diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Tubes / Valves (http://www.diyaudio.com/forums/tubes-valves/)
-   -   Tube filaments (heaters) power supply (http://www.diyaudio.com/forums/tubes-valves/105121-tube-filaments-heaters-power-supply.html)

matejS 12th July 2007 01:47 AM

Tube filaments (heaters) power supply
 
Hi, all!

Everyone makes its out version of the PSU for the filaments (heaters) - 100 designs for one problem. Let's gather "powers"/experiences...
Let's say we are dealing with ECC88 (6922) - 6.3V and we have 2 tubes, using DC.
What's the "best" way to do it:
- in series (both get the same current)
- in parallel (both get the same voltage, but current is higher - harder to stabilize DC)
- separate (both get... well different - this is why I do not like this, since one channel can play louder due

How to regulate power? CRC? Or the best - battery :) Is battery worth trying? :)

Thanks,
Matej

astouffer 12th July 2007 02:33 AM

I would say a dedicated regulator for each tube. A 7806 would give 6.0 volts and still be within tolerance. An LM317 could be set for exactly 6.3. Constant current does not have any advantages over constant voltage. Sure it stops turn on surge and extends warm up time but honestly when was the last time anyone experienced a heater failure?

HollowState 12th July 2007 02:52 AM

Hi Matej,

Yes, there sure are many designs for DC filament supplys. Some are simple and others needlessly complex. (as in look what I can do) Since everything in electronic design is a choice of compromises, it often comes down to what you have to work with. Meaning space, resourses and goals.

For small signal tubes like you mentioned a regulated DC supply need not be complicated. My personal choice for a pair of 6922s, (let's say at line level) would be solid state rectifiers (either bridge or full wave) to a large electrolytic capacitor (10,000ufd) and then a three terminal 12 volt regulator, with a smaller bypass at the output.

This, of course, would be for a series connection of 12.6 volts. If you wanted the full 12.6 volts (not entirely necessary), insulate and float the three terminal regulator with a single silicone diode between the common and ground.

In most cases a battery supply isn't necessary and often proves to be more if a PITA then it's worth. Unless you're working with extremely low level signals, even then a filter with more sections and/or chokes will usually do the job.

Victor

matejS 12th July 2007 12:55 PM

Quote:

Originally posted by astouffer
I would say a dedicated regulator for each tube. A 7806 would give 6.0 volts and still be within tolerance. An LM317 could be set for exactly 6.3. Constant current does not have any advantages over constant voltage. Sure it stops turn on surge and extends warm up time but honestly when was the last time anyone experienced a heater failure?
I used 7805 w/ 2 sillicon diodes connected to the GND pin to get 6.3V.
I got 6.3V for one and 6.2V for the other tube -> one channel plays a little but louder.

Matej

ilimzn 12th July 2007 02:56 PM

Quote:

Originally posted by matejS


I used 7805 w/ 2 sillicon diodes connected to the GND pin to get 6.3V.
I got 6.3V for one and 6.2V for the other tube -> one channel plays a little but louder.

Matej

There should not be anything close to that sort of difference for only 0.1V difference in heater voltage. Have you moved tubes around between the channels?
Keep in mind that you need at least 3V higher LOWEST input voltage to the 78xx compared to output voltage, for proper operation. Check this out with a scope - a DMM is not good enough, as it shows you the peak (maximum) value of the ripple waveform, you should be looking at the minimum.

Oh yes - regarding heater supply rectification in general, if you need a voltage drop somewhere in circuit, be sure to provide some of it by a dropping resistor BEFORE the filter cap. Remember, you have a low voltage, high current supply, which rectifies into a typically very large filter cap (nx 10000uF is not uncommon). Diode and transformer winding peak currents will be high, reverse recovery will therefore be a problem, and the current waveform will be very ugly - and a transformer works both ways. Some of this bleeds into other windings in the same transformer. Schottky diodes help, but peak currents are still high, so any wires going around there are prone to induction.

matejS 12th July 2007 06:50 PM

Quote:

Originally posted by ilimzn


There should not be anything close to that sort of difference for only 0.1V difference in heater voltage. Have you moved tubes around between the channels?

I investigate a little but more and found out that the problems was on DAC PS.

Quote:


Keep in mind that you need at least 3V higher LOWEST input voltage to the 78xx compared to output voltage, for proper operation. Check this out with a scope - a DMM is not good enough, as it shows you the peak (maximum) value of the ripple waveform, you should be looking at the minimum.

Now I have 8.5V to 6.3V using 7805. I lowered it intentionally to reduce the heat on IC.
Voltage drop fpr LM7805 is 2V, so I guess this is enough. I tried with 12V SLA and the difference is still there. Maybe there is a difference between 7805 and 7905 or drop on 1N4008.

Quote:


Oh yes - regarding heater supply rectification in general, if you need a voltage drop somewhere in circuit, be sure to provide some of it by a dropping resistor BEFORE the filter cap. Remember, you have a low voltage, high current supply, which rectifies into a typically very large filter cap (nx 10000uF is not uncommon). Diode and transformer winding peak currents will be high, reverse recovery will therefore be a problem, and the current waveform will be very ugly - and a transformer works both ways. Some of this bleeds into other windings in the same transformer. Schottky diodes help, but peak currents are still high, so any wires going around there are prone to induction.

Thanks for this tip. Currently I am using CRC w/ schottky. I will replace schottkies with SiC diodes ("zero" reverse recovery) to have it more clean.

So in general for the heaters: SLA or CRC + LM317 voltage regulation (should be better to LM7805 due to better ripple rejection, etc).

Thanks,
Matej


All times are GMT. The time now is 06:59 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2