looks like common failure. I have schematics from 5500, it's similar but there are some changes, e.g. 65V+/- supply have zener diodes connected to SCOM, but in 5520A they are connected somewhere I don't know. First issue is that -65V is -80V. I checked all transistors, zener diodes and everything in regulator, but everything is ok there. PCB have more than 2 layers so it's little annoying to follow the route. That's the reason why I need your help. Is there someone, who know something about circuitry in 5520A? Or someone, who had the same issue?

I guess that the problem is at A8 (voltage) board. DDS looks OK - but there may be circuitry for measure current in voltage mode for overloads. Sometimes after pressing operate in ranges >33V calibrator restart itself 

Any idea, note, hint, word, will be very helpful. Thank you so much!

[Analogdevice]

Sooo, first update. +/-65V supplies are OK by now.
There was improvement against 5500 in 65V supplies with AD822 opamp at regulation part, which was dead. I replaced it by LT1013 and ordered AD822.

Good news are that errors still occurs. But first step is done, I'm going to next step.

[RYcal]

Following as I have seen this before....And never found a fix.

[Analogdevice]

So I'm here with some news.
With little help of 5500 schematics, I definitely determined source of fault - DDS CH0 and CH1 are not working correctly. Suspected IC's are D/A converters, flip-flops, some of opamps and of course CPLD (inguard CPU). At this point I will try to replace D/A, which may take about 1 week (I must order it from mouser).

Where problem is: voltage error out comes from sense buffer circuitry to first DA (common for DDS CH0 - Voltage channel), to the CH2 of DA to reference (DDS have overall 4 DA channels - 2 IC's, each have 2 channels). Output of this 2nd channel comes to next DA to reference input, which generate output to error amp. Because first DA is not working correctly, there is no correct voltage (especially when A8 board tun on voltage transformer) and this voltage trips the relays (it generates FAULT signal, which stops the program cycle and hardwarw reset all relays).

So that's all yet, I will try to fix this problem and then A7 current board is waiting with several burned components. I'm so excited.

[Analogdevice]

Update #2...

I replaced D/A for DDS CH0 and CH1. Still not working, but it's more "cultured". Err 104 appears only when 1kV is dialed, otherwise we have Err 1503 (output current limit exceeded). When e.g. 35V is dialed, we have 0,6V at the output.
On DDS D/A converters we have N/A state at WR input (for write input must be logic 0 to write, it's negated input). This signal comes from CPLD. But there's a little chance. Both D/A's not working and use separate CPLD output, so there's smaller chance that CPLD is kaput. Both WR commands are in N/A state (CH0 is between 0 and 1, CH1 is high impedance).

At the side of CPLD we have CMOS SRAM (AS7C4098 from Alliance) routed to CPLD. I'm guessing that this is wavetable RAM, where waves are stored for the DDS.
Symptoms:
 - I/O's are at high impedance state (because of 3-state logic)
 - RAM configuration is in read mode (according to the truth table from datasheet)
 - row and column selections are working
 - after power-up instrument, CPLD sent words to I/O's to be stored
 - every I/O is at HIz state, so may be common failure
 - this may be irrelevant, but it was soldered, there are stains from flux (but boards are cleaned, when comes from factory).

According to internal logic diagram of SRAM it looks like I/O buffer is not working correctly (or is not working). The biggest problem is that I don't have schematic and 5500 was different, there was more components, but at 5520 looks like it was reduced to CPLD. And I'm guessing that CPLD doesn't need RAM for working - besides that RAM is at HIz state and program is working correctly, so this must be wavetable RAM...

What do you think? May I be right? Anyway, I'm ordering new SRAM from mouser so next update will be here in about 1 week.

Let me point out that every idea, opinion, hint etc. is still welcome. Just don't be shy, friends 

[Analogdevice]

Meanwhile I do some "meterings".

