Cooler recifier diodes?

[Herald1360]

Will it though? Some interesting calculations or a bit of simulator playtime might be needed to be sure.


Is there any sign of thermal damage from the hot running? How hot are the various components actually getting? If they're not actually being overrun, just upgrading the more tempereature sensitive parts (the electrolytics) may be all that is needed for a long and happy life!

[paulsherwin]

Originally Posted by snowman_al

Any space for a small computer fan in there?

Even a very small one underrun would help a lot.

[Top Cap]

Paul is right! Just a small fan MOVING the air will do great things.
You won't need much, no noisy forced air, just a gentle movement of air from the heat sink fins and the other components will work wonders.

[Radio Wrangler]

Quite right about the reg dissipation

So how about bolt-down bridge rectifiers screwed to some cool metalwork with wire leads to the PSU board. It'll get several watts removed from the board.

David

[mole42uk]

Originally Posted by Herald1360

Is there any sign of thermal damage from the hot running? How hot are the various components actually getting? If they're not actually being overrun, just upgrading the more tempereature sensitive parts (the electrolytics) may be all that is needed for a long and happy life!

Thermal damage includes dry joints and capacitor degradation, revealed by spurious stoppages of the logic board.

My first will be to replace the electrolytics with 105º versions and new 1N5401 devices, standing off the board by 20mm, as well as evacuating and re-soldering all the hot connections.

I have some small fans which I may try and see if they are quiet enough - I'm familiar with under-running a fan so that it moves air quietly.

Longer term testing will reveal if I've been successful in making the instrument more stable.

[Silicon]

Have you compared the power dissipation in the regulators with the dissipation in the rectifiers?

Off the wall thought, dunk the whole thing in a tin with fins on it full of oil? Baby oil is nice and clean, but you might get an odd look buying a few bottles.

[mole42uk]

Definitely an idea, might be okay but the oil might leak on stage?

[turretslug]

Curious that they went to the effort of speccing a nice, high quality, long life 125 degree C Rifa PEG124 electrolytic- and then completed the job with run-of-the-mill 85 degree types. It could always have been a case of analysing necessary ripple ratings in each case and only spending the money where it was needed, but when the whole shebang has a "standing room only" air to it, you'd think someone would have felt that it would be wise to have elevated-temperature type elkos all round. I suppose the "it'll last just beyond the warranty" approach applies in many electronic devices.

[McMurdo]

Might be worth losing those little blue Philips capacitors. They are particularly susceptible to hot environments. I wonder if changing some of the regulators for switching ones (the drop in replacement type) might buy a few less degrees in the enclosure?
I've always found Nord stuff to be pretty well made.

[jamesperrett]

Originally Posted by McMurdo

I wonder if changing some of the regulators for switching ones (the drop in replacement type) might buy a few less degrees in the enclosure?

I was just about to suggest the same thing. The Traco TSR-1 series worked well when I used them in a PSU design although I've not used them for audio.

[trobbins]

Perhaps a single custom heatsink plate that touches the middle body section of each diode, and is as tall as the transformer case. Glue the plate to the transformer case, and use some thermal paste to achieve conduction from each diode to plate. With such a cramped location, any diode is going to have a hard time transferring heat to ambient.

Perhaps beef up any pcb traces from diodes to themselves and to nearby parts like the transformer, as a way to improve thermal conduction away from the diodes, and to average out the diode junction temps.

There was also no indication of 'how hot'. If the diode leads are under say 70-80degC then that would be no concern for the diodes themselves.

[joebog1]

After reading the whole post nobody has yet asked:
How old is the unit?.
with diodes running hot I would first suspect leaky electrolytics. Even the PEG124 series ( which I love) will dry out if "cooked" over years.

Caps will cost a couple of quid at best, and even very high quality low leakage high temperature types are very cheap today.
Long leads on a new set of 1N5404's will set you back all of four bob.
SO!! even if the unit has always run hot, a cap change for better specked ones, a new set of diodes.
Cant "hurt" anything, and I doubt it can make anything worse.

Sorry for my rant

Joe

[Radio Wrangler]

If you could sneak in something like a couple of these:

https://uk.farnell.com/on-semiconduc...gbu/dp/1700187

mounted on some metalwork off of the board, with wired connections, you can export some of the heat off of the board and improve the reliability.

It looks like the whole PSU was designed with a prime intent to cram it all into the smallest possible space, and to hell with worrying about the temperatures components ran at. There isn't necessarily a fault in the unit, but there is pretty definitely a design fault.

Going to a lower drop diode is possible and will save you some of the heat from the rectifiers, but as said, will increase the heat in the regulators. Going to remotely mounted bridge rectifiers will save you all of the rectifier heat and without upping the heat in the regulators.

I had a good smile at the suggestion of immersing it all in baby oil. When buying the large quantity of baby oil and getting an odd look, you could helpfully explain that it was for use in a stage performance Expect an even odder look, or a request for tickets...

David