Dual pos. regulators after an SMPS

[redrabbit]

I've been surfing online for an answer...cannot find one, so I'm now here.

Straight to the point....

I have SMPS +43v out.

Can I put linear regulators (LT1084cp) on it's output......in the (often seen) wiring method of using 2 positive regulators to get V+/0v/V- ??

The reason I am unsure, is that usually this method of using 2 positive regulators requires dual/split secondaries on a Xformer. But I am connecting it to the SMPS DC output, and it's single ground point.

....or are my worries unfounded because it's DC already, i.e. no rectification ?

Use...powering a chipamp
Why?...to reduce the SMPS to a more reasonable +35v

thanks

 

[redrabbit]

Here's more details if needed......

I'm building the amp with one SMPS per channel....so it's a dual-mono configuration, but housed in one box.
I've read the "discussions" on linear vs SMPS power, and the cost issues as well. I happen to already have these 2 SMPS units...been sitting there for years....and have most of the parts to add post-regulation.
The purpose of this amp....is to be portable(with a handle)...to bring to the jobsite. So huge Xformers, I'd rather avoid.

I load-tested these identical SMPS units at full power (48 ohm/160 watt resistor load)...and hardly got identical output voltages.
Unit A) +35.8v at full load.
Unit B) +40.5v at full load.
Though not scientific, to even-things-out, I reckon post-regulation should shoot for the lower voltage unit (~35v)
If they did not differ so much, I could just filter the outputs and maybe drop the voltage by almost 3v by adding 2 diodes in series....and call it a day.

Mock-up...w/active fan cooling:. (that's the inner case, the outer case is from a tektronix waveform monitor !!)

[/IMG]

 

[Budgie]

I think it would work (I have done similar successfully) - but why not use a positive and a negative regulator instead of two positive regulators?

 

[redrabbit]

Thanks for replying.
I suppose I could use LM317/LM337 with some power transistors....but to answer your question....

Ease of use.
LT1084 is an LDO.
Lack of a 7 amp negative counterpart.
Lower parts count (and space) than adding extra power transistors to boost a lesser rated regulator.
I already have the LT1084's.

 

[Sorento]

one problem though:
what if current drawn from the positive supply is bigger than from the negative ?
since these regulators can't sink current
so, where's the return path for the current from the positive regulator ?
I had a similar case and because of that decided against ...

are voltages under low load also different ? if so I'd rather adjust one of the PSUs voltages; usually there's a voltage devider in the feedback loop often with a pot ...

otherwise, if one PSU sags from 40v to 35 under load and the other doesn't, then I'd suspect that there's something wrong with it or the wiring needing investigation ...

 

[redrabbit]

They both sag under load. I think that's typical from most SMPS.

Yes, the units I have do have a pot on them, on an attached daughterboard......so are regulated on the output. However, the trim is not wide enough to bring these units close enough together (at least to my liking).

I adjusted the trimpots, to see how close I could get them.
Tests across the +43 and -43 pins....

..............800 ohm load........150 ohm load..........50 ohm load
unit A)..........91.......................85.5....................80
unit B)...........93......................82.5....................73


One unit sags 11v (5.5v per rail)
but the other sags 20v (10v per rail)

You can the trim is effective at a very light load, in fact, it brings unit B, "high".
Under heavier load though, things change.

 

[redrabbit]

I could always try the mods as explained on the page I linked, which bypasses the reg. daughterboard...and lowers the voltage.

Since unit B seems to be weaker under load (volts)...if the discrepancy is NOT in the reg. board.......I won't know until I've done a fair amount of work, on both units.

 

[redrabbit]

Finished the mods on the Lambda SCS-200 supplies.
Now they are within 2v of each other, throughout a variety of constant loads.
No need to post-regulate, I believe.
If I want to get picky....two 6 or 10 amp silicon diodes off the rails of the higher voltage unit will bring it down another volt or so.

 

[AndrewT]

most dual polarity circuits do not draw equal current from the supply railsat the same time
Generally they draw current alternately from each polarity.
The return current from the load is usually via the Power Ground trace/wire back to PSU.

You need to take care that you trace out the route the current takes on the whole journey from +ve PSU back to Return of PSU.

 

[macboy]

You absolutely cannot use two positive regulators to regulate a center-tapped dual supply, which is what you have. You can only do that when you have two independant supplies, such as two secondaries on a transformer, two transformers, or two SMPS's.

You need to use one positive and one negative regulator.

 

[redrabbit]

You absolutely cannot use two positive regulators to regulate a center-tapped dual supply, which is what you have. You can only do that when you have two independant supplies, such as two secondaries on a transformer, two transformers, or two SMPS's.

You need to use one positive and one negative regulator.

Is this so ?
Please reply if you agree or disagree with macboy's reply !!!

 

[AndrewT]

You absolutely cannot use two positive regulators to regulate a center-tapped dual supply, which is what you have. You can only do that when you have two independant supplies, such as two secondaries on a transformer, two transformers, or two SMPS's.

You need to use one positive and one negative regulator.

Macboy is correct.
This answer has been given very many times to very many repeated questions about this same topic.

 

[redrabbit]

Macboy is correct.
This answer has been given very many times to very many repeated questions about this same topic.

Where ?
I could not find it....on this forum, or elsewhere. I search this forum for quite some time, and this is the reason I started a "new topic".

You (absolutely) need to use one positive and one negative regulator.

This was my worry, as I stated in my first post.

I'm surprised though, that if the answer was an emphatic "no"...that no one else bothered to say it as such, until now.
Sounds like this is a "will produce smoke" issue.....and being that my question is in the tittle of this thread, I would have thought a direct answer would have come sooner.

I could very well be saying right now "Turns out, it didn't work because I just tried it !!".
I'm glad I didn't do that.
I nearly did.

Don't get me wrong, I do appreciate the help thus far...greatly.
I realize there's no guarantees here, and other people are not responsible for what I do.
Not everyone clicks on, and reads every thread. Knowing that, I specifically worded the title to include "red flag" info, if there was any.

I'm not upset, just disappointed......


The Bystander Effect

Although not an "emergency", David McRaney (author, "You Are Not So Smart") describes this idea rather simply, and I think it applies to asking for help on a forum (especially if there may be be smoke or sparks involved):

The Misconception: When someone is hurt, people rush to their aid。

The Truth: The more people who witness a person in distress, the less likely any one person will help.

If your car was to break down and your cell phone have no service, where do you think you would have a better chance of getting help – a country road or a busy street?

To be sure, more people will see you on a busy street. On the country road, you might have to wait a long time before someone comes by.
So, which one?
Studies show you have a better chance on the country road.
Why?
Have you ever seen someone broken down on the side of the road and thought, “I could help them, but I’m sure someone will be along.”
Yeah, everyone thinks that. No one stops.

This is called the bystander effect.

CREDIT:Fair use, educational
ISBN-13: 9781101545355
Publisher: Penguin Group (USA)
Publication date: 10/27/2011
Sold by: Penguin Group