Hi All,
Came across this article on 300B filament supplies. Can you share your thoughts pls. Me still learning about DHT
https://www.decware.com/cgi-bin/yabb22/YaBB.pl?num=1649817393
Thks
Came across this article on 300B filament supplies. Can you share your thoughts pls. Me still learning about DHT
https://www.decware.com/cgi-bin/yabb22/YaBB.pl?num=1649817393
Thks
I mostly disagree with the premises:
1) I have never seen a DHT suffering from stripping. So far, I have only seen DH tubes that can be supplied with both AC and DC or tubes that can be supplied with DC only (battery tubes). Maybe there is some uncommon type that can be AC heated only but that's not the 300B for sure. Done it many times with different types of 300B and never had problems. It is also clearly stated in the datasheet.
2) Supplying 5V/1.2A is not scary at all. High plate voltage is definitely scarier.
3) Best possible DC heating is not just about hum but it is also about getting rid of intermodulation distortion. Conventional 50-60Hz AC heating is the worst regarding intermodulation and hum. The best components in the world will not change that. Conventional well done DC heating will result in somewhat lower intermodulation distortion and lower hum. Intermodulation basically disappears if the DC heating also isolates the filament from the supply. Coleman's regulators and Tentlabs modules do exactly that.
I have seen people using high frequency AC heating but I have not seen anything about lower intermodulation. Using high frequency heating is no guarantee of intermodulation absence at audio frequency.
The only valid reason for using AC heating is that is simple. But performance follows accordingly.....
1) I have never seen a DHT suffering from stripping. So far, I have only seen DH tubes that can be supplied with both AC and DC or tubes that can be supplied with DC only (battery tubes). Maybe there is some uncommon type that can be AC heated only but that's not the 300B for sure. Done it many times with different types of 300B and never had problems. It is also clearly stated in the datasheet.
2) Supplying 5V/1.2A is not scary at all. High plate voltage is definitely scarier.
3) Best possible DC heating is not just about hum but it is also about getting rid of intermodulation distortion. Conventional 50-60Hz AC heating is the worst regarding intermodulation and hum. The best components in the world will not change that. Conventional well done DC heating will result in somewhat lower intermodulation distortion and lower hum. Intermodulation basically disappears if the DC heating also isolates the filament from the supply. Coleman's regulators and Tentlabs modules do exactly that.
I have seen people using high frequency AC heating but I have not seen anything about lower intermodulation. Using high frequency heating is no guarantee of intermodulation absence at audio frequency.
The only valid reason for using AC heating is that is simple. But performance follows accordingly.....
Last edited:
After having had some, and customer experience, i would look at anythig DEckard does with a big bag of salt.
dave
dave
Potential customers who have forgotten their high school physics generally view these essays by Deckert as technical documents when they are actually just marketing documents. It is not uncommon on some forums to have people incorrectly argue actual engineering principles, having taken these essays as accurate engineering documents.
In this particular case he has intentionally drawn the two schematics that accompanies the essay incorrectly. Deckert states that, "The right side (on the left below) seems to be the standard method ." The two resistors offer no useful purpose and the cathode resistor is incorrectly connected. The correct way to implement AC heating along with a cathode resistor and hum pot is shown below on the right. The two resistors, when correctly positioned, serve to reduce the current through the hum pot wiper, lowering noise by providing a lower resistance path from ground to the cathode of the DH tube.
In this particular case he has intentionally drawn the two schematics that accompanies the essay incorrectly. Deckert states that, "The right side (on the left below) seems to be the standard method ." The two resistors offer no useful purpose and the cathode resistor is incorrectly connected. The correct way to implement AC heating along with a cathode resistor and hum pot is shown below on the right. The two resistors, when correctly positioned, serve to reduce the current through the hum pot wiper, lowering noise by providing a lower resistance path from ground to the cathode of the DH tube.
It's simple. You want 5VAC for each 300b heater if you want to build a HiFi single-ended amplifier.
There are many fancy ways to do it. You can also just use LM317 with a big heat sink. 😉
There are many fancy ways to do it. You can also just use LM317 with a big heat sink. 😉
I think you meant 5 V DC. You can't create AC with an LM317...
I've used the LM22673 and LMZ22003 for powering 300B filaments. That can be done without a heat sink.
Tom
I've used the LM22673 and LMZ22003 for powering 300B filaments. That can be done without a heat sink.
Tom
Simple CRC filter is a big step up from AC. 5VAC transformer with sufficient current rating, Low drop diodes followed by 10mF cap or bigger, then small dropping resistors to set the final voltage to the filaments, then another 22mF cap and done. Not as good as a regulator, but easy and effective.
I also do not see the need of added resistors if, as he says, one has to buy a high quality potentiometer. I still have some high quality 22R and 50R pots, rated 3.5W, that I bought many years ago exactly for this use. No need for added resistors and they were not expensive (few euros each). The pot will only "see" the anode current that is 80-85 mA max in the 300B case. That means 0.36W dissipation in the worst case with 50R.
Actually Deckert has only the cathode bypass cap between the wiper and ground. The cathode resistor is between filament transformer center tap and ground.
My interpretation is that Deckert does not like or trust using the wiper as the center point of the cathode since it's a mechanical contact so he added two resistors and solder points to solidify the contacts. I believe, as Palustris suggested, there are errors in his drawings, missing some connection points like below. His preference of using filament transformer center tap allows the pot's only duty is for hum balancing. Right or wrong, I don't know but that's my interpretation of his thought process.
Sure, but I have lots of cheap LM317's and free HQ heat sinks from junked computer power supplies here...
I used AC on 300b filaments using an excellent quality 5V transformer that had a center tap. This was good but there was residual hum.
Going to CLC was much better. Cheap regulator circuit was heaven. If I didn't have so many LM317's left over in my stock, and salvaged computer Heat Sinks, I would buy a few of those regulators that tomchr suggested.
Naturally going to regulated 5VDC means you probably need a different Heater transformer...
Going to CLC was much better. Cheap regulator circuit was heaven. If I didn't have so many LM317's left over in my stock, and salvaged computer Heat Sinks, I would buy a few of those regulators that tomchr suggested.
Naturally going to regulated 5VDC means you probably need a different Heater transformer...
It doesn't make sense to me. If contact is bad hum will be bad. High quality pots guarantee contact for sure. It's always their main virtue.
And what's wrong with adding some resistors for better contact? I don't trust pots either. We get scratchy noise from Daven pots so we add high value resistor all the time like 1M from wiper of a 100K pot to ground to ensure the grid is not floating when the pot is open intermittently. He might have added some flowery tech terms for marketing but I don't see the harm of adding couple extra resistors.
AC filaments, versus DC filaments and a 300B tube:
1. Start with 1.25 Amps AC in a 300B filament wire.
Each section of the W filament (or M filament if you look at it that way) has an AC magnetic field, like the magnetic field of a wire of an AC electric motor.
The magnetic field causes the filament wire to be attracted to the steel plate of the 300B.
No problem, right?
Sometimes OK, sometimes Not OK.
Did you ever wonder why there was a small amount of residual hum from the 300B, no matter how carefully you adjust the hum balance potentiometer?
Did you ever wonder why some 300B tubes have more residual hum, than from other 300B tubes, no matter how you adjust the hum balance pot?
I have what often is the cause of the different levels of residual hum in different 300B tubes:
** If the filament wire is closer to one side of the steel plate than the other side of the steel plate, guess what . . .
The filament wire moves toward the closer side of the plate, and away from the further side of the plate.
It does this at a 60Hz rate (50Hz for some of you).
Perfect and Equal spacing of all the W (M) filament wires to both sides of the plate, reduces the resisdual hum of the 300B.
Your 300B tubes with AC filaments: Your residual hum may vary.
## But there is an additional problem with using AC filaments on 300B:
There is Intermodulation distortion on each tone / music frequency, of the test tone / music note.
It does this at a 120Hz rate (100Hz for some of you).
2. Start with 1.25 Amps DC in a 300B filaments wire:
It completely eliminates the effect of ** above.
And with DC powered filaments, the intermodulation of 120Hz (100Hz) effect of ## above, is also completely eliminated.
Just my opinions.
I have designed 300B amps both with AC, and with DC powered filaments.
Try it out and prove it for yourself. You will need a test tone, and either a spectrum analyzer or an FFT to prove it.
"You should make things as simple as possible, but no simpler" Alfred Einstein
1. Start with 1.25 Amps AC in a 300B filament wire.
Each section of the W filament (or M filament if you look at it that way) has an AC magnetic field, like the magnetic field of a wire of an AC electric motor.
The magnetic field causes the filament wire to be attracted to the steel plate of the 300B.
No problem, right?
Sometimes OK, sometimes Not OK.
Did you ever wonder why there was a small amount of residual hum from the 300B, no matter how carefully you adjust the hum balance potentiometer?
Did you ever wonder why some 300B tubes have more residual hum, than from other 300B tubes, no matter how you adjust the hum balance pot?
I have what often is the cause of the different levels of residual hum in different 300B tubes:
** If the filament wire is closer to one side of the steel plate than the other side of the steel plate, guess what . . .
The filament wire moves toward the closer side of the plate, and away from the further side of the plate.
It does this at a 60Hz rate (50Hz for some of you).
Perfect and Equal spacing of all the W (M) filament wires to both sides of the plate, reduces the resisdual hum of the 300B.
Your 300B tubes with AC filaments: Your residual hum may vary.
## But there is an additional problem with using AC filaments on 300B:
There is Intermodulation distortion on each tone / music frequency, of the test tone / music note.
It does this at a 120Hz rate (100Hz for some of you).
2. Start with 1.25 Amps DC in a 300B filaments wire:
It completely eliminates the effect of ** above.
And with DC powered filaments, the intermodulation of 120Hz (100Hz) effect of ## above, is also completely eliminated.
Just my opinions.
I have designed 300B amps both with AC, and with DC powered filaments.
Try it out and prove it for yourself. You will need a test tone, and either a spectrum analyzer or an FFT to prove it.
"You should make things as simple as possible, but no simpler" Alfred Einstein
Last edited:
Nothing wrong but that is usually done if the pot hasn't got enough power handling or is cheap.
If it is cheap, what's the point? Spend significant amount of money to buy 300Bs, good transformers etc and then save few quids on cheap pots. Never heard of Daven.
Dude, why are we still on this? Cheap or not cheap, an analog pot is a mechanical operation with the wiper. True to its name, it wipes on a circular path just like a car wiper and over time it builds up gunk or dry out or loosen tension or whatever and ends up losing contact integrity. If adding couple resistors aids in operating more smoothly, why not? I can't speak for the Deckert guy as I simply stated my observation. If you have beef or ax to grind, go to his forum and give your comment.
And a vintage Daven pot certainly ain't cheap! Look it up. Internally it's made of multiple tiny resistors in stepped rotation that in essence it's a multi-position selector so, again, a mechanical operation that leads to intermittent contact loss over time. But adding a resistor avoids the scratchy noise. And if you tell your client to throw it out he would kill you!
Are you are using DHT tubes?
Then be sure to use a DC supply for the filament; that gets rid of the hum balance pot.
Oh, and speaking of using DC filament supplies . . .
1. It gets rid of the 120Hz / 100Hz intermodulation (upper and lower sidebands on each test tone and on each musical note).
2. DC supplies also get rid of the residual 60Hz / 100Hz hum (that a hum balance pot can Not completely get rid of).
[Post # 17, try it yourself.]
AC powered DHT filaments?
Then be prepared to have fun using your FFT or spectrum analyzer . . . you will see all those upper and lower sidebands on the test tones.
They are there on the music tones too, but are harder to see, unless the instrumentalist holds the note for a long time.
Just my experience.
Whatever tube amp you use, Enjoy the Music.
Then be sure to use a DC supply for the filament; that gets rid of the hum balance pot.
Oh, and speaking of using DC filament supplies . . .
1. It gets rid of the 120Hz / 100Hz intermodulation (upper and lower sidebands on each test tone and on each musical note).
2. DC supplies also get rid of the residual 60Hz / 100Hz hum (that a hum balance pot can Not completely get rid of).
[Post # 17, try it yourself.]
AC powered DHT filaments?
Then be prepared to have fun using your FFT or spectrum analyzer . . . you will see all those upper and lower sidebands on the test tones.
They are there on the music tones too, but are harder to see, unless the instrumentalist holds the note for a long time.
Just my experience.
Whatever tube amp you use, Enjoy the Music.
Last edited: