New problem with trusty old D-50

ElecticGeek · Joined: Aug 16, 2008 Posts: 7 Location: Melbourne, Florida

My 10 year old D-50 has led an easy life here at home and has always worked flawlessly. That is, until I powered it up a few days ago and found that the tones were being triggered on the key release action and no sound was emitted for the key down action. This problem seems to affect all patches and all keys in the same way.

I checked the owners manuals and noticed for the first time the recommendation that the backup battery be replaced every 5 years. Good grief! I bought my D-50 in 1987 and the memory backup battery has NEVER been replaced. Since I am an embedded electronics designer, I immediately removed the bottom cover of the D-50 and located the CR-2032 cell. It was easily replaced after cleaning up a bit of minor corrosion. Afterward I replaced the bottom cover and powered up the keyboard and saw scrambled characters on the VFD. Ah, memory loss. So I copied all the patches from the standard ROM card to the internal memory. The VFD now showed normal characters everywhere.

But the key issue remained. So I once again removed the bottom cover and methodically de-mated and re-mated every connector on the Main Board, and then retested. Same problem.

Some time ago I saw some postings from a former Roland service technician. If you're reading this now, I would appreciate your thoughts on this problem.

The problem seems to be detecting the key down event for all keys. I have yet to check to see if the MIDI interface still works properly.

Anyway, I am fully able to troubleshoot digital and analog electronics and have a 100 MHz mixed signal digital oscilloscope here. Of course, without a service manual it will be very hard to troubleshoot.

Any help would be appreciated. This synth is in perfect physical condition and this problem has shown up "out of the blue."

EdisonRex · Posted: Mon Aug 18, 2008 1:15 am Post subject:

I don't have a manual for a D-50, but it sounds like something in your keyboard scanning logic has gone awry (stuck high or low). you might check the (probably discrete) chips at the end of the keyboard's connector.

Can you see the same problem with MIDI out from the keyboard? That would tell you if it's local to the keyboard hardware.

Also try MIDI into the D-50 and see if you get proper note operation.

ElecticGeek · Joined: Aug 16, 2008 Posts: 7 Location: Melbourne, Florida

I appreciate the suggestion about the keyboard scanning logic being stuck. There may be something to that. I've been busy with flood problems here in Melbourne due to TS Fay, but was just now able to try the D-50 using MIDI. The good news is the the note on event is handled correctly and the patch sounds fine when using another keyboard as a MIDI controller.

I was lucky enough to find an online source for all of the D-50 board schematics and also instructions for invoking the built-in self-tests. The aftertouch self-test fails, which might be due to the Dyna Scan PCB not reporting the key on event. There is a custom gate array chip on the Dyna Scan PCB and also an obsolete RAM chip. I may be able to replace the RAM chip but they're rare. I may have better luck finding a Dyna Scan PCB to replace the one I have.

I plan to pull out my oscilloscope to check the aftertouch circuit to see what's up with that. It's not too complex - some op amps and an analog data selector that feeds into a CPU A/D input (I think).

So, there is still some hope I may be able to fix this keyboard. The VFD is nice and bright and everything else seems to be just fine. Oh well, it is over twenty years old.

EdisonRex · Posted: Thu Aug 21, 2008 1:52 pm Post subject:

Stuck circuits are usually a discrete component. Check caps, especially, and clean the board as there might be tin whiskers or something silly shorting it.

ElecticGeek · Joined: Aug 16, 2008 Posts: 7 Location: Melbourne, Florida

Hmm, I've cleaned the Dyna Scan PCB using solder flux remover and looked for obvious silly problems. None noted. I didn't see any caps used except for bypass/decoupling, so those didn't seem appropriate to suspect.

The schematic is here:

http://home.arcor.de/richardon/richy2001/d50/sch4.gif

See what you think and I'll consider any additional suggestions. I can still try using the D-50 MIDI out. I haven't tried that yet.

Thanks for your suggestions!

EdisonRex

ElecticGeek · Joined: Aug 16, 2008 Posts: 7 Location: Melbourne, Florida

Yes, I used a very stiff brush (more stiff than a toothbrush) but non-metallic, to clean the scummy flux off the board. It was not stiff enough to damage the traces - I've used it many times to clean up PCBs.

Could you possibly offer some hints please as to how to interpret the keyboard switches in the schematic I linked to? I see 32 switches and obviously the keyboard has 63 keys so there's something else I need to know about how the keys are wired.

Hopefully I'll be able to do the MIDI out test tomorrow, after which I'll report my results.

