Linear Power Supply outputting AC only! ACK!

[jaunty]

I beg for your mercy -this has been driving me absolutely BONKERS.

I have a synthesizer I've mostly resurrected from the dead (we're talking vintage here - analogue - 1970s style) - at any rate  - I got nearly EVERYTHING working except SOMETHING happened to the power supply I simply do NOT understand at all - I've been over everything... thinking maybe I messed up on the AC feed wire - or even that my DVM was bad (Tek TX1) or my test leads were messed up because NOTHING is making ANY sense to me.

Basically I'm only getting AC (about the right voltage but 60Hz - yes i'm in north america) out of the power supply. I figured once I'm past the diode bridge (and esp first set of reservoir caps) I'd be getting at least SOME DC and a bit of ripple.. but i'm getting ZERO DC.

I'm using the case ground as my ground reference for measurements - and the heat sink on the PS shows continuity to the case ground.

Can anyone give me a nudge in the right direction?? I've already checked the diodes of the diode bridge and they seem FINE - as do the reservoire/smoothing caps. Could this be some weird ground fault??

the second picture on this page shows precisely the power supply in question. thanks in advance...

http://www.keithrobertmurray.com/articles/arp-omni-rebuild.html

Jonathan

[David_AVD]

Are you sure the 0V wire of the power supply is connected to the case?  If not, your measurements will be invalid if using the case as the -ve point on the DMM.

[jaunty]

Triple and quadruple checked. I get about 13.5V @ 60Hz - no matter the ground reference - straight off the heat sink on the PS OR the case - or via the ground coming into the case.

I get the same reading off my lines that should be +5v,-5v, +15 and -15v... which says... what?

Are you sure the 0V wire of the power supply is connected to the case?  If not, your measurements will be invalid if using the case as the -ve point on the DMM.

[Brumby]

Is the heatsink actually connected to the circuit?  If the devices mounted on it are isolated, it might not have any clear connection to the circuitry.

I suggest you find a clear reference point on the circuit itself.  Only then can you be sure you are actually measuring the right thing.

[jaunty]

Yes sure.
the heat sink is bolted down through the bottom of the case. as mentioned - i tried multiple points for ground reference. It's totally befuddling. I checked continuity between the heat sink and the case ground and i get a measurement of 0.2 ohms.

if it's at all helpful - the DC measurement I get out of the thing starts at 0.5mv and then dies to zero over several seconds...

Is the heatsink actually connected to the circuit?  If the devices mounted on it are isolated, it might not have any clear connection to the circuitry.

I suggest you find a clear reference point on the circuit itself.  Only then can you be sure you are actually measuring the right thing.

[Joule Thief]

I would measure DC and AC voltage across the electrolytic caps in the power supply.

DC should be at least 3 - 4 volts above the expected rail voltage and the AC voltage content should be minimal - less than a few hundred millivolts max.

If you get elevated AC readings, suspect dried out electrolytic filter caps.

[jaunty]

thanks Joule Thief (ha - good name by the way) - do you mean 'across them' (as in a probe at each end - i presume not) or from cap output to gnd? If that's the case then yeah i'm getting only AC.

It's POSSIBLE - although it happened so suddenly - this condition - I'd be a bit surprised to see caps go bad within seconds... but I guess i'll swap them and go from there. I was planning on rebuilding the PS anyway... but its just irksome to not know exactly what's up...

I would measure DC and AC voltage across the electrolytic caps in the power supply.

DC should be at least 3 - 4 volts above the expected rail voltage and the AC voltage content should be minimal - less than a few hundred millivolts max.

If you get elevated AC readings, suspect dried out electrolytic filter caps.

[Joule Thief]

measure with DVM positive probe on the cap + lead and the DVM negative probe on the cap - lead.

at those points, any pulsating DC voltage would be filtered to a near steady state DC voltage.
you could be getting very high AC component if the circuitry following the filter caps are putting a high current demand on the caps- meaning components after the filter caps could be shorted / miswired.

are these electrolytic caps the original ones to this circuit board?

check the date code on the caps. if they are 20 - 30 years old, they could be "dryed out".

[Joule Thief]

looking closer at the image you pointed to for the power supply, the yellow cased electrolytic caps have a date code of 1977 and are highly suspect given your symptoms. The 2 smaller tantalum caps  just to the left of the yellow caps are also candidates for replacement. Their failure mode can be much more spectacular than the failed electrolytics.

