Critique my UCD400 linear PS

[badman]

Heya:

Any feedback on this schema? Specifically, the use of the resistors. I'm figuring as long as I'm not running high power sinewaves, there should be minimal voltage drop. I'm not going to be running this high power often (if at all) and if I find it to be an issue I can always bypass the R.

If it matters to y'all, diodes are HexFreds and heavily heatsinked, caps are quality cornell dubliers, with poly bypasses, resistors are big wirewound dealies (95W 8 ohm, 2 in parallel). Oh yeah, little 100nF ceramic bypasses on the diodes are in there too.

I'm thinking I'm going to drop the change on a hypex soft-start, unless someone else has an easy inexpensive solution (not series thermistors). Also, I'll fuse the rails.

[luka]

Yes this will work, since light music (even if loud) will not draw much power...
why do you need those resistors in place?

[wintermute]

Hi Badman, You *might* get more ripple reduction with a CRCRC with 2 ohms for each R rather than CRCC with 4 ohms.. I'd have to sim it to see... What sort of current are you looking at? if you are drawing 1amp you are going to get 4V drop....

Also wondering what the 50 ohm resistors in series with the 1uF caps on the output are for??

Tony.

[wintermute]

OK Just did a sim, and as I suspected ripple is quite a bit less with CRCRC instead of bigger R and CRCC

blue is with CRCC 4 ohms, green CRCRC with 2 ohms for each resistor.

Load was around 900mA. (Sorry don't know what your likely actual load is).

Tony.

[luka]

music...

[badman]

Originally Posted by luka

Yes this will work, since light music (even if loud) will not draw much power...
why do you need those resistors in place?

The resistors are there for ripple reduction- no cap bank can prevent all ripple voltage from getting to the load without some resistor/inductor/regulator to help it along.

Originally Posted by wintermute

Hi Badman, You *might* get more ripple reduction with a CRCRC with 2 ohms for each R rather than CRCC with 4 ohms.. I'd have to sim it to see... What sort of current are you looking at? if you are drawing 1amp you are going to get 4V drop....

Also wondering what the 50 ohm resistors in series with the 1uF caps on the output are for??

Tony.

Yes, I certainly would. The only challenge is that this is a high-power amp and thus resistors are big and expensive. I'm planning to use what I have on hand: 8 ohm big resistors (like 9" long) in parallel for the job. Dual 2 ohms... I'd probably run out of resistors and fill up the chassis something fierce (8 resistors per channel, as the supplies will be dual mono).

As far as current, I want it to support up to about 4A, giving me most of the 400W potential into 4 (continuous, naturally short peaks are reliant on the cap bank).

The resistors on the 1uF are just for damping, prevent any resonance in the supply.

[luka]

but then again, why not go with industors? resistors are only good coz they are cheap, not coz they are the best... even better would be regulator

[BZed]

What class amp is this?

If its a class "A" amp then there isn't going to be much load induced voltage change of the supply output. The current will idle at about 2 amps and swing from 0 to 4 amps under dynamic conditions. The amp should be able to live with that.

If its a class "B" or "AB" then this type of supply will have the voltage drop across those resistors causing a lot of change in output voltage under dynamic conditions. Up to 16 volts of load induced voltage change in output voltage for each rail. Not a good idea. You could have stability issues, big time.

I would go for a series regulator. A much better option.

BZ

[luka]

its class D

[badman]

Originally Posted by BZed

If its a class "B" or "AB" then this type of supply will have the voltage drop across those resistors causing a lot of change in output voltage under dynamic conditions. Up to 16 volts of load induced voltage change in output voltage for each rail. Not a good idea. You could have stability issues, big time.

I would go for a series regulator. A much better option.

BZ

It is indeed class D, and thus the supply will have minimal long term drag, with real world music. I'm thinking the time constant for voltage drift would be long enough that instability would not be an issue, but I'm certainly open to options- that's why I'm posting here!

Got a series regulator in mind for +/- 60V DC and 4A? I'm poking around a little bit, but to be honest my regulator experience is exclusively in 3 leg LM317 etc., simple cheap little low-power circuits.