How to make a dead quiet PS for tube ?!?

[12afael]

I`m finishing a tube pre. I have some troubles but I think is not problem of ground loop. I think the problem is on the HV filter. I use a 220volts transformer rectified with4 1N4007 in bridge ,100uF for first filter , then a 4,7kOhm and 50uF  I think is sufficient filter.
I have seen some schematics(fender) that put a small capacitor in paralell with the diode (.47nF or less) could that reduce the hum in some way?

If I have a transformer without screen caps can y put a copper layer or maybe aluminium paper?
I have an aluminium chasis I always fear that not conduct ground well, so I put a wire between star grounds...  :oops:


is posible to build a field "receptor" to locate the electromagnetic field that produce the problem? maybe a coil and a diferential amp? it could be usefull for find oscilations too.

[Doug_H]

I have seen some schematics(fender) that put a small capacitor in paralell with the diode (.47nF or less) could that reduce the hum in some way?

I

That's a snubber cap. See one of my posts above.

Doug

[fikri]

I have seen some schematics(fender) that put a small capacitor in paralell with the diode (.47nF or less) could that reduce the hum in some way?

I

That's a snubber cap. See one of my posts above.

Doug

Where  :?:

[Doug_H]

[\

That's a snubber cap. See one of my posts above.

Doug

Where  :?:

2nd page, 2nd post from the bottom. (Sigh...) Unfortunately I think it was crushed between R.G. and J.C. when they were whipping it out and grabbing their measuring tapes... :lol:

Doug

[fikri]

Aaah... That one !  

You know what, i almost felt guilty about writing this thread !  :lol:

[RedHouse]

I don't think that PSUD is much help at all, two years ago I looked into it, joined the forum and tried to rally up everyone into posting some baseline *.psu files. Eventually I got frustrated over a year of trying and left the forum.

What good is making up a design if the software can't even emulate a known power supply? I was hoping to use the program to arrive at which Hammond transformer I neede to do a design.

A quick look over there just now it seems over the last 2 years they STILL haven't got anywhere with it the standard stuff I proposed, the usual suspects:

Fender Twin Tweed
Fender Twin Reverb
Fender 5F6 Bassman
AC30
Hiwatt
Marshall Plexi
Mashall JMP 2203/4
Marshall JCM800 2203/4


http://www.duncanamps.com/psud2/psud_list.php

Look at the link and one can see they've still never got going on the real-world stuff, but there is a Fender Champ 5F1 (woo hoo) and an AX84 *.PSU file ...but none of the stuff most folks want to work out or clone.

Another problem with that Duncam PSUD was one needs parameters that are hard to come by like the impedances of the transformers for instance and the impedance of a capacitor's node to get the correct circuit emulation and that (to me) kind of defeated the whole purpose of the program, one is supposed to be able to "arrive at" the parts needed.

If one deasn't have the transformer and built circuit to measure these parameters then the emulation isn't realistic. If one has that, one doesn't really need the Duncan PSUD program as much as ...say... a good DMM to measure things.

Just my 2¢ worth.

[Doug_H]

I have found psud real useful for helping me design amp power supplies. But I usually determine the output stg operating points using  a load line and have a good idea what the current draw will be at the voltage I want to operate at, and etc anyway. The obstacle I see to cloning vintage amps is that the fender, marshall, etc schems that are available usually don't include voltage measurements. Couple that with the fact that the only description of an xformer is usually just a part no. (instead of something useful like "275-0-275") and it's anyones guess what they are running at. That kind of info seems to circulate among amp techs who have experience opening them up and working on them. Anyway, you can't design a power supply unless you at least know your target voltage first.

Lack of data is always a problem with any simulation. But I find the default  psud values for coil resistance, ESR and etc to be accurate enough. I'm actually kind of surprised how accurate it is. 10 or 15, sometimes 20 volts one way or the other won't break the bank for me. Line fluctuations contribute to that anyway, depending on if the fridge or A/C is running or someone down the street is running a hair dryer. I'm always going to tweak it to exactly what I want anyway, but psud gets it close and starts it off sanely.

I don't agree that most people just want to clone stuff. It's a lot more fun to design something from scratch!

Doug

[12afael]

Doug I have a 100hz oscilation(50hz line) I put a spectral analizer . can you recomend me a  value for that snub cap?and type? I suppose that the voltage should be high as the supply or not?

a friend tell me that some kind of diode (switching) don`t need that cap .

[Doug_H]

I haven't used snubbers before. I used FREDs one time when I was having problems with hashing noise.

But here's an example of snubbers:

http://195.178.239.50/ax84/media/ax84_m261.pdf

You can get more help with them at the ax84 forum if you have questions.