[jaunty]

thanks for the measurement advice. makes good sense. The picture supplied was just one with the same topology as mine but NOT mine. Mine are all original equipment yes - and almost that old - but they were working just fine until recently. I replaced my tants with new tants (I know i know - but i kind of like them - at least for decoupling caps - as long as they have a well behave voltage source they are very well behaved)

looking closer at the image you pointed to for the power supply, the yellow cased electrolytic caps have a date code of 1977 and are highly suspect given your symptoms. The 2 smaller tantalum caps (blue "blobs") just to the left of the yellow caps are also candidates for replacement. Their failure mode can be much more spectacular than the failed electrolytics.

[jaunty]

seem to be getting a very strong 36 or so VDC off the caps - their condition is great (cosmetically) but are rotated around on the board into a position not allowing me to read their date code. WEIRD thing I have noticed is that - relative to the black Ov conductor all the other pin voltages are EXACTLY correct and totally in spec...

so gosh - maybe there IS a crazy ground problem somewhere... time to start snooping around....

thanks mucho for the help so far...

[David_AVD]

You seem to be missing the important point.  Don't use the chassis, heatsink or whatever as the DMM -ve.  Use the actual 0V point.

There's no reason the heatsink would definitely be connected the the PSU 0V.  In fact I would expect that they supply is floating with respect to mains earth / chassis.

[jaunty]

well thanks! oh okay - that's new to me actually - i always thought it was 'best practice' to use the ground point most central or far UPSTREAM as possible in order to avoid reading potentials that were referenced too far away trom something like a 'true ground'... don't ask me where i got that - I really can't recall...

You seem to be missing the important point.  Don't use the chassis, heatsink or whatever as the DMM -ve.  Use the actual 0V point.

There's no reason the heatsink would definitely be connected the the PSU 0V.  In fact I would expect that they supply is floating with respect to mains earth / chassis.

[Joule Thief]

well thanks! oh okay - that's new to me actually - i always thought it was 'best practice' to use the ground point most central or far UPSTREAM as possible in order to avoid reading potentials that were referenced too far away trom something like a 'true ground'... don't ask me where i got that - I really can't recall...

You seem to be missing the important point.  Don't use the chassis, heatsink or whatever as the DMM -ve.  Use the actual 0V point.

There's no reason the heatsink would definitely be connected the the PSU 0V.  In fact I would expect that they supply is floating with respect to mains earth / chassis.

and this is true for more subtle problems
we are still in the stage of getting to know each other, and synchronizing our troubleshooting philosophies to search for major faults

the 36 VDC sounds fine for primary unregulated voltages. i would check the AC voltage just so we can confidently move past the caps as being a major issue. check your dates on the caps and understand they are not getting any younger.  quality 1000uf caps can't be more than $2 ea. tops.

are there designated test points marked on the PCB for your regulated voltages?
if so, with your DVM ground lead connected to the negative lead of the electrolytic cap in the positive 35 VDC circuit (brown lead of the power transformer most likely).
measure all your regulated test points.

its 5 in the morning so i'm out of here in the next few minutes

good luck

[jaunty]

Thanks - very kind of you - the caps in question (the main filter caps at 1000µF) are actually identical to the ones on the web page above - I ordered some decent 1400µF caps rated at 63vdc instead of the 'spec' 1000µF at 50... so i guess they should be decent.

As for test points - no - sadly this one doesn't have any. I have a number of other keyboards i'll be needing to work on (more early 80s vintage) that DO however.

Yes - funny how even in a 'niche'community like this - certain cultural and linguistic barriers must be overcome.... well i thank you for your patience... i'm one of those long time hobbyistss - been screwing around with electronics since i had a radio shack 75 in 1 kit in the 1970s - but never got deep enough into it to really 'learn' beyond just soldering stuff up here and there - i did lots of electronics related stuff in in high school and university physics courses - but i'm not sure calculating the behavior of LCR circuits is really so useful when you're trying to troubleshoot analog audio circuits full of op amps and 4000 series logic gates. But I'm trying to make a run at it.

can you (or anyone) recommend a great book that might kind of fill in the gaps between the practical basics (which i have a good handle on) and the more theoretical stuff which isn't so useful for me (?)

