Just a quick power supply capacitance question

[AndrewT]

Hi Markie,
re post 57:
the return line of the top reg and the feed line of the second reg get commoned as you show.
This common line becomes your 0v line.

The return line of the second reg is at a voltage below the feed line of the same reg.

If the feed line is @ 0v then it must follow that the return line is @ some negative voltage with respect to the feed line (0v).

It seems that

When I simulate it, the bottom one doesn't work correctly as the 0v line 'turns in to' the regulated -ve voltage?

is reporting the correct voltage.

However, I return to my earlier advice.
Keep your first one simple.
Build a temporary monoblock, get it working and understand the input connections, the output connections, the power grounding, the signal grounding and the safety earth. Particularly, how these interact.

The step up to a stereo pair and still getting a quiet output when zero signal is fed in is a big jump.

When you have finally got a stereo pair operating quietly then go for the system you want to run with all the extra bits and bobs. But going for this first and trying to debug it will prove extremely difficult if you don't understand what is happening.

[markiemrboo]

Originally posted by AndrewT
Hi Markie,

Hi Andrew,

re post 57:
the return line of the top reg and the feed line of the second reg get commoned as you show.
This common line becomes your 0v line.

The return line of the second reg is at a voltage below the feed line of the same reg.

If the feed line is @ 0v then it must follow that the return line is @ some negative voltage with respect to the feed line (0v).

It seems that is reporting the correct voltage.

I think I might just about understand that. Sort of creating a virtual ground in a way, right?

The simulation sort of works. I get output voltages from both supplies, and they're what I expect them to be. However... trying to change the output voltage of either of the two 'negative' outputs.... doesn't work quite work. If I adjust one negative side, it does nothing until I also adjust the other negative side. Given that they're 'linked' as they are, I guess this is expected.

I think it was 't.' that said that it was OK in "the real world", but I haven't tried adjusting either of the negative rails on the actual real circuit to see if they will adjust independantly, so that I can have different negative rails on each supply (say, the first supply being -24v and the second being -26v).

The positive sides are OK in the simulation, and I can adjust these to be different on each supply. I can see why these are still OK though.

I had a feeling this might be why one of my regulated supplies gets much hotter than the other, which is also why I came here to ask about it, but everyone got lost in misunderstanding, likely due to my lack of proper terminology or something

However, I return to my earlier advice.
Keep your first one simple.
Build a temporary monoblock, get it working and understand the input connections, the output connections, the power grounding, the signal grounding and the safety earth. Particularly, how these interact.

The step up to a stereo pair and still getting a quiet output when zero signal is fed in is a big jump.

When you have finally got a stereo pair operating quietly then go for the system you want to run with all the extra bits and bobs. But going for this first and trying to debug it will prove extremely difficult if you don't understand what is happening.

Well, I already have a pair of mono blocks working fine, quiet with no signal / cable unplugged and all that. The only problem I seem to have is mobile phone interference. I'm hoping this might be something to do with the fact I am taking the speaker return and input ground to the PCB's, rather than to the starground point.

I've pretty much redone it all on new boards. I might attempt replacing it all tomorrow, but this time taking the speaker returns and input grounds back to star ground instead, as you suggested to me so long ago

Anyhow, thanks.

[Nordic]

Reasonably sure I saw a datasheet schematic where they used 1 pot to adjust multiple regulators....

yep bottom of pg18 LM117.pdf

[CarlosT]

How about just using fixed resistors? The trimpots are convenient but are they necessary?

[markiemrboo]

I fear we may be getting in to my poor explanations / misunderstanding territory again

I think I may have properly figured out what I wanted to know from Andrew's last post though. With any luck it will work as expected and I will have learnt some more. Probably wise to try with a couple of 9v batteries or something first though I reckon

[markiemrboo]

I am back.



I have now built and wired together the regulators, like this:



It is no different in the real world than it is in the simulator That is, the -ve rails are all 'linked' and you can't change them independant of each other, e.g the preamp supply actually has +15v on the positive side, as expected, but -28v on the negative side, from the power amp regulators.



Go ahead and call me stupid etc etc etc, I am honestly half expecting it



Looks like I will have to go to one regulated supply for both channels, and try and fit a separate preamp transformer in there somehow.
I don't feel comfortable running one supply per channel on the same transformer now. I have convinced myself this is also why one supply got much hotter than the other.


Unless someone can explain to me what I have done wrong, as 't.' reckoned it should work "in the real world"?

[AndrewT]

Hi,
I don't understand why you get +15-28Vdc after your regs.
It almost seems as though the negative side is not regulating.
That would hint at a mis-wire somewhere.

BUT,
I would seriously ask you to consider a separate rectifier for each regulator, yes, share the secondary windings across all the rectifiers.
For the sake of cost, just use standard bridge rectifiers, (4 diodes with metal cooling jacket for about $1.50).

[markiemrboo]

Originally posted by AndrewT
Hi,
I don't understand why you get +15-28Vdc after your regs.
It almost seems as though the negative side is not regulating.
That would hint at a mis-wire somewhere.

Sadly no mis-wiring. I quadruple checked. It worked exactly as the simulation was showed after all

BUT,
I would seriously ask you to consider a separate rectifier for each regulator, yes, share the secondary windings across all the rectifiers.
For the sake of cost, just use standard bridge rectifiers, (4 diodes with metal cooling jacket for about $1.50).

I was thinking of trying this. I've simulated it though, and it didn't work right either, though it may have "not worked" in a different manner. I can't remember... already

[t.]

Originally posted by markiemrboo


Sadly no mis-wiring. I quadruple checked. It worked exactly as the simulation was showed after all



I was thinking of trying this. I've simulated it though, and it didn't work right either, though it may have "not worked" in a different manner. I can't remember... already

Lots of people have built this circuit me included and can assure it works! I've built both Pedja's LM338 reg and his discrete regulator and both work spot on.

Maybe posting a picture of the actual circuit you built will give people a better idea where you are going wrong

[markiemrboo]

Originally posted by t.


Lots of people have built this circuit me included and can assure it works! I've built both Pedja's LM338 reg and his discrete regulator and both work spot on.

Maybe posting a picture of the actual circuit you built will give people a better idea where you are going wrong

Do note that *individually* they work fine, I have never questioned this. It is only if I try and run them all off of a single transformer "in parallel", as in my drawing, they don't.

Have you specifically done this and it has worked, or are you speaking only of a single supply?