diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Analogue Source (http://www.diyaudio.com/forums/analogue-source/)
-   -   Extremely noisy DC (Phono Stage PSU) (http://www.diyaudio.com/forums/analogue-source/218927-extremely-noisy-dc-phono-stage-psu.html)

lordvader 2nd September 2012 09:56 AM

Extremely noisy DC (Phono Stage PSU)
 
1 Attachment(s)
Hey all.

I'm currently trouble shooting noise problems on my tube phono stage (each channel contains two triode strapped C3gs, loaded by a CCS). The power supply is a schottky diode rectified with an RCRCRC filter (R = 270ohm, c = 50uF, except the last one, which is 100uF).
The power supply is in a seperate enclosure, and via the umbillical wire, I'm carrying B+, 0V, earth, and two 6.3VAC heater supplies (so 4 wires there). Each wire is tightly twisted with its opposing pole, and then covered in a shield (connected to earth).
Anyway, when I connected the DC to a scope (the probe was on B+, the reference was the 0V line), the output looked like the attached file (this was taken from the phono chassis, not the PSU chassis).

Any clues ? If I zoom in, there is ripple, but it looks modulated on some really low frequency noise.

grommeteer 3rd September 2012 10:51 AM

Hello,
is that 2s/DIV on your scope? That would mean you have a supply that varies rather than a noisy one.
Is the varaition larger close to the rectifier or close to the amplifier? Meaning: do we deal with a variation in mains or in load?
Your load should be a very constant one as it uses a CCS.
Have you compared the time constant of your filter chain to the "slew rate" of the voltage changes?

lordvader 3rd September 2012 11:14 PM

2 Attachment(s)
Well I took a trace of the PSU unloaded (close to the rectifier), and it looks the same (I also added a zoomed in trace, showing the actual ripple)
And yes, that's a 2s div.

My next question was going to be is if this is what motorboating looks like ?

And what do you mean by time constant of the chain, to slew rate of change ?

danielwritesbac 3rd September 2012 11:29 PM

Quote:

Originally Posted by lordvader (Post 3148994)
The power supply is in a seperate enclosure, and via the umbillical wire, I'm carrying B+, 0V, earth, and two 6.3VAC heater supplies (so 4 wires there). Each wire is tightly twisted with its opposing pole, and then covered in a shield (connected to earth).

I notice two things.
Umbilical = big inductive resistor, so do you have at least 220u right next to the tubes (aka got caps in the amp enclosure)? I ask because, with inductance disabling the caps at HF, your amp has no defense against the unfiltered AC that is wrapped up with your DC power cable. Same flaw exists in most USB sound modules--dirty power in the same cable as signal line.

Perhaps, reg the heater supplies (clean DC heater?) within the power box and put some power caps in the amp box. Maybe that will get rid of the HF blast. I do not know.

lordvader 3rd September 2012 11:59 PM

I don't have any filter caps in the phono chassis, so that's something I'm planning on trying.
I'm also thinking about a filter cap next to each CCS (as there is a bit of distance between the B+ lug, and the CCSs).

The above images though (yellow traces) are of the PSU completely unloaded (well, only loaded by a 100K bleeder, and a 100K|33K voltage divider to raise the heaters).

kevinkr 4th September 2012 12:52 AM

:cop: Since this is trouble shooting a noisy phono stage it seems to me that this might better go in the analog source forum.. I'll retitle to reflect that it is a phono stage PSU..

I have a number of thoughts which I will post separately.

kevinkr 4th September 2012 01:02 AM

I suspect the low frequency wander is probably slow cyclical variations in your mains voltage.. I see the same thing here...

You definitely need local decoupling at the CCS, perhaps even a few uF would be enough to assure that they are not misbehaving..

I use DC to heat the filaments in phono stages as this is just one less source of noise to couple into the audio circuitry, and more to the point variations in heater voltage will cause LF wander in the output of your phono stage.

Please post the design of both PSU and phono stage and we'll help you get it sorted. (See my Muscovite thread for the extremes I went to with the power supply - of course I used a cascode front end with zip for PSRR.)

lordvader 4th September 2012 01:17 AM

2 Attachment(s)
I'm attaching the PSU schematic, and phono schematic (note, that rather than 210 for B+, B+ it about 250V. The CCS drops the voltage to about 180V). The grid stoppers are also 1K rather than 120.
Also, the diodes are schottky diodes, so there are no snubbers on them.

The single B+ powers both channels, with each stage having its own star ground (ie, the first tubes share a star), and the stars then joining up to a single 0V point.

To decouple the CCS, would 10uF suffice ?

I'm thinking about DC for the filaments, but as they're already 6.3VAC, I don't think I can get 6.3VDC from them (well, get a nice 6.3VDC anyway).

Also, thanks for everyones input so far ! I'm been DIY'ing for a while, but this is my first "sensitive" project :)

danielwritesbac 4th September 2012 01:54 AM

Quote:

Originally Posted by lordvader (Post 3151146)
To decouple the CCS, would 10uF suffice?

The inductance of the cable and the tracks does what any inductor will do and that's disable the remote caps from being useful at higher frequencies. The amount of patchwork needed relates to this, but I don't know the exact amount, since it is layout specific. However, the device is highly sensitive, so the most effective approach could be tuning it by ear (it is, after all, a device made to please an ear). That works as long as you pick the right spot to apply the fix. Perhaps someone smarter than me could name the correct value or plausible range of values for it?
Quote:

Originally Posted by lordvader (Post 3151146)
I'm thinking about DC for the filaments, but as they're already 6.3VAC, I don't think I can get 6.3VDC from them (well, get a nice 6.3VDC anyway).

Ordinary prefab bridge rectifier and caps will up the voltage to about 9.2vdc and then a regulator can work. You might need a schottky bridge rectifier (high amperage low voltage schottky specs are lowest loss) and a low dropout regulator though since I don't know how much rail droop is present.

lordvader 4th September 2012 02:06 AM

I am planning on adding a 100 or 220 uF capacitor in the phono chassis (there are currently 0.1 and 0.01 uF bypass caps there).

I'll have a go at DC rectifing when I get a chance, but there are about a million options out there when it comes to regulation !


All times are GMT. The time now is 04:54 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2