and this is true for more subtle problems
we are still in the stage of getting to know each other, and synchronizing our troubleshooting philosophies to search for major faults

the 36 VDC sounds fine for primary unregulated voltages. i would check the AC voltage just so we can confidently move past the caps as being a major issue. check your dates on the caps and understand they are not getting any younger.  quality 1000uf caps can't be more than $2 ea. tops.

are there designated test points marked on the PCB for your regulated voltages?
if so, with your DVM ground lead connected to the negative lead of the electrolytic cap in the positive 35 VDC circuit, measure all your regulated test points.

its 5 in the morning so i'm out of here in the next few minutes

good luck

[Brumby]

You seem to be missing the important point.  Don't use the chassis, heatsink or whatever as the DMM -ve.  Use the actual 0V point.

There's no reason the heatsink would definitely be connected the the PSU 0V.  In fact I would expect that they supply is floating with respect to mains earth / chassis.

This is exactly the point I was (trying to) make - but I didn't know how to re-state it without sounding stroppy.

I'm glad we've got that sorted.

[jaunty]

Yes Brumby - thank you so much - it didn't quite hit home until the next guy said it - i can be a bit thick sometimes but I had an idea of what you meant.

A great lesson learned for me. It's amazing the bad habits we pick up and what an impediment they can be to success!

At any rate - I needed a good excuse to recap the board fully. All is well and operating well now.

You seem to be missing the important point.  Don't use the chassis, heatsink or whatever as the DMM -ve.  Use the actual 0V point.

There's no reason the heatsink would definitely be connected the the PSU 0V.  In fact I would expect that they supply is floating with respect to mains earth / chassis.

This is exactly the point I was (trying to) make - but I didn't know how to re-state it without sounding stroppy.

I'm glad we've got that sorted.

[Brumby]

All is well and operating well now.

That's what we like to hear. 

[David_AVD]

Good to hear that it "clicked".  I'll cancel the wet-fish-slap I had scheduled for you.   

[Joule Thief]

good work jaunty 

i was hoping to come back tonight and continue troubleshooting.

the entire issue was dried up caps?

you may want to consider purchasing one of the $20 component testers so popular on ebay. the tester would have identified all the  bad caps on your board within a few minutes

[jaunty]

If you mean an ESR meter - i actually have one. But I haven't broken it out of it's package yet. Since recapping was in the original game plan anyhow I just went ahead and did it. I'm not sure if it was the caps for the loose screw that was pinned under the PS board when i pulled it out that was causing the problem... 

but all is working well now.

good work jaunty 

i was hoping to come back tonight and continue troubleshooting.

the entire issue was dried up caps?

you may want to consider purchasing one of the $20 component testers so popular on ebay. the tester would have identified all the  bad caps on your board within a few minutes

[Joule Thief]

i would encourage you to use the component tester to evaluate the caps removed from the power supply board.

it is certainly a great feeling to get the pcb up and running but don't overlook the additional learning experience from checking the characteristics of the failed components

check the resistance of the caps with your DVM, check ESR and capacitance value with the component checker.  get that experience in your mental toolbox.

build a simple half or full wave low voltage supply using a dried electrolytic cap.

probe it, scope it,

learn

[jaunty]

Yeah - well it's one of those seven dollar jobs from china (the ESR meter) so I had a bit of trepidation about using it - i already checked the resistance of the caps with the DVM and it seemed high enough to be plausible that they weren't completely useless - but of course that will only tell you part of the story. But yes I should get used to using one - would definitely help a thing or two.

sorry one thing though - why using the dried up caps to build a power supply? Is there some 'old trick' to that I don't know about - or you mean build it to watch it's failure?

[Brumby]

He means to set up a faulty circuit, check it with your meter and see how it looks on an oscilloscope.

Learn the traits of such a circuit, so that you will be more familiar with how it behaves - and be able to recognise that when you encounter the same problem in another piece of equipment.

[jaunty]

yes of course! i'm learning all i can... of course i'm onto much harder circuits now - in other keyboards i have - but I'll have the benefit of using service manuals and known typical failures to go by... but those will be necessary scope material  - even more than a lowly linear power supply. thanks again..   i surely do appreciate everyone's input!