Modern Components for an Old LT PSU?

[Bazz4CQJ]

I am working on a 1970's PSU configured to supply a fixed 24VDC at 20A (from a multi-tapped transformer) for which I don't have a circuit diagram and it seems to be working (in that limited role) ok, as-is.

On the DC side of things, there are 6 off 4,200uF FILMCAP PROSEC 85 (63VW) and 1 off SPRAUG 10,000uF capacitors, each one being 35mm diameter and 120mm long. There are 12 off 2N3055 all in parallel doing the controlling (another 9 discrete transistors and a 741 op-amp do the clever stuff). I would like to re-build this, getting it down in size/weight from its present passively-cooled, large 19" rack format in to a chassis of 300x230x230mm (which I have) and which would have fan-cooling. It also needs the capability of running at 12-13 volts. Does the team think....

1. This is working; leave those old electrolytics and 3055's alone (ain't broke - don't fix).
2. This could be 'compacted' (and perhaps improved in electronic terms) by using some alternative caps and replacing the 3055's with some more recent components (cost could come in to this )
3. Something else...

I don't have an ESR meter so cannot comment of the performance of the old caps. I do have to work on it to get the working voltage to something useful, so some surgery is inevitable. Your considered responses welcomed.

B

[Ed_Dinning]

Hi Bazz, lets have a spec of what you are trying to achieve, then we can give better advice on where to go.

Note that the loss across the series pass elements will be the same no matter what these devices are if:
They are in the linear mode (no SMPS conversion)
Input and output voltages remain about the same.

Changing to FETs will not help here, so you might as well stick with 3055's.

You may want to improve the regulator chip and add some sort of overcurrent trip or regulator.

If you want to draw 20A at a low output voltage you will need more transistors and a bigger heatsink.

Ed

[Bazz4CQJ]

Hi Ed

remember I am a very amateur amateur and may be missing some of the more subtle issues that come in to play here.

I have never owned anything that needed more than a nominal 12 (max 13.8) volts (until we get up to 63 volts HT). That being so, the transformer has 4 off 9V secondaries, 3 off 3V and 3 off 1V, so if that was configured to give 14V, do we think the 12 3055's will let me run between 12-13.8 V at a maximum current of x amps? Typically, the most demanding use in terms of "purity" and safety will be HF/VHF transceivers.

The terminals on the transformer are nut & bolt, so if I ever need significantly different voltages, it would be easy to 'bolt on' an external module.

B

[robin0577]

MY thoughts are as follows..

Measure the unregulated DC output straight from the rectifier with the present configuration. Subtract the regulated output voltage from this to give the voltage drop across the regulator. Provided you adjust the transformer taps so that the voltage drop across the regulator remains the same or a fraction lower, no more heat will be generated than is being generated at present.

With a suitably designed heatsink that has a fan blowing through it you may be able to reduce the physical size using transistors of equivalent rating in a smaller package size, i.e. TO220 instead of TO3 but you'll need to check the heat dissipation carefully.

Modern capacitors of the sam rating will be fine and probably quite a bit smaller.

[Bazz4CQJ]

Hi Robin

in the case of the 3055's these are currently mounted in groups of 4 on thick aluminium panels 19" wide and about 8" high, there being 3 such panels.

Very passive and quite large!

B

[Ed_Dinning]

Hi Bazz, OK I'll reduce the complexity.

You have a transformer that is presently set up for 24v out at X amps. If you series /parallel the transformer for 12v out you can have 2 * X amps out
Note that we are only talking about the transformer here.
To get a regulated voltage of say 24v at the output the transformer will feed about
28/30v to the rectifier. This allows for losses in the rectifier and a minimum loss across the 3055's as the control element. This is the voltage you will see on the large caps.

If you reconfigure the taps on the transformer so you get about 1/2 the volts on the caps, that will be the transformer side sorted out.
With this configuration you will not get twice regulated current out as there will not nopw be sufficient transistors for the higher current. Tjey will give you the old level of max current though.

The chip regulator circuit will then need to be modified to set up the new required voltage. That might be the difficult bit if you are not used to working on power supplies, but is basically quite simple. If you manage to work out the circuit, post it here and we should be able to help.

A picture of the unit and the control board would be helpful.

Mail again if you need more info.

Ed

[Bazz4CQJ]

Pictures as requested! The board is one of two, the other being tucked inside so impossible to get a picture. The rectifiers are on the mid-section heat sink and the 12 3055's are visible on the 3 forward panels. Transformer is temporarily removed, but you can guess where that sits.

Thanks

B

[Bazz4CQJ]

It has transpired that Bill WME posted a handbook for this supply under a different designation in post #7 here https://www.vintage-radio.net/forum/...d.php?t=103475 I have not re-posted to keep safe re any copyright issues.

