Repairing a late 80' stereo amplifier

[HooRide]

I picked up an amplifier on the cheap (well, relatively) because it had a slight hum in both channels. Turns out, I think I found the (a?) problem (see the pic).

My question is, would having the main filtering cap doing basically nothing affect the smaller filtering caps upstream (for example the one in the second picture)? I can't at all easily get to them to measure due to the design of the amplifier. (I'd have to desolder 14 TO-3 and 12 TO-220 packages per channel.)

Thanks!

P.S. Also, if anyone knows off the top of their head if Cornel Dubilier's CGS or DCMX is a better replacement for the Sprague, that'd be useful too.

[tatus1969]

you could check the effectiveness of these by connecting them to the rectifier and applying a known load to them. look at the discharge slope with a scope and calculate capacitance from that. if you have a switcheable load you can measure ESR as well.

[David_AVD]

I'd simply replace those main caps and retest.  No need to overthink it too much.

[SteveK]

That looks like it is a Mark Levinson amplifier.  If it is, yes the 1900uF capacitors could cause a hum so I would try to replace them first.  I prefer United Chemi-Con capacitors -- only issue is you usually have to special order them from the manufacturer in quantities.

Yes it is a major pain to replace the capacitors on the heatsink board.  Not only is it a time consuming process, you have to be very careful as the PCB can be easily damaged when removing parts (e.g. damaged plate holes and lifted PC lands.  Not only that, you could have some corrosion damage caused by leaking capacitors.

[HooRide]

you could check the effectiveness of these by connecting them to the rectifier and applying a known load to them. look at the discharge slope with a scope and calculate capacitance from that. if you have a switcheable load you can measure ESR as well.

That's the issue, I have no way of disconnecting them from the circuit without undoing the whole thing. The ESR meter doesn't give an accurate reading in this particular case because of the other components unfortunately.

I'd simply replace those main caps and retest.  No need to overthink it too much.

That is sound advice, not too hard to install replacements and test.

That looks like it is a Mark Levinson amplifier.  If it is, yes the 1900uF capacitors could cause a hum so I would try to replace them first.  I prefer United Chemi-Con capacitors -- only issue is you usually have to special order them from the manufacturer in quantities.

Yes it is a major pain to replace the capacitors on the heatsink board.  Not only is it a time consuming process, you have to be very careful as the PCB can be easily damaged when removing parts (e.g. damaged plate holes and lifted PC lands.  Not only that, you could have some corrosion damage caused by leaking capacitors.

Good eye! It is a Mark Levinson, a No. 23 to be exact. I've removed one channel of the PCB before to replace some caps that literally exploded (ERO KP1834 to be exact (how critical is it to replace that with similar kinds?) and had to re-plate some of the holes (for the output transistors). Luckily, no corrosion.
*Any* advice you could give on getting this beast back to spec would be very greatly appreciated.

[tatus1969]

you could check the effectiveness of these by connecting them to the rectifier and applying a known load to them. look at the discharge slope with a scope and calculate capacitance from that. if you have a switcheable load you can measure ESR as well.

That's the issue, I have no way of disconnecting them from the circuit without undoing the whole thing. The ESR meter doesn't give an accurate reading in this particular case because of the other components unfortunately.

please read my suggestion again. my proposal is meant to be done without disassembling anything. just put it together and measure, all you need is access to the dc bus voltage.

[HooRide]

you could check the effectiveness of these by connecting them to the rectifier and applying a known load to them. look at the discharge slope with a scope and calculate capacitance from that. if you have a switcheable load you can measure ESR as well.

That's the issue, I have no way of disconnecting them from the circuit without undoing the whole thing. The ESR meter doesn't give an accurate reading in this particular case because of the other components unfortunately.

please read my suggestion again. my proposal is meant to be done without disassembling anything. just put it together and measure, all you need is access to the dc bus voltage.

My apologies, I didn't understand the first time around. I do have all the test gear for that; I'll look into it more this evening. Thanks for explaining.

[SteveK]

Probably easiest to just scope the 1900uF capacitors while the amp is on to see if they are bad.  If they have a lot more than a few Vpp ripple, they are bad.

It is usually best to replace all the electrolytic capacitors, all the 0.01uF yellow PP capacitors,  the green or red Wima capacitors around hot components, and any heat stress resistors while you have the amplifier PCB removed.   The yellow PP and wima capacitors have a tendency to short out and blow up  -- sometimes spraying molten metal around causing other things to blow up.  No self healing here with the currents involved. The heat stressed resistors have a tendency to get noisy.  Most of the resistors are mil spec metal film resistors of 1% tolerance. I usually curve trace all the heatsink transistors before I put them back in because it it is not atypical to find one that it breaking down at normal operating voltages. You don't want to have to take it back apart after a few month or even before you are done repairing the amplifier.

With no solder mask on the board, it is very easy to get a solder bridge.  So inspect the board very carefully before putting it back in.

The black Dale 10 ohm 10 watt resistors should be replaced with non-flamable types otherwise the amplifier can literally catch on fire and cause major damage if something goes wrong elsewhere causing them to overheat.

[Seekonk]

If you don't have a scope, measure the AC volts on a meter.  It would be best to measure that through a 1 uF capacitor.  Some meters will measure the DC volts and display that as a value.