This MUST be wavetable RAM, because after powering UP, on display it shows "loading waveforms", RAM is in write mode and data are going to inputs. Then RAM is switched to read mode, after waveforms are loaded.
Column and row selections are working only when 33 and upper range at DCV is dialed or function selected to ACV or ACI - when DDS needs to be initialized. In every other calibrator mode col and row sel. are at low state. So chance for help is now higher. Now I know that RAM is kaput, but this doesn't mean that CPLD is 100% OK. We will see after RAM replacing.

[SQ9MT_PL]

Do you have a service manual for this? (include sch)

[ddrl46]

For faults on the A6 and A7 assembly the Fluke 57LFC/AN calibrator service manual (https://xdevs.com/doc/Fluke/57LFC/57lfc.pdf) may be of more help than the Fluke 5500A service manual.
The A6 assembly is pretty much identical to the later Fluke 5520A A6 boards, the A7 board from the 57LFC/AN is also very similar as far as the protection circuitry, low/mid current sources and current shunt section are concerned.
The A8 board from the 5500A is very similar to the A8 board from the 5520A with the exception of an additional transformer which is used for the higher frequency AC voltage capabilties of the 5520A.

For troubleshooting your voltage amplifier problems I would recommend that you temporarily remove your A7 current assembly just to make sure it isn't causing any additional problems.

You might also want to check the lefthand side of your A6 board for damage, a while ago I had a Fluke 5520A which suffered some similar abuse. In my case however the FPGA on the A6 board was damaged (got no more in-guard communication) by some arcing between internal layers on the A6 board. Attached is an image of the damage I found, if I remember correctly the high-voltage feedback line arced to one of the control lines coming from the FPGA.

[Analogdevice]

Thank you very much!
All tests here was done with A7 assembly removed from calibrator. A8 board has 1 fault at 65V PSU, this was done at 20th October.
I will check left part of A6 board, but analog hardware is working, so I think this is OK. Only DDS is not working, but RAM for waveforms is not working (HIz output at read mode), so this is next step. This part is stocked at Mouser, so it took about 1 week. A6 repairing is now waiting.

Today I will start with A7 assy with burned components which was ordered month ago. Because DDS is not working, A7 may be tested only at DCI and retested for ACI when A6 will be repaired. Besides that A7 is burned, after placing to the machine it activates permanent Err 104. Plenty of components are burned (e.g. OPamps - 3pcs, PTTC - 2pcs, 2 surge arresters, 2 DG444DY, relay etc.) - I'm talking only about visible faults. So this will be very very juicy...

[Analogdevice]

Hello,

after some time I'm here again with news.

After last post I replaced RAM at A6 board. After replace, waveform DA's was generating waveforms of selected f and waveforms, bud scaling DA's was not working because write inputs of both DA's was in undefined state. These contrl inputs comes from altera flex 8k FPGA.
I also replaced altera, but now I have Err104 permanent with relay cycling.

Altera (U5) is configured by configuring device EPC1. I checked all config pins and this looks good. As you can see in picture below, when nCONFIG goes high, CLK is activated (altera's internal generator) and DATA is transfered from EPC1 EPROM to altera (via DATA0). When configuring id done, the CONF_DONE goes high and reset FPGA to user mode via MAX705 supervisory circuit. Some signals are present, e.g. strobe for DAC reference (600Hz), which means that altera was configured succesfully (also nSTATUS is still low, so during configuring there were no err's). But when "loading waveforms" appears at display, there are no data flow to RAM. The only problem is that I really don't know, where are data of waveforms stored - from where are coming to RAM. I thought that it's stored at outguard CPU, but serial communication is working (watchdog is not tripping and also there is no err message that inguard is not responding).

Any ideas, ddrl46? Thanks.

[Analogdevice]

Okay, so calibrator is running. It was poor flux quality I used, after cleaning the board voltages are running correctly, both AC and DC functions.

