Component Replacement in Vintage Synth Repair

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

thermionic

Well-known member
Joined
Jun 3, 2004
Messages
1,671
Hi,

Over the 20-odd years I’ve been DIY-ing, I’ve built many amps / preamps / EQs / PSUs etc. However, the one thing I’ve never tackled is synths…

I have some decent test gear + tools, including a digital Weller iron + De-non desoldering station (ideal for DIL ICs and caps).

A couple of years back, I picked up a Prophet 5 rev 2 at a bargain price. I was told it didn’t work, and I ushered it away before the owner could change his mind. The synth has been living in storage ever since and I haven’t attempted to turn it on. This may sound cautious (it is), but my moral is that it should be easy to replace the obvious perishables, i.e. electrolytic caps, look for any visual damage, and then try and turn it on via variac… I think this makes sense as the PSU in the P5 is notorious for problems (some people cite its inadequacy as being part of the early P5’s sound), and if it hasn’t been on for years, turning it on with caps that pass DC could do a lot of damage.

Anyway, according to my synth geek friends, this guy knows his stuff, and he’s published an article here on his website: http://www.oldcrows.net/~oldcrow/synth/tips.txt

Now, I know that replacing all the electro caps will appear conservative to some of you – you might argue to replace them only if you can tell they’re gone, but if you read the link, he recommends going way further than just replacing the wet caps…

Adding additional PSU bypass caps to ICs… A good move IMHO, if they’re not there in the first place (I’ve seen what op-amps do without bypassing, it surprises me that the synths lack them…)

Replacing all 4000-series ICs on the spot… The reasoning makes sense, so maybe I’m missing a point here, but if an IC has worked ok for 30-odd years, isn’t there an argument to let sleeping dogs lie?

Swapping out all tants for electrolytic caps Tants do get leaky after a few years use and can fail in a multitude of ways (I’ve seen burning ones fly across the room…), but surely, if the original OEM specified tants when they could’ve fitted electrolytics, couldn’t one argue the tants are there for a reason? My impression was that tants were developed to get a lot of capacity into places where electrolytics wouldn’t fit, so I can’t see an issue, aside from authenticity.

I don’t want to appear to be doubting the engineer who wrote the link – I’m sure he knows much more than I do about synths, but replace every 4000-series IC? Am I naïve here, or is this uber-conservative?

Thanks in advance for your thoughts,

Justin

edit - I hope this is in the right forum... If the moderator wants to move the thread, I won't be offended - I do think some of the questions relate to 'design', so I posted here.
 
4000 stuff only if it bad and if you know how to work with static sensitive stuff.

Some of the 4000 stuff gets hurt easy plus you have no idea what shape or how the the chips were handled before you get them. You might do all you can to be static safe but the chips are an unknown. Do you have a mat and wrist strap and know how to handle static sensitive chips?

Learn logic and use a scope.

Tants I don't know if I would change them for Al often they fail as a short and you can sometimes find them with a meter the problem is if they are on the power lines then you would need to pull them one by one to find what one(s).

Machine sockets the chip WILL walk out with time temp changes and vibration and the metals are dissimilar they do allow multable remove and replace. The cheaper ones often hold the chips better and have similar metal. This is from more than two decades of fixing stuff. Some chip legs will tarnish in gold machine sockets.

15 year life is that 15 years or 15 year on time?
 
I too question the wisdom of arbitrarily replacing 4000 series chips if they are working fine. If worried about obscure parts going obsolete, buy some to have on hand JIC.

Tantalums were often specified for low leakage and low HF impedance, more than density in those old designs. While you could probably get equivalent leakage performance from modern low leakage aluminum electrolytic, and decent HF impedance from caps designed for use in switchers, there are other small differences.

If the capacitor is used in an envelope circuit where it is charged and discharged at different rates, the DA (dielectric absorption) "could" make a difference. This will be extremely subtle but to be true to the original sound character perhaps a consideration. For power supply decoupling replacing tantalum with aluminum and a ceramic disc is probably OK.
------

I once used a tantalum capacitor in a time constant circuit for an old consumer vinyl NR system (CX) when I learned that the master encoder circuit used a tantalum in that location to master the records. Were I involved in the encoder design I would have tried to mirror the typical playback path (cheap aluminum electro) more closely. The medium failed for any number of other good reasons, RIP.

JR
 
Thanks :guinness:

I once worked at a firm where you would lose your job if you failed to test your wriststrap and sign a declaration every morning!

Replacing PSU bypass tants with electrolytic makes sense.

It's a pertinent warning to stock up on obsolete ICs...must make a note. I'm glad I'm not the only one who finds the idea of replacing all the 4000-series CMOS arbitrarily somewhat OTT.

BTW - Dolby A cards use tants for coupling, don't they? I know someone who converts them into a limiter! I don't think he'll be forthcoming with a schematic, but I'm sure the Brains Trust here could come up with something.

Justin
 
I can remember a post here about pro's and cons about tant's.
If I remember correctly tantalum capacitors work fine even for longer periods but get leaky when you put them in places where they're reverse-biased even for short periods of time, for example during power-up when one side of a bipolar psu ramps up faster than the other side or on the peaks of an audio signal.

See PRR's post here:
http://www.groupdiy.com/index.php?topic=3317&highlight=tantalum+reverse+bias

Search a bit. I can remember someone posting more about this.

Some people here replaced the tantalum coupling capacitors on the inputs of their pm1000 channelstrips with elco's and even film caps and didn't like it so tantalums certainly have a 'sound' :grin: .
 
[quote author="Mendelt"]I can remember a post here about pro's and cons about tant's.
<snip>
Some people here replaced the tantalum coupling capacitors on the inputs of their pm1000 channelstrips with elco's and even film caps and didn't like it so tantalums certainly have a 'sound' :grin: .[/quote]

Dunno about a 'sound' but they sure as hell have a 'smell'. A client has a small Neve mixer which was expiring smellily every few months with yet another shorted tant. After a few visits we decided to ditch the lot (hundreds of them) and replace with ali caps. No complaints after that and no reports of changed sound. Perhaps tantalum was best used for the anodes of type 833A tubes
M
 
Indeed tantalums don't fail gracefully. Perhaps the newer low impedance aluminum electrolytic designed for switching supplies may make a fair drop in for PS decoupling. Perhaps ironically I've seen several short circuit failures in those low impedance aluminum caps too...

IMO if the cap expresses a sound, the circuit could be better designed to be less sensitive to whatever error the cap is suffering from. Perhaps easier to say than do.

JR
 
Bumpa de bump...

I finally got enough time away from the daily grind to get the P5 open...

Some of the sights that greet you when you open a P5 are beyond mind-boggling... 8-pin ICs glued on top of one another. Most scary of all: packages potted in black compound, amalgamated into the tops of 14-pin IC packages, with numerous passives and leads emanating from them to other places...Heath Robinson would be an understatement!

(I would like to make it clear that I hold SCI in absolute highest regard - one of the most innovative synth OEMs ever, but jeepers, these things really were hand-built)

So, anyway... Following John R's advice, I have replaced all the tants with like on the Voice + Logic boards (it won't make a difference if I sub higher voltage caps - as long as the capacitance is the same, right?)

The PSU is due to have all its tants replaced (I found a shorted one, which *could* explain the P5's problem...fingers crossed...). Each Vreg seems to have a 2.2uF tant to ground on its output leg. I'm thinking that a 10uF electrolytic will be fine here?? I won't go to town with bigger sizes, but that won't stress things too much from what I can figure.

In the link in the 1st post, the author suggests replacing all Mylar film caps with poly caps on P5s... Is this really necessary? There are hundreds of the blighters - what's the problem with mylar caps? I'd prefer to have a life if possible...


Thanks in advance,
Justin
 
[quote author="PRR"]> I'd prefer to have a life if possible...

Then get out of synth repair.[/quote]

Ain't that the truth! Baptism of fire would be an accurate description. The only Synth Tech I trust is around 400 miles from me... Considering that the fault *could* be in the PSU (I haven't powered it up yet...but a PSU tant *did* have a short), it didn't make sense to lug a rather fragile old synth all that way if there was a chance I could do the repair myself. Not only that, but Tony-the-synth-guru is booked up until August...

