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

[J M Fahey]

You can NOT repeat NOT feed the rectifier 5V filament from that power supply, nor from 6.3V used by others either, a couple days ago somebody at DIY Audio tried that, adding a 1 ohm resistor in series with filament , big sparks flew.

5V was chosen on purpose, to force you to use a dedicated well insulated rectifier winding.

FWIW Hammond tests their insulation with 2000VAC or 3000VDC, go figure.
They don´t call it HIPOT test for nothing

[tnmgcarbide]

i get it, thanks. i'm fully aware of the directly heated 5v tube ..

[tnmgcarbide]

so ... i am sure i'm not the first-nor the last to use some non-standard approach . if that power supply had worked..it would have been the
common thing to do-40 years ago.. that's why i like to build and experiment . test it and see why it failed to work.. it is how i learn.


so- looking through my ac- transformers, i have several 120/240 to 24v 6.5A out ,
which could drive an equal transformer 24v 6.5a INPUT to 120-ct-120

if i double up all fhe 12v heaters into 24v, , and /or use a voltage doubler diode/cap array. output.. this should provide .enough
voltage and power for 6l6 or 6ca7 outs. and 12v6 alike.

diodes or recto tubes...this chassis will still be a lead-sled. even more reason to separate the PS ..

[Chuck H]

i get it, thanks. i'm fully aware of the directly heated 5v tube ..

And this was tnmgcarbides recognition earlier as well after Helmholtz post on the matter. He has already acknowledged that if a tube rectifier is to be used it will need to be an indirectly heated type.

[Chuck H]

so ... i am sure i'm not the first-nor the last to use some non-standard approach . if that power supply had worked..it would have been the
common thing to do-40 years ago.. that's why i like to build and experiment . test it and see why it failed to work.. it is how i learn.


so- looking through my ac- transformers, i have several 120/240 to 24v 6.5A out ,
which could drive an equal transformer 24v 6.5a INPUT to 120-ct-120

if i double up all fhe 12v heaters into 24v, , and /or use a voltage doubler diode/cap array. output.. this should provide .enough
voltage and power for 6l6 or 6ca7 outs. and 12v6 alike.

diodes or recto tubes...this chassis will still be a lead-sled. even more reason to separate the PS ..

My only concern is that the voltage doubled circuit would make it harder to get a pure/flat DC voltage to the filaments. Without extra filtering measures it's possible that artifacts could present on the signal through the cathodes. Full bypass on cathode circuits helps but is hardly a guarantee since the impedance @ frequency for electrolytic caps is less than fabulous.

This could certainly be tried though. Peavey runs series DC filaments in some popular models I think. I've read here but haven't actually worked on any.

[tnmgcarbide]

as with the smps- i'll do a mock-up circuit and see what happens ..

sure- i could easily just buy a hammond or classictone PT and be done with it. that's no fun. why not just build a kit ? or i could sell my car and buy
a couple mono-blocks , hand built by Saul Marantz himself?

it would be tempting.

this week when i have some time i'll do a mock-up fender deluxe style circuit ,powered by the 240v transformer with and without the multiplier .

if it doesn't work out, and is noisy- 12v6 runs just fine @ lower voltage. it is rated for much less distortion @ 250v vs. 300v . 8wpc vs 10wpc . is not a big deal . into an EV or University corner speaker, 3-4 watts is often too much.

i'm not going to bother with dc heaters. that was just an afterthought of the smps interest. they'll be in series but still ac.

and-12aq5 has a max plate voltage of 275vdc. that could work too.

[SoulFetish]

If you wanted to bias the amp in class A or even cathode biased AB, you could run a pair of the 12A*7 heaters in series to achieve a 24VDC cathode voltage for a pair of 6L6s like Silvertone did in some of the amps.

[J M Fahey]

And this was tnmgcarbides recognition earlier as well after Helmholtz post on the matter. He has already acknowledged that if a tube rectifier is to be used it will need to be an indirectly heated type.

Maybe, but as soon as I saw that I jumped to answer, didn´t keep reading expecting "oh well, SOMEBODY must have risen that point"

Double warning?
Not bad in a potentially deadly situation.

[Chuck H]

Maybe, but as soon as I saw that I jumped to answer, didn´t keep reading expecting "oh well, SOMEBODY must have risen that point"

Double warning?
Not bad in a potentially deadly situation.

Not a bad point to make twice at all. Also noteworthy, and addressed by you above AND worth pointing out twice is that any supply used for a directly heated rectifier filament also needs to have winding insulation suitable for HV.

[tnmgcarbide]

ok , before my bench became too cold, voltage divider circuit of120v-24-24- 120v - 1n4007's + panasonic 47uf 450v caps produced 340v dc. i added 1000 ohms load resistor,
and the voltage dropped under load to 285-290v . the resistor was cookin' so i ended that , but seems to have been a success. the 230v
tap was more of a challenge . it produced 620v dc, then settled down .

[Chuck H]

ok , before my bench became too cold, voltage divider circuit of120v-24-24- 120v - 1n4007's + panasonic 47uf 450v caps produced 340v dc. i added 1000 ohms load resistor,
and the voltage dropped under load to 285-290v . the resistor was cookin' so i ended that , but seems to have been a success. the 230v
tap was more of a challenge . it produced 620v dc, then settled down .

Did that 620V occur with the loaded main supply test? And was the 230V node loaded? Screens do draw a small current and that will probably be enough to keep the voltage from peaking so high on that supply node. You should probably use 500V caps or a pair of 350V caps in a totem arrangement anyway just to avoid potential problems with inrush current and voltage peaking doing damage to the power supply caps.

[Helmholtz]

. the 230v tap was more of a challenge . it produced 620v dc, then settled down .

I wonder how that was wired. I don't see how you could get 620VDC short time without using a bridge rectifier or a voltage doubler.

[tnmgcarbide]

i was using the voltage doubler from the 240v coming from the output winding . i shut it down , fearing the 450v caps would burst in my face. from the
runaway voltage i witnessed. further testing will include 450v caps in series.

[Helmholtz]

Sorry that's getting too confusing for me.

Seems you're talking about different projects and PTs in one thread..

if it doesn't work out, and is noisy- 12v6 runs just fine @ lower voltage. it is rated for much less distortion @ 250v vs. 300v . 8wpc vs 10wpc . is not a big deal . into an EV or University corner speaker, 3-4 watts is often too much.

What is rated for much less distortion?

What does 8wpc mean?

[tnmgcarbide]

different voltages on the plates- different data per mfg. tung sol, ge, rca , have to put into data to really know.