Universal power supply for your projects!

Hi!
This is a universal powersupply I just draw! Still untested but the pcb is made and as soon as I tested it I will send the gerbers to Gustav!
Except for the Negative part I use this layout in most of my units.

Needed an universal supply for some units I'm racking right now so I came up with this little project! Seen some requests lately for this kind of supply.
Any suggestions just PM me and I'll see what I can do!

http://www.vintagedesign.halmstad.net/diy/schem_mods.htm

Will add more small helpful gizmos to this page soon!
Hi Z driver is my own version of Peter Purpose driver!
VU / Headroom meter is a LM3915 based 5 led driver with inputbuffer
BA226 is a modified Neve unity ampstage modified to drive 600 ohms load. Last time I checked it it was -0.05dB 10hz - 48kHz

Hi TK,

Great idea!

It looks like a simple and universally useable circuit indeed.

But..

..Won't the 10uF capacitor at the reference pins at the 317/337's mess up regulation at higher frequencies? It helps only a little at stabilizing the DC levels - I think I would leave them out alltogether..

Or does they have a function that I don't see..?

Jakob E.

Jakob,

The LM317/337 datasheets recommend them to help improve ripple rejection. The Nat Sem app sheet says that with a 10u cap from Adj to Ground the ripple rejection will be 80dB. The diode from Vout to Adj is then needed to stop the cap discharging through the 3*7's internals during a short circuit condition at the output.

I think this is why the LM3*7's are regarded as an improvement over the 78xx and 79xx series of regulators. I use LM3*7's in all my PSU's now, because if you buy them in bulk they become as cheap as 78xx and 79xx and you have full control over your output voltages.

Tekay, thanks for this PSU package!

:guinness: :sam:

Mark

This cap's function was mentioned in the regulators data paper. Thats why I put it there...

"Improved ripple rejection (80dB) can be accomplished by adding a 10uF capacitor from the adjust pin to ground. Increasing the size of the capacitor to 20uF will help ripple rejection at low output voltage since the reactance of this capacitor should be small compared to the voltage setting resistor"

I see - thanks for sorting that out.

It just seems to me that it's regulation is affected a lot at high frequencies, as the "error-signal" is shunted to gnd by the cap. But if it's in the datasheet, I would certainly not complain!

:razz:

Jakob E.

Now all we need is someone to find the RS part numbers and we're made :grin: :grin: :grin:

[quote author="gyraf"]I see - thanks for sorting that out.

It just seems to me that it's regulation is affected a lot at high frequencies, as the "error-signal" is shunted to gnd by the cap. But if it's in the datasheet, I would certainly not complain!

:razz:

Jakob E.[/quote]

You can think of the adjust pin as an input, so any noise on the pin can get back into the regulator.

There was a good article on power supplies in TAA that included the jung regulator stuff, but it also had graphs of the noise of the 317 with and without that cap, and the graph looked better all the way across the spectrum with a cap on the adjust pin.

> seems to me that it's regulation is affected a lot at high frequencies, as the "error-signal" is shunted to gnd by the cap.

The 317 series senses error between Vout and Ref. Not between Ref and Ground as in many other regulators. So the cap from Ref to Gnd actually improves high frequency regulation.

The downside is that when you drop a screwdriver on the output, Out and In go down very fast but Ref is held up by that cap, which discharges through some junction backwards and pops the chip. That does not always happen in real life, but I thought you were supposed to use some protection diodes if you used that cap. (I'm too lazy to check.)

TK,

This is very useful. Thank You for starting this project.
Did you say you can use to independent 24V transformers
to feed it?

The other TK :grin:

Yeah, there should be a diode from the adjust pin to the output.

It seems that you have it included on the +48 but not on the +v and -V rails.

You're right Brad! I will add the diod to the layout and schematics! This diod wasn't needed when you're using the LT317! That's why I fogot to add it!

tk@halmi! I made it to be feeded with two separate 24V AC windings but one of the layouts shows how to connect it with a centertapped 24 V transformer!

[quote="PRRThe 317 series senses error between Vout and Ref. Not between Ref and Ground as in many other regulators. So the cap from Ref to Gnd actually improves high frequency regulation. [/quote]

Thanks, PRR!

That clears up the confusion I had about this circuit..!

Jakob E.

bump

Hmm, I have one on my bench right now, powered with 28vac secondary. Odd behaviour: when AC is plugged in, everything is OK and the phantom is 48 volts. Voltage before the rectifier is some ~75 volts or so. However, it quickly starts dropping and reaches 45 volts in about two minutes, which obviously is not enough for the regulator.

Layout checked many times, can't see any errors. I made the two 100uF caps larger (200uF) and the 220uF/100v cap in my version is 330uF/100v.

:?

I'm not sure why you used two bridges instead of one.. using LMXXX negates a lot of reasons for using two bridges IMO. I would place all of the diff rails on one bridge and then use the other for 48v. 5v could be derived from either. but if it works good then :thumb:

It would be cool if the pcb was constructed that if someone only need the + power, we only need to populate that side. :thumb: