I would like to try to heat my 300B's DC. Now there seem to be several possibilities to tackle this. I could use a bridge rectifier and a couple of 10.000 uF caps. Or two schottky diodes with the same caps. Or two caps and a choke. Those are the possibilities I've seen so far. Which of those would be the best sonically ? Or should I make it a regulated supply ? Could someone point me to a schematic for this ?
Thanks for your help.
Martin
Thanks for your help.
Martin
Without actually addressing your question (I apologize; I don't know the current requirements of 300B filaments), I would like to throw one other question out to whomever picks this up:
I suspect the best performance is probably had from a regulated filament supply.
But does it make sense to voltage regulate it or current regulate it?
I would imagine that an advantage of a current regulated supply is that it would automatically limit inrush current drawn by cold tubes, and it might compensate better for aging.
http://ken-gilbert.com/images/tubes/lm350.gif
http://ken-gilbert.com/images/tubes/DC_current_regs.gif
I suspect the best performance is probably had from a regulated filament supply.
But does it make sense to voltage regulate it or current regulate it?
I would imagine that an advantage of a current regulated supply is that it would automatically limit inrush current drawn by cold tubes, and it might compensate better for aging.
http://ken-gilbert.com/images/tubes/lm350.gif
http://ken-gilbert.com/images/tubes/DC_current_regs.gif
So if AC is good for filaments, because it promotes even heating, should it be 50 Hz or 60 Hz, or should it be some high frequency >> 20 kHz, which is inaudible along with all of its harmonics and all of its low order intermods with anything audible?
Of course, building a 100 kHz sinewave generator perhaps isn't practical, but might it not be theoretically better?
Of course, building a 100 kHz sinewave generator perhaps isn't practical, but might it not be theoretically better?
I have seen this asserted before, but I don't understand how it could work! Why would one end be hotter than the other, assuming the filament is a uniform resistance the whole way?
If one side was hotter than the other, you would be able to see it- one end would be white hot and the other would not be glowing!
If this was the case, the local resistance in the white hot region would increase until it reached equilibrium, and vice versa.
I imagine that if you could measure the voltage along the length of the filament it would be a regular drop along it, and the power dissipation would be constant over the entire filament.
Anyone ?
If you put dc on a directly heated tube, because the voltage drop along the length of the filament starting from one end varies from zero to dc supply, the grid bias along the effective length of the filament varies by the same amount. This would mean that one end of the filament gets more bias and so less emission, while the other end gets less bias and so more emission so it would wear out sooner. See if the knowledge of *that* ugly fact enhances your listening pleasure. 
Nothing like an indirectly heated tube, I think.
Nothing like an indirectly heated tube, I think.
Circlotron said:
This is just plain wrong. Bias has no effect whatsoever on filament emission. If you put 5V into the filament, a certain amount of electrons are released, period. It makes no difference to the filament if the tube is biased "off" or "on", or somewhere inbetween those states.
I'm afraid it does.
Consider a valve operating in a condition where -10 grid volts causes a certain anode current, and 10v DC is across the filament. Where do we measure the grid voltage relative to? One end of the filament or the other? Obviously the centre point would be logical, and it all evens out.
But, one end of the filament actually is only 5 volts above the grid, so emission is great, whilst the other end of the filament is 15 volts above the grid, so much less emission takes place.
Which is just another way of describing what circlotron said..
Now, whether it affects the sound or not is another argument
Is one side of a resistor in series with a load less hot? Stick your finger on it and report back to us.
DC voltage can reduce life by approximately 50%-70% due to the "sawtooth" or notching effect. Notching, also known as electromigration, is a change in the molecular structure of tungsten due to the DC voltage. The results are hot spots that accelerate the thermal condition, the evaporation rate and embrittlement.
Notching can also be caused by the "soret effect" which results from a temperature gradiant. Notching, due to the soret effect, takes place next to the mounting posts and support wire and occurs in both AC and DC applications.
http://www.oshinolamps.co.jp/english/product/limit.html
However the above is with lamps..and Victor Khomenko put it all in perspective in a post at the audioasylum..as far as I am concerned...
*******his is due to Metal Migration, a sort of electroplating process where the metal moves from Positive to Negative end of the filament and eventually get so thin that it breaks. The same would apply for Tube Filaments.
Not necessarily. Metal migration is a very strong function of temperature, and the temperature difference between the indicator lamps and tube filaments is huge. Filaments run at much lower temperature.
***As far as the sound is concerned should make no difference, except for AC Hum on Low Level circuits such as Microphone or Phono Pre-Amps.
There is another consideration - the start up currents. In DC circuits you pretty much have a complete control over this, you can ramp the voltage over several minutes, if you wish. This is much harder to do with AC, so there the thermal stress is usually much greater.
Net result? In thousands upon thousands of tubes used over the years (used both with AC as well as DC filaments) I don't recall a single filament failure. Number of tubes that failed for OTHER reasons: noise, microphonics, arching, breakage, etc. is so much greater than this whole AC/DC discussion is really largely academic.
Notching can also be caused by the "soret effect" which results from a temperature gradiant. Notching, due to the soret effect, takes place next to the mounting posts and support wire and occurs in both AC and DC applications.
http://www.oshinolamps.co.jp/english/product/limit.html
However the above is with lamps..and Victor Khomenko put it all in perspective in a post at the audioasylum..as far as I am concerned...
*******his is due to Metal Migration, a sort of electroplating process where the metal moves from Positive to Negative end of the filament and eventually get so thin that it breaks. The same would apply for Tube Filaments.
Not necessarily. Metal migration is a very strong function of temperature, and the temperature difference between the indicator lamps and tube filaments is huge. Filaments run at much lower temperature.
***As far as the sound is concerned should make no difference, except for AC Hum on Low Level circuits such as Microphone or Phono Pre-Amps.
There is another consideration - the start up currents. In DC circuits you pretty much have a complete control over this, you can ramp the voltage over several minutes, if you wish. This is much harder to do with AC, so there the thermal stress is usually much greater.
Net result? In thousands upon thousands of tubes used over the years (used both with AC as well as DC filaments) I don't recall a single filament failure. Number of tubes that failed for OTHER reasons: noise, microphonics, arching, breakage, etc. is so much greater than this whole AC/DC discussion is really largely academic.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.