AC irregularity on a DC control voltage - Peavey PC1600

[a.d.a.m. baby]

My PC1600 midi fader box has begun playing up. Specifically the faders are displaying scrambled values. The other data entry methods - switches and an encoder wheel - are behaving normally. It's been on the bench for a few days while I've been scoping it out - to cut to the chase, I've found a 50 hz "dip" in the CV coming from each fader. Basically the faders are voltage dividers of a 5 volt rail. These voltages are then passed to the mux chips (4051's). Today on powering up the dips in this divided voltage were absent and lo and behold - the faders behaved normally! After a minute, the problem returned, and the dips had returned.
So I believe I've found the problem behaviour - can someone suggest a possible cause?
Power is supplied from an external 16 volt/1 amp step down AC supply (Swapped out - no change). The 5 and 12 volt regulators and the filter caps/diode bridge seem to be displaying the expected voltages and rectified waveforms - moving the faders doesn't cause any dips there.
My feeling is that this is PS related - but could it be elsewhere?

[Futureman]

Not being an expert, I have no advice.

Are you sure your PS has enough Juice (Amps) to not sag under load?

You could replace the electrolytic caps in the PS for the hell of it.. maybe they have gone south?

When you see the 50hz 'Dip', is your earth on your scope connected to the earth on the Peavy chasis? What I'm trying to say is that if not, you might just be seeing the 50hz 'all around us' sort of thing on your scope.

Hell, the 4051's are prob about $2 jaycar, It couldn't hurt to replace them too.

Advice from a non expert.
Mike

[no-fi]

never seen inside one, but look for a series of diodes or a bridge rectifier near the PSU input. change it/them.
also, look for big electrolytic caps near the PSU input. change them.

[rob]

yep, chances are it is the main filter cap(s), prior to the 5V regulator IC, swap any caps immediately after the regulator whilst you are at it.

[davek]

I'm more inclined to suggest dirty faders/pots are what could be causing the random fader values... Have seen it numerous tiems before in various pieces of gear where a 0-5v voltage is fed into a A/D converter for controlling a digital fader, etc..

What is the amplitude of the 50HZ ripple on the output side of the DC regulator .. ? A dirty power supply would cause more that just the fader values to go screwy.

[a.d.a.m. baby]

Many thanks for the input all.
Each of the 16 faders exhibits the exact same amplitude dip, so I thought a physical problem such as contamination would be less likely.
Are there any in-circuit tests I can do to check whether the filter caps are up to spec? It seems a bit shotgun to just go replacing them willynilly..!!

[rob]

an ESR meter would likely identify bad caps, but not always

since you have noted ripple on the 5V rail, this would cause your problem. Since the sliders are being scanned and their DC level which will vary from 0V to 5V as the slider is taken from one extreme to another, the ripple will cause the scanned value to change depending on the timing of the scan and the point in the ripple that exists as the reading is taken. This would appear to generate a semi random value.

If you can spot the 5V regulator IC, get a data sheet for it ( possibly it is common variety 7805 ), identify the input leg and trace back to where it comes from. You should hit a largish value cap pretty quickly. Perhaps in the order of 1000uF to 10,000uF and say 16V. Note the orientation of this sucker, pull it out and put a fresh one in.

then see what happens.

[a.d.a.m. baby]

What is the amplitude of the 50HZ ripple on the output side of the DC regulator .. ? A dirty power supply would cause more that just the fader values to go screwy.

Good point - I thought the external transformer was running way too hot, and may have a shorted turn, causing current loss, however replacing it with a higher specced AC supply (1.5 amps) made no difference. There is no discernable ripple at the regulator (at least on my 40 year old scope!!).

And the regulator volts are absolutely fine - however I haven't done any current tests, which of course are more of a PITA, and the normal or expected current values would be unknown.
I should add I can't find the service docs/schematics anywhere either, so what I've outlined is based on observation and in-circuit tests.

[rick]

how any of these things are you going to fix in your lifetime ?
i am all for the shotgun approach when it comes to 30 cent components
its a much better use of your time usually
unless your time is worth 30 cents an hour

[a.d.a.m. baby]

an ESR meter would likely identify bad caps, but not always

since you have noted ripple on the 5V rail, this would cause your problem. Since the sliders are being scanned and their DC level which will vary from 0V to 5V as the slider is taken from one extreme to another, the ripple will cause the scanned value to change depending on the timing of the scan and the point in the ripple that exists as the reading is taken. This would appear to generate a semi random value.

That's exactly what I thought - I can see the random effects on the out put of the 4051s.

If you can spot the 5V regulator IC, get a data sheet for it ( possibly it is common variety 7805 ), identify the input leg and trace back to where it comes from. You should hit a largish value cap pretty quickly. Perhaps in the order of 1000uF to 10,000uF and say 16V. Note the orientation of this sucker, pull it out and put a fresh one in.

then see what happens.

Yep, it's a 7805. I'll do as you say and get back to you,

cheers
Adam

[a.d.a.m. baby]

how any of these things are you going to fix in your lifetime ?
i am all for the shotgun approach when it comes to 30 cent components
its a much better use of your time usually
unless your time is worth 30 cents an hour

I seem to be replacing filter caps every week with all the old crap in this studio!
I meant it only from the point of view of furthering my education with this stuff - of course, often a bit of trial and error is the only way forward..

Anyways, there's two 470uF/25v electros as filters, but I only have 1000uF/35v. Any big deal if I use them? Obviously bigger current draw on power up, but any negatives otherwise?

Incidentally, I 've just noticed that there is a lot of ripple on the input leg of the 7805 - I was looking at the outputs...

[no-fi]

AC ripple in your faders was why I was telling you to look at the power input circuits... ie, diodes and caps....

The reason why you're finding you have a lot of big PSU caps to replace is that in pretty much all gear, these are the parts that have massive current movement in and out of them 50 times a second... so they get hot and dry out and lose capacitance and stop doing their jobs way before anything else craps out.


I'd be tempted to try the bigger caps in place of the old ones, as long as they fit. Those peavey power supplies are pretty damn beefy. I doubt they'll complain about a bit of inrush current. One thing - everything else the same, bigger value caps tend to have worse ESR values, so I expect they will heat up and die faster than the same series parts with correct capacitance, but we're talking you having another 10 years good working time with any old basic electros anyway.... so maybe not an issue for you.

if you're getting replacement caps especially though, go for lowest ESR and high temp rated caps in this position. they cost a bit more but will last longer.

[a.d.a.m. baby]

Excellent, thanks Julian, I'll give it a shot.

[a.d.a.m. baby]

Thanks to all for your time and expertise - problem solved.
It was a power diode.
I can't see any major difference in the before-and-after voltages or ripple, so maybe a shotgun approach is better in these situations - shoot the PSU first, ask questions later.

BTW Rob, I just read your "Mr (Can't) Fixit" article. It reminded me of when recently my MOTU 828 mk2 went belly up after a few dummy spits. I took a look inside - voltages seemed OK, but with all the big chips in there and SMDs, it was out of my league. Took it to Musiclab in Brissie (who used to fix my old 8 track and loads of stuff - very experienced) to be told that MOTU don't make the schemas/service docs available - to anyone!! Not even these guys!!?? Their policy is to swap it for a good one.
Luckily the oz agent had one in stock. What happens next time?
I hate software, and now hardware is coming to this!
I might take up gardening.

[no-fi]

I'm not sure 100% how diodes die (cooked outright from overheating from too much current I suppose??) but yeah I've seen a few dead PSU rectifier diodes in my small amount of poking about in malfunctioning gear.

so - did you swap the caps as well?? may as well.. they'll need it soon enough anyway.
:-)

[Futureman]

I have an Oberheim DMX that didn't work when I got it, Was a busted rectifier... Nice when it's a $0.30 repair.

[a.d.a.m. baby]

Yeah I put the caps in first, and when that didn't fix it I moved on to the diodes.

And now ha ha it seems I may have a RAM problem - when trying to load sysex into it to restore the old presets, I get "error in edit buffer"... Nice. I seem to be able to load in one preset at a time however, so that will just have to do!

[no-fi]

that sounds like a weird problem to develop from fixing a PSU circuit.... have you ever had bulk receive working with it before?? maybe it's a data overload thing? not enough time between presets, or something.

[a.d.a.m. baby]

Yeah it's odd - never had a bulk sysex problem that I can recall.. had it since '92!
But I did do a factory ROM reset in the course of things, so maybe I need to go in and tweak some data packet times or something.. there's a bit of finessing possible.