I still can't believe how a firm such as SCI could have such a deep knowledge of electronics as to design the audio and control circuitry, yet ignore many of the basic principles even a DIY-er would observe in regards to the PSU and general construction - I guess they just weren't turned on by the more mundane aspects to gear design.

If I followed all the advice in that link in the 1st post, I'd pretty much be reworking the entire Prophet! It's 2-layer PCBs throughout, so fun, it would not be.


Justin
 
Ok - FWIW...

After replacing all tants and battery, I powered it up gently with variac.

At about 60v across the primary, most of the switches light up. As soon as I wind it up, the lights go out and there is nothing - absolutely nothing.

I can't get a 'scope probe in to look at the PSU under load... All waveforms look ok with no load. The "+30v" rail drops from +28 (no load) to around +24v under load. The other rails stay the same. Nothing smells / smokes.

I found 2 shorted tants. Is there a chance the one in the PSU could damage a Vreg? I thought monolithic Vregs were short-circuit protected... Aside from the +30v rail dropping (I suspect all early P5s do this), the PSU appears ok, although - as previously said - I have yet to see rails under load.

I don't know what the hell to do to be honest... Tony the guru is in Scotland - it'll cost several hundred pounds just to get the P5 there and back, let alone the wait until August...

:mad: :cry:

Justin

edit - I've been told that the machine won't boot up at all if an SSM2020 in the VCO section is faulty... If the LEDs come on when the voltage across the primary is 60v, that suggests the 7805 is coming up, but then shutting down??

Why am I measuring +28v from the LM320T (no load - waveform very clean unloaded), when the manual says it should output +20v? Do they drop out of regulation and go over-voltage?

Either, when I ramp the variac, a circuit is 'waking up', registers a faulty SSM2020 (for example) and shuts things down - or - one of the Vregs is dropping out of regulation under load, which explains the lights coming on and then going off????
 
This is from the lm7805 datasheet on national semiconductor's site..

Safe area
protection for the output transistors is provided to limit internal
power dissipation. If internal power dissipation becomes
too high for the heat sinking provided, the thermal shutdown
circuit takes over preventing the IC from overheating.

Looks like you're playing find the short.

Check to see if they're getting hot.
 
[quote author="PRR"]> I'd prefer to have a life if possible...

Then get out of synth repair.[/quote]

I've been doing it for well over 20 years, it can be a difficult vocation sometimes in a country with a small population. I only do "vintage" gear, but the few other service centers down here do all kinds of new warranty stuff as well to keep them going, I guess they have to, but I just don't have any interest in it.
 
I don't think that it is a good idea to replace all CMOS under normal circumstances, particularly in a big polysynth. As for the bypass caps, Old Crow is mostly referring to the Yamaha CS-80, because Yamaha did not put any bypass caps inside this machine. I sold one of my CS-80's a couple of years ago, and before I did I swapped out all the CMOS and added the caps, and I am about to do the same thing on the other CS-80 which I kept, and there is another one coming in soon for the same treatment.

As for other machines, particularly ones with lots if non-socketed CMOS, I suggest that unless you are doing critical live work where you need the machine to be as absolutely reliable as possible, then just leave it be and let the chips live out their lives and just replace them as they fail.

I fix rev 3 Prophets every couple of months, rev 2's come in much less frequently, and when they do they usually are smoking (in the bad way).

Last one had a shorted EPROM of all things. I ALWAYS replace ALL tantalum capacitors in old machines as a matter of course, particularly in machines that they are very prone to failure, ie; all ARP synths and AMS reverbs and delays.

I usually put tantalums back in, if the originals lasted 20-30 years then I have no problem putting the same in again. Sometimes there is an exception, such as the audio decoupling capacitor on ARP filter boards, where I do replace the tant with an electro. I replace electro's if the customer can afford it, and if I think it is important such as if it is in a filter circuit.

Replacing all the tantalums was a good idea, do you have a schematic? Be very careful of voltage marking in Prophets, they are often wrong, PSU boards were often modded and interchanged without the markings being changed to reflect the new voltages. According to the PSU schematic there are: +22, +15, +12, +5, -5 and -15.
 
Does this look like it may be the correct drawing?

psu2.jpg
 

Latest posts

Back
Top