Power supply filtering

[Craig405]

Hi

I have measured the AC component on my power amplifier PSU (which is not switch mode) rails and it is massive, a lot of jagged looking high frequency noise at 5Mhz.
What is causing this?
I have 20000uf per rail plus 100nf caps on each rail and another 400uf of caps per rail on my circuit board.

????????

Cheers
Craig

 

[still4given]

How old are your caps? Perhaps you would benefit from reforming them.

Good article on reforming caps

Blessings, Terry

 

[Craig405]

The caps should be ok, the large ones were reformed for a few hours on a bench power supply before use and all current readings were very close to each other once stabilised.

Thanks for the suggestion.

 

[still4given]

Maybe the bridge is failing. They are cheap enough, try replacing it.

Blessings, Terry

 

[Cortez]

RLC-resonance from parasitic Ls and Cs.
Search for the expression "snubber" at diyaudio!
Wait for your results after the mods !

 

[Leolabs]

U may try choke filtering too🙂

 

[Craig405]

Cortez, thanks for that

Ok i am convinced it is parasitic inductance in my wires as some are 1 foot long!
Oh well they are very thick wires (rated for 30amp) i thought they would be ok.

I have looked into snubbers on the search and found out what they are. I think i should add 2 snubbers per rail (+/-), one at the heatsink for my power transistors and the other on my circuit board power rail tracks.
I only need to calculate my R's and C's now.

Can anyone send me this document ?http://www.hagtech.com/pdf/snubber.pdf.
Its a dead link now but people used it on the threads.

Cheers everyone
Craig

 

[Geoff]

Can anyone send me this document ?http://www.hagtech.com/pdf/snubber.pdf.
Its a dead link now but people used it on the threads.

Email me. I have a copy I can send you.

 

[Danko]

Can I get one of that file too? 🙂

I can upload it to my server, and link the file to here.


Thanks in advance, Danko

 

[jan.didden]

Hi
I have measured the AC component on my power amplifier PSU (which is not switch mode) rails and it is massive, a lot of jagged looking high frequency noise at 5Mhz.
[snip]

Craig, what level?

Jan Didden

 

[Craig405]

Jan
it is not as big as i initially thought, 150mV pk pk per rail and looks like an AM signal sometimes.

Although it is small it still gets through to my amps output and messes up my signals.

Putting caps accross the rails attenuates it at bit but it is still about 100mV.

I think i will have to make snubbers and hope putting it in a case will clear it up some more 🙁

Cheers
Craig

 

[jan.didden]

Jan
it is not as big as i initially thought, 150mV pk pk per rail and looks like an AM signal sometimes.

Although it is small it still gets through to my amps output and messes up my signals.

Putting caps accross the rails attenuates it at bit but it is still about 100mV.

I think i will have to make snubbers and hope putting it in a case will clear it up some more 🙁

Cheers
Craig

Right. You may want to see if it is stronger on the supply side or on the amp side. If it is stronger on the supply side, you may try an inductor in the outgoing supply lead, like 20 windings, 1mm or more mag wire, followed by the snubber.

Jan Didden

 

[Craig405]

The HF noise is pretty equal from the rectifier output all the way to the furthest piece of wire.

I actually have some sort of snubber caps on my rectifier bridging each connection with 22nf, so the rectifier shouldn't contribute much to this noise.

Thanks for the PDF Geoff.


Craig

 

[Eva]

Are you sure that your amplifiers are not oscillating?

Also, it may be common mode noise from mains line.

 

[bmcevers]

PS Layout

Try placing .1uF caps (of appropriate Voltage rating) across each diode in your bridge, this should cure almost all of this "RF hash" (big bridge diodes are notorious for being noisy) you might also try a .1 to .47uf across the AC mains, some call it an "xcap". If the amplifier is in the standard range of gain (20-23dB) and metal chassis with shielded input cables, I would find it hard to believe that you are picking up stray RF.

Good Luck.

 

[multi-volti]

while on the topic of noise

If the offending signal was accidentally discovered (not a symptom you sniffed out as an audible problem), consider whether there is anything in your test setup that is introducing the noise...long scope ground, floating scope ground, etc.

I am not surprised by Zener noise now, but was caught off guard by it once in a friend's integrated amp. I slept on the couch one night, turned on the stereo out of boredom the next morning & kept the volume low. I kept noticing a really annoying hiss. I found 1 uF electrolytics across Zeners in the power supply and changed those to non-descript film caps. Noise gone...

I then took a look at a 'cheater' power supply I made to run a pilfered preamp board (I got the circuit board but had no chassis); it was a 5 V 3-terminal regulator with 27 V Zener in the ground lead to give me 32 V (and a corresponding circuit for -32 V). Now I was embarrassed that I had been listening to this & not notice the noise!

 

[Craig405]

Thanks for all your ideas guys, the gain is about 13 (22dB) so this isnt too large, I am going to put it in a grounded steel case and see what happens. It is all open air at the moment.

My test leads are all ok apart from my signal input, which is unshielded for about 15cm but i get similar noise on my output with the inputs shorted to ground.

I guess im just going to have to play around with casing, capacitors on the diode bridge and snubbers until it goes away.

Thanks again
Craig

 

[jan.didden]

Thanks for all your ideas guys, the gain is about 13 (22dB) so this isnt too large, I am going to put it in a grounded steel case and see what happens. It is all open air at the moment.

My test leads are all ok apart from my signal input, which is unshielded for about 15cm but i get similar noise on my output with the inputs shorted to ground.

I guess im just going to have to play around with casing, capacitors on the diode bridge and snubbers until it goes away.

Thanks again
Craig

A quick one that has saved me many headaches: Just connect both the probe leads (both hot tip and gnd lead) to the gnd wire where you are measuring. Don't be too surprised if you STILL measure the hf junk...

Jan Didden

 

[destroyer X]

The signal level seems too much big to be RF pickup.

Eva, told a more strong possibility.

regards,

Carlos

 

[Craig405]

My amplifier was oscillating, Eva thankyou for suggesting this!!

This was solved by substituting a different (slower) transistor in my power stage drivers.
Now oscillation is almost all gone, less than 5mV on one rail under 1mV on other.
Now the oscillation at my output corresponds roughly to that on my supply rails and only appears on positive peaks above 300mV and is invisible once the wave gets larger.


Anyway, its pretty much solved! thanks everyone