Why are they building -15v regs

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Member
Joined 2006
Paid Member
Hi

I am studying the layout for a +-15v PSU for my new pre amp.

Untill now, I needed to use dedicated +15 and -15v regs like 317 / 337 with the known shortcomings.

Following this article: http://www.pedjarogic.com/gc/supplies.htm I understand that a simple topology adjustment enables me to use a positive reg for the negative rail. Quote"Negative regulators with necessary current ratings are neither easy to find nor are cheap these days, but if each secondary has its own rectifier, the positive regulators can be used both for positive and negative rails."

As I am not at ease regarding these matters, I need your advise on the possible shortcomings of such a design.

Why is it not used more frequently ?

Do the (-) regs shortcomings justify the savings by using only one rectifier bridge ?

Regards

Ricardo
 
OK... I'll way in on this. Please take what I say with a grain of salt and remember all advice is free so you get what you pay for...

In general, you can adjust your topology to drive from a single rail. The problem comes when you need to operate at either extreme of the amplifier limit. Meaning anytime you approach ground or the Vcc rail. Remember, all circuits need headroom to operate near the rails. That is what makes +/- voltages so easy to work with. Your signals are alrady working around the midpoint (GND) so everthing stays nice and semetric.

If I am wringing the last few pennies from a design, the negative rail is the last thing I drop because biasing everything to the correct point and making sure everything has the proper headroom can be a pain. Besides that, the parts that operate well near the rails are relatively expensive (rail-rail op amps) as opposed to using generic LM339/LM393 type parts.

I hope that helps...

Tony
 
This schematic uses LM338 which is a 5A positive floating regulator. It's true that 5A negative regulators are hard to find, but do you really need that much current? For a preamplifier?

Consider the LM317/LM337 pair, which is rated at 1.5A and widely available.

Only special regulators, like high current, high voltage, micropower or low dropout ones, are hard to find in its negative version.
 
Member
Joined 2006
Paid Member
Thank you Eva

Apart from the benefits of using two bridges, if I build a +-15v psu using two LM317 , should I expect any improvements regarding LM317 / 337 setup ?

Have positive regs better specs than negative ones ?

I sometimes experience instabilities while using 337 regs.

Regards

Ricardo
 
Member
Joined 2006
Paid Member
dtproff said:
That is what makes +/- voltages so easy to work with. Your signals are alrady working around the midpoint (GND) so everthing stays nice and semetric.

Thank you Tony

I agree that I must use two rails, I also know that two bridges are better than one but I wonder if LM317 equals LM337 in performance ?

Regards

Ricardo
 
Hi,
even those that design and build discrete regulators sometimes use two positive regulators rather than design a + & - pair.
The only down side to this twin positive regulator is the requirement for dual secondaries and dual bridge rectifiers.
Most will accept the extra cost because of the technical benefit accruing from +ve regulators.
 
Administrator
Joined 2004
Paid Member
AndrewT said:
Hi,
even those that design and build discrete regulators sometimes use two positive regulators rather than design a + & - pair.
The only down side to this twin positive regulator is the requirement for dual secondaries and dual bridge rectifiers.
Most will accept the extra cost because of the technical benefit accruing from +ve regulators.


Exactly what I do with my bi-polar regulators, and for diy it has the additional benefit of not requiring the debug of two new and potentially unstable designs. (Mine are a mix of discrete and op-amps somewhat along the line of the Jung/Sulzer super regulators.)
 
AX tech editor
Joined 2002
Paid Member
Speedskater said:
Somewhere in Walt Jung's articles in AudioXpress magazine. He writes the the LM317 is better than than the LM337 and suggests going the dual bridge route.

It's in one of the Walt Jung pdf's.

http://www.audioxpress.com/magsdirx/ax/addenda/index.htm


The 317 definitely has better specs than the 337, due to the use of a fast NPN series pass transistor in the 317 (the 337 by necessity uses a slower PNP pass transistor).
Ideally, one would want to use two isolated, center tapped secondaries for each polarity, but isolated single-ended secondaries will also work.

There's another advantages to going with separate secondaries. You can tie the two ground/return lines from each polarity together at a nice star point to avoid ground loops. Using a single secondary for both polarities will inevitaly combine ground currents of both in a single transformer return (like the center tap), much more difficult to avoid ground loops.

See fig 6 here: http://waltjung.org/PDFs/Regs_for_High_Perf_Audio_3.pdf

Jan Didden
 
Member
Joined 2006
Paid Member
Thank you so much jan

You made it perfectly clear now. (My suspicions where right after all ... NPN rules).

This might seem an exageration but I believe that the PSU quality is determinant in any sound aplication. (My CDP already has 6 independent TX and I would like to add another one in the RF amp to isolate it from the rest of the circuits).


This topology (I hope I can find a TX with independent output windings) will be used in my future pedja preamp.

Best regards

Ricardo
 
as a negative regulator i suggest you to use the positive one as a reference for a AOP + PNP transistor regulator.
 

Attachments

  • alim.png
    alim.png
    2.3 KB · Views: 288
Administrator
Joined 2004
Paid Member
It is just a tracking voltage regulator using an op-amp and a couple of transistors to generate the negative regulated voltage. All of the red "+" are connected to the output of the LM317 which provides the positive half of the supply output. In addition you need a negative voltage to feed to the negative half of the regulator which I suspect should be around -20V or so. It's quite clever and will work with either a conventional CT transformer and bridge or independent windings with separate bridges for each rail.

The op-amp uses the positive supply from the LM317 as a reference and is configured as a unity gain inverter, it in turn drives a non inverting transistor buffer which provides the actual load current. The whole thing works by comparing the negative output to the positive regulated supply and makes the negative rail track the positive. Q2 and R4 provide current limiting and short circuit protection by reducing the base drive to Q1 under overload conditions. R4 samples the output current and causes Q2 to turn on when the voltage across it exceeds 0.6 - 0.7V depending on the transistor used. For example a 56 ohm resistor at R4 would limit current to a little over 100mA.

Note that this a little more complicated than using the floating LM317 mentioned in earlier posts.. FWIW I would probably stick with that topology as you just build two identical regulators that are completely isolated from each other and then ground the "+" out of one to produce the negative supply and the "-" out of the other to produce the positive supply.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.