diy dual lab power supply

[Meshuganah]

Have a look at them LM317T and LM350T devices.
You can build a great PSU around these. put in a big transformer so that you can double up on series pass power devices later.

They are really the 'gainclone' of PSU chips.

Let us all know how you go. I've done it and forgotten about it.
Never gives problems.

Always decouple input and output on these devices.....otherwise pretty well bullet proof.


Meshuganah

[Christer]

Carlos,

It would be more interesting to see AC sweeps showing load- and line regulation vs. frequency and a plot of loop gain and phase vs. frequency for the regulators. We don't want our PSU to start oscillating, do we?

[morpheus82]

thanx to all...seems to become interesting...right, my idea was to use something like lm317 and lm337 with pnp bjt to achieve more power and a current limiter...but by the datasheet i don't know how to use both the configurations at the same time...does anyone have a schematic?and...can i use a mosfet instead of pnp?in that case is simple to realise the current limiter with a current mirror!

[destroyer X]

And many counter oscilation things.

This schematic shown is the basic from the most basic things.

Fast to introduce into the simulator, small to publish, easy to understand.

The ones interested, can go deeper.

Related oscilations...well.... 35 years running in many homes, and connected to oscilator friendly equipments (radio frequency transmiters)...it was hardly tested.

regards guys...LM317 is fine...also 78XX series and many others.... long life to the electronics hobby!....nice thing.

regards,

Carlos

[Meshuganah]

Yes, LM317 and 337 will work but the nim. voltage (can't go to zero. Only 1.2v or something) I think the min volts on the 337 is not the same, but the LM350 was closer......better check which one is better 350 handles more current though.

You really can not go better......'gainclone' power supplies now.

Good luck Meshuganah

[Christer]

Carlos, I just wanted to point out the importance of stability. I haven't even tried to analyse the schematic you posted, so I am not making any claims about its performance. It could be very good. What I wanted to say was perhaps rather this general observation on lab PSUs.

Line regulation and load regulation are ususally very important aspects for any PSU. Perhaps load regulation is the most important for a lab PSU since it cannot be known in advance what loads it will be connected to. However, for the same reason stability is also very important, since it might happen to be hooked up to a very nast load. I personally think one should rather sacrifice high frequency load regulation to assure stability under all load conditions.

[dangus]

There's a schematic of a lab power supply regulator in the LM10 app notes. http://www.national.com/pf/LM/LM10.html
(Adjustable voltage and current, and output down to zero volts)
http://cache.national.com/ds/LM/LM10.pdf#page=13

Or here's a similar application using the LM723 (it probably needs some tweaking to make it suitable as an adjustable supply):
http://www.national.com/ds/LM/LM723.pdf#page=8

Here's a dual tracking supply using three-terminal regulators. If you used the HV version of the 317 family you might get to 50V, but you'd also need some kind of high-voltage op-amps.
http://www.national.com/an/AN/AN-103.pdf#page=10

There used to be an MC1466 chip that was ideally suited to a lab power supply; it needed only a handful of external components. http://www.datasheets.org.uk/search....466&sType=part
Searching reveals that Gary Lecomte has created an upgraded discrete version of it, and there's a PC layout too:
http://www3.telus.net/chemelec/Proje...466/MC1466.htm

One problem with an adjustable supply with a top-end of 50V is that if you short it and it current-limits at 1 A, there will be smoke unless you've kept SOA in mind when selecting the power transistors in the regulator. The MC1466 datasheet talks about stuff like this; it's worth reading even if you don't use the '1466 or Gary's clone.

[Meshuganah]

Christer,
Its about time to get real!
Read the application notes on these devices!
You will find optimum stability unless you wish to draw max amps at max volts. Obviously then you need a welding transformer and a big tail wind.
LM 317/337/350 is the best you are going to get! 723 regulators are/were (more like it) good but lots of parts count and unreliability....beleive me I have built lots.

Chuck out all the academic rubbish and get on with it. You will not be dissapointed.

These circuits are SIMPLE and WORK.....try them!

I am not called Meshuganah for nothing.

Regards and good luck, Meshuganah

[Christer]

Meshuganah,

I was not referring to IC regulators at all, but to discrete and/or more complex designs, such as the one Carlos showed. If you buy an IC regulator, you get something that is usually well designed and well proven to work. If you build a regulator from scratch, you must analyse it and consider its properties, just in the same way as if you build an amplifier from scratch. If you buy an IC regulator, the manufacturers have already done this for you.

I also did not in any way mean to claim that the particular schematic Carlos showed was wrong or doesn't work. I haven't even tried to analyse it. I just wanted to point out that a just a few a voltages from measurements or simulation doesn't say much about its performance. I also stated what I think are important considerations for lab PSUs.

[Christer]

BTW, the 317s and similar are not great performers thermally. They have low max Tj and high thermal resistance. Take for instance the LM317T from ON semi. It is rated at 1.5 A but you can't drop many volts over it at those currents.

The threadstarter wanted a 50 V 1 A supply. The LM317T has a max Tj of 125 deg. C and Rjc of 5 deg C/W. Let's be optimistic and calculate with an ambient temperature of 25 deg. C. In the theoretical optimal case of a perfect infinite heatsink, this means the regulator can dissipate 20 W, that is, at 1 A we can drop at most 20 V over it. Such heatsinks do not exist in reality, and one should also calculate with a higher ambient temperature to be on the safe side. Calculations must be done for the particular heatsink and conditions to be used, but say that we might realistically be able to drop at most 10 to 15 V at 1 A. That would mean that the PSU would only be able to output 1 A for output voltages down to 40 or maybe 35 Volts. Unless he plans to use it only for such high voltages, just a single 317 or similar seems not advisable. Using one of the standard designs with a 317 and an external pass transistor might do the job, though. Power transistors typically perform much better thermally.