using 230v mains voltage directly for high voltage

[black_labb]

hi guys, i have been planning to build a ac15ish amp and have ordered some transformers, but was trying to calculate the value dropping resistor to get 315 v or so for the high voltage (im using 250 v ac output into diode rectification), and was pondering that for a while, wishing that i had 230 or 240, as that would be closer. then it hit me that the mains voltage in australia is around 230 v ac, which comes to 325v dc.

would there be any issues in running the b+ off the mains? would a choke be more nessisary than in other cases for some reason?

i cant see any flaws in the idea, but i thought id ask incase there was something im missing.

on a similar note, is there an safe dc voltage that could be used for the heaters on the tubes? or do you need different tubes (i saw that some sovteks are designed for 12v dc filaments, how are the filaments different)

edit: just had a bit of a read, seems that 6.3v dc is fine for the filaments. if i end up doing this ill either use dc or ac, which ever is cheaper (depending on wether a 6.3v tranny is easy enough to find in australia, or if 5v or 4.5 for rectification would be easier to find/cheaper).

[black_labb]

just found some relevant information (regarding hot chasis'). sounds like an isolation transformer would be needed to keep it safe (along with another tranny for the fillaments). seems that its more trouble than its worth, especially considering ive got a perfectly good tranny that will be here tomorrow.

whats the best way to determine the value for a dropping resistor? would it be to determine the resistance the tube creates and use ohms law? how do you find that value? using an ef86, there is a connection between pin 8 ( i think) and the cathode, which complicates it more, as that must affect the resistance. is there a value in the datasheets of the tubes for this purpose? if so, what is it called?

[Don Moose]

Originally Posted by black_labb

just found some relevant information (regarding hot chasis'). sounds like an isolation transformer would be needed to keep it safe (along with another tranny for the fillaments). seems that its more trouble than its worth, especially considering ive got a perfectly good tranny that will be here tomorrow.

Good move! Your safety trumps convenience (or should).

Originally Posted by black_labb

whats the best way to determine the value for a dropping resistor? would it be to determine the resistance the tube creates and use ohms law? how do you find that value? using an ef86, there is a connection between pin 8 ( i think) and the cathode, which complicates it more, as that must affect the resistance. is there a value in the datasheets of the tubes for this purpose? if so, what is it called?

If you run a Windows, hop over to Duncan Amp's Software Page and grab a copy of the PSU designer. Figure on between 2 and 4mA per triode for the load, and I'd guess about the same for an EF86.

The dropping resistor is partly about reducing the voltage the preamp tubes see, but it also forms a pi (3.14159) filter with the caps ahead of and behind it. That ripple reduction is an important thing, too.

Hope this helps!

[black_labb]

thanks, that calculator seems to be a great tool. now to figure out how to use it.

[black_labb]

im amazed at how much ripple is left after the filter caps are in place on a system without a choke. what kind of ripple would be suitable, how much hum would say, 1v ripple produce if it was going to the first stage of the preamp? is that 1v of ripple going to be the same as a 1v 50hz signal going into the preamp?

makes me want to get a choke now

[Don Moose]

Before you choke your circuit, consider going to a multi-section pi filter - divide the series resistor into three convenient values that add up to what you want and stick another filter cap in each of the two new nodes.

That first triode is the sensitive spot, and 1V ripple on the plate supply is going to be like having (1VAC/(that stage's gain)) on the input. Somebody smart needs to opine about how bad that would be.

Hope this helps!

[black_labb]

Originally Posted by Don Moose

Before you choke your circuit, consider going to a multi-section pi filter - divide the series resistor into three convenient values that add up to what you want and stick another filter cap in each of the two new nodes.

That first triode is the sensitive spot, and 1V ripple on the plate supply is going to be like having (1VAC/(that stage's gain)) on the input. Somebody smart needs to opine about how bad that would be.

Hope this helps!

thats a great tip, didnt think of that.

in my mind i was thinking that the high voltage ripple was going to be amplified, in the first stage, so you can imagine my shock when i was thinking 1v x 150 or so (mu of an ef86). i realised what is happening after i saw your formula, and now 1v seems alot less threatening now, but ill still try to get it as low as possible. the amp i built based on a 5e3 has plenty of hum, and id like to minimise it for this build, and possibly go back into the 5e3 and minimise it again. wouldnt be very suitable for recording.

thanks for the help and advice. have found it very helpful