Basic Linear Power Supply Troubleshooting

[cincin]

I'm working on an old Kurzweil K250 sampling synthesizer (1984!) that has a massive 60Hz buzz on the audio out. I traced the buzz up all the way to the output of all 12 DACs (for the 12 voices), and found that buzz was present as a ripple on the +/- 15V DC rail feeding the DACs. So I chased it up the power rail all the way to the power supply.

This is my first linear power supply troubleshoot...

I reviewed my basic understanding of how linear power supplies work, and it seemed to me that if 60Hz of some AC was going through it must be either a faulty diode on the rectifier, a faulty filter cap after the rectifier, or a faulty voltage regulator after that. There's obviously many more parts involved, but if these work the power should be pretty flat. The diodes being the more likely culprit given the size (1.4Vpp on the + and 0.6Vpp on the -)  and chopped-sine shape of the ripple. But all these components check out fine. So I'm somewhat stumped.

The power supply is a Lambda LOD-X-152-40102, the main PCB says HAL-02-087 R-A I could not find a user manual, service manual, or schematics for it. I'm including photos of both sides of the PCB in case it's useful.

My main questions are: How can there be such big square ripples on the OUT when the rectifier and filter caps are working as expected? (obvious answer is it's coming from somewhere else, so:) Where else can the ripple be coming from? Ungrounded ground-plane?

I'll add that this power supply comes out of the POD part of the synth, which includes a separate 5V power supply, which also has a bit of a ripple. But both the 15V and 5V power supplies feed off the mains in separately, so the problem cannot be connected. But perhaps both ripples are the result of a common mishap.

Thanks for help, I learn everytime I post here!

[Wallace Gasiewicz]

Does this PS have "SENSE" terminals and are they connected??

[cincin]

Good question! No.

Well, it has SENSE terminals, but they are bridged to the outs in this use.

[ledtester]

Does your power supply come with the large silver capacitors as shown in this pic?

[cincin]

Does your power supply come with the large silver capacitors as shown in this pic?

It sure does, I removed them for testing prior to the photo. In my photos the rectifier diodes +1 have been lifted at one leg, the filter capacitors have been removed, and the voltage regulators which were outside the enclosure, on the bottom of the PCB (you can see the outlines on the PCB) have also been removed (those were not easy). I should've specified.

Cheers!

[ledtester]

The TO-3 devices which are screwed to the case are likely the pass transistors.

The voltage regulators are IC1 and IC21 with the LAS723B marking - they are "LM723" regulators.

I found this manual for a different model Lambda supply - but still a linear supply. On page 13 it contains a trouble shooting guide:

http://www.bitsavers.org/pdf/lambda/Lambda_LCD-4_Instruction_Manual.pdf

[cincin]

The TO-3 devices which are screwed to the case are likely the pass transistors.

The voltage regulators are IC1 and IC21 with the LAS723B marking - they are "LM723" regulators.

I found this manual for a different model Lambda supply - but still a linear supply. On page 13 it contains a trouble shooting guide:

http://www.bitsavers.org/pdf/lambda/Lambda_LCD-4_Instruction_Manual.pdf

I'll have to read up on pass transistors. Thanks for pointing that out. And great find on the manual! I will read that and dig deeper. Thanks!

[shapirus]

Now, as a side note, that's an actual MTBF chart printed on the PSU itself. Go figure!

[Kim Christensen]

the filter capacitors have been removed,

Before you disassembled the unit, did you measure the voltage and ripple across those two filter capacitors? The DC voltage should have been 20Vdc or higher so that the regulator circuitry had enough headroom to regulate and suppress ripple.
Did you do an ESR test on those capacitors after removing them in addition to measuring their capacitance?

LAS723B datasheet attached:

[cincin]

the filter capacitors have been removed,

Before you disassembled the unit, did you measure the voltage and ripple across those two filter capacitors? The DC voltage should have been 20Vdc or higher so that the regulator circuitry had enough headroom to regulate and suppress ripple.
Did you do an ESR test on those capacitors after removing them in addition to measuring their capacitance?

I did not measure ripples specifically there, but I recall finding about 20V at that stage. I will likely have to reassemble it for more testing. Will check then.

I did check the ESR, but with one of those cheapo LCR-TC2... I got 37Ω and 40Ω.

[Kim Christensen]

37Ω and 40Ω, is terrible. If correct, those capacitors are faulty and most likely your problem.

[cincin]

I was just reading up on that. Seems I should be getting 0.01Ω to 0.1Ω??

Those will be replaced before I test anything else. Cheers!

