View source for GS1restoration

== GS1 Restoration Page! ==

This is a page for keeping track of the project of restoring the CCRMA Yamaha GS1 prototype.

Will keep track of info here as the project progresses.

First, the initial state of things:
The console unit connects to the GS1 to do the initial programming of the 16 presets on the unit. The console is one of the coolest looking pieces of electronics I've ever seen.
There are four monochrome green screens on the console. When I first turned it on two of them had no horizontal sync. This was very easy to fix by opening up the unit and adjusting the pots. Note that this still happens sometimes when you first turn it on but it clears up after things warm up a bit.

The console connects to the keyboard with a ribbon cable. There are three connectors on the console where the cable can plug in. I was concerned about getting this wrong so I opened up the unit to see which was which. Turns out the two that are right next to each other aren't connected to anything at all. So, it's the one by itself, you can't mess it up.

The cable has a broken connector on one end. One must be extremely careful when unplugging and replugging the unit. Note that one of the lines on this connector was shorted to the housing when I first unpacked it. I haven't found where this connects yet, but that part, whatever it is, is worthy of come scruitiny.

The connector is a centronics type 24 pin, like a 24 pin version of the 36pin printer connector. I saw it in digikey and it's NOT called centronics. We should order a set to make a nicer cable that is longer and possibly shielded.

Now, about the unit itself...

The keyboard, when it's first powered up, is in some kind of undefined digital state. All the keys seem to produce random noise.
If you reset the console, or even press ANY key on the console, the keyboard goes back to a default tone that is pretty nice. It plays on all the keys except the bottom two octaves. THe bottom octaves seem to be in order but have a very strange collection of tones. THis may be a data bus error where some important bit is not getting through. Not sure. This particular problem seems reasonable to solve. But, this must come after the reset problem.

Previous work on the unit shows that a fuse has been improperly replaced in some kind of butcher job. We'll fix that soon enough. Next, it looks like someone had soldered a kynar wire between the two bus tranceivers on the CPU board. One end was off. By the look of the solder, it seems that someone was jumping 5V to the ~EN line on the adjacent chip. Seems like they were troubleshooting it. This is fine except that it may have overloaded whatever drives the ~EN line.

First things first, the power supply needed to be checked. THis one is odd. It outputs. +15, -15, -12, and -7. The 5V for the logic appears to be using the -12 and -7 lines. SO the logic ground is on the -12 line. (This is still not confirmed.) Anyway it appears that the voltage across the main CPU chip is 4.49 volts. This seems low. I am not sure what the tolerances are for these parts. It might not be a problem. But, I like to see something a lot closer to 5V. This could be caused by loading. If there is a TTL logic short somewhere that would probably cause this. It seems likely based on the other problems.

I checked all the clocks. They are very interesting. The main clock is an LC circuit and it makes a 2.22 MHZ clock. There are some gates to divide this, and there is another clock at 1.1MHZ. These seem to be working. The vibrato seems to feed right into the main clock! It modulates the whole thing. That's kind of neat. There are also some 300KHz and such clocks that also seem to go all over the system. I checked these and they looked ok, but I got nervous hooking the scope to this thing because of the ground reference. It seemed to mess up the unit a little. Not sure what the offset is between the grounds. There does not seem to be a ground reference on the power supply, so it should not matter too much, but I'm going to have to watch this. Maybe plug the unit into an isolation transformer just to be safe.

The main clock on the CPU seems to be working. It's a 6MHz crystal. It looks like it's been worked on, at least the loading caps have been changed. It is making a nice 6MHz signal. Also, it seems that the selector switches on the front of the unit are connected to the CPU, and it is the CPU that changes the lights when the buttons are pressed. I have not confirmed this. If this is the case it means that the indicator lights change because the software is running properly.

I don't know much about the MPX board. But there is a rc oscillator connected though a comparator. The comparator's output is fed to it's input using the RC rig for a cheap clock source. This is IC8 on the MPX board. This shows evidence of having been resoldered and I have not confirmed that it is working at this time.
 
'''The plan going forward:'''

So, my first plan is to find out why the system does not come up in a reasonable reset state, and why the programming console can reset it when it can't reset itself. The first thing is to get the system to turn on in the reset state without the console.

Next order of business will be to find out why the lower octaves are acting strangely. This will involve making sure those keys are talking to the keyboard unit properly, or at least not significantly differently than the others.
The next trick will be to find out where the beginnings and ends of the data busses are. Then, one bit at a time, check the line at the source and at the destination and make sure they are doing the same things. 

After that, we should be able to play the whole range of keys. Then it will be interesting to see if the console is actually changing the voices, I have not confirmed that yet.

Another interesting thing is the magnetic strip reader. Right now it clicks when the unit powers on. It does not seem to sense when a strip is inserted. Nothing happens. If a strip is in when the power is put on, the motor in the reader runs continuously, but it still does nothing. Not a priority at this time but I will have to take the unit out and confirm that the various parts are all working, then see what the main board is doing. That will have to wait though.

'''Wish List'''

So the wish list would be:
* A complete schematic. The last row of pages for the schematic is missing. That's a problem.
* More docs on the chips in the system. I have great pin diagrams but I don't have any info like, if you send this, the chip will do that. That kind of info would be very helpful.
* A logic analyser! I might break down and buy one but they're pretty pricey! And that's for a lousy one. Might be necessary for this project.