Half-wave rectifier for filament voltage

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

JanusRec

Well-known member
Joined
Oct 25, 2008
Messages
66
Location
Berlin, Germany
Hi,

for my low voltage tube preamp I need a simple way to get 6,3V out of a 18VAC trafo.

I thought of half-wave rectifying one of the trafo terminals and driving a Lm317 with the resulting 11 or 12V DC (with charging caps), so I have 11V -3V (consumed by the LM317) and 1,7V for the LM to burn = 6,3V

Anything wrong with that?

Or are there better designs to get 6,3V without a DC-DC converter? I didn't find anything in the PSU meta or using the search function...
 
JanusRec said:
for my low voltage tube preamp I need a simple way to get 6,3V out of a 18VAC trafo.

I thought of half-wave rectifying one of the trafo terminals and driving a Lm317 with the resulting 11 or 12V DC (with charging caps)

If your transformer has a single 18V winding (ie: only two wires/terminals coming out of it), half-wave rectification will still give you ~24-26V at the filter cap; you'll just have more ripple.

Can you reconfigure the tube filaments for 12V operation?

Tube filaments are a pretty predictable load. Consider a C-R-C filter before the regulator, with the R calculated so that it leaves ~10V at the input of the 317 at full load. Check if the 317 is still happy with the (larger) load offered by a cold filament. It might be better/simpler to just pick an R that drops enough to have 6.3V at the filaments and forget about the 317, though.

JD 'regulated DC filament supplies: Just Say No' B.
[or an LC filter after the rectifier, but you did say simple]
 
Svart said:
Why not just fully rectify it and just use a bigger heatsink?

Yes, but why should I do that, if there isn't anything wrong with the half-wave rectification? I would need some space for the heatsink on my pcb and besides it's a waste of energy - on the other hand I could turn down the heater in my studio  ;)

Seriously, if there are any problems with my idea, I will do what you proposed.

jdbakker said:
If your transformer has a single 18V winding (ie: only two wires/terminals coming out of it), half-wave rectification will still give you ~24-26V at the filter cap; you'll just have more ripple.

Can you reconfigure the tube filaments for 12V operation?

My trafo is a 18-0-18 center-tapped one, so one rail will give me about 12VDC.

How can you reconfigure tube filament operation? I have two pentodes à 6,3V/300mA and one double triode with 6,3V/300mA. So the latter drwas 600mA which is why I can't daisy-chain the filaments.

How do I calculate the C-R-C filter (planed to use a simulater to find out how to starve the LM317 a little). I know how to calculate a R-C though... Can't I just use a load capacitor a little too small and leave the R-C part?
 
JanusRec said:
My trafo is a 18-0-18 center-tapped one, so one rail will give me about 12VDC.

If it's an 18-0-18 transformer, it's 36V center-tapped, and two-diode rectification will give you 24-26V at the capacitor.
If it's an 18V center-tapped transformer, it's 9-0-9V, and two-diode rectification will give you 11-12V at the capacitor.
Which is it?

JanusRec said:
How can you reconfigure tube filament operation? I have two pentodes à 6,3V/300mA and one double triode with 6,3V/300mA. So the latter drwas 600mA which is why I can't daisy-chain the filaments.

What are the tubes you're using?

JanusRec said:
How do I calculate the C-R-C filter (planed to use a simulater to find out how to starve the LM317 a little).

You calculate the R with Ohm's law, then you pick a C which gives you acceptable ripple for an acceptable price.

Simulators are great at giving you the wrong answers, especially for boundary conditions. Does your simulator accurately model a 317 in deep dropout? Does your sim offer a model for the resistance variation of a tube's filament with temperature? Safe bet is 'no' to both questions.

JD "I really hate this damn machine/I wish that they would sell it/It never does quite what I want/But only what I tell it" B.
 
The trafo is 18-0-18, so you're right with two diodes it will give 24-26V at the capacitor. But my idea was to use half-wave rectification, so only diode, which should give me 12V compared to GND/center tap.
I just wanted to know, if it's bad for the rest of the PSU (+-18) and I read something about the trafo becoming magnetized, if you use half-wave rectification for the whole thing. But I'm also going to have a conventional full-wave rectifier, so I guess, that's not I probelm. But I'm sure, that's why I asked.

jdbakker said:
What are the tubes you're using?

EF98 and ECC86


jdbakker said:
Simulators are great at giving you the wrong answers, especially for boundary conditions. Does your simulator accurately model a 317 in deep dropout? Does your sim offer a model for the resistance variation of a tube's filament with temperature? Safe bet is 'no' to both questions.

I guess not  ;) So do you think it is better to have a series resistor to keep the current low, while the filaments are still cold?
 
JanusRec said:
The trafo is 18-0-18, so you're right with two diodes it will give 24-26V at the capacitor. But my idea was to use half-wave rectification, so only diode, which should give me 12V compared to GND/center tap.

It doesn't.

I believe what you suggest is using only one half of the transformer, connecting a diode to one of the 'outer' terminals of your 18-0-18 transformer, and having the reservoir cap between the other end of that diode and the transformer's center tap.

This arrangement will give you 24-26V at the capacitor.

(If you have any other arrangement in mind and you still believe you'll get 12V, please show a drawing or link to the circuit you're planning to use).

JanusRec said:
EF98 and ECC86

A quick glance suggests you could put those heaters in series (twin triodes don't draw double the heater current specified in the data sheet).

JDB.
 
jdbakker said:
I believe what you suggest is using only one half of the transformer, connecting a diode to one of the 'outer' terminals of your 18-0-18 transformer, and having the reservoir cap between the other end of that diode and the transformer's center tap.

This arrangement will give you 24-26V at the capacitor.

Now, rethinking it: You're right, of course. I will get 24-26V, only with a lot more ripple. Gosh, I'm glad I asked.

So I have to use the R before the LM317

jdbakker said:
JanusRec said:
EF98 and ECC86

A quick glance suggests you could put those heaters in series (twin triodes don't draw double the heater current specified in the data sheet).

JDB.

Ok, in the data sheet of the ECC86 it says "parallel supply" and on the pictogram it looks like two filaments. But I'm not too experienced with tube data sheets, so...

I can simply regulate my 24VDC at the cap to 19VDC and daisy-chain the three tubes then, right?

And thanks for your help, you saved me from burning my tubes!
 
No, you can't daisy chain the tubes, as although the heater voltage is the same, the current required is not. So they would not share the voltage equally.
 
Back
Top