Doug

[brett]

Hi.  Great thread.  Good reminder about the snubbers.

I'm a little ashamed, but here's what I did that made for lots of hummmm in my 18 watter. :oops:

Instead of buying that $180 Hammond power tranny to get 240V out from a 240V line in, I put two cheap transformers back-to-back.  It worked for me previously in a low-power amp (a firefly).  Buth there's no centre tap, so I made a "synthetic" one by connecting diodes from the power star earth each AC line.  To also made a synthetic centre tap for the flment transformer by using the 2 x 100ohm resistors to ground method.

Reading RG's stuff above, I *think* I've fouled up by connecting the "half bridge" to the star ground instead of the -ve side of the 1st filter cap.  Maybe??  I'll also put snubbers across those diodes.

Or is there's something obviously wrong with what I've done?

PS All this will be hypothetical in a few days when my GDS/Heyboer transformer arrives, centre taps and all.

[R.G.]

Lots more on diode snubbers for power supplies at http://www.hagtech.com/pdf/snubber.pdf.

[R.G.]

Oh, yeah, almost forgot.

I was prepared to go full bore with snubbers and/or FREDS on the Workhorse line. I don't know whether to claim luck or skill, but the snubbers and/or FREDs made no difference whatsoever, so I left them out.

[Doug_H]

Oh, yeah, almost forgot.

I was prepared to go full bore with snubbers and/or FREDS on the Workhorse line. I don't know whether to claim luck or skill, but the snubbers and/or FREDs made no difference whatsoever, so I left them out.

I have found that most of the time I don't need them. The one time I used FREDs was in a real cramped space modding an existing amp. With proper planning, layout, experience, etc I don't think they are really needed.

Doug

[fikri]

Yeah, all my projects havent dealing with snubbers till now.

Oh, and for brett, you cant use back to back transformer for power tubes ! they need a LOT more current to work !

[brett]

WRT back-toback not working with a largish load, I'm using 2 x 72VA for an 18W amp.  The trannies which give me 235V unloaded (my supply is about 250V) and 220V with a 20W resistive load.  I thought this might work, but I get hum and some odd distortion when turned up (maybe the pair of transformers are slower to respond to changing load??  I suppose the load on the first transformer is very complex and inductive compared to a normal single transformer).

[12afael]

I understand that snubber are for hi frecuence noise. I have mainly 100hz so. I will try to exorsice this beast on other way.

I did not know freds. I am going them to use from now.

thanxs

[R.G.]

Snubbers *are* for high frequencies in almost all cases. The problem they solve in rectifiers is that the turn-off transient in a standard diode is fairly violent, and it makes the inductance/capacitance of the diode, transformer, and leads ring. You get one of these ringy-dingies every AC half cycle, so you get ringing hash chirps at 120Hz or 100Hz depending on your power suppply frequency.

The chirps can ride through circuitry wiring just fine and appear on the output as a buzz that appears to be at twice the power line frequency. The human ear interprets this as a hum/buzz. This is the same reason that light dimmers and fluorescent fixtures can cause a buzz   - they make ringing hash at power line frequency.

Kill the ringing hash and the buzz goes away. If it was ever there.

[Finom1]

hi, here is an article I translated from a Japanese site which might be helpful:

http://www.maximoaudio.com/forum/viewtopic.php?t=220

Does anyone have a working link for this information?
Thank you!

[tubegeek]

maybe try the Internet Archive "Wayback Machine" - great tool for dead links.

EDIT: ....nope....

[amptramp]

Nice necropost!  I never saw this thread before.  I am also on some sites that hotrod FM tuners and they use some of these techniques plus others have not been mentioned.  FRED's and Schottky diodes are popular because the reverse recovery pulse on a typical rectifier diode is a pulse with a fundamental of 25 to 35 MHz and the third harmonic often sits somewhere in the 88 - 108 MHz FM band.  They also like paralleling electrolytic capacitors with film caps.  One thing they often do that I have not seen mentioned here is drilling out the transformer mounting holes and adding rubber grommets to reduce magnetic coupling into a chassis.  Some ground the core directly with a wire from the mount to the core and others leave this ungrounded.  Grounding the core through a resistor or parallel RC circuit may make more sense.

For tube amps, one thing that is often seen is the use of voltage regulator tubes which are made in 75, 90, 105 and 150 volt varieties.  You have to be careful about parallel capacitance because they can oscillate if capacitance is connected directly across them.  They start at a higher voltage than their operating voltage and the octal ones regulate from 5 - 40 mA and the 7-pin miniature ones regulate from 5 - 30 mA.  These are shunt regulators, so some power is wasted.  The large octal ones use argon gas, so the pink colour looks cool.  Some of the small ones use hydrogen, so they glow orange.