Kit bench power supply 240V to 1V/12V

[gildasd]

Before posting this I did a check of past topics with the search function, but found nothing suitable.
I also researched Jameco, Sparkfun, Vellemans and Gotron.

Ideally, this would be a kit that is fully NON soldered together.
- Work on 240V Euro mains.
- Go as low as 0.7V and as high as 12V, but I can live with 2.5V to 9V range.
- Have plus/minus voltages.
- 1A is more than sufficient, 0.25A ideal. I'll modify a PC PSU later for bigger loads.
- Be in a inclosure.
- Have potentiometers that can be used regularly (problematic with the Jameco kits)
- I don't need voltmeter, but if it does have it, I'll take it.

The closest I've found are:
- http://www.jameco.com/Jameco/Products/ProdDS/20626%20Manual%20Revision%203.pdf?CID=CLB
But it's meant to be set up and then dedicated to a device.
- http://www.electronics-lab.com/projects/power/001/index.html
But it seems to have been discontinued (like many things in Greece...).
- Dave's mythical USB power supply...
But then I also want a Unicorn and a free energy source.

I know it's easy to find one new or used for less than 80€ delivered, but that's not the point.

Edit:
- http://tuxgraphics.org/electronics/201005/bench-power-supply-v3.shtml
Seems oK but there's no neutral line... (and is that important?):
- http://www.amazon.com/Elenco-Triple-Output-Power-Supply/dp/B0002DT0GU
Is it 240? Looks perfect! (Re-edit - it's not DUH!)
Could I modify the kit with a different transformer without risking magic smoke? Using any of these for example:
                    http://www.gotron.be/en/printtransformator-1-8va-15v-120ma.html
                    http://www.gotron.be/en/open-frame-transformator-15-0-15v-1000ma.html

[mariush]

Why don't you MAKE one?

Don't stress yourself over the 240v, most transformers designed for 230v AC will run at 240v AC without any problems.

Get enclosure, get two potentiometers, get a couple linear regulators. 
Get a transformer with two secondary windings or a center tap, get a bridge rectifier or four diodes ... Set it all up.

There's lots of online stores that ship to Greece (Farnell, RS-Components,  tme.eu in Europe, Mouser and Digikey in US so a bit more expensive shipping, Tayda Electronics in Asia, eBay etc etc)

A 24v AC transformer with center tap or one with 2 12v windings will work, four 1n400* diodes will do for a rectifier, LM317 for positive linear regulator,  LM299 or LM337 for negative regulator... datasheets for these chips have example circuits you can use to build everything.

[gildasd]

I'm not in Greece, but Belgium, however, the manufacturer/distributor of one of best kits is...

This is the first build for me in over 15 years, so I wanted a nice, good looking "all in one" kit to faf around to ease back into it.

In desperation, I was looking at using  multiple stages (7812, 7809, 7805, LM3940-3.3 etc) based on an universal open frame transformer.
But not having done this in a long time, I'll probably be still procrastinating over data sheets and breadboards this time next year.
With a kit, your foot is already on the ladder.

[mij59]

Just buy one, don't turn it on, take it apart !
There's your kit.

[gildasd]

Just buy one, don't turn it on, take it apart !
There's your kit.

Good idea, but not my cup of tea.

Anyhow, I've downloaded the schematic for the ELENCO XP-720:
http://www.elenco.com/admin_data/pdffiles/XP720Kweb.pdf Go to page 18.
I'm trying to reverse engineer it...

But a big problem from the get go: the transformer...
I think this is purely an ELENCO part. I can't find affordable substitutes with 3 windings giving 4 secondary voltages.
The outputs are 12.6AC, 6.3AC and I have no idea for the 5V/3A output part of the circuit.

Could this single transformer be replaced by two run of the mill open frame parts?
http://www.gotron.be/en/transfo/draadtransfo/open-frame-transformator-6-0-6v-100ma.html
http://www.gotron.be/en/transfo/draadtransfo/open-frame-transformator-15-0-15v-1000ma.html

Or am I barking up a dead end street and should start with another schematic better adapted to "off the shelf" parts and my lowly level?

This is rather critical, I will probably be building this while on duty on a vessel: I'm pretty sure that Digikey does not deliver to the middle of the North Sea...

[mariush]

That is a  transformer with 2 windings ... 1 windings of about 24v AC each with a center tap and one winding that's maybe 6-8v AC.

The last winding is just for that fixed 5v @ 3A - you can omit that last section completely and the power supply will work just fine with the two adjustable sections.

You can get 1 transformer designed for 24v ac with center tap or with two separate 12v secondary windings and you can also get a separate transformer for that 5v dc @ 3a if you want.

Such transformers aren't that expensive, for example here's one that should work : http://uk.farnell.com/pro-power/ctfcs50-6/transformer-50va-2-x-6v/dp/1780889

To be able to do that 1.25v - 15v  @ 1A, each winding of the transformer needs to be capable of 15v * 1A = 15w. So it needs to have a rating of about 15 / 0.62 = 24 VA or better. Since you have  two separate windings, it means your transformer should be at least rated for 2*24 = 48VA. 
Here's an example: http://uk.farnell.com/pro-power/ctfc50-12/transformer-50va-2x-12v/dp/1780857

http://www.gotron.be/en/transfo/draadtransfo/open-frame-transformator-6-0-6v-100ma.html
^ as the link says, this one is rated for 12v 100mA or about 1.5 watts. it's crap.

http://www.gotron.be/en/transfo/draadtransfo/open-frame-transformator-15-0-15v-1000ma.html

This one is a 30v transformer with center tap.  Yes, it will work, but since it's only 30VA, you're only going to get about.. maybe 0.5A out of each output.

A transformer with 15v on each side will be better but the regulator will be much hotter all the time, as it dissipates the difference between input voltage and output voltage as heat. The transformer with 12v windings is a bit lower than 12.6v but you can counteract that by raising the capacitors from 2200uF to 3300uF or 4700uF (they need to be rated for 25v or higher)

[gildasd]

THanks, I start with that.

I'll do Kicad schematic and post it here with a bill of materials.

[gildasd]

I've tried to listen to the recommendations and got to this point:

The objective has changed, mostly because I can only get a "shit" transformer for now
(I'm currently a student thus no VISA, thus only shops in Belgium - for a all in one kit, I could bend this rule).
This is "trying to be a 1.2V to 12V, 0.5A bench supply - no more no less.
It's missing a couple of basic Voltmeters that I'm trying to source.

I have big doubts, most notably about the resistor values...
I'll save money by using a re purposed enclosure and CPU heat-sinks...
The resistors and capacitors will be in assortment boxes to troubleshoot the magic smoke...

Be gentle, this is my 1st attempt at KIcad.

[Jebnor]

If you are looking for something to get you started, You can look at converting an old (or new for that matter) computer power supply into a bench supply.  You'll get -12v, -5v, 3v3, 5v and 12v. Lots of Amps on the positive rails.  300W computer power supplies are cheap and you can probably pick one up locally for not that much.

A couple of people on youtube have made videos about it.



I don't know if that is what you are looking for, but it might work until you can get the parts you need.

[gildasd]

THe PSU mod was going to be project N°2

On that subject, is it possible to add voltage regulation to individual positive rails (if I need 4.5V or 9V for exemple)?

[Jebnor]

On that subject, is it possible to add voltage regulation to individual positive rails (if I need 4.5V or 9V for exemple)?

Absolutely. Not a problem what so ever.  It's a common practice.  Try looking up circuits of the lm317.  With that you can have any input voltage higher than the output voltage + ~2V.  so, for 9v output, a 12V input will work just fine.  The 4.5V output is a little trickier, as you'll need at least 6.5v to get it to work, so the 5V rail isn't viable, you'll have to again use the 12v rail. You'll just have to watch out for heat -- a heatsink will be needed.

See also: Zener Regulator, lm7805, lm7809.

[dentaku]

I've tried to listen to the recommendations and got to this point:

The objective has changed, mostly because I can only get a "shit" transformer for now
(I'm currently a student thus no VISA, thus only shops in Belgium - for a all in one kit, I could bend this rule).
This is "trying to be a 1.2V to 12V, 0.5A bench supply - no more no less.
It's missing a couple of basic Voltmeters that I'm trying to source.

