Power supply chokes

Design and use of Chokes for PS, anode and filament use.
Post Reply
anton.dorrestijn
Posts: 2
Joined: Sat May 14, 2005 8:49 pm

Power supply chokes

Post by anton.dorrestijn »

Hello guys,

This is my first post here and I want to talk about chokes. I don’t have much experience with building tube amplifiers, but I’m having a lot of fun discovering how things work. From the Tube DIY asylum I learned that when designing the PSU, you have to keep the capacitance moderate, most prefer values below 100µF, even 50µF. On most designs, the values of the chokes are between 5H and 15H.

This leads me to the PSU design for my 300B amplifier as a CLCLC filter, build with GZ37 – 4µF – 10H – 20µF – 15H – 100µF. Nothing wrong with that.

Now let’s go crazy. Let’s design the filter as GZ37 – 1.5µF – 30H – 10µF – 40H – 40µF. This filter still gives good ripple rejection (even better), and it allows me to play with the first capacitor in order to get different voltages from the PSU.

So, the question is; What is the drawback in using “huge� chokes in the PSU filter?

Anton Dorrestijn
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Re: Power supply chokes

Post by dave slagle »

anton.dorrestijn wrote: I don’t have much experience with building tube amplifiers, but I’m having a lot of fun discovering how things work.
sounds like you might have found the right place.
This leads me to the PSU design for my 300B amplifier as a CLCLC filter, build with GZ37 – 4µF – 10H – 20µF – 15H – 100µF. Nothing wrong with that.
if you were to build this supply, i would expect you would be able to remove the 10H 20uF filter section leaving just a 4uF-20Hy-100uF filter and not notice any subjective change in ripple, however many have noted a distinct improvement in the sound from a "simpler" filter. So the simple answer is to build it as planned, then remove one LC section and see what you hear.
Now let’s go crazy. Let’s design the filter as GZ37 – 1.5µF – 30H – 10µF – 40H – 40µF. This filter still gives good ripple rejection (even better), and it allows me to play with the first capacitor in order to get different voltages from the PSU.
i like the idea of tuning the first cap to tweek the output voltage, but 1.5uF is on the upper limit of what i find sounds good. I also find this cap must also be an oil cap, since plastic caps just sound... well plastic in this situation. I personally have no problem with the tuning cap and am not as fanatical about pure choke input as most. The tuning cap lets you listen to a few different situations and possible learn something. a dedicated B+ transformer on a large variac really helps for these types of listening comparisons.
So, the question is; What is the drawback in using “huge� chokes in the PSU filter?
traditionally caps were much cheaper than chokes, but given unlimited resources, i'd still wonder about using those sized chokes in a PS.

in a perfect world, the 30 and 40hy chokes would not be an issue. In reality they add DCR and capacitance to the filter pretty quickly. Some have noted that for some reason low DCR chokes sound better, and adding shunt capacitance across the choke acts as a low impedance path for junk on the line.

think of it this way. Many use a simple LC filter of a 10hy choke and a 47 mic cap and report wonderful results (and no hum) adding the tuning cap before will drop the ripple substantially, and simply doubling the choke to 20hy will cut it in half again. This would seem a good safe starting point from which you could play and try to learn what matters to you.

I guess i am suggesting a little hands on listening (invite some friends over) and a few swaps might teach you something about what is important to you. Then based on that you can proceed in a logical manner to get the sound you like.

dave
Guest

Post by Guest »

Hi Dave,
I would expect you would be able to remove the 10H 20uF filter section leaving just a 4uF-20Hy-100uF filter and not notice any subjective change in ripple

Many use a simple LC filter of a 10hy choke and a 47 mic cap and report wonderful results (and no hum)
Okay, sounds as good advise, keep it as simple as possible. But, when you run the GZ37-4uF-20Hy-100uF by PSU Designer, you will end up with 270mV ripple.

When using a 300B Ultrapath with 3K4 transformer and using the 8 Ohm secondary, it will calculate to 51mV ripple (270 / SQRT( 3400 / 8 ) * 3.9).

That does not look solid, or am I missing something?

Thanks,

Anton
Post Reply