Page 1 of 1

Roland DR670 malfunction-Memory and sound

Posted: Tue Sep 11, 2018 1:25 am
by coldwar
I recently got a DR670 that has some kind of strange problems. It powers up fine and the first couple of times worked normally in that it output sound and would recall patterns. So here are the symptoms and any help or ideas would be appreciated. I have attached the service manual with schematics to help troubleshoot.
Upon power up it displays the RST MEM OK? at which point you press Enter and it goes into main mode with patterns ready to play. Sometimes it freezes after this point and won't play any patterns. The only factor seems to be leaving it sitting at the RESET MEM OK screen a couple of minutes before pressing Enter.
Regardless, if it does boot past into main mode, a horrible screeching tone is the only thing that comes out of the headphone jack or line outs. It sounds like a system data buss rather than feedback if that makes sense to yall.

Things I have tried: System reset. Loads presets back no problem. Upon reboot (with mem battery inserted) it still shows the RST MEM OK? prompt. If the memory battery is taken out, then it still goes to RST MEM OK but then after pressing Enter, it shows BAK BATTERY denoting the missing battery (just in case you were wondering). I have also run the system tests listed in the service manual, checking the voltages of the unit, backup battery, and main ICs:CPU, Gate Array, Mask ROM, and RAM and according to the system all check out okay.

When checking voltages things seem within range. It splits at the beginning to 8V and 4V (multimeter showed 8.08V and 3.57V) a lil wonky but within range as far as my experience has shown me. Most of the major ICs I have tested have 3.37V feeding them. The only weird finding was 1.66V at C52 (C53 right next to it showed 3.3V) and D2 (on the mainboard not jackboard)tested as backflowing on Diode test. I also originally replaced D1 on the jackboard as it tested sporadically failed, just to be sure ya know.

I think that covers the main bases. Does anyone out there have experience with this machine or have an idea about the problem? I feel like I am starting to chase my tail with this thing. Many thanks as always!
dr670 service notes

Re: Roland DR670 malfunction-Memory and sound

Posted: Tue Sep 11, 2018 2:03 pm
by Rasputin
coldwar wrote:Upon power up it displays the RST MEM OK? at which point you press Enter and it goes into main mode with patterns ready to play. Sometimes it freezes after this point and won't play any patterns. The only factor seems to be leaving it sitting at the RESET MEM OK screen a couple of minutes before pressing Enter.

Regardless, if it does boot past into main mode, a horrible screeching tone is the only thing that comes out of the headphone jack or line outs. It sounds like a system data buss rather than feedback if that makes sense to yall.
Random Thoughts and Wild Guesses:

Board flex problem centered around the SRAM. Try taping something non-conductive to the top of the SRAM in such a way as that it will be compressed when you put the machine back together. A penny or two wrapped in electrical tape? Depends on the thickness you need.

It seems like there's an intermittent but frequent fault with the SRAM, such as that it can be R/W about 50% of the time. So you might be able to access it enough to RESET MEM here and there, but all contents are invariably lost regardless.

Perhaps the screeching sound is the system hanging during or before the sound circuit can be initialized.

The DR670 has a fairly comprehensive self-test so I'm puzzled as to why you can't get any problems reproduced in self-test mode. I'd spend some time running and re-running self-tests under various conditions to see if you can ever get it to fail.

Re: Roland DR670 malfunction-Memory and sound

Posted: Wed Sep 12, 2018 12:39 am
by coldwar
Rasputin, thanks as always for your help :) The MC303 is still going strong so thanks for that again!
You are most likely correct in your assessment of the board flex issue. I did some more probing around today and ended up reflowing the RAM chip as a few pins in the middle just looked kinda smooshed down compared to the rest. Adding to the abuse/flex damage possibility was the fact that a lot of screw lugs on the face panel had cracking in that same suspicious NW quadrant of the board. That seemed to get it to behave a little more normally but still beeped when playing presets. I started probing around IC 10 and checking in and out for voltages when I checked the resistor R23 above it. When my probe touched it, the beeping would stop and it would reset. I tried to dab a little solder on that side of the resistor but I guess I didn't get it good. Like an idiotic mad scientist I decided it needed a lower biased diode attached to that side and it seemed to work but it turned out the bit of diode wire 'foot' for the resistor solder pad was really all it was smdh. I wanted the nonsense to be right though lol.

So as of now, it boots and plays presets and is stable enough to drum on a bit. The MEM RST OK? message still shows up every friggin time of course but I can live with that for now. Best to let the poor lil almost fried electronics chill out a bit I feel. I decided to pull the panel buttons to wash and epoxy the cracked screw lugs while everything cooled down. Will cold boot test it tomorrow morning after reassembly and get back to you/update part numbers :)
Thank you!

Re: Roland DR670 malfunction-Memory and sound

Posted: Thu Sep 13, 2018 1:57 pm
by coldwar
Well, that was short lived lol. So things worked for a while yesterday until it started hanging at boot and doing the squealing thing again. It looks like IC10 and R23 are attached to the RESET line, which is my next suspect (though bad RAM chip is also in the works) outside of the obvious crack of some kind in the motherboard. It may be within the layers of the board as I cannot see a crack at all and flexing the board from either side provides only a temporary fix until it stalls again and reboots. The oddest thing tho is that everytime I touch the right side of R23 with my meter probe, it loses the squeal and behaves normally smdh. Wish I could just hotglue the test probe in there to it lol. The components in this area are all SMD (which I can repair basics with my iron) but I am starting to weigh whether or not it is worth too much more time if the issue keeps presenting itself and is indeed an interlaminate crack which I could never repair.
Any ideas, Rasputin? Do you think the Reset circuit IC is possibly the culprit or have any ideas of elsewhere to check? I did try penny stacks and even a C clamp trying out some flexing for a workaround fix, but the machine acted pretty much the same. I have had that work before tho, so good suggestion :) Thanks :)