Gainclone Power Supply Design *cont* (sorta)

[officeboy]

Well continuing and hijacked I guess

I have just acquired a 750Va Plitron 31-0-31 transformer, that will power basically a BPA200 amp. ($29 on ebay )

Of course there is the obvious trouble of supply voltage being too high. I could always and had planed on just unwinding a bit to drop a few volts.
But with all the talk of regulated supplies in the thread whose name I’ve stolen, would that be maybe a better way for me to go? Problem is can I parallel up regulators for more current, or use one per channel? And is the almost 44V DC to high for the regulator? I find info on 35V being the max, and then it says; “Since the regulator is “floating” and sees only the input-to-output differential voltage, supplies of several hundred volts can be regulated as long as the maximum input to output differential is not exceeded"
So any suggestions or opinions?

[azira]

Look at Rod Elliots page (sound.westhost.com) for his Capacitor Multiplier circuit. It's just as complicated as a linear regulator circuit but won't have current limitations like a standard linear reg. I have a transformer that puts out 66VCT also that I plan on doing this with. You'll have nice smooth rails coming your way... not that a GC cares that much.
--
Danny

[carlosfm]

I have 47v before the LM338s.
It's on the limit, but no problem.
The amp has been playin' for 3 days and it didn't the house.
I'm regulating for 31v.

The LM338 has lots of current, check the datasheet.
But for a BPA-200 I'd say at least 3 in parallel per rail.

[Cro maniac]

Now that You guys mention BPA200, I have few questions about it.
I'll be making one BPA200 soon for powering my active subwoofer.
Can somebody recommend me how high supply voltage must I use to get max power available from each chip. I have really massive heatsink with vent mounted on it so cooling is no problem to me. The speaker is 4ohms.

Also, i don't want to use servo circuit. Is matching resistors to <0,1% enough to keep the offset voltage low ? Any experiences?

And one more thing. Can I use feedback and input resistors about 6 times greater than recommended in the BPA200 application note ?
Rf=120kohms
Ri=5,6kohms
That way I could use much smaller (easy to get) Ci and still have very low low frequency cutoff which is important when powering subwoofer.
If I change this values I don't know what should be the value of Rb. Can I use the same value as Ri (5,6k) ?

I know this thematic has been discussed many times, but still hope somebody will help me.

[officeboy]

Well 30Vac 42Vdc is what the BPA200 plan calls for. And it's also the max for the lm3886.

Bridged and paralleled into 4ohms, seems like it could give right around 500W, but thermally I don't think the chips could take that. The Design guide spreadsheet seems to think that would be .40 c/w with the T package.

If you know you will have a 4ohm load, you could add a second set of parallel chips (I may do this too) to ease up the load on each chip. You may be able to get up to 700W from that.

As for the other question, I don’t know, but I don’t plan on having any caps other then a input DC filter in my signal path. And I don’t plan on using servo’s either.

[Cro maniac]

officeboy, thank you for replying !

If you know you will have a 4ohm load, you could add a second set of parallel chips (I may do this too) to ease up the load on each chip. You may be able to get up to 700W from that.

Yes, I already have two 8ohm 12" speakers which will be wired in parallel. But I don't really need 500W power so 4 chips would do just fine for this project.
I think I'll use 2x29V transformer to be on the safe side .

As for the other question, I don’t know, but I don’t plan on having any caps other then a input DC filter in my signal path. And I don’t plan on using servo’s either.

Which schematic are you using ?

[officeboy]

Something like a minimal NIGC, and I'll have an inverted and non-inverted signal fed to the banks of chips to "bridge" them.

I really haven’t though about it till now, and since it's Friday and I don't wanta work I tossed this together. It probably won't work as I’ve shown it, but it's probably not far off from working either.