Any reason to not use an atx power supply for dc to filament/heaters ?

[tnmgcarbide]

i'm working on a project w/ a very nice Thordarson 300-c-300 HT power transformer . i plan to have the power supply remote from the pre and power
output . i'm spec'ing (2) 12ax7s, (2)12au7s, (4) 12v6 , and (2) 5y3s for a stereo pair .. that's about 1.6A for 12vdc
, 4.0A @ 5v . i have several ATX style power supplies with at least 4A and 12A respectively . they would be in a remote enclosure w/ the B+ transformer. that's the plan anyway.

is there any reason not to use this power supply that i'm missing here?

[Chuck H]

Discreet filament transformers are pretty common. But is the power supply you propose using capable of 12V@4A and 5V@12A simultaneously? Or are the specs 12V@ 5A or 5v@12A?

[Helmholtz]

The 5V heater supply for the rectifier will be floating on B+.
I wouldn't trust the power supply to provide enough insulation for that kind of operation.

Of course 5V and 12V heater supplies must not have a common ground.

[tnmgcarbide]

the legend on most state the output of 5v + 12v must not exceed ___# of watts, which is typically 200 to 350 ..

[tnmgcarbide]

ahhh.. so there it is. a 5v rectifier is a directly heated device...and needs ac at the filament , but the indirectly heated audio tubes would be fine with dc.

..that's why i asked here . thanks for the tip. a 4A-5A AC transformer is no big deal .

or i could go with an indirectly heated 12x4 recto ? it was used in cars with the 12v6. that coud be much simpler.

[Chuck H]

The HV caveat didn't occur to me either just glancing at the matter. Thank you Helnholtz!

And... Why not just skip the rectifier tubes? Using an AXT power supply is hardly "vintage" anyway. So what nostalgia is to be regained by using recto tubes? I know they would be showy, but beauty follows function.

[tnmgcarbide]

if money were no object i'd restore a nice pair of Eico hf-50's and be done with it...a tweed Fender twin while we're dreaming would be great , too.
diodes are fine, but i already have a bunch of oddball tubes i can use.

i enjoy the sound of tube audio. not the nostalgia .

[glebert]

they would be in a remote enclosure w/ the B+ transformer. that's the plan anyway.

So you are going to have an umbilical going from the transformers to the amp? I guess I would be a little concerned about the B+ high voltage (AC or DC) going down a cable that is "out in the open."

[tnmgcarbide]

maybe just a foot or so away... to keep unfiltered AC away and avoid magnetic hysteresis with the OTs ..simple enough. i might use flexible conduit for
shielding .
i'm not building some snooty bouquet-of flowers here. if i thought it would sound quieter- i'd put the whole shebang inside a galvanized trashcan as a
Faraday cage.

[g1]

Noise from the SMPS could be a concern?

[tnmgcarbide]

more so than unfiltered AC from the wall? i guess i'll try it out first with an existing device .. atx are isolated from the line and ground , the 12v iwould not be
in circuit and not referenced to ground-just running heat lamps. the switch -mode frequency is likely 100khz or higher ....maybe some rogue oscillations
might occur?

we shall see ...................... i have owned several 12ax7 preamps powered by wall-warts, but as starved plates at 30v or so. i never had a notion to put the output into my scope, but i'll look into it.

i would think a car's alternator is pretty noisy?

[tnmgcarbide]

ok, i did some experiments .i tried several different smps. i wired a 9 pin socket to pins 4,5 ,put in a 12au7 and waved it around in my HiWatt 50 amp at about
1/2 volume.. most were incredibly noisy-mostly the older specimens .a cable-box 12v ps was seriously noisy, but i had two modern ,kinda' expensive
medical smps that could work.
i roughly checked the noisy ones on my scope, and they had jagged ripple all over the place :@ around 60, 1k , 9-10k,, 80-100k .whatever.
more artifacts than egypt..

. i have a dozen or so 12 and 24v ac transformers i could rectify for dc filament use, we shall see. i could still run 6x4s in series , so that would at least let me

get some funky tubes into the mix.

thanks for the replies

[The Dude]

I'm with Chuck. I see no reason to use rectifier tubes. If this build is for hifi, you probably want a stiffer supply anyway. 1N4007's are about 15¢ a piece and are more reliable than a tube rectifier. Also, IMO, less build hassle than installing and wiring up a socket for a tube rectifier. The rectifier isn't in the audio path, so you'll still have the "tube sound".

[Chuck H]

I have to admit now that I really like tnmgcarbide's attitude on the matter. He has a bunch of funky stuff around and wants to build something unique and useful. If it seems a little weird so much the better. As long as it PERFORMS. That's just fun and I would do more things like that if I had the stuff and the time.

But yeah. Even with this ideology in mind in a pinch (if I didn't already have stuff that could be made to work) I would just default to diodes for practicality.

[tnmgcarbide]

yeah, i have a bag of 1N4007s ..and if you buy 100 them from Mouser, they were more like .08 cents apiece.
that said, i still have to light-up the driver,phase inverter and power tubes . i've a bunch of options @12 and/or 24 volts. 6bq5 or 6aq5 in series, 12aq5 , 12v6 just to name a few outputs. 6av6 and 12av6 are some other sleepers that make for fine phase inverters.