The last one is burned current board, but we have some more job to do, maybe at christmas vacation. Every burned component was replaced, now it's time for measurements.
The biggest problem will be with Z2 resistors net, one of the resistor is burned, so I had to obtain new or used net or replace burned part for some precision resistor with low tc.

Summary:
A8: -65V supply unregulated - replaced OPamp at regulation loop
A7: several components burned - replaced, running to 20A at every settings (looks like problem with feedback loop, IDK)
A6: Err104 trip at >3V3 range, no AC voltages - replaced SRAM and Altera with dead output for scaling DA
A5: no err's
On the PSU for A7 brd also 7912 dead - replaced, OK.

Preliminary tests for accuracy - OK, 24h stability - OK.

[ddrl46]

That is some good news, looks like it's slowly starting to gain more functionality .

I've noticed that the following components are almost always dead after a serious overload on the A7 board:
Current shunt switches U23 and U28 (DG444)
U53, U48 (OP27E)
One or more of the protection diodes, especially around CR17, CR24 and CR25.
It's always worth checking the output transistors.

A while ago I also had one of the big 75N06 MOSFETs (Q31, 32, 34, 35, 38, 39) fail. In my case this caused the positive output currents to be correct, but the negative output currents in the 33 and 330mA had an offset at the high end of their range.

I haven't yet managed to have Z2 fail.

One area you might want to pay attention to since your A12 filter PCA had a defective regulator is the area that monitors the output compliance voltage and sets the output voltage of the A12 filter PCA accordingly (U19, U20, U21, U47).

Something else which may be worth checking is the settling time of the A8 voltage amplifier when programmed for 33V DC in the 330V range. If the calibrator takes a long time to settle, but eventually does reach the programmed value, U13 (LF356) may have failed on the A8 assembly.

[Analogdevice]

Well, after a short pause, I'm here again. I have no time during vacation, so today I looked at the A7 current assy.

Based on advantages from ddrl, I checked circuits listed in contribution:
U23 and U28 were replaced before, they were burned (visible);
U53 and U48 were replaced today, because Z4 was hot during operation;
Q31, 32, 34, 35, 38, 39 were checked and they are OK;
all transistor mounted at aluminium heatsink were checked and they are OK;
CR24, 25 and 17 were checked and they are OK;
U31 was replaced preventively (because both 10R0 resistors located between K3 and K2 were smoking ago);
Z2 was repaired by mounting 40R/1% resistor (30W, TO220) as a temporary fix, I must obtain Z2 or some precision thin film resistor.

Besides that, following components were replaced - some of them was burned, some of them causes malfunction, just like 74HC374 which gives Err 104:
U19, U20, U21, R116, U16, U3, U4, U5, U43, RT2, RT3, E1, E2, K16, K20 (these relays were OK, but they had mechanical damage due to burning of E1 and E2) and CR18 (I hope I don't forget anything).

Situation is following:
330uA range, AUX output: OK (both AC and DC)
330mA range, AUX out: OK (AC/DC)
3A range, AUX output: NOK (8A permanent!)
20A range, 20A output: NOK (18A permanent!)
AUX voltage 0-7VDC: OK
AUX voltage 0-5VAC: OK (tested 50Hz and 1kHz)

Well, question is, if there is something, which has 3A range and 20A range in common or they are two independent faults. I'm working with schematics from 57lfc and 5500, but they are different. It takes some time, but it's on the good way, completely dead board, many faults were reduced to 2 also thanks to ddrl with good hints, which were helpful. Later I will also post some pics from A7 (and other boards).

ddrl, if you have something to tell about my problem, I will be very happy, I like happy endings

[saturnin]

Nice to see the calibrator slowly rising from the ashes (literally) 

While posting pictures of the affected board(s) later might definitely help someone in future, it is a common practice here to document the current state of a repair by posting photos during the repair process as it might provide useful hints for potential contributors what it is going on...
Without any schematic and/or additional information available, provided data about replaced components (e.g. U19, U20, U21, R116, U16, U3, U4...) is basically useless for further online debugging...

