Monochron display problem

[Stephanie]

This is a longish story, please bear with me.

I assembled my Monochron kit about a week ago. Discovered that the RTC didn't work, as per this thread. I replaced the DS1307 with a Chronodot and that solved the ticking problem, all was well. Then after running fine for about 36 hours, the display went wonky. The top half of the display was blank, and everything was displaying only on the lower half. It looked like an old TV that had lost vsync.

In case it was a software problem, I reflashed the firmware. I tried multichron and ratt, but the problem persisted.

I figured there might be a bad solder joint or two, and while trying to find the problem I learned all about water soluble flux and had a nasty education, as per this thread. I removed my modifications to the board, removed the ICs and battery, and gave the board a thorough clean.

The problem remained. Wondering if the problem was with the screen itself, I hooked that up to an arduino and tried the example sketch in the ks0108 library, and found the display is working perfectly.

Back to the monochron pcb. I figured that maybe the cleaning I had done, did not remove all the traces of water-soluble gunk. It was possibly still hiding underneath the 20-pin connector for the LCD, or maybe hiding under the dip28 connector. I removed them both, thoroughly cleaned the board of water-soluble residue, and replaced both connectors. The problem remained.

Finally, wondering if the Atmega328p was itself the culprit, I replaced it with the one from my arduino. I used a usbtiny to flash it with the bootloader and the fuses, then used the ftdi friend to flash the ratt firmware. And the problem with the display remains. The top half is blank, the bottom half has both upper and lower images appearing.

To ensure I'm not going nuts, I took the Atmega328p out of the monochron, put it back in the arduino, burned the arduino bootloader back onto it, wrote the ks0108 sketch back to it, and verified that the display itself is still fully functional.

So at this point I have:

Cleaned the board
Tried different firmwares
Verified the display itself is functioning
Replaced the dip28 and 1x20 sockets
Replaced the atmega328p

And through it all, the monochron continues to have a borked display.

What is left for me to try / test / replace? At this point the only thing left is the PCB. Is it multi-layer? Could there be damaged traces inside that I can't see? The traces I can see are not damaged.

The fact that it's only the top-half of the screen should be a clue - I'd have thought maybe it's the chip-selects that is buggered, except they control left-side / right-side, don't they? If one of the /CS lines were stuck or shorted, wouldn't that cause the screen to be messed up horizontally?

The monochron is such an attractive clock, I don't want to give up and scrap it. But I don't really want to have to replace the whole thing. Could I order a partial kit, i.e. everything except the display and case?

Any help would be most appreciated!

Thank you & cheers!

p.s. I can provide photos of the display problem, pcb, whatever else might help figuring this out. I've been at it all day and it's left me stumped, frustrated, and tired.

[adafruit]

post a photo, it could be the LCD in which case its easy to fix

[Stephanie]

Thanks for the quick reply!

Attached are two images, 1st is the display as it shows up on the monochron. 2nd is my test of the display with a boarduino & the ks0108 library example sketch. I figured if the lcd worked with the arduino it must mean the lcd is ok?

edited to add: hmm the images appear in the reversed order, obviously #1 is the boarduino test, #2 is on the monochron.

[adafruit]

wow that is some great debugging! yes you've determined the display is fine. so that good. can you post up a photo of the kit PCB front and back?

[Stephanie]

Thanks!

Attached are photos of the pcb front and rear. Some notes:

As explained, in my debugging I removed and replaced the sockets (dip and lcd connector). I didn't have another dip28 so I used a pair of dip14s. I have verified that there are no solder bridges or shorts; I've inspected the entire board as best I can using a 5x jeweler's loupe.

The chronodot is not connected electronically at the moment, though the plan is to replace the DS1307 and incorporate temperature display along with time.

Final note, the 32.768 crystal is not there, I've got one of the faulty crystals. Not an issue for me, as I plan to use the chronodot. For testing purposes I've been using the DS1307 -- although it doesn't keep time, it makes the firmware happy to have it there during testing. (And the display did work normally for the first ~36 hours, regardless of the status of the DS1307 / crystal.)

Thanks again for the help!

[adafruit]

OK, the best thing we can suggest to do is full continuity testing. make sure each pin of the LCD is not shorted to its neighbors. then, check that the LCD pin is continuous to the AVR pin it should be. maybe give that board a good scrubbin with alcohol and an old toothbrush?

[Stephanie]

I just finished verifying every line again. No shorts or bridges between any of the LCD connector pins, nor any between any of the 328's pins. And I verified continuity between every one of the LCD pins and the corresponding pin on the 328.

Thought for a second I had found the culprit with DB0 till I remembered the 1k resistor - which verifies as well on my DMM.



It's like when 2+2 comes out as 5. Everything indicates that it should be working correctly, yet it doesn't work.

Trying to think it through logically - it's consistant about being the top half of the display that remains blank and the bottom half appears as in the picture. If I hit the menu button, the text can be seen only on the lower half and appears 'messed up', i.e. the letters are malformed and stuff appears overtop of other stuff...

Other thoughts... is it possible there is something wrong with the LCD display itself, which for some reason does not manifest during the boarduino test? A timing issue? Text & graphics together? When I test for shorts on the LCD display on its own, I don't get any dead shorts but I get a low reading between DB3 and DB2 - only about 80 ohms between those two pins. I'll have a closer look at that...

Edited to add: Ok I think this is it. If I put it on the breadboard and test it witht he boarduino, and short pins 9 and 10 of the LCD (db3 and db2) I get almost the exact same failure!

I must have spent like 4 or 5 hours fussing over the pcb yesterday, but the fault is almost certaily on the display board -- but it was only manifesting when the display was on the monochron, and not on the arduino test.

I'll give the connectors on the display a very thorough cleaning and a bit of a redo then test it again.

[Stephanie]

It works! It works! And there was much rejoicing!

Thanks for all your help & guidance, and patience!

There was some solder / flux residue underneath the black plastic of the 20-pin male header, like between the plastic that holds the pins together, and the display PCB.

For whatever reason the 80-ohm 'soft short' was enough to scramble the display on the monochron, but not enough to scramble it when I was testing it on the breadboard. What a debugging ordeal it's been... but I guess it's another learning experience! "Just because a part checks out on the test rig doesn't mean it's perfect." Or words to that effect.