Have some parts .. need suggestions..

[Tsanford]

I work for a company that makes amplifers for life alarm systems. They are terrible sounding with low response and lots of distortion 70v/25v transformer. I have several dead ones to use for parts and would like some suggestions, since transformers, big caps and heatinks are some of the more expensive parts, would like to use these. I could also savange up the 3055 NPNs, However these tend to be blown so no matter if the wont be used.

Transformer 22-0-22
Heatsink holds 2 TO-3s
Caps are 64vdc 5kuF


I would like to keep my first project relatively simple, I am an electronics tech. ,but havent done component level troubleshooting in at least a decade. any help would be greatly appreciated.

[destroyer X]

You will see a lot of amplifiers you can make with 22 plus 22 volts and 2N3055.

I strongly suggest you to start with the best one i heard, some people in the forum agree with me.

Its a class A amplifier, using 2N3055 output

The name, is Death of Zen

have fun and welcome

Carlos

[Dozuki]

Just be careful of the brand of 2n3055. There are lots of really cheap ones available made specifically for switching and I've not had luck using these to make anything that sounded remotely good. As you said, the PS Transformer, caps and heatsinks are the biggest investment. Use these but find some different output transistors (IMO). Transistors are cheap comparativley speaking. As DX pointed out, Rod Elliot's site is a good starting place. Many DIYers cut their teeth on his designs. Also The Zen Amp is another good place to start. (www.passdiy.com).

-Dozuki

[djk]

You probably want something like this:

http://sound.westhost.com/project3a.htm

22-0-22AC translates to about ±33V after you rectify and filter it (with a 7% regulation transformer). Unless the 22V is no load, in any event you should be OK.

How many watts was the 70V/25V amplifier?

You will probably need one transformer per channel, and two filter caps. If you want to drive a 4 ohm load I would use four filter caps per channel.

With TO3 heatsinks I would use MJ21193/94 for outputs, MJE15029/30 for drivers. These are Motorola (ON Semiconductor) part numbers and available in small quantity from Digi-Key.

[sreten]

Hmmm.....

I suppose it depends on how the transformers output voltage
is measured / specified. With 22-0-22 V no load I make it ~
29V rails with loading.

Neverless its far too too high for the DoZ and also the heatsinks
are unlikely to big enough for class A operation. The capacitors
will be a little large (physically) for a split rail design being 64V.

Ripple ratings need to be taken seriously so I'd suggest a minimum
of 2 per rail, more for lower impedance speakers, 3 or 4.

I'd suggest investing in D.Selfs Power Amplifier Design Handbook,
there is a wealth of useful information in here, even if you don't
agree completely with his approach.

Whether you need one or two transformers is down completely
to transformer size, but seperate rectifiers and smoothing per
channel is always nice, but not strictly necessary.

sreten.

[djk]

"Ripple ratings need to be taken seriously "

Not really.

Ripple current ratings are based on temperature. The bigger the can size, the higher the rating.

If you were running RMS current at a 100% duty cycle, then the ripple rating would need to be taken seriously.

Does that describe audio?

[sreten]

Originally posted by djk
"Ripple ratings need to be taken seriously "

Not really.

Ripple current ratings are based on temperature. The bigger the can size, the higher the rating.

If you were running RMS current at a 100% duty cycle, then the ripple rating would need to be taken seriously.

Does that describe audio?

Well if you consider conduction only occurs at the peak of the
half wave cycle, and currents are much higher than the current
drawn from the supply some would beg to differ.

But getting any sensible information on this is difficult.

sreten.

[djk]

"Well if you consider conduction only occurs at the peak of the
half wave cycle, and currents are much higher than the current
drawn from the supply some would beg to differ."

You're not thinking heat.

Let's take your example to its logical absurdity.

I have a need for a 1A average current.

If I use a 1000µF cap the average current and the peak current will be about 1A.

If I use a 100,000µF cap the peak current will be 100A, so I need to buy a cap with a 100A ripple current rating?

With adequeate ventilation the main filter caps seldom fail.

In general I replace more power transformers than I do filter caps.

10µF~100µF electrolytic bypass caps? They need replacement on a regular bassis.

"It ain't got no punch no more"

Take off the cover and four pencil eraser sized 10µF cans fall out.

[sreten]

Thinking about it further 64V capacitors should have higher ripple
capability than the 35V capacitors that would normally be used in
this application, so I admit my concerns about increasing the
number of capacitors are unfounded.

sreten.

[Tsanford]

This was my first post, so I thank you for your responses.
Typo in original was a combination of thoughts .. "life alarm" = life safety system and fire alarm all thought to fast.
The original amp was rated at 100w for a 70/25v system however i would be suprised if it was designed much pass the 75% mark of components because of its life safety use. These things weight 60 pounds in the box.
I will try to link a picture again, didn't make it last time for some reason.
Transformers have a 3.5A breaker on the primaries, safe to guess they are probable near 300va?Im in the US on 115vac.
Is it possible to safely series a couple of these transformers for more voltage? They are by no means matched some are even different makes, but will have same specs. and will be pretty tight speced.
yes, 22-0-22 is unloaded right off the xformer, and each amp has 2 big 5k caps i got 5 amps so far tho.
Heat sinks look small for class A to me .. Im new at this tho.
May need a good class AB design using a pair of TO-3s as outputs.I should surely be able to get the 100Ws the factory got out of it.