B

[llama]

Hi Bazz,
Something to consider is the use of a couple of Dell server switch-mode supplies available from many vendors on ebay. I bought a couple for less than £30 the pair and they each deliver 12V at 57 A. My colleagues on the WS-19 forum tell me they are electrically very quiet. There are instruction on the Internet on how to electrically isolate the 2nd one so it can be seriesed-up with the first to get your 24 V.

Worth considering as a quick, cheap and small solution to what you need.
Graham

[Anthony Thomas]

I was given an ancient power supply some years back that came from an equaly old computer used at the local Mercedes Dealership, it also had a huge transformer that must have weighed about 20 kilos. The power supply also had six 2N3055 connected in parrallel and used as a rectifier as near as I can deduce and these supplied the 5 Volt supplies. There were three 723 regulators that regulated the various plus and minus 5 Volt supplies. The 12Volt power supply had its own rectifier diodes (that were burnt out and the reasoni that the computer was replaced with a modern unit).

By comparrison I would guess that a SMPS could supply the power of this monstrosity quite easily apart from the 110V for the two cassette stations.

Tony

[Bazz4CQJ]

Yes, maybe I should try and define more precisely what I am working towards here. At the time I found this PSU in a dustbin, SMPS's were somewhat less prevalent than now. I do have a strong feeling though, that if I parted with that great transformer (in all senses of the word), then I would live to regret it

B

[Ed_Dinning]

Hi Bazz, looks like the PSU has some adjustment for output voltage in the form of a 10 turn pot, you could see what range that will give.
If you can get the circuit that would be more helpful.

Ed

[Richard675]

Bazz

You say your unit has been designed for 24vDC output which would mean the transformer unregulated DC input voltage is probably over 30v. A lot of volts will need to be "lost" to take it down to 13.8V or whatever. This will have to be done in the 3055's which will run quite warm.....

I would expect a 13.8v output supply to have around 20v DC input as a rule of thumb.

Are there any lower tappings on the transformer secondary? If so then some modding of the regulator board could follow on.

If you can, take a look at the output waveforms before condemning the existing C's etc.

This article might help you:
http://warc.org.uk/?page_id=404

HTH
Richard

PS I once re-wound a large transformer to achieve something similar, but I wouldn't recommend it!!

[Diabolical Artificer]

I'd stick with the 3055's they're good and reliable and 12 should supply 20A ok.

See schematic below which with a bit of alteration will be a less complex PSU. I'd put some diodes on the OP for protection (see LM317 data sheet) and you won't need the HV diodes in the schematic. you can replace the 5k pot with a fixed resistor, but I'd use a fixed resistor of the right value and a 1k pot so you can adjust the voltage when under load.

Andy.

[kellys_eye]

Forced air cooling tends to create hotspots on the base of the heatsink (device/sink junction) which may not be handled by the devices in question. To make a forced air flow cooling system work - assuming the 'doubling of dissipation' you may need to use a more effective heatsink (more fins) along with the air blower.

[Bazz4CQJ]

Hi Guys; thanks for your inputs. Just to re-iterate my post above, it has turned out that the circuit diagram is shown in post 7 of this thread https://www.vintage-radio.net/forum/...yright issues.. It looks like Kingshill made a range of LT PSU's probably based upon just 2 or 3 common core units, and additionally, they would sell you a core unit and you could customise it to suit your needs - and that's what I think I have.

The magnificent ALBION transformer (28lbs ~13kg) has huge (2BA ?) nut and bolt terminals and has 4 off 9V secondaries, 2 off 3v and 2 off 1V and Kingshill advice to customisers was to find a combination of secondaries which just gave you the max voltage you needed, so the 3055's could then take it easy. In the case of this unit the four 9V secs are wired in 2 parallel pairs in series (giving 18V) and then in series with both 3V and 1 1V secs, hence 25VAC(nominal) giving 24VDC.

I've had this unit for ages and never done anything more than test and repair it, but it is now under scrutiny as the dreaded "clearing out prior to a house move" era starts . I think I've come the view that the transformer and the 12 3055's are definite keeps, but if I really wanted to bring this beast back in to operation, then the control circuitry and the big caps could probably be replaced to advantage.

B

[Herald1360]

Don't forget you can also wire some of the secondaries in series opposition to get lower voltages- eg 9V - 3V = 6V. Might help for your 12V application.

[Bazz4CQJ]

Yes, Chris, you could have quite a lot of 'fun' trying out all possible combinations. Apart from obtaining the optimum voltage, do you know if including a reversed winding offers any other benefits, e.g. spike suppression?

B

[Herald1360]

I wouldn't think so. Ed?

[Ed_Dinning]

Hi Gents, assuming that the spikes are proportional to the winding's voltage, then series aiding/ opposing connections would do the same to the spike voltages, so no gains there.
In all probability the major form of spike coupling will be capacitive, so if one end of the winding on top of the pri can have one end earthed this may give results similar to an earthed shield, but I could be wrong!

Ed