[vk6zgo]

I'm working on an old Kurzweil K250 sampling synthesizer (1984!) that has a massive 60Hz buzz on the audio out. I traced the buzz up all the way to the output of all 12 DACs (for the 12 voices), and found that buzz was present as a ripple on the +/- 15V DC rail feeding the DACs. So I chased it up the power rail all the way to the power supply.

This is my first linear power supply troubleshoot...

I reviewed my basic understanding of how linear power supplies work, and it seemed to me that if 60Hz of some AC was going through it must be either a faulty diode on the rectifier, a faulty filter cap after the rectifier, or a faulty voltage regulator after that. There's obviously many more parts involved, but if these work the power should be pretty flat. The diodes being the more likely culprit given the size (1.4Vpp on the + and 0.6Vpp on the -)  and chopped-sine shape of the ripple. But all these components check out fine. So I'm somewhat stumped.

The power supply is a Lambda LOD-X-152-40102, the main PCB says HAL-02-087 R-A I could not find a user manual, service manual, or schematics for it. I'm including photos of both sides of the PCB in case it's useful.

My main questions are: How can there be such big square ripples on the OUT when the rectifier and filter caps are working as expected? (obvious answer is it's coming from somewhere else, so:) Where else can the ripple be coming from? Ungrounded ground-plane?

I'll add that this power supply comes out of the POD part of the synth, which includes a separate 5V power supply, which also has a bit of a ripple. But both the 15V and 5V power supplies feed off the mains in separately, so the problem cannot be connected. But perhaps both ripples are the result of a common mishap.

Thanks for help, I learn everytime I post here!

Linear power supplies almost universally use full wave rectification, so the hum (not "buzz") you would normally expect from a faulty supply would be 120Hz.

Earlier (pre mid 1960s) supplies often used a  centre tapped transformer secondary & two diodes to obtain full wave rectification, but any lowish voltage supply you are likely to encounter would use a plain secondary winding & a bridge rectifier.

If a diode in one leg of the bridge rectifier failed, it would effectively become a "half wave" rectifier & you would see 60Hz hum, as the filtering would not work as well under those circumstances.

Do you have an oscilloscope?
If so, can you provide us with screenshots?

[cincin]

Do you have an oscilloscope?
If so, can you provide us with screenshots?

I do, and I will.
Give me a few days to get some new filter caps in those specific specs and I'll reassemble the thing for more testing and some screenshots.

[Kim Christensen]

If you're ordering parts, you may as well replace all the electrolytics on that board.

[cincin]

If you're ordering parts, you may as well replace all the electrolytics on that board.

Good call.

[Wallace Gasiewicz]

I just repaired a Lambda Lab PS. This is not a Lab Grade PS and it is simpler.
We do not have a schematic but the board is helpful. It has little in common to the schematic in the LCD 4 manual. 

An Mentioned already, the two diodes in each PS indicate center tapped transformers. Can you check if the transformers are OK? Is there continuity between the leads of the transformer and the center tap. Is the Center tap connected to GRND on both transformers?
Please start here, if you do not have good transformer, it will not work. Maybe the  lead on one was loose.
The four big diodes at the top of the board are the rectifiers. You said you checked them

As Already mentioned, The ten lead  723 are LM 723 type volt regulators . If they are intact, they should get rid of most ripple. If they are bad they can produce nice square waves, since the op amp inside it can do that quite well.

The two cans with the heat sinks are the transistor drivers for the final pass transistors, which are usually mounted on a heat sink nearby. You can check all of them with a diode checker, in place.

If one of the 723 regulators is bad you could easily see something on the other PS output, The output with the big square wave is the culprit. 

I think one of the 723 regulators is bad. Or, less likely,  one of the components around it is bad.  These are mostly resistors.

https://www.ti.com/lit/ds/symlink/lm723.pdf?ts=1702220856089

Although others may disagree, I think the ESR of the big caps is irrelevant.   

Wally

[cincin]

Thanks! I will check on that this morning and report back.

Cheers!

[cincin]

Can you check if the transformers are OK? Is there continuity between the leads of the transformer and the center tap. Is the Center tap connected to GRND on both transformers?
Please start here, if you do not have good transformer, it will not work. Maybe the  lead on one was loose.

Alright, my apologies for my utter ignorance, but I'm not sure which lead is which. I've attached a photo of the wiring before I touched anything. The leads that come out of the top (secondary) of the transformer from left to right: orange, red, red, green, green, yellow. The orange and red leads have 0.3 Ohm between each other, and the same for the yellow and green leads. There's no connection between one side and the other (between yellow green green, and orange red red).

So which one is the center tap?

On the primary side there are 3 connectors, the center one is not used. The ground lug is on a corner screw.

