Fading segments on DMM. Recovery after >1 hour off time.

[marcel22]

I have been given a bench DMM ( Metex MXD-4660A )  for repairs but Im fairly new to LCD technology so looking for advice. Im also looking for repair manual or any documentation that might help me with adjusting the DMM as it has a drift in measurements.

When I turn the DMM on the display seems okay, after about 15 minutes or so after the dmm heats up a bit some of the segments of the display start fading away. The display itself is not using zebra stripes, just a simple pin connections.

Im wondering what would be the best approach to find an issue here.  From what I know about LCDs they are about 5V AC driven, but thats about it.

Any suggestions on what to measure and what the probable cause may be? Possibly caps that drift too much with rising temperature?

https://www.dropbox.com/s/bjzh0vfw7r0q58o/2021-08-16%2017.54.50.jpg?dl=0
https://www.dropbox.com/s/v41n6jnnzqizubo/2021-08-14%2011.17.50.jpg?dl=0

Thanks,
Marcel

[wraper]

Probably just a bad contact with conductive rubber strips Between PCB and LCD. Heating may cause some small deformation and contact disappears completely. You could apply some pressure on LCD or backside of PCB and see if it fixes the issue.

[marcel22]

Thanks for a reply, but as mentioned in my first post the DMM is not using zebra rubber strips. Ist standard metal 2.54 pins.

[wraper]

Thanks for a reply, but as mentioned in my first post the DMM is not using zebra rubber strips. Ist standard metal 2.54 pins.

Missed that. Then it may be a bad connection between LCD and terminal. They are basically spring contacts covered with adhesive, and sometimes contact goes bad. Or LCD controller might be dying,

[marcel22]

Yeah, but wondering how to measure / confirm this. I can attach a scope to each pin and try to measure the characteristics right after power on and then later when the segments start to fade, but I was hoping to get an inspiration for a more practical approach

[wraper]

I guess you could measure signals on LCD pins with oscilloscope before and after the issue happens. If they are the same, most likely LCD is bad. If they are different, there is something wrong with LCD controller.

[Runco990]

I'd start simple.  Use a stick to push on connections, and try a little freeze spray.  That should localize the problem to a particular area of interest.  Even the metal pin LCD's can have poor connections, there is also a "bias" voltage that adjusts contrast.  Maybe check that as well.

[Gyro]

That's a multiplexed LCD. Unlike the simple eg. 4 digit non-multiplexed LCD that you mentioned. They rely on switching the LCD pins signals up and down through various threshold voltages in combination to 'light' individual segments.

These bias voltages are generated by the driver IC,  but there may be external decoupling capacitors and external contrast adjustment. This might be a pot, a fixed resistor divider (maybe with temperature compensation thermistor) or even a signal from a microcontroller.

The best thing to do is to try to find the driver IC datasheet. That will give you a much better idea of how it operates, and any external parts that might be influencing the display contrast. I wouldn't start putting mechanical or thermal strain on the LCD pins at this point!

Internal photos would help too.


P.S. The Service Manual...  _digital_multimeter_sch.pdf/download.html]https://elektrotanya.com/metex_m-4660m[1]_digital_multimeter_sch.pdf/download.html

EDIT: It looks as if the resistor chain down the right hand side sets the bias levels. Probably the first place to look...

[wraper]

It does not look like there is something wrong with bias voltages. Part of the LCD appears to work fine, so voltages should be fine too.

[Gyro]

Maybe. Thermal and reversible / repeatable is good anyway.

[marcel22]

Pics of the display PCB

https://www.dropbox.com/s/4gt930lipcm2cvq/2021-08-16%2020.55.14.jpg

https://www.dropbox.com/s/5qnf77gx8aflyot/2021-08-16%2020.55.24.jpg

I could probably replace the black blob with a new HD44780 driver, but Im kind of thinking the problem is somewhere with the passives that start having drift and fade several segments.

[Gyro]

At least it's nice and modular, no nasty flex-prints.

I Just noticed that the manual I linked is for the 4660M so the driver chip could be different (irrc, the HD44780U is a dot matrix driver) but I'm not clear on the two DMM models / variants.

It's certainly worth checking for bad solder joints  - I notice that there's a Zener too... worth monitoring.

Worst case, yes, you can replace the blob with the appropriate IC on the pinout. Just a warning though that the epoxy is really tough, you'd almost certainly need to desolder the LCD before you could safely attack it.

[negativ3]

"It's certainly worth checking for bad solder joints  - I notice that there's a Zener too... worth monitoring."

This as a first pass. A nice coating of flux and touch-up all the LCD pins with a dab of new solder.

[marcel22]

Ill do that,  Its not very intuitive try tough as the display works great for the first hour...Ill give it a go and report back.