I have big doubts, most notably about the resistor values...
I'll save money by using a re purposed enclosure and CPU heat-sinks...
The resistors and capacitors will be in assortment boxes to troubleshoot the magic smoke...

Be gentle, this is my 1st attempt at KIcad.

I just got a 10 turn 2K linear potentiometer and some 4700uH 35V caps but I'm waiting for some stripboard to build a power supply similar to yours.
Mine is going to have an analog panel voltage meter.

[gildasd]

I like what you did with the "coarse" and "fine" adjustment, I think i'll snap that feature up 

But there's a few things I don't understand:
- What is the purpose of the D2 D5?
- Is R5 sufficient? Is it because you put it after the regulator so it only lights on when the voltage is high enough to trip it?
- Did you look at something else than a 317K 337K combo? The voltage drop is pretty big at 1.8V to 1.6V...
- Why don't you have a neutral out?

[mariush]

D2 and D5 protect the regulator from voltage going back into the regulator.

See datasheet, page 11: http://www.ti.com/lit/ds/symlink/lm117.pdf   or this other datasheet, page 7: http://www.onsemi.com/pub_link/Collateral/LM317-D.PDF

R5 is used to limit the current going into the LED, and just for that.  LED has a voltage drop of about 1.8-3v, let's say 2.5v  so with 5v out from lm316 the led will have  (5v - 2.5v)/1000 = 2.5mA going through it.  With ~ 20v on the regulator, you have about 17.5mA.

So not quite a smart idea, because at 2-3mA you may not see the led light up, it depends on led - if you use a 3v blue led then at 17mA it may be a flashlight

No reason to use the K versions (assume TO-3). Maybe you'd pick them for better heat transfer, but really nowadays to-220 packages are just as good or the heatsink can be cooled easily.

Schematic wise, I'm bothered by C10, I'm not sure why 10uF is used.
Also not sure why there's a 1n4007 and a 1n4004 - why buy two separate diodes when you can buy one type in volume and save money. Maybe it's a typo.

The 2k and 500 ohm seem a bit badly chosen... at the maximum, you'd have Vout = 1.25 ( 1 + 2500/120) = ~ 27.5v  which is more than the transformer's output voltage - voltage drop in rectifier - voltage drop in linear regulator.

[gildasd]

In my course work (in french) the C10 is part of the filter. no idea what it does...

4007 and 4004 are different, it's in the datasheet, but I've forgotten, because it makes not difference for this kind of usage.

I'll check your remark about the 2K and the 500, but it's bed time here!

Cheers!

[mariush]

The main difference between 1n4000 series diodes is the maximum voltage. The other differences are not relevant considering the usage in the circuit. So might as well use 1n4007 everywhere.

I don't see a reason for that C10, because there's already a 4700uF that does the job. I could see a 0.1uF capacitor there to filter some noise passing through the large capacitor, but the 10uF would be pointless.

[dentaku]

D2 and D5 protect the regulator from voltage going back into the regulator.

See datasheet, page 11: http://www.ti.com/lit/ds/symlink/lm117.pdf   or this other datasheet, page 7: http://www.onsemi.com/pub_link/Collateral/LM317-D.PDF

R5 is used to limit the current going into the LED, and just for that.  LED has a voltage drop of about 1.8-3v, let's say 2.5v  so with 5v out from lm316 the led will have  (5v - 2.5v)/1000 = 2.5mA going through it.  With ~ 20v on the regulator, you have about 17.5mA.

So not quite a smart idea, because at 2-3mA you may not see the led light up, it depends on led - if you use a 3v blue led then at 17mA it may be a flashlight

The LED is just there because I needed something to indicate it was working while it's on my breadboard. I don't know if it's terribly useful in real life and yes, it dims allot when working at lower voltages

No reason to use the K versions (assume TO-3). Maybe you'd pick them for better heat transfer, but really nowadays to-220 packages are just as good or the heatsink can be cooled easily.

