Just a quick power supply capacitance question

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
While making a preamp I decided I would pretty much rebuild my first chip amp (LM3886) on slightly neater and probably better laid out, though probably still not particularly "good", PCBs.

Currently I have 4400uF per rail, per channel (so 17,600uF in total). It's a dual secondary 500va transformer, with a single 35A metal bridge rectifier on the secondaries, then the filter caps are sort of... in parallel I from there. That is, there's two wires coming off each bridge rectifier "DC" tab, one goes to one set of filter caps, the other wire to the other set of caps. Forgive me for being a bit rubbish at explaining, but I hope you get the idea. Each channel is also regulated separately. One regulator PCB per filter cap PCB.

I'm pretty much just wondering whether it might actually be better to just have one big lot of filter caps for both channels instead?

I've made the filter cap board a bit bigger to accomodate up to 20,000uF per rail this time, but I actually only currently have 2200uF caps. Three of these will fit on the board, per rail, so that would be 6600uF per rail, and 26,400uF in total. I would still like each channel regulated separately, but I think I can still do this "in parallel"?

Or I could keep it separate and have 13,200uF per channel... will it really make much of a difference? If not, I am tempted to just have one big lump of caps for both channels instead, as it should be slightly tidier.
 
Nuuk said:
What is the difference in sound between 1000 uF per rail and what you have now?

I don't have any 1000uF's to try, so I don't know :) Supposedly, because it's regulated, massive capacitance before the regulators doesn't make the highs and things disappear so much? That's what I read anyway... By the chip I only have 100uF per rail.

That said, to test the new amp board I used 3x680uF per rail, with 13v rails. I haven't plugged speakers in to that though. It clips dead easy, and a square wave from my sound card (which isn't the best of square waves to start with admittedly) looks really quite evil! :) I'm hoping the square wave will look more respectable with more capacitance at the PSU though. It should do. I think.

Is there any particular reason why I might want to stay with separate capacitor 'banks' per channel, rather than just one big lot for both channels?
 
Nuuk said:


That's down to whether you want a dual mono supply or not. I prefer dual mono myself. ;)

Ok.

Another question!

I currently have it set up (found the picture I was looking for on your site, better than my description) like http://myweb.tiscali.co.uk/nuukspot/decdun/gc/reg_psu_options.gif the second one on there. Suppose I keep it like this, would I be able to hook up a third regulator to the same set of two bridge rectifiers, but this third regulator regulating to a lower voltage than for the power amps (as it's for a preamp)?

I tried, rather crudely, to simulate this in LTSpice / SWCAD III, but it seems like I can't do this, as the 0v from the bridge appears to "turn in to" the negative regulated rail? I'm confused again :(
 
Nuuk said:
I hope this doesn't sound snobby but in my opinion we are talking hi-fi here. As such, I wouldn't run a pre-amp/filter supply from the same supply that is being modulated by the power amp.

A separate supply isn't that expensive and will pay dividends! ;)


I actually have a smaller transformer, but having it powered off the power amp supply would mean.... less wiring mess and much more space. Looks like I might have to though.... I can't figure how I could possibly get it to run off the same supply.

For that matter, I'm now even confused as to running multiple regulators, regulating to the same voltage, off the same supply. Like in the diagram I linked to, the second configuration, is that correct (not trying to question your knowledge or anything, just confused!)? Paralleled from after the bridge rectifiers, like:

An externally hosted image should be here but it was not working when we last tested it.


Notice the two negative rails are joined? :xeye: This is how I currently have it. Maybe this would explain why one of my regulators boards gets hotter than the other too? Wouldn't it be better to parallel before the bridge rectifiers, so that there's four bridge's instead of two? One set per regulator board? Or does it actually not make any difference? :xeye:


EDIT: sorry, made the screenshot more readable.
 
Oh right. Hm.

Well that's how I currently have it (like the second picture in the diagram from your site), and I think that would be equal what I tried to simulate (see screenshot). It doesn't really seem right, but I dunno. I was hoping someone here would be able to explain it to me.

I can't seem to find much on google regarding two "bipolar" regulated supplies running off of just one dual secondary transformer either!
 
Hi Markie,
If you had asked before you bought this hardware for your first project, I would have advised that you do not include a regulator/s for supplying a poweramp.

For a first, keep it simple.

Unfortunately the transformer without a regulator may supply too high a voltage into your chipamp.

What exactly do you have:-
transformer, chipamp, speaker impedance?
 
About the size thing... i tried the prereg - regulator path to avoid a second transformer, after some headscratching, I remembered how easily a 9V battery drove similar circuits in headphone amps, so it doesn't use too much power... 1 hour later I had a small preamp with a its own PSU made out of a small PCB mount transformer... turned out much smaller than what I tried first to prevent the extra PSU...
 
Nuuk said:
How are you doing the regulated supply then? With only one bridge you would need a positive and negative regulator! :confused:

Maybe you should update that diagram on your site, or add some text under it stating this :eek:

An externally hosted image should be here but it was not working when we last tested it.


Talking about that one. There you have one transformer, one set of bridge rectifiers, and two regulated supplies. This is how I currently have it, and while it does work (apparently, too scared to turn my amp on now)..... it appears even you know that it is 'wrong' [for this particular regulator layout i.e two lm338's)?

AndrewT said:
Hi Markie,
If you had asked before you bought this hardware for your first project, I would have advised that you do not include a regulator/s for supplying a poweramp.

For a first, keep it simple.

Unfortunately the transformer without a regulator may supply too high a voltage into your chipamp.

What exactly do you have:-
transformer, chipamp, speaker impedance?

Well, it's not so much of a problem if I can't run two sets regulators. I'll just have to "downgrade" to a single regulator board for both channels :) This is still a bit of a shame, as I was planning on perhaps putting four channels in there, which would really probably need at least two sets of regulator PCB's eh?

Speaker impedance..... well, all I know is 8 ohm nominal. I opened them up (original Celestion 7's), but the drivers are unmarked so I don't have any graphs etc :( Shame really. Transformer is 25VAC dual secondary 500va. From what I remember this gives me about 37V DC.

I really can't afford to spend yet more money on it (hence just downgrading to a single regulator PCB for both channels). I've spent uh... way more than enough on it already!

Is there absolutely no possible way to have two regulated (with lm338's) bipolar supplies off of a single transformer? Rod Elliot has a schematic for two unregulated bipolar supplies off of a centre tapped transformer, but this isn't quite the same obviously :)


Nordic said:
About the size thing... i tried the prereg - regulator path to avoid a second transformer, after some headscratching, I remembered how easily a 9V battery drove similar circuits in headphone amps, so it doesn't use too much power... 1 hour later I had a small preamp with a its own PSU made out of a small PCB mount transformer... turned out much smaller than what I tried first to prevent the extra PSU...

Fair play :) I'll have to just use the littler transformer I have and somehow make it fit then.


Apologies for the delay. We've just had a power cut... that lasted....about 9 hours!
 
Maybe you should update that diagram on your site, or add some text under it stating this

Talking about that one. There you have one transformer, one set of bridge rectifiers, and two regulated supplies. This is how I currently have it, and while it does work (apparently, too scared to turn my amp on now)..... it appears even you know that it is 'wrong'?

Sorry, I'm getting old (after the hundreds of hours work putting that site together) and all I can see is options with two sets of four rectifier diodes - making two rectifier bridges.

Referring to the page on regulated Gainclones, isn't this diagram clear enough?

An externally hosted image should be here but it was not working when we last tested it.


However, I do apologise for this shocking service! Would you like a refund? :)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.