making a dual rail +/- power supply with two LM317 and no LM337

[dentaku]

I know this is a naïve question but I'm going to ask it anyway
I've been looking through all my salvaged parts and lots of the stuff I recently got from an old man's collection of parts and I have two LM317 regulators. I wish there was an LM337 in there somewhere, but there isn't.

Question?
Is there there a way to make a dual rail +/- 15V power supply for powering dual supply op-amps with two LM317s? or is there no way around needing a negative voltage regulator?

[ElectroIrradiator]

There is, but you need a mains transformer with two isolated secondaries. Given that, then you can make two isolated/regulated LM317 supplies, and simply 'stack' them. If won't be as efficient as using an LM317+LM337, as you need 8 rectifier diodes instead of 4, but it will work.

[expertmax]

It's not worth the trouble in my opinion. Just a quick drive to radioshack or eBay and you can get the LM337 for under a buck, free shipping.

eBay auction: #http://www.ebay.com/itm/2-x-LM337-LM337T-ADJ-Negative-Voltage-Regulator-IC-1-5-/260815926833?pt=LH_DefaultDomain_0&hash=item3cb9d73631

[Paul Price]

All you need is a PNP power transistor, two small-power PNP  transistors, a current sense resistor,(.01 ohm 1-Watt) and a few 10K resistors, and an op-amp (LM358 will do) configured as an inverter with a gain of 1. You will need either two power transformers or else something like a 24V AC 70VA rated (3 Amp capable) center-tapped transformer and you can make a dual-tracking +- 15V 1.5 A power supply.

[dentaku]

As I suspected, it's rather impractical

By the way, kits like this http://www.jameco.com/webapp/wcs/stores/servlet/Product_10001_10001_20626_-1 come with a transformer and everything is self contained but I'm assuming if you where to make one that uses a wall wart and no transformer inside your enclosure you would only be able to make a positive voltage adjustable power supply.

Basically it would be a plugged in version of this battery powered adjustable supply http://www.ladyada.net/library/equipt/diypsupp.html ?

I think I'm going to start simple and do that first before I try building a proper one.

[Paul Price]

If you use two wallwarts that are isolated from each other and build two identical regulator circuits, then you can have two floating output supplies  and each one could be used either as a positive or a negative output supply, depending on whether you ground the positive or the negative output terminal.

The circuit I suggested in not that complicated! The Jameco dual output supply only goes down to 5V and it has 48 components. Mine has about half that many.

[c4757p]

On that note, old AC wallwarts are a decent source for line transformers. People throw them away all the time. If you don't want a big old wart hanging off your wall (or two! in this case), Dremel them open and take out the transformers.

[TerminalJack505]

If you use two wallwarts that are isolated from each other and build two identical regulator circuits, then you can have two floating output supplies  and each one could be used either as a positive or a negative output supply, depending on whether you ground the positive or the negative output terminal.

The circuit I suggested in not that complicated! The Jameco dual output supply only goes down to 5V and it has 48 components. Mine has about half that many.

Your schematic looks like a tracking supply.  I'm not sure if the OP wants the two outputs to be individually adjustable or not.

[dentaku]

If I was to take two 12 volt wall warts and connect the positive of one to the negative of the other...
1 - is this safe, or will it just make one get really hot?
2 - it looks to me to be the same as snapping together two 9 volt batteries (- of one battery into the + of the other) right?
this is the easiest way to get +9/0/-9 out of 9V batteries I've ever used

[alm]

Yes, as long as the output is not connected to ground (eg. if it doesn't have a grounding pin), then you can safely connect them like this. Don't do it if both of the supplies are grounded, however, then you'd be shorting them.

The presence of a ground pin does not mean that the output is connected to ground. You can measure (while unplugged) the resistance between ground pin and the positive and negative output to see if there's a low resistance path between the output and ground.

[dfmischler]

Yes, as long as the output is not connected to ground (eg. if it doesn't have a grounding pin), then you can safely connect them like this.

Aren't you concerned about capacitive coupling to hot or neutral in a switching mode wall-wart?  I have an ungrounded  12V supply here that shows 48V AC from the DC return to AC hot or neutral, although I'm sure there wouldn't be much current.