[Analogdevice]

Some little update, U36 (AD823) was in positive saturation, this opamp is common for both 3A and 20A ranges. Replaced, 3 and 20A ranges are OK.

Next step: Fault during DVZS occurs when zeroing calibrator (it crashes at 2nd DC zeros step). Looks like something around U43, I guess.

What is really bad is that Mouser doesn't have AD823 in stock and it will be stocked at ca. july this year. I used my last piece to U36. I feel nervously 

[Analogdevice]

DONE

I'm exhausted, so very quick:
Fault during DVZS, also Fault during Trim1V was caused by MIC5800 malfunction.
Fault dunring Trim1 was caused by bad optocoupler for snubber. 

[tavisho]

Hello man, i am experiencing some trouble with my fluke 5520 A that keeps showing error 104 and an unstable VCD output. Any ideas where we can start looking for the problem? Anyd schematics you can share?? Any kind of help is welcome.

[chinapp]

DONE

I'm exhausted, so very quick:
Fault during DVZS, also Fault during Trim1V was caused by MIC5800 malfunction.
Fault dunring Trim1 was caused by bad optocoupler for snubber.

where is the optocoupler ? A6 ?

[Ugur]

Hello ,

What error messages does the device give in the Diagnostic test and Zero tests?

NOTE: Don't forget to check the 22V varistors on the A3 Motherboard. RV1-2-4

[Jetecnet]

Hi Analog, I hope you are still monitoring this thread. Interesting repair.
I have a 5520A that has a problem with the thermocouple read/zero. I believe I'm missing a "SRC" signal direct from the Altera chip. From your post I'm see you replaced the Altera? and I think you are saying that EPC1 (U12?) will reprogram a replacement? Is it that simple? Just replace the Altera chip? Or do I need to somehow tell EPC1 to do the reprogramming? Not very familiar with this FPGA stuff.
Thanks for any reply. JC

Quote:
I also replaced altera, but now I have Err104 permanent with relay cycling.

Altera (U5) is configured by configuring device EPC1. I checked all config pins and this looks good. As you can see in picture below, when nCONFIG goes high, CLK is activated (altera's internal generator) and DATA is transfered from EPC1 EPROM to altera (via DATA0). When configuring id done, the CONF_DONE goes high and reset FPGA to user mode via MAX705 supervisory circuit. Some signals are present, e.g. strobe for DAC reference (600Hz), which means that altera was configured succesfully (also nSTATUS is still low, so during configuring there were no err's). But when "loading waveforms" appears at display, there are no data flow to RAM. The only problem is that I really don't know, where are data of waveforms stored - from where are coming to RAM. I thought that it's stored at outguard CPU, but serial communication is working (watchdog is not tripping and also there is no err message that inguard is not responding).

[Ugur]

Hello,

general information,
The problem you mentioned about the Thermocouple is that the R128 10 ohm located on the A6 board is faulty.

[Jetecnet]

Hi, Thanks for the response, I'm pretty sure we checked all the components on the board but I will check again. JC

[Ugur]

What error messages do you get in the device diagnostic test and zero tests?

[Jetecnet]

Zero and diagnostics fail on TCMEASZ, Error 308 constant outside limits.
I don't see a switching waveform at pins 8/9 on the ADG511 on the thermocouple board. This is high at power on and never changes even on RESET.
I see switching at pins1/16
JC

[Ugur]

It's not a bug I'm experiencing, but more information may be needed.
What was the first fault of this device? How did this error occur and which parts did you replace? Or were you getting this error from the very beginning?
Now it can't complete the reset test but does it give a correct output?

I think you checked the A10 board first and it's solid.

Can you check this?

A5 board
TP14 : +3.75V with respect to FCOM
TP13 : -5V with respect to FCOM

Have you checked these parts on the A6 board?
U13,U6,U82,U96,U22, CR18,CR19 (I don't think U5 will be a problem)
Did you check the regulated voltages during operation?