High Power Switching Power Supply

Thread Starter

Ilmarj

Joined Jan 6, 2012
1
Hi.

I have a need to construct a high powered switching supply for my 500W solid state linear amplifier. The supply needs to have a 120VAC input and a pretty clean 40VDC 20A output. Has anyone dabbled with such an animal or can lead me to a circuit I can modfify for my needs? Your help is greatly appreciated!
 

PaulEE

Joined Dec 23, 2011
474
re-edit: At first, I suggested hooking a few ATX computer power supplies together in series to get your output needs, and then I flip-flopped, deleted that comment that used to be here, and suggested a linear supply below. Upon receiving the next comment by bountyhunter, I flip-flopped again after thinking the ATX supplies through in a separate response.

------------------------------------------------

Big transformer, big rectifiers, big caps. Designing a switch-mode power supply that can safely and reliably deliver 800 watts would be quite an undertaking. Any stereo receiver rated for wattages that high usually has a FUSED unregulated DC supply in it as well...at least, they used to.

Alternatively, you could go this route:
http://www.antekinc.com/details.php?p=270

or something similar.

Good luck!
------------------------------------------------
 
Last edited:

bountyhunter

Joined Sep 7, 2009
2,512
Hi.

I have a need to construct a high powered switching supply for my 500W solid state linear amplifier. The supply needs to have a 120VAC input and a pretty clean 40VDC 20A output. Has anyone dabbled with such an animal or can lead me to a circuit I can modfify for my needs? Your help is greatly appreciated!
That's an 800W power supply. We built them and I can tell you that you can buy them way cheaper than build them. You could use a half bridge but peak switch currents would be very high. I would use a four FET full bridge offline design if I had to build one. The idea of kluging one together from PC supplies might work (?)
 

PaulEE

Joined Dec 23, 2011
474
That's an 800W power supply. We built them and I can tell you that you can buy them way cheaper than build them. You could use a half bridge but peak switch currents would be very high. I would use a four FET full bridge offline design if I had to build one. The idea of kluging one together from PC supplies might work (?)
I liked the ATX supply idea because:
1) They're completely idiot proof (which is good for me; I always inevitably short the outputs accidentally)
2) They're mass-produced and are quite affordable
3) They're already designed and come in a neat little box
4) They've proven themselves to be reliable in computers for years
5) It's an excuse to order from Newegg or TigerDirect (har har har)
6) They're pretty darn quiet (noise on output power), considering they power computer components
7) They're easily mountable with pre-drilled standard holes
8) They're not a heavy, bulky, expensive linear supply

I would still hook one or several giant capacitors on the output; I don't know how a computer power supply would act under an abruptly high load...similar to abrupt bass in audio...computers tend to change their load requirements slower...or, at the very least, they're more constant with time than an audio amplifier supply load would be.

Enjoy
 

kubeek

Joined Sep 20, 2005
5,794
PaulEE: there are also some cons of ATX´s:
You need to modify them to make the outputs floating, so that you can safely connect them in series. Not a hard task, but still..

As for the large capacitors on output, how do you think the PSU will react on startup? Most likely the overcurrent protection will kick in before the caps even start to charge, so you would need some soft-start to let them charge before you turn the amp on.
 

PaulEE

Joined Dec 23, 2011
474
PaulEE: there are also some cons of ATX´s:
You need to modify them to make the outputs floating, so that you can safely connect them in series. Not a hard task, but still..

As for the large capacitors on output, how do you think the PSU will react on startup? Most likely the overcurrent protection will kick in before the caps even start to charge, so you would need some soft-start to let them charge before you turn the amp on.
All good points. I'm angry with myself for not thinking through the capacitor bank downfall. Soft-start would solve that, though.

In my head, I was modeling the supplies as batteries, one after the other. Can you describe how you'd "make the outputs floating"? As far as I know from taking them apart, they're just buck converters, which, in the simplest case, has the switching circuitry, inductor, and capacitor. The cap is what has the output voltage impressed upon it...not sure what the issue is with simply hooking them up in series.

Anxious to hear your thoughts.
 
Last edited:

kubeek

Joined Sep 20, 2005
5,794
In my head, I was modeling the supplies as batteries, one after the other. Can you describe how you'd "make the outputs floating"? As far as I know from taking them apart, they're just buck converters, which, in the simplest case, has the switching circuitry, inductor, and capacitor. The cap is what has the output voltage being onto it...not sure what the issue is with simply hooking them up in series.
The issue is that the ground on the secondary is typically connected to the metal case and that is connected to mains protection earth. You try to connect two in series, you short the lower one through the earth loop.
To make it work and still be safe you need to disconnect the ground-protection earth connections on all but one supplies, so that the total series combination is still safely grounded at one point.
 

PaulEE

Joined Dec 23, 2011
474
Excellent. Thank you for that.

And,

"being onto it" was supposed to be "impressed upon it" in the post before last.
 

thatoneguy

Joined Feb 19, 2009
6,359
First, get a 120 to 40V step down toroidal transformer with center tap, or It needs to be rated at 500VA.

Then rectifier, center tap is ground, add several large capacitors to +20V and -20V rails.

I found ±25V from 120V, 500VA That would filter to a bit over ±30V, but give you more dynamic headroom if the amp is built properly.

Reason for toroidal is to keep EMI/60hz "hum" from crossing over to audio section.
 

bountyhunter

Joined Sep 7, 2009
2,512
I'll speak the heresy: IMHO, the power rails used in an audio amp don't need to be regulated, so they can be generated the old fashioned way: a big honking transformer followed by rectifiers and giant capacitors.

I also don't see why any power supply running an audio amp would need to be able to supply rated power CONTINUOUSLY since average power of music is typically about 1/20 or less of peak power.

just my opinion
 

kubeek

Joined Sep 20, 2005
5,794
I'll speak the heresy: IMHO, the power rails used in an audio amp don't need to be regulated, so they can be generated the old fashioned way: a big honking transformer followed by rectifiers and giant capacitors.
true for typical class AB setups and similar. D-class is more sensitive and some have poor PSRR.

I also don't see why any power supply running an audio amp would need to be able to supply rated power CONTINUOUSLY since average power of music is typically about 1/20 or less of peak power.just my opinion
Depends on what the amp is used for, for example my bass guitar amp on a gig could be under full load 50-80% of the time I am playing. On the other hand a home stereo surely doesn´t need fully rated supply at such powers.
 
Top