So when I engage poly 8 only and work the keys from left to right I will run into a key that makes no sound. Eight keys later that key will make no sound but the original key will now work. It does that pattern all the way down the keyboard. every 8 keys. there are alot of 8 bit chips in there and think it may be related to that. It does some kind of keyboard scan in groups of 8 that I read.
I read that the CD4051 8 channel multiplexer chips give all kinds of problems. It also looks like the keyboard is processed in 8 key sections through LS138 Chip(s). Also here of problems with Nicad battery retention circuit leaking and causing some strange problems. At this point just trying to figure out how to get inside the thing. Service manual doesn't cover tear down instructions.
So the unit has 8 voices and one is dead. It isn't the keys, they all work. The key assigner just "assigns" each new keystroke to the next available voice. SO you should find eight voice circuits all alike. Somewhere between the key assigner output to the one bad voice to its output to the signal mixers, it isn't working.
A 4051 might be involved, but I have no knowledge of them being more problematical than anything else in there.
A 4051 might be involved, but I have no knowledge of them being more problematical than anything else in there.
So I have located blk diagrams on location of Channels(Voices) on pg.12. Circuit diagram on pg13-14. Reading of program on pg 32 describes circuit operation of voices and other associated circuits. So I first need to get this thing open and will get going. I will attach SM so you can see what I'm talking about. Doh! To big to attach. I'll see if I can PM you.
See page 9.
> Is the Voice a chip?
Could be, but isn't.
Sorry for recap-
An oboe has one reed-set and one filter-set. It is monophonic. The 1st-generation Moog/ARP had only a few oscillators and filters, and were commonly played monophonic per keyboard (ARP had a trick to work 2-voice on one keyboard).
A piano or organ has (at least) a set of resonators and filters for every key. A "voice". They are normally played polyphonic.
When this synth was designed, a "voice" was fairly expensive (in part because a voice had not been reduced to a chip, not to mention a universal chip is not totally satisfactory). Instead of 88 or 40 or 25 voices one-per-key, they noticed most people have just 10 fingers and do not use all of them; when they do lay 10 fingers down the sound is so complex that you could leave a few out and most listeners would not complain. 8 is a nice binary number. They made 8 voices and work them like a temp-agency. When a finger lands, a master brain finds an idle voice and assigns it to this job. More fingers, more voices assigned.
These voices are not "human voices". Just because you assign them a job does not mean they will show up. Aside from the too-many-keys problem, there is the problem that after 30 years a voice may crap out. A human temp-agency would notice and send another worker. This system doesn't.
As you say the 8 voices may be all on one board. But identifying the dud one is too much for my tired brain.
> Is the Voice a chip?
Could be, but isn't.
Sorry for recap-
An oboe has one reed-set and one filter-set. It is monophonic. The 1st-generation Moog/ARP had only a few oscillators and filters, and were commonly played monophonic per keyboard (ARP had a trick to work 2-voice on one keyboard).
A piano or organ has (at least) a set of resonators and filters for every key. A "voice". They are normally played polyphonic.
When this synth was designed, a "voice" was fairly expensive (in part because a voice had not been reduced to a chip, not to mention a universal chip is not totally satisfactory). Instead of 88 or 40 or 25 voices one-per-key, they noticed most people have just 10 fingers and do not use all of them; when they do lay 10 fingers down the sound is so complex that you could leave a few out and most listeners would not complain. 8 is a nice binary number. They made 8 voices and work them like a temp-agency. When a finger lands, a master brain finds an idle voice and assigns it to this job. More fingers, more voices assigned.
These voices are not "human voices". Just because you assign them a job does not mean they will show up. Aside from the too-many-keys problem, there is the problem that after 30 years a voice may crap out. A human temp-agency would notice and send another worker. This system doesn't.
As you say the 8 voices may be all on one board. But identifying the dud one is too much for my tired brain.
Attachments
I recall some old Oberheims that had LEDs at each voice circuit, so you could just look and see each voice being assigned. Too bad we don't have that here. But it is generally clear from the schematics where the trigger for each voice comes from, then a scope or even a logic probe can check that point on each to see when it is assigned. And the one that gets an assign pulse but no sound results is your dead voice.
SO the dead voice...
isn't getting its assign trigger
waveform generator doesn't,
VCF and VCA stages getting faulty input from envelope filters.
etc til we come out the other end.
SO the dead voice...
isn't getting its assign trigger
waveform generator doesn't,
VCF and VCA stages getting faulty input from envelope filters.
etc til we come out the other end.
Ok thanks guys. I think I have found some dead voices. Channels per the manual terminology. It started as just one voice only when the poly 8 button was on. I noticed now that it does it also with the poly 4 button also. I can see at least 3 voices that do not look like they aren't doing anything compared to the ones that are working. A couple of opamps and a SSI2044 chip. So in the morning going to start backtracking.
So looking at page 14 in SM. I'm assuming that the the connector with SYN42 and G0-G7 are the voice assignment inputs?? Out of curiosity why do you feed the output of an and gate into the input of an opamp. Is that some kind of threshold detect circuit. Also do you know what the "EG" designator means.
So I monitored the output of each voice and found three that do not seem to work. A real world observation is that playing across KB it aligns with three keys from group of 8 that do not tone. It also shows up that 2tones fail and then the next sounds and then the next fails and this aligns with the physical positioning of the failing voices with one that works in between. So I monitored the G0-7 and verified that there is a control pulse or square wave to assign all 8 voices like it should. Also looked at the input from DCO's to verify each voice is being commanded to sound out and they are. Can also see the pulse string increment and decrement as you go up and down in frequency as it should. So that means the controls section feeding each voice is intact and that the problem lies within the analog signal processing for the voice. So next tasks are to troubleshoot through the opamps, 4 pole filter and switching that makes up each channel. Absolutely fascinating that engineers designed something in the analog domain that works like this. A very sophisticated system from back in trhe day.
Ok a long overdue update. So this has been a head scratcher for sure. The good news is that it appears that I have all the voices back by replacing with the new version SSI 2144 VCF. When I replaced the last one it did not get the voice back. Looking further into it the sawtooth gen and waveform converter IC was hot along with the transistor QA in feedback loop. I replaced the opamp with a LM4562 that I had along with a new transistor. now the voice works but it does have a bit different tone than the other channels. I also notice that the opamp gets hot too. I have looked for oscillation on scope but do not see it. I went ahead and added bypass caps to each power supply to ground right at the opamp power legs and it still gets hot. I also disconnected the output pins, lifted them from the socket so there was no connection to circuit. the chip still runs warm but not so hot about 125 F. I then lifted all the input and output legs with just power to chip and it still runs warm. I have tried multiple opamps and they all get hot. Not sure at this point how to troubleshoot this opamp. I have service manual and the analog channel is page 14. Upper left corner IC opamp A.