Thanks,

Dave

EdisonRex · Posted: Fri Aug 22, 2008 12:13 am Post subject:

They've abbreviated the keyboard circuit - the dashed lines and the lines going to nowhere, along with the telltale key markers "C7" and "F6" on the left side (top range of the keys) and "C2" and "G2" on the right. BR0 and MK0 on the right of the Dynascan board correspond to the lowest octave, repeat the sections above for BR1/MK1 through BR3/MK3 and you can see the progression, likewise on the other side you can see the same abbreviation. No sense in printing too much of those arrays. Each BR/MK combination is 8 notes, so BR1/MK1 are for G#2-D#3, BR2/MK2 are for E3-B3, etc. Looking at the switches and the nomenclature, I am intuiting that the switches in the keyboard are "break/make" and BR stands for "break" and MK stands for "make". ie when the key is up, one switch is engaged (BR) and when pressed, it disengages and the second switch is engaged (MK).

Anyway there are two switches per key. I can see how one stuck switch in the entire array could cause trouble, and given you're dealing with a 20 year old mechanical switch array, likely there are opportunities for it to crud up.

It looks to me that the dynascan looks for the combination of T and BR/MK lines to determine what notes are being pressed. Notice that there are extra BR/MK lines to the left (pins 18-24) tied high, looks like the chip can address 24 more notes (to get to 88 keys...)

So I'm suspecting that there's a stuck switch up in the keyboard, either not making, or not breaking. Visual inspection, if possible, would verify it. Certainly you should be able to find it with a meter (disconnect the keyboard at the connectors to the dynascan and measure each BR/MK versus the T lines)

Maybe somebody else can add more or knows an easier way to troubleshoot, but that would be my next step.

ElecticGeek · Joined: Aug 16, 2008 Posts: 7 Location: Melbourne, Florida

I appreciate the excellent explanation of the schematic. I believe I can now troubleshoot the keyboard itself. If there's a key stuck open or closed I should be able to find it.

I'll resume troubleshooting probably tomorrow. I've got real work to do now to pay my bills!

Thanks again..

ElecticGeek · Joined: Aug 16, 2008 Posts: 7 Location: Melbourne, Florida

I had a chance to check the MIDI output from the Roland D-50 and it worked but was not quite right. It sounded like some note-on events were being sent multiple times or something. I didn't have a chance to use a MIDI display program to see exactly what was being transmitted. Interestingly, some of the keys were transmitted intermittently (i.e. with drop outs when a key was held down), such as: Bb2, F#3, D4, Bb4, F#5, D6, Bb6. This problem suggests a keyboard decoding issue perhaps?

Today I had a block of time to disconnect the keyboard connectors and to "buzz out" all of the keyboard switches with a DMM. Because of the isolation diodes, the DMM lead polarity was critical.

Interestingly, the MKn and BRn notations on the schematic are somewhat misleading; all key switches are normally open. The BRn switch closes first when a key is pressed and the MKn switch closes when the key is pressed almost all the way. I believe the timing between the two closures is used to compute the key velocity. Anyway, all of the keys checked out perfectly.

So at this point, my suspicion continues to be focused upon the Dyna Scan board. I will reconnect the keyboard and use my scope to look at the qualities of the keyboard scanning lines to see if some of the scanning signals are messed up. I am hoping to see a pattern that explains the MIDI note transmission problems.

Any thoughts?

Thanks - Dave

EdisonRex · Posted: Mon Aug 25, 2008 2:23 pm Post subject:

ElecticGeek · Joined: Aug 16, 2008 Posts: 7 Location: Melbourne, Florida

Ok,

T2 is used to detect all of the notes I mentioned before. However, a check with a scope shows the T2 strobe to look pretty much the same as the other lines, with or without one of its corresponding keys pressed. The timing between T2 and T3 looks stable. About the only unexpected thing I noticed so far is that the normally stable T line scanning becomes erratic when a (any!) key is held down.

I may take a quick look now at the aftertouch circuit - this section failed the built-in self test. The problem with it could be totally unrelated, or not. As you probably know, multiple random failures at once are pretty unusual, unless a surge or other fault condition is involved.

Maybe I will get lucky.

EdisonRex · Posted: Mon Aug 25, 2008 3:33 pm Post subject:

hm

go for it. All info is good info.

niteowl · Joined: Nov 05, 2011 Posts: 2 Location: Texas

I had the same problem described here. After poking around and
verifying that the cabling and keyboard itself was okay, I narrowed
the problem to the dynascan board. I re-soldered all the pins
going from the Gate Array (63H149) to the SRAM (6116) and the
keyboard started working again.

It's likely that the D-50 was bumped a bit too hard and cracked one
of the solder joints on the 63H149.

Check out this for details:
http://music-electronics-forum.com/t21165-post175932/

EdisonRex · Posted: Sun Nov 06, 2011 9:11 am Post subject:

niteowl and thanks for the info!

niteowl · Joined: Nov 05, 2011 Posts: 2 Location: Texas

Well, my 'fix' lasted a day, and then the D-50 went back to its failure mode
of only sounding on key releases. Desperate, I removed the 6116 SRAM
from the Dyna-scan board. Oddly, without the 6116 the keyboard behaved
exactly the same. I ordered a replacement chip IDT 6116SA15TPG from
Avent.com and soldered it in. NOW it works again.