Mettler FP80 broken

[Eheran]

I had this thing sitting around for some years now...
Its used to measure melting points and other stuff.
When turned on, it displays "error 12". In the manual this means:
"A/D-conversion error (overflow) - maybe measureing cell not connected or pins bend"

As can be seen on the first picture, the cell is connected. I measured from the sensors in the cell to the pins (pic 2)  and they are all fine.
The voltages from the transformer (pic 3) are all fine. All the marked Voltages on Board 1 (pic 4+5) are okay. I replaced a broken tantalum capacitor on board 2 (pic 6), tho i just didnt have the exact value so just used another one (original 6,8µF 35V).
I lifted all socketed chips up and put them back down to deal with potential corrosion there.
The silcscreen (pic 7) is in a very bad shape and seems to come off. However i was unable to see broken copper.

Any ideas how to proceed?

[Vgkid]

Lm399, nice. Have you checked your dc rails?

[Eheran]

On (pic 4) all of the voltages marked on the right side connector as well as the bottom left +-12V, the +-15V a litte right of these and +5V middle bottom are okay. I just found another 6,4V AC bottom right that i have to check, but since its AC it should be okay.


Semi-offtopic: These blue things, like that in (pic 5) bottom right with "100nS", are foil caps? Why is the unit nS?

[Kjelt]

Make a copy of the eprom's content and refresh the charge in them (overwrite them with same code) while you are at it 

[Eheran]

How would i do that? Any recommendations for the hard/software?

[Kjelt]

Oeps you have a 2516 that is an old one and has a different pinout than the more common 2716.
Also (check just saying here) they might have a hard to find Vpp requirement like 25V or 21V which was only used for the first generation of eproms and most modern (and affordable) eprom programmers do not support it.

If you're eprom programmer does not support the 2516 directly than you should build an adapter prototype board with socket and pins to a 2716 and at least read the data out and store it for later.

Anyway knowing it is a 2516 it is really time to back them up, this could well be a cause of your problem that the firmware is starting to go corrupt. I had a Bally pinball machine with 2532's that started to go weiry end of the 90's, I was just in time to back them up but luckily the firmware was well available on the internet.
Try searching finding the firmware of your device on the net or ask here , maybe someone has an equal device.

[Kjelt]

Yeah as I expected (see attachment) this is a 25V Vpp eprom.
Difficult to find any device that can handle this, old eprom programmers from last century might handle them but you should find someone in your vicinity that has one.
Other possibility is that you only read out the data yourself using a modern microcontroller or modern eprom programmer and adapter socket.
Than you at least have the data. Than you need to find some 2516's and someone who can program them for you and sent them to you.
Or which is more future proof is to modify the device sockets so it will take 2732's or even better 27256 which can still be bought new and anyone can program them.

[Eheran]

So a EPROM programmer can (always?) read the data... but the device itself cant?

How would one modify the sockets? What would have to be done?

[Kjelt]

So a EPROM programmer can (always?) read the data... but the device itself cant?

No if the eprom is damaged and really lost some data than also the programmer can not recover it, although I have heard that some programmers can change the operating voltage a bit and read out multiple levels so to determine if it was a '0' or '1' more intelligently. But I have to check on that.

How would one modify the sockets? What would have to be done?

The pin layout is different so if you have the original content backup'd you have to choose a new different target device that prefferably can be programmed with a normal programmer with 12V or even 5V if you choose a eeprom (28 series) or perhaps a flash rom.
 Then you have to look at the differences in pin layout between the original 2516 and the new device and make an adapterboard or adaptersocket.

[1kostik1976]

Hello.
 In my experience very often faulty temperature sensor.

[Eheran]

But I have to check on that.

That would be great. I dont like to throw money away after all 

@1kostik1976:
U mean a faulty Pt100 sensor? 
Its at a little over 100 Ohm, so pretty sure to be okay.

Do u have experience with this device?

[1kostik1976]

Hello.
 with such a device, I did not work. Most temperature gauges are arranged roughly the same ... the digital part is probably OK. you need to check resistors and capacitors in the input circuits and the chip SN74LS251. also need to verify the source of the reference voltage (unlikely problem).

[1kostik1976]

 remove and insert back TBP18S030. possible poor contact in the panels ...

[Eheran]

LM399: 6,982V

check resistors and capacitors in the input circuits and the chip SN74LS251.

To check them... dont i have to unsolder them? How would i test caps otherwise?
And how to check the SN74LS251? Its the only one i saw btw and only on the connector. The Problem is exists with or without the 2nd device connected. So im not sure if that matters anyway.

The Testpoints are labled as follows:
1: -15V
2: +15V
3: UAD
4. Y
5. X
6. ZDO
7. C-CONT
8. -U
9. AGND
10. 1/7UREF
11. Z
12. H-CONT
13: 166,7/150kHz
14. ST.-B.-CL.
15. +5V
16. DGND
17. RESET

Many are clear, but what does ZDO or C-CONT or ST.-B.-CL. etc. mean?
And what do you think about the reset? Should i try that?
Sorry for all the questions... 

[Eheran]

Recommendations for a Programmer?

some programmers can change the operating voltage a bit and read out multiple levels so to determine if it was a '0' or '1' more intelligently.

Sounds interesting.

[Kjelt]

To come back to this, I can not find directly a programmer type that performs this operation during read. I found a Batronix that does a multi voltage read and compare , see picture, they call it a multi-pass comparison, however that is not exactly what I meant , although it is a good test to see if the eprom will fail soon. If it fails this pass than you know the internal charge is starting to wear out/leak.

But what I meant is that you can adjust the voltage during read yourself and view the results. I found a few descriptions of people that made an adapter and seperated the Vcc pin to be connected to a lab voltage supply, thus applying a lower voltage to the eprom whoose checksum was incorrect. One guy applied downto 4,1V to get a good checksum.
So with any eprom programmer and powersupply you can do this yourself manually.

[Eheran]

Alright, cool. But i still need a programmer with 25Vpp.
Willem PCB6.0 (and PCB5.0E) doesnt list the TMS2516.
Some SIVAVA Willem lists it and seems to be the same.
TNM5000 for 300$ list it.
There are also several ELNEC products for $$$ that list the Chip as compatible, which i would always prefer over those like the TNM5000.

Any other ideas or suggestions?