Vintage Synth Repair

On Saturday 19 April 2008 16:00, eggwheatis wrote:
> If it were me and I know nothing about this synth by the way...I'd
> start by looking at the PSU with a scope....

Yup,  I looked through the diagrams pointed at, and the behavior described can 
only be a matter of the last couple of them in there...

First thing I'd do is to check the power supplies.  The +5 needs to be within 
a quarter volt of +5 or things won't quite work right.  A scope may show 
*some* hash on the power supply pins,  depending on how good the actual 
design is,  and some of that may be normal.  Not if there's a lot though.

> then assuming this is perfect I'd be looking around the CPU to see what
> activity I have going on again with a scope or logic probe referring back to
> the schematics, reset lines..clock, ports etc etc. I'd then check the
> circuit around the rom and ram etc..chip selects etc etc write enable
> blah blah..theres so much it could be..you only need one dead node in
> the logic circuit and the whole thing won't work.

Yup. A scope is *much* better for this sort of thing than a logic probe.  I'd 
start out by scoping the data bus pins,  DO-D7 on the Z80.  Look for valid 
logic levels and transitions between them,  but not much time spent in the 
middle (though I have seen designs like the c64 where stuff regularly did 
spend short bits of time in the middle.  Nanoseconds...

If there's a bad chip on the bus somewhere and it's messing things up to give 
the symptoms described this will be one way to find it.

Next is scope the address bus pins.  Depending on where the software is taking 
things there may or may not be activity on all of them.  A logic probe with 
the ability to "catch" very short pulses might be handy here as well.

Here's a trick you can try with that Z80.  Temporarily solder a bit of wire 
across all of the data bus pins and to a ground point somewhere.  This is a 
whole lot easier if there's a bus buffer in the system,  but the diagram 
doesn't show one.  What this does is it feeds the chip NOP instructions,  and 
it just keeps on cycling through _all_ of the addresses and keeps 
on "fetching" them.   You should then see a square wave on each address bus 
pin,  with each time you go to the next higher bit the frequency is going to 
be half of what it was.  If you don't see this somewhere then there's a 
problem,  as something ks making one of those lines get "stuck".

I don't know these instruments and don't know what's likely to be socketed or 
not.  From what I can see you should be able to remove the 68B50 and maybe 
the 8253 as well,  as the former is just for MIDI and I'm not sure what 
they're doing with the latter.  You should also be able to remove one or both 
RAM chips and see if there's any change in behavior -- if there is then you 
have a problem pinpointed.

If you don't find anything in particular this way,  start looking at the 
outputs of the decoders,  the memory address decoder especially -- a "stuck" 
output on any of the decoding circuitry can cause some bus contention,  check 
the outputs of the 74LS138s and also the outputs of the gates that connect to 
those outputs,  and figure out what they _should_ be.  Like that ROM select 
line shouldn't be low _all_ the time unless the CPU never gets there.  
The "feed it NOPs" trick above will help you get there.

> I'd say the rom is ok if you still have a problem with the test rom..

Maybe.  Or maybe there's a fault which has done something to one of the bus 
lines and has ended up trashing them both.  One hopes not.  :-)

> I don't see any benefit in bypassing the digital side even if it were
> possible..what with the garbage on the screen there is quite clearly a
> logic problem.

Yes,  and in an instrument like this I don't see how it's possible,  or at 
least not practical.

-- 
Member of the toughest, meanest, deadliest, most unrelenting -- and
ablest -- form of life in this section of space,  a critter that can
be killed but can't be tamed.  --Robert A. Heinlein, "The Puppet Masters"
-
Information is more dangerous than cannon to a society ruled by lies. --James 
M Dakin