Don't take the exact model names to be to important. I just used what I could get for the simulation. In real life I'm going to use slightly different TO-220models.

Schematic wise, I'm bothered by C10, I'm not sure why 10uF is used.

I guess it came from a discussion we had here https://www.eevblog.com/forum/projects/which-type-of-0-1uf-cap-should-i-use-with-an-lm317/msg458115/#msg458115

Also not sure why there's a 1n4007 and a 1n4004 - why buy two separate diodes when you can buy one type in volume and save money. Maybe it's a typo.

Oops, I guess I missed that but I doubt it would make a difference anyway.

The 2k and 500 ohm seem a bit badly chosen... at the maximum, you'd have Vout = 1.25 ( 1 + 2500/120) = ~ 27.5v  which is more than the transformer's output voltage - voltage drop in rectifier - voltage drop in linear regulator.

I don't actually need the 500ohm pot anymore because I recently got a 2K 10 turn pot for the positive voltage but since the front panel is already made and the graphics printed/laminated I'm going to use it anyway

Interesting comments though...
I don't want to de-rail the original posters discussion anyway. I'm VERY beginner and this is still only on a breadboard at the moment. I just wanted to show gildasd that I was in the same situation.

[gildasd]

I don't want to de-rail the original posters discussion anyway. I'm VERY beginner and this is still only on a breadboard at the moment. I just wanted to show gildasd that I was in the same situation.

Derail all you want, it's very interesting.
I'm trying to get dead PC power supplies and yank out the transformers until I find a viable one, so I might be changing my layout a bit...

[mariush]

Don't waste your time.

The transformers inside pc power supplies are designed to work at high frequency (20kHz to 100-150kHz), not your rectified mains of 100-120 Hz.  The high frequency is the reason why they're so small and capable of "transforming" so much power, compared to classic transformers.

You can use a whole pc power supply as an input to your linear regulators, but that somewhat defeats the purpose of using linear regulators, which is having as little "noise" as possible. Some of the noise produced by a switching power supply (like a pc power supply, laptop adapter etc) can not be filtered by the linear regulators.
Now, for microcontrollers, leds, powering various stuff like mp3 players etc, the output of your linear regulators (with input from a switching psu) will be good enough, only for some few noise sensitive applications you really need classic transformers.

You can tweak regular pc power supplies to output more than 12v and 5v by playing with the feedback mechanism and tweaking the overvoltage protections. But above 13.5-14v, you'd have to disconnect the fan and power it from the 5v (which will also be higher) and above 15-15.5 you'd have to replace the 16v rated capacitors in the power supply with 25-35v rated ones.  18-23v can be easily achieved with lots of power supplies.
But since you're struggling to understand how this circuit works, it's better to stay away from pc power supplies that can kill you if you screw something up.  Classic transformers are much safer and a good starting point.

[gildasd]

Ok, I'll avoid dying 

[dentaku]

You'll have much better luck finding useful transformers in old stereo equipment.

[gildasd]

If I can't get 1 good 18V-ish transformer, why not use two 9V ones? One of the two is plugged with the line/neutral reversed. Will that create a negative current?
Or will the winding not being reversed annul any advantage? I could always hand wind one?
Or take out the 9V side and flip it over 180deg?

Of course the result is a +6/0/-6 power supply... (I chose 9V just for the feasibility study, if this works, I'll use a different rating)
The updated schematic is below.

Will this kill me/burn my house down?

[dentaku]

If I can't get 1 good 18V-ish transformer, why not use two 9V ones? One of the two is plugged with the line/neutral reversed. Will that create a negative current?
Or will the winding not being reversed annul any advantage? I could always hand wind one?
Or take out the 9V side and flip it over 180deg?

Of course the result is a +6/0/-6 power supply... (I chose 9V just for the feasibility study, if this works, I'll use a different rating)
The updated schematic is below.

Will this kill me/burn my house down?

I've done that with two AC plugpacks/wallwarts before.
Hlfspace made a nice diagram of it here. https://www.eevblog.com/forum/beginners/two-400ma-adapters-versus-one-center-tapped-transformer-of-unknown-ma/msg466774/#msg466774