Looking at various wiring for center-tap transformers is not much help. It almost looks like the one I have here is wired as 2 completely separate sides (is that what you meant by "both transformers"?) , 60V each, and then each of those is split in 2 with a center tap as return, with each red / green wire providing 30V 180 degrees out phase. If that's the case I should have 2 center taps? (yellow and orange) Again IF that's the case, then the 2 center taps are not connected to the ground lug on the corner screw. If that's what you meant?

I will make sure to clean the connectors when I put it back together once I get the filter caps.

Testing the 10-pin LAS7238, assuming they are setup like LM723 I get pretty wild results, so those look like part of the problem. Questionable results marked with ?:

What I was expecting from diode tests on the various pins (I may be wrong):
10 to 9 = .5V diode drop
9 to 10 = ol
10 to 1 = .5V diode drop
1 to 10 = ol
9 to 7 = .5V diode drop
9 to 6 = .5V diode drop
6 to 9 = ol
7 to 9 = ol
5 to 8 = .5V diode drop
8 to 5 = ol

1st LAS 7238:
10 to 9 = 1.4V ?
9 to 10 = 0.75V ?
10 to 1  = 0.7V
1 to 10  = .9V ?
9 to 7 = 0.8V
9 to 6 = 1.3V ?
6 to 9 = 1.3V ?
7 to 9 = ol
5 to 8 = 0.5V
8 to 5 = ol

2nd LAS7238:
10 to 9 = 0.7V
9 to 10 = 1.3V ?
10 to 1 = 0.7V
1 to 10 = 1.2V ?
9 to 7 = 1.2V ?
9 to 6 = 1.5V ?
6 to 9 = 1.4V ?
7 to 9 = ol
5 to 8 = 0.5V
8 to 5 = 1.4V ?

Can I replace those with actual LM723's like these:
https://www.digikey.ca/en/products/detail/texas-instruments/LM723CH/5056006

Thanks again for all the help!

[cincin]

Checking the RCA FBN L109 driver transistors, I see no continuity between any of the 3 legs on either of them. I couldn't find info on what kind of transistor it is. I'm assuming I should have continuity on one pair of legs?

The diode test on them shows a 0.6V drop in one direction and 1.2V in the other direction, on a pair of legs, the third leg (GND?) doesn't respond to diode tests. Both FBN L109 give similar results.

The 1.2V drop is apparently ok for a "reverse biased diode junction"?

[Wallace Gasiewicz]

Center taps should be connected to neg on the board, Sorry.  Each one connected to Neg near where their Pos are connected to the board. The low and high voltage parts  of the board are probably entirely separate. The two PS are probably "Floating" and not connected to GND.
BI Junctional Transistors should measure about .6 volts for each PN junction, so .6 volts from base to emitter and .6 volts from base to collector. but the FBN 109 is a unijunctional transistor, Perhaps someone else knows how to measure it.I was wrong, they are not driver transistors. Read about them on WIKI. One way they should measure .6, the other way I don't know, but should be much higher. Maybe too high for a normal diode checker to detect.
Too bad we do not have a schematic.They could be there for protection against too much power draw. They may shut off the power when there is a short on the output. Without the schematic it is hard to guess.
It would be interesting to see if there is that square wave on the outputs of the PS if you had a proper load on them, not just the rest of the circuitry.....You did not happen to measure the PS with it disconnected??

[cincin]

You did not happen to measure the PS with it disconnected??

I did not.

So you're saying the LAS7238s might be fine?

The center taps (yellow and orange) were connected to the negative terminals on each side.

[Kim Christensen]

RCA FBN L109 is an NPN transistor. See attached.

Unless you see a short circuit, diode checks on the LAS723B are almost useless...

[Wallace Gasiewicz]

Sorry 

NTE says it is an NPN also
NTE makes lots of equivalent transistors and you can search for them here: 

https://www.nteinc.com/search.php

sometimes very helpful when you cannot find a datasheet.

[cincin]

Good news, everybody!

(did you hear that in Farnsworth's voice?)

I received the new filter caps, and other replacements for the electrolytics, so I proceeded to do the swap and rebuild the power supply for further testing, taking care to clean the contacts.

Well wouldn't you know it, C2 was bulging at the bottom (sealed tops) and tested 100% bad. As you can see in my original photos the condom was peeling down a bit, pulled by the bulging.

Upon rebuilding everything, the ripple, and the hum, are gone! The synth sounds great again.

I think I got lucky, and got away with a cheap repair on an expensive supply full of expensive parts.

Thank you all for your help. I learned many things and gained some experience points, thanks to your guidance.