Stumped by odd LM7815 output -- request help

$10 for a piece of sheet ali? Wherever you're buying it from is steaming you.

NT

** With a CT transformer and a bridge rectifier, both filter caps always charge together at switch on.

But with a single winding and two diode voltage doubler, the caps charge during alternate half cycles - so at switch on, one rail can be momentarily fully up while the other is at zero.

Some PSU makers take extra measures to prevent asymmetrical DC rail conditions.

In the above, Tr1 & TR2 will pull either +/-17 volt rail down unless the other is operating too. Otherwise, a simple fuse failure would result in damage to the PSU and load which consists of lots of op-amps and tantalum caps.

.... Phil

The current was going to be 2A, and the headroom 10V to 20V; you need to dissipate 20W, maybe more. It's a big piece of aluminum, and yes, everyone who sells heatsinks applies a high markup.

I don't agree at all. If the heat sink were small enough for the junctions to track, there's no way you could be dissipating 80W.

Yup

Nope. It's possible in the spherical-cow universe, but the real one has too many uncontrolled variables, such as somebody putting a book on top of the supply. Natural convection is really easy to screw up, which will make the temperature gradients in the HS change all over the place.

Yup.

Agreed. But this is a one-off.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

Do let us know if, in all this ranting, you have found the information you're looking for.

Personally I'd be tempted to use a 723, assuming that they're still in current production. I may still end up using a high-voltage version of the 7815 -- but the temptation would be there.

I'm not sure if anyone else has really tried to ram this home, but if you have the flexibility to choose a different transformer next time, do so. You want to choose a transformer that'll just keep the valleys of the ripple voltage of the raw supply above the dropout voltage of the regulator at your worst-case power line sag. This takes some tedious computation and even component selection to figure out, but you end up with a supply that burns a minimum of power in the regulator; even if you don't care about power consumption it means the thing is smaller and cooler overall.

Or, just buy a 15V switcher from Farnell or DigiKey or whoever is in your neck of the woods.

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com

Or just pre-regulate each raw side with a power-transistor-resistor- zener so most of the dissipation is in the power transistors. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142     Skype: skypeanalog  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.

From the description it sounds like the OP is making a +/- power supply with 0v or ground at the center. They'd need a 7815 and a 7915 to pull that off. Nobody else has brough this up. Am I reading the question wrong?

Right. It seems the OP _is_ using a center-tapped transformer. It's unclear if he has the bridge rectifier connected up properly to yield plus and minus raw supplies. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142     Skype: skypeanalog  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.

precisely, so buy your ali elsewhere

NT

I forgot what the project was, but I was about to whip up the same thing- a +/- 15 volt power supply. I said screw it and just bought one of those "international" style linear power supplies. Surprise, it was made with a

7815 and a 7915. It was a center tap half bridge design, no doubt to eliminate the cost of two more rectifiers and those elevated solder stud turret things/heatsinks and assembly costs. The transformer is way oversized, but may have been the smallest size they made that would keep the bobbin from flopping around. Cost is no doubt a huge factor in how they designed those things. Pretty sure mine as a silkscreen copyright of 1990.

I'm missing something here. If it's +-15, you need both halves of the bridge as well as the CT. Both rails are full-wave rectified.

Or is it half-wave rectified? (eeew)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

I don't think he mentioned a negative output, but it seemed like he was describing a full wave bridge on a center tapped transformer with 30-0-30 output windings. That seems a poor choice to begin with so it seems that he is a hobbyist using something from the junkbox and without much basic knowledge of electronics and measurement techniques.

I made a three-output linear supply with +/-15V and 5V using an 8-0-8 (16VCT) transformer, with a FWCT 2-diode connection for raw 12 VDC for the

7805, and doubler circuits using two capacitors and two diodes each for raw +/-24 VDC for the 7815 and 7915. The doubler circuit provides a "soft" raw supply that drops as more current is drawn and thus limits the power dissipation of the regulator, and also behaves well if short-circuited.

Paul

Was this design based around what parts were available, or was the doubler part of the original plan?

The first version of this circuit used a PCB-mounted 40VCT transformer that came from an even older design and I had a few of the parts already. This provided raw +/- 26 VDC which was OK for the low current +/-15V analog section, but required a series power resistor (in lieu of a heatsink) for the 5V logic power supply. I also had a huge quantity (100+) of 16VCT chassis mount transformers and I came up with the idea of using the doubler configuration. It works well and has proven reliable after 20 years of operation in the field, and now is my preferred design for such applications.

Paul

ler

at

er

Reminds me of a story of a product with the wrong transformer fitted. Tryin g to remember details... the guitar amp worked but overheated rapidly as th e PSU V was about double what it should be. Rejig the rectifier so both sid es of the transformer half-wave fed a single + rail instead of creating dou ble the total V_out. Return the amp ground to a pair of power zeners across the psu - I forget what the power zeners were circuit-wise. Make sure the fusing was good! Icky but worked happily enough.

NT

That's amusing- all parts of the story.