Got transformer for bench supply now...

[throbscottle]

I just pulled the mains transformer from an amp that I found. It has something like 40-0-40 + 0-15 windings. Amp says on the back is 120W so assuming the transformer is maybe 150VA.
So I now have something to base a linear bench PSU on, maybe +/-0-35V range. Thinking about putting in a couple of (pairs of) buck converters so I can create staged voltages internally, so the pass element doesn't have to take so much punishment at low outputs. So assuming I can safely draw 3A, create stages at around +/- 25v and +/-10v, maybe give them some adjustment span of their own. So linear pass element never has more than 45W to dissipate
I know I've visited this idea before, but coming back with some actual hardware now
Thoughts, anyone?

 

[atferrari]

I just pulled the mains transformer from an amp that I found. It has something like 40-0-40 + 0-15 windings. Amp says on the back is 120W so assuming the transformer is maybe 150VA.
So I now have something to base a linear bench PSU on, maybe +/-0-35V range. Thinking about putting in a couple of (pairs of) buck converters so I can create staged voltages internally, so the pass element doesn't have to take so much punishment at low outputs. So assuming I can safely draw 3A, create stages at around +/- 25v and +/-10v, maybe give them some adjustment span of their own. So linear pass element never has more than 45W to dissipate
I know I've visited this idea before, but coming back with some actual hardware now
Thoughts, anyone?

Hola throb

Not actually replying your questions.

I did many attempts to build a PSU to put a +/- 38V DC trafo I got years ago, to good use. All my attempts failed because, basically my attempts of a switching preregulator of the negative rail were never feasible (locally cannot get anything more sophisticated than a 2N3904 or a 555).

My attempts to make it variable by controlling the reference voltage of an LM337 and LM317, had problems like unexpected oscillations something I do not know how to deal with.

I will follow this with interest.

 

[KeepItSimpleStupid]

Without doing anything 70 V CT gives +-50 VDC; So you have +-56 V to deal with doing the calcs.

 

[throbscottle]

Eeek! First thing I do might be to re-wind it then! I can put lots of taps in if I do that, save mucking about with buck converters Made somewhat harder by the fact that it has a mounting frame welded across the laminations. Ho Hum.

OTOH, all the I's and all the E's are each in a solid slab, welded together, so at least it's easy once I've cut the welds. Well this is going to be interesting!

 

[KeepItSimpleStupid]

Why do I know the 70CT to +-50? An amp I built with a custom transformer. The transformer should have been bigger current wise. 4x35V@3A each is too small for 100 W into 8 ohms. But with a sine wave voltage regulator and 50, 000 uF of capacitance, it sounds pretty good. At one point, I had an 18 A single center-tapped secondary constant-voltage transformer, but the transformer hummed too much. The article I used as a basis for the design, used a single center-tapped winding. I doubled it, and it was too small.
e.g (0.62*3A) * 8 ohms is < 100 W. ((V-x)^2) / 8 ohms is much better, where x is the minimum drop across the output transistor,

 

[crutschow]

A 150VA transformer would be good for about 150VA/80 = 1.875A from each of the 40V outputs.

 

[throbscottle]

I never thought of it like that - I was taking it as half a winding. Duh. Definite case for re-winding then! Especially I have no use that I know about for +/- 56VDC.
Suppose I have a few taps, and arrange some kind of automatic switching so the bridge can take inputs from any pair, what's the best way of switching? I think I can do hysteresis so there'll be no hunting around the transition point. I do have an isolated extra winding on there which might help. At the moment it looks like a choice between triacs and relays ((for tv burp viewers which is best? there's only one way to find out...).

KISS - what do you mean by a "sine wave voltage regulator"?

 

[KeepItSimpleStupid]

You can take some voltage off the primary too. If you took 120 V off the primary (240:120) xformer in anti-series you would have half. Using two xformers like a 2x? could take the same voltage off the secondaries too. 2x? torroidal xformers should be easy to find.

==

I have a Sorenson ACR500 https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=11&ved=0CCAQFjAAOAo&url=https://www.teknetelectronics.com/DataSheet/SORENSEN/SOREN_acr200044311.pdf&ei=ZioHVY6DJcKnNuWigvAE&usg=AFQjCNEpAnM_yJBeEGGDz1ElnKECAO8D7Q&bvm=bv.88198703,d.eXY&cad=rja

that I found in a surplus store when in NYC. It's like 29 lbs or 13 kg and I was lugging it back to the car blocks away. No meter and it still doesn't have one. I bought it for like $100 USD, Looks like they are selling for around $1500 USD today. https://en.wikipedia.org/wiki/Canal_Street_(Manhattan)

 

[throbscottle]

Interesting device.
I'm rather confused about what you mean regarding taking some voltage off the primary. If I connect another tranfsormer in series with it that will just be like a big choke, but what are you saying should be in anti-series with what? A diagram would help. I just want to use one transformer that I already have (possibly re-wound), if I can pull some useful tricks using other things I already have, great, otherwise I may as well just buy the proper transformer and be done with it.

 

[KeepItSimpleStupid]

You can put transformers in series. They will either add or subtract. So, you hook them up in series and determine the polarity.

With a 2 winding toroidal transformer, you can take an equal amount off the center tap. You can also take some voltage off the primary in the same way. So instead of a 230 V primary, you could have a (230-24) Volt primary or even a (240-120 V) Primary. Yea, 2 transformers takes up more space. and high voltages aren't common.

Anyway, example, (1.4*35 V= 50 V). Secondary DC current after FW rectification is about 0.62 of the AC RMS of the secondary current.

 

[throbscottle]

Ahh, that makes more sense

 

[dr pepper]

Watch out for overated amps, the trans isnt necessarily 150va, you can tell by the size and weight to an extent.

 

[throbscottle]

Yeah, I should measure the core and work it out really, but lack of time + laziness means I took the wattage figure off the back of the amp, added a bit and called it VA! Extremely bad of me and I didn't realise I'd added quite so much. Anyway I've an inkling Dave Jones may have done a switch mode pre-regulator so I should try and find that before rewinding, since it's what I was originally thinking of.

 

[dr pepper]

I have a transformer calc, if you know the core cross sectional area I'd be able to ogive a ball park va.
Its possible to get more va by adding more copper, however there is a common ratio of copper to iron.

 

[throbscottle]

That might be useful, however I'll try the excellent spreadsheet here: https://ludens.cl/Electron/trafos/trafos.html - always worth reading again and again! For today I can't do anything 'cos I'm in a hotel...

 

[throbscottle]

Been having a look at switching pre-regulator designs. Actually a lot simpler than I thought. I didn't think buck converters worked too well if you try to vary the output but apparently this isn't the case. A few people have been using LM2576-adj variants, so I might get a couple of HV ones to play with.

 

[dr pepper]

Homo's site is ok yes.
So your having a mains trans followed by a switching reg and a postreg, sounds interesting and ought to be reasonably efficient.
You'll also get more current at lower voltages.

 

[crutschow]

To minimize power in the output linear regulator you can have the switching regulator track the output but stay a few volts above the output voltage to provide the minimum required drop across the linear regulator.

 

[dr pepper]

Theres a post on this site that demstrated a tracking pre reg, not so long back, it might have been a lm25xx series too.

 

[throbscottle]

Ok, next question.
The transformer has a centre tapped output. Would I be better to use it as it is, where I can have:

• positive rail, adjustable buck converter, linear regulator, and;
• negative rail, adjustable-(boost converter configured as negative buck)-converter, linear negative regulator ( **broken link removed** but I'd use LM2586, possibly sit it on a zener to accept the higher voltage input or else I'd need pre pre-reg!)

OR I can
Make the transformer create a single positive supply, and make a split supply converter, which would be simpler, though I would need to wind a small transformer for it. The big advantage is that the + and - outputs of the converter would track perfectly, and easily based on the + output. The big disadvantage is removal of the possibility to adjust them independently, except by independently adjusting the linear section for that rail.

Which is best? There's only one way to find out....