Regulated bipolar supply from single secondaries

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

living sounds

Well-known member
Joined
Jul 26, 2006
Messages
3,931
Location
Cologne, Germany
I've got a couple SCAMP modules I'd like to rack up. They need +/- input voltages which get regulated down via L791x/L781x.

Now I'd like to use a wall wart single ouput AC transformer to power the modules. In the past the schematic attached provided by PRR has always worked well used by itself, now I wonder if it can be combined with the regulators (with the voltages adapted for the application).

Stupid question probably, but I think I tried this once before and may have gotten sparks or smoke. Or voiced differently, is there a compelling reason not to do it like this, since technically the ground is derrived directly from one of the secondary connections.

Thanks!
 

Attachments

  • simple-PSU.gif
    simple-PSU.gif
    1.8 KB · Views: 69
No smoke will be released, but half wave rectification means 2x the ripple voltage so you will need larger capacitors than with full wave rectification.

While 12V ac will nominally make +/- 18v peak, I wouldn't count on a solid +/- 18v DC.. while results will vary with wall warts. I used to use a 16vac 1a wall wart that worked OK to generate +/-15v regulated rails as long as you didn't try to get the full amp out of it. 

JR
 
I've noticed that with some of the other linear PSUs I've build this way. I'll be using a 24V wall wart to get +/- 24V which will be regulated down to +/-15V. And big capacitors.

Thanks!
 
> I wonder if it can be combined with the regulators

Why not?

> I think I tried this once before and may have gotten sparks or smoke

A clever builder can release sparks and smoke from ANY build. That's what makes this hobby fun!
 
a problem with regulators for this design is the possibility of latch-up, if one side comes up before the other (and it will here), current can flow through the consuming circuit, and pull the opposite-side voltage regulator to less-than-zero, which will latch up the regulator preventing it from turning on.

Me asking National about this some 10 years ago:
http://wwwd.national.com/national/PowerMB.nsf/8178b1c14b1e9b6b8525624f0062fe9f/FB19F9D1B2E1D9B888256DE2003DB534?OpenDocument

From the National data sheet, the official and only way around this potential problem is adding a germanium power transistor as protection diode. Expensive and weird solution.

Jakob E.
 
gyraf said:
From the National data sheet, the official and only way around this potential problem is adding a germanium power transistor as protection diode. Expensive and weird solution.

Wouldn't it be a more simple and cheap solution to crudely pre-regulate with zeners?
 
gyraf said:
a problem with regulators for this design is the possibility of latch-up, if one side comes up before the other (and it will here), current can flow through the consuming circuit, and pull the opposite-side voltage regulator to less-than-zero, which will latch up the regulator preventing it from turning on.

Me asking National about this some 10 years ago:
http://wwwd.national.com/national/PowerMB.nsf/8178b1c14b1e9b6b8525624f0062fe9f/FB19F9D1B2E1D9B888256DE2003DB534?OpenDocument

From the National data sheet, the official and only way around this potential problem is adding a germanium power transistor as protection diode. Expensive and weird solution.

Jakob E.

I thought that was fixed some time ago... One old school tweak was to put a simple silicon diode in series with the ground leads, that way the latch up was prevented, but regulated output was one diode drop higher, and regulation specs a little worse.

JR
 
gyraf said:
a problem with regulators for this design is the possibility of latch-up, if one side comes up before the other (and it will here), current can flow through the consuming circuit, and pull the opposite-side voltage regulator to less-than-zero, which will latch up the regulator preventing it from turning on.
...
From the National data sheet, the official and only way around this potential problem is adding a germanium power transistor as protection diode. Expensive and weird solution.

JohnRoberts said:
I thought that was fixed some time ago... One old school tweak was to put a simple silicon diode in series with the ground leads, that way the latch up was prevented, but regulated output was one diode drop higher, and regulation specs a little worse.
Du.uuh.  I've used this simple PSU a lot in da old days and never encountered a problem.  But I was also putting a diode in series with the ground lead.  This was to get equal +ve & -ve clipping on 5532/4s and squeeze another 0.5dB of overload margin.  Wouldn't bother today.

Can't remember which supply to put the diode.  Are 79xx OK?

Gyraf, have you a link to the datasheet which recommends the Ge power transistor?  I'm looking at the TI/National LM340/LM78xx datasheet http://www.ti.com/lit/ds/snosbt0h/snosbt0h.pdf  No one seems to have answered your question either.
 
I'd be inclined just to try if it works but not sure if components might not be damaged if only one side fires up. Some of the transistors are hard to get:

http://www.groupdiy.com/index.php?topic=2430.0
 
ricardo,

It was in the application notes, not in the "standard" data sheet - that's why I had to ask at National.

datasheet-search engines now pollute googling for the application notes, so it's hard to find application notes. I did a search, but gave up after >30 google pages of misleading results.

Jakob E.
 
I saw the germanium arrangement in LM340 series app note

Page 10, figure 16

http://www.ti.com/general/docs/lit/getliterature.tsp?literatureNumber=snoa660&fileType=pdf
 
gyraf said:
From the National data sheet, the official and only way around this potential problem is adding a germanium power transistor as protection diode. Expensive and weird solution.

A couple of years ago I also had problems with 78xx/79xx latching in my GSSL. I solved using a schottky diode (1N5819) instead of germanium.
 
ricardo said:
gyraf said:
a problem with regulators for this design is the possibility of latch-up, if one side comes up before the other (and it will here), current can flow through the consuming circuit, and pull the opposite-side voltage regulator to less-than-zero, which will latch up the regulator preventing it from turning on.
...
From the National data sheet, the official and only way around this potential problem is adding a germanium power transistor as protection diode. Expensive and weird solution.

JohnRoberts said:
I thought that was fixed some time ago... One old school tweak was to put a simple silicon diode in series with the ground leads, that way the latch up was prevented, but regulated output was one diode drop higher, and regulation specs a little worse.
Du.uuh.  I've used this simple PSU a lot in da old days and never encountered a problem.  But I was also putting a diode in series with the ground lead.  This was to get equal +ve & -ve clipping on 5532/4s and squeeze another 0.5dB of overload margin.  Wouldn't bother today.

Can't remember which supply to put the diode.  Are 79xx OK?

One engineer who worked in my mixer group religiously put diodes in both ground legs. At less than a penny a piece I didn't complain.

I used many simple 3 terminal regulators (78/78xx) with half wave rectified wall wart PS with no extra diodes and no latch up problems at all.  FWIW most of my applications were modest current draw, and lots of rail capacitance that may have influenced the latch up behavior.

While this is a more subtle issue, when using a half wave rectified supply for bipolar circuits, turn on transients can be very much affected by which leg comes up first. That same engineer who used diodes on every regulator went as far as to add a trick power supply in at least on product where it waited for the second rail to come up, before applying power to both. I think I only saw him do that in one product.

JR
 

Latest posts

Back
Top