DIY Benchtop PSU

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Yep, this has been up before. Without any result.

Here's my thought: we all have a storage of powerful transformers, heat-sinks and some transistors.

I need a bench-top PSU and have almost all the above materials to make one, but I need the right idea/schematic.

The problem (at least for making amp's) is that you would like something in the neighborhood of +/- 60V/2-3A.... of course variable and protected against abuse. With a linear PSU that's over 400W of wasted energy in the worst case (lov voltage/ High current). Even with massive heat-sinks 400W is a lot.

In these eco-friendly times that seems wrong.

Solutions

1) Several taps on the xformer will drop the waste in a "class-G" (or is it "H")- sort-of-way. That works, but not many xformers has that many taps.

2) SMPS. Should work. Could be tricky to get stable and noise-free over a 0-60V area.

3) SMPS + lin. reg. With a switch-mode preregulator one could ensure that there were only a 5-10W drop over the pass-transistor. That would lower dissipation a lot.

In lack of SMPS-knowledge I'm tempted to go the crude way, - meaning massive amounts of cobber, iron and alu.

Any ideas? Are there variable SMPS-schematics out there? Perhaps even simple ones?

Yes, this might never get to be a "project" as I have many other things to do, but the subject should interest a lot of people?

Regards TroelsM
 
Hi,
the commercial ones use variable tappings (listening to the relays clicking as I ramp the voltage up or down). It works well.
You could be clever and use FETs for silent switching.

A secondary on a toroid could fairly easily be rewound with multiple taps. Even 25%, 50%, 75% 100% would be a big help.
Look, those 50% and 100% are already there. What about tapping into the midpoint of each dual secondary and soldering in 25% and 75% without having to rewind.
 
I am using a large low-voltage lighting toroid for my psu.
It has ten (10) windings at 10 amps of 12vac.
I bought it for the price of a pint from some shopfitters who were doing a makeover. All the old lighting was going in a skip!

Recycling and Reusing - very ecofriendly !

Andy
 
Hi Andrew

Yes, I know that relay-switching is widely used and yes it is effective. My problem is that I don't like messing around with the windings and that its hard to do on a EI-core.

Maybe one big xformer that converts from 230V to say 60V and then a (un-regulated for simplicity) switchmode with multiple secundary windings would do the trick? Then it would also be possible to have multiple isolated outputs from one PSU? But then again.. that would require a lot of fast diodes and low-esr-caps in the smoothing-stage.. bad idea. -and costly

I still like the idea of a switchmode stage that (crudely) tracks the output-voltage and makes sure that there is only a small voltage over the pass-transistor of the linear reg.

Regards TroelsM
 
I know that the shipping to DK is expensive, but there have been a number of Eico 1078 variable A.C. power supplies on EBay -- quite inexpensive from what I can determine -- thus you can dial down the transformer on your power supply. Probably uses a Variac.

My big Lambda supply (600V 300mA) uses a Variac as well.

Note that there are some Heath supplies (IP-27) which use a transformer with many taps.

Surprisingly, the Heath and Eico supplies do show up on the European EBay sites.
 
I would go switchmode with gentle output filtering. There is no need for post regulation if it is done that way, and 300W may be easily packed in the size and weight of a standard PC power supply (0-60V 0-5A). Switchmode voltage regulation is quite straight to achieve, although adjustable current limiting is somewhat more complex.
 
Hi Eva

You´re well known as being the switch-mode-guru and of course you're right, that it's "quite straight to achieve", but I don't think many of us would be able to make it work. Sadly.

That's why I was thinking in a combined switch/lin-solution. I had a "feeling" the demands to the switch-part would be a lot easier if there were a lin-post-reg that could do the actual regulating and current-limiting.

Without a reliable current-limit i would be very scared of connecting my new amp to 60V/5A.. that's alot of juice if somethings not working....

Could you recommend any dedicated switch mode-chips to the pre-reg-job? Any hints?

Regards TroelsM
 
+1 for using a variac in front of a standard linear setup.

That said, I have a big linear regulated job bolted to the bottom of the bench for 0-40V (small analog meters) and currently in-construction is a 0-2V setup for maybe up to 8A, with digital voltage readout and 10-turn voltage setting control. Different equipment for different needs.

Yes, the low-voltage supply is on a custom-wound (by me) secondary.
 
the german magazine elv published a regulated smps supply a few years ago using a sg3526 (600 watts output power).
if you want to do it discrete i would say use a transformer with middletap and build 2 powersupply's for instance based on the good 'oll lm723.
this way you have the possibility of using them in serie up to 74 volts .
depending on youre skills you could choose to make a sort of hybrid ,mosfet in the powerline before the bufferelco and have a mcu keep a 5 volt difference between youre output and the elco.
(sorry design and building this will have to wait till i have time for it)
good luck
 
Hi.

Any links to this article? Or perhaps a more precise guesstimate of the publishing-date?

I tried google without any luck. The local library could probably get anything, but it seems a bit crazy to order the last 5 years :) of ELV.

I had actually thought about the pre-reg-mcu-idea before. If one puts a MCU in a PSU there seems to be a lot of funny options... It wouldn't be out of my reach, but it's a big project, - especially because one (me) needs to restrict oneself (myself) and not go crazy with MCU-control, feedback, PC-control and the like....

Regards TroelsM
 
AndrewT said:
Hi,
the commercial ones use variable tappings (listening to the relays clicking as I ramp the voltage up or down). It works well.
You could be clever and use FETs for silent switching.

A secondary on a toroid could fairly easily be rewound with multiple taps. Even 25%, 50%, 75% 100% would be a big help.
Look, those 50% and 100% are already there. What about tapping into the midpoint of each dual secondary and soldering in 25% and 75% without having to rewind.
A bench power supply I built once (0-18v,3A) used a center tapped transformer and a SCR for switching. The configuration is just like a bridge configuration, except there is a SCR between the bridge rectifier and filter capacitor and an additional diode from the center tap to the filter capacitor. When the SCR is off, it works like a 2 diode full wave rectifier. When the SCR is on, it works like a normal bridge rectifier.
 
Simple suggestion :-

Find a Farnell H60-25 bench reg linear PSU. Superb design and build with forced air avionics style cooling.

0-60 volt at upto 25 amp. Uses scr's in the TX primary to control dissipation in the pass reg.Used at less than 10 amps there would be negligible heat. Fine piece of equipment !

They can be picked up on ebay for about £50 max !!! I have 2 & didnt pay more than £40 for either ! Also have their big brother the H30-100 , yes, thats 0-30 at upto 100 amp !! Weighs 90 kg !!!
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.