[alm]

Safety regulations usually dictate very low limits for this current. It will be shunted to ground during normal use (eg. plugged into a grounded computer), so I don't see any harm in shorting it. Grounded equipment will often have a very similar arrangement: one Y cap from hot to earth and one from neutral to earth. With even larger caps than permitted for double isolated equipment.

[dentaku]

I just built two identical LM317 circuits on a little breadboard an connected them together like snapping together two 9V batteries + to -.
One is powered by a 12V UPS battery and the other from a 13.8V wallwart. and it works. I set them both to 5V and one side shows 5V and the other -5V.

It's not terribly practical but for a beginner it's nice to see something work

[jakeisprobably]

All you need is a PNP power transistor, two small-power PNP  transistors, a current sense resistor,(.01 ohm 1-Watt) and a few 10K resistors, and an op-amp (LM358 will do) configured as an inverter with a gain of 1. You will need either two power transformers or else something like a 24V AC 70VA rated (3 Amp capable) center-tapped transformer and you can make a dual-tracking +- 15V 1.5 A power supply.

  Ancient Post, I know, but this thread came up on a google search for a dual tracking regulated power supply without a LM337. After seeing Mr Price's hand drawn schematic I knew I had the spares and junk parts to build his drawing. I put together a schematic on Fritzing to try out their software. I thought it was worth sharing with the next person to come across this thread. Thanks.

[Alex Eisenhut]

Where in Canada are you? I might have a 337 to spare.

[jakeisprobably]

I like to see what I can recycle and build from junk I already have as much as possible. The second I go to order something online, a digital black hole develops in my shopping cart, I blackout, and somehow manage to spend 100 times more than the one little chip I needed. I'm just a beginner with more time than sense, or money.

[rglover]

All you need is a PNP power transistor, two small-power PNP  transistors, a current sense resistor,(.01 ohm 1-Watt) and a few 10K resistors, and an op-amp (LM358 will do) configured as an inverter with a gain of 1. You will need either two power transformers or else something like a 24V AC 70VA rated (3 Amp capable) center-tapped transformer and you can make a dual-tracking +- 15V 1.5 A power supply.

  Ancient Post, I know, but this thread came up on a google search for a dual tracking regulated power supply without a LM337. After seeing Mr Price's hand drawn schematic I knew I had the spares and junk parts to build his drawing. I put together a schematic on Fritzing to try out their software. I thought it was worth sharing with the next person to come across this thread. Thanks.

Hi. I came across this old post, and looking at the hand drawing it specifies two 1N4184 diodes which are 82v zeners. This seemed odd and I wondered if they are supposed to be 1N4148.

[Paul Rose]

I wondered if they are supposed to be 1N4148.

   Yes, almost certainly.

[rglover]

I wondered if they are supposed to be 1N4148.

   Yes, almost certainly.

Thanks Paul. After 50 years working in and with electronics I didn't think I had lost the ability to spot a simple transposition error.

[David Hess]

As I suspected, it's rather impractical

There is nothing impractical about the suggestion to make two isolated supplies and stack their outputs.  It was commonly done with a transformer that had dual identical secondaries.

[S. Petrukhin]

What about the virtual midpoint?
Or a single-period rectifier just that needs more capacitors.

[Mechatrommer]

poorman transistors based adjustable and tracking dual rail +/- power supply with one (or no) LM317 and no LM337

[MarkF]

What about the virtual midpoint?
Or a single-period rectifier just that needs more capacitors.

The circuit appears to have a lot of issues:

  See https://www.eevblog.com/forum/beginners/virtual-ground-push-pull/msg2343744/#msg2343744

     

[S. Petrukhin]

What about the virtual midpoint?
Or a single-period rectifier just that needs more capacitors.

The circuit appears to have a lot of issues:

  See https://www.eevblog.com/forum/beginners/virtual-ground-push-pull/msg2343744/#msg2343744

     

I even used this circuit to power an amplifier, but that was a long time ago, back in my youth, and I didn't have any good instruments to test. I looked with a simple oscilloscope, there were fluctuations. It seems to me that the question here is the speed of the operational amplifier and the ability to control the output transistors. I used composite transistors at the output.

[David Hess]

What about the virtual midpoint?

It is possible but more complex and difficult because it requires a class-AB output stage for equivalent performance.