Help with board level D-50 issue

[jimk]

Help.... I've owned this D-50 for 30 years and its never given me any s*** until now and I've been unable to find anyone locally who has any experience with them (though they're happy to dig into on the clock with no promises), so I blew the dust off an old educations and dove in. It all started a few weeks ago when I opened it up to clean a few key switches. No problem. Have done this many times before. While open, I decided to replace the battery, knowing full well I'd have to d/l the patches again. Got her all back together and fired 'er up. Got the right screens (address info) for only a few seconds before the display started to flicker and then.....right as I was watching it......it went blank.....except the top row of the display in which all the pixels are lit. No button pushing (including the LCD test in the Service Notes) has any effect. I've checked the obvious stuff like the power supply and in an effort to get "anything" on the display, disconnected all the connectors on the main board and booted from there and no matter what I do, the display just gives me this row of dots (see attached picture). Also, in the recent past, I've had power supply issues (low Vdd) which turned out to be a bad regulator. Replaced it and "recapped" the PS so now its doing fine. Also, had an issues with several of the DRAMS pulling over .3A each which is what caused the PS to melt down. Replaced them and all seems good. Now....with this display situation, I'm in over my head. I don't want to just start replacing stuff and the service notes makes a cryptic remark about the row of dots indicates a possible problem with the related IC and then turns around and says if I do the LCD test and it doesn't go out for a second, its the display..... at this point, both my D-50 and I need help. Thanks everyone. Jim

[AncientGeek]

Assuming everything is working except the display...

1. Check the -12V supply for DC value and ripple.
2. Check VR2 on the main board which is a trimpot for setting the LCD display contrast. It's next to CN11 which goes to the LCD. Check that the voltage on the wiper is adjustable and negative with respect to ground. A bit of contact cleaner (*NOT* WD40!) on the trimpot followed by exercising it might allow if to bring the contrast back into range.

If this doesn't help let us know.

Cheers,

Steve

[jimk]

Steve, Thanks for the quick response. -12 supply is present (-12.23) and "clean." VR2 is clean and functions well, reporting -2.0 to about 3.6 -Vdc. The display responds correctly from dark screent to full contract. What next?
Thanks for your help. Jim

[jimk]

I wrote my original inquiry late at night while exhausted and didn't explain things well, so here's a better chronology of events. When I put the D-50 back together and fired it up (now without sounds because of losing memory from the battery replacement), I got the correct display (bank addresses, etc.). Next, as a matter of routine (I wasn't expecting to find anything), I entered diagnostics mode and began stepping though the process which is when the display flickered a few time and went blank, but for the dots as seen in the picture. After my initial panic, I began general diagnostics at the power supply. I quickly realized that the +5 volt Vdd line was only +3.5V, so I pulled the PS and on the bench, the problem remained. Meanwhile, plus and minus 12 and the 5v mute lines were fine. The +5 regulator was toast, putting out 5 V with no load, but even a 2K load pulled it to under 4 volts, so I replaced it and "just cause" recapped the entire supply. All outputs now tested perfect with the 5V regulated output hold right at 5 all the way to 1 amp and no noise (on the scope). I reinstalled it and the MB was still pulling the +Vdd to 3.5, so I removed all the connectors on the MB but the PS connections and got the same result. Hooked up a current meter and sure enough, the MB was pulling near 2 amps on the +5 regulated line. I pulled the MB and systematically starting lifting a leg on each of the inductors until isolating the problem at the bank of 4416 DRAMS. Sure enough, (on a breadboard) 4 of them were individually pulling over .3A each. I replaced all 6 (this time in sockets) and fired 'er up. Smooth sailing on the MB right at 4.99Vdc, but still, no display, but for the dots (as pictured above) and no response to button pushing. The only light on the front panel is the "Upper Bank" light. Other tests I did that did not produce any changes: Believing that the timing of this was very suspicious, I needed to start by assuming that something happened that was caused by my activity, so I started by analyzing the entire board under a magnifier for breaks, bridges, and artifacts that might be causing a problem, then touched off the plus/minus 12vds solder connections of the display back light, checked all the connectors for continuity and possible bent pins, tested all the button switches, touched off the solder joints of the switch matrix lines on the switch board and checked the 27C512 PROM and it's programming. And last, just because I had a couple extras around, flashed a fresh PROM using a v2.1 bin file I found on line and also flashed a separate PROM with a copy of the existing data in the PROM from my D-50 (also 2.1). In the D-50, neither made any difference. I did find 13 random, scattered single value differences in the two files which I assume are little more than the beginning of gradual data corruption in my factory PROM after 32 years. Without any guidance, my next step was to try to determine if the display was actually working or not. The service notes are not clear (to me) on how to accomplish this. While surfing the internet looking for information that might help, I discovered the "Roland Clan Forum" which I didn't know existed. So here I am. Thanks again to Steve and anyone else who is able to help. Jim

[AncientGeek]

Jim,

The LCD is, as usual, an HD44780 (or compatible) type. Do you have an oscilloscope or access to one? If so, then it's be worth looking up the specs for the interface and 'scoping it out. Compatible displays, albeit not EL backlit, are easy to obtain so it may be worth just trying a replacement.

Here's a site showing how to do it on a D-550: http://hacks.slashdirt.org/hx/D550/

If you're unlucky, the problem could be with the IC that drives the LCD which is a custom gate array. If it were me, though, I'd find a cheap LED backlit display and give it a try.

Best of luck.

[jimk]

Yea.....<sigh> I do have a scope buried deep in storage...if needed. I was hoping one of you would see that I overlooked something obvious (I am NOT smarter than a 5th grader) and I'd get this thing singing its wonderfully distorted 1987 D/A converter sounds soon. Anyway, I just purchased a 4002 display online and expect it here in a few days. I'll keep my fingers crossed that its not the gate array... I've already looked around...nada...and much rather modify the synth for a thicker display and 5vdc backlight than try to find a proprietary IC. I'll report back once I get the new display installed. Thanks again for your input. One other question. No button pushing anywhere on the board causes the red status LEDS to change. Only the "Upper" LED is lit. Does this condition suggest anything that warrants separate investigation or would you recommend I just stick with chasing down the LCD issue? Thanks again for all your help. Jim

[jmbe]

Got the same issue with my D-50 that I recently bought (as defective), the only difference is that mine is stuck on the Local Edit light.
After checking and recapping the PSU I started poking around with a logic probe. It turned ot that a couple of the address lines running between the ROM ant other IC:s were stuck in high (or low, don't remember, but they should show activity). I pulled out the ROM and then the lines came to life (not the rest of the synth, of course). Have ordered a new ROM (V2.22) and will see if that does any difference.
Any success on your part?