• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

DC filaments setup

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
playing around with PSUDII the other night I decided to try setup a 6.3V PS for filaments. playing around and trying different setups including many from headphone amps on www.headwize.com, I couldn't get 6.3V or even close on the output, i tried different filter setups including just a capacitor. also the AC ripple seems somewhat enormous. I really doubt that if most of these designs are actually used, that they provide these low voltages. So my question, is PSUDII not able to function accurately at low voltages? Or perhaps I am doing something wrong?
 
Ex-Moderator
Joined 2003
I find PSUD tends to be a bit pessimistic about output voltage (and yes, I use as accurate data as I can), it's not much of an error, 5% at worst. However, why worry, why not design for 9V, and add a regulator such as a 317 to give precisely 6.3V, even when the mains moves up and down?
 
diyAudio Senior Member
Joined 2002
BEING MY USUAL SMART ***...

Hi,

While I agree that the LM317 is a fine regulator...sometimes a good regulator with limited bandwidth can be used to advantage:

A prime example is the 78/79xx series of regs, their limited bandwidth has a positive effect in that it keeps mainsborne/rectifier crud away from the tubes/valves.

While it doesn't allow for a spot on 6.3VDC/12.6 voltage (well not without the use of a zener noise source anyway) but gives you a 6/12VDC output this will not affect performance in any way...

Unless EC8010 is onto something I'm not aware of yet?

Cheers,;)

P.S. I hear people getting fine results by regulating both filament voltage AND current...the why for this is still beyond me but like everyone else I'm learning.
 
I have read various things on AC versus DC on heaters... does anyone know the real deal that can explain it? I would think that using a perfectly stable DC voltage appropriate to the tube would be best, but I keep reading that AC is better for the tubes or some such thing (and it sounds better). Also, if it is so much better to use a different transformer for the heaters from the B+ transformer, why is it not used more often? Just cost? Is there a benefit from running different heaters from different trannies? I guess what I am saying is why not use 1 tranny for B+, and a seperate one for all the heaters using solid state rectification and regulation?
 
Ex-Moderator
Joined 2003
Differential vs common-mode

I've always found 78/79xx series to be noisier than 317. If you look at the manufacturer's data sheets, they show poorer power supply rejection as a consequence of reduced (internal) op-amp gain-bandwidth product. Whether this translates into any real difference, I don't know, after all, we're talking about noise at the 20mV level as viewed by a something MHz oscilloscope on a heater that would otherwise have had 6.3VAC on it.

I think that when it comes to noise, regulation is a bit of a red herring. Regulators reduce differential noise, to which the valve is not particularly sensitive. However, the valve is sensitive to common-mode noise.
 
diyAudio Senior Member
Joined 2002
The Cost Of Living.

Hi,

Also, if it is so much better to use a different transformer for the heaters from the B+ transformer, why is it not used more often?

Cost it is.

Is there a benefit from running different heaters from different trannies?

Yes, especially when current demands are varying wildly.Capacitive coupling once again but we're splitting hairs here.

I guess what I am saying is why not use 1 tranny for B+, and a seperate one for all the heaters using solid state rectification and regulation?

SS regs for the heaters, you don't have much of a choice here anyway.
Tube regs for the B+ and make any stage as independent as possible from the other unless you don't mind IMD.

Not to mention that you need independent PSUs for each channel too of course.

Boy, do I love PSUs...;)
 
Ex-Moderator
Joined 2003
Mach Y,

your question comes directly back to the issue of common-mode and differential noise. Regulators imply rectification, which generates noise. Thus, if you were unlucky, and used a poor transformer with noisy diodes and poor layout, you could generate more common-mode noise than if you had simply used AC.

In theory, DC shortens the life of directly-heated valves because one end of the filament/cathode emits more electrons than the other. AC heating balances this out. However, if you wire the filament one way round on one amplifier, and the other way round on the other, you can simply periodically swap the valves over and rebias. For a push-pull amplifier, it's even easier.
 
diyAudio Senior Member
Joined 2002
ON PAPER...

Hi,

I've always found 78/79xx series to be noisier than 317.

Yes, they are but not significantly so to abolish them for heater use.

Tip of the day: balanced heaters using a set of 7806/7906s on a 12.6 centertap?
This can be used on any heater as long as a balance is obtained for maximum CMR...useful for DAC stages.

SET folks are you listening?

Back to you EC, ;)
 
Ex-Moderator
Joined 2003
Re: ON PAPER...

fdegrove said:
Tip of the day: balanced heaters using a set of 7806/7906s on a 12.6 centertap?

Interesting idea. But negative regulators are almost always noisier than positive ones (I can't remember the details, but it's to do with the fact that it's easier to make NPN transistors than PNP, so the differential pair at the input of the op-amp is implemented with NPN, and their common-mode performance suffers).

However. One possibility is to use a 317 for the rail to rail regulation, and add a rail-splitter to couple each rail to ground. This is all becoming a bit theoretical because it's hard to reliably measure these sorts of problems. I suppose one could measure it by fitting a 100R, 100R 0.1% potential divider across the rails, and then measure the noise from the centre tap to ground.

Back to you, Frank
 
diyAudio Senior Member
Joined 2002
RE:RE:ON PAPER.

Hi,

It's been a while since I had so much fun on a Sunday night here... :)
Must be me...

But negative regulators are almost always noisier than positive ones (I can't remember the details, but it's to do with the fact that it's easier to make NPN transistors than PNP, so the differential pair at the input of the op-amp is implemented with NPN, and their common-mode performance suffers).

Correct, and that's why you need to balance both halves for two reasons: imbalance in the heater and imbalance from the regs.
Once done it equals a battery supply in performance.
So, we need to correct two basic sources of imbalance here:
The heater and the regs.

Is all of this worth the effort?

Well, I use this technique in a MC headamp since 1985 and it's been earning its salt there big time...
It's only when I noticed so many people having hum problems with SET amps and even simple preamps that I thought of this little beauty again.

Cheers,;)
 
Frank,

The active regged split rail heater setup you mentioned for DHTs was also written up by John Camille in one of the early Sound Practices (SP5?) with apparently very good results in a 211 SET. I'm contemplating it for the big momma 813/GM70(soon) PP that's currently singing in my system. I need to source a different powertrans to try it, and I'll probably use LT108x regs as I have some around here somewhere. Currently I'm using a passive LCLC (branch to each channel) LCLC setup, and whilst the noise and hum are fairly low, it's massive with those huge chokes (~3kg ea), so the active version has the potential to be a lot smaller.

Cheers
 
OK, EC8010 says, and I have read this from others, that DC on the filament of a DH tube (from triodes to pentodes, etc) causes uneven wear on the filament.

My contention is this:

From this picture, would not the schematic on the left show that the voltage of importance, the B voltage, is evenly distributed across the filament, inasmuch as it is in the center of both ends via the resistors? Can 2.5 or 5 volts actually impede current from a 300-500 voltage potential that much to cause the kind of current crowding you and others refer to? Whereas the circuit on the right is obviously going to cause one sided current flow. However, there doesn't seem to be a problem with any of my battery tube radios, let alone the heaters of the AC-DC radio tubes.:scratch:

Gabe
 

Attachments

  • dcfil.jpg
    dcfil.jpg
    8.2 KB · Views: 383
Ex-Moderator
Joined 2003
Tes, it can.

Remember that we are not talking about the voltage between the anode and cathode, but the voltage between the control grid and cathode. DC filaments mean that one end of the heater/cathode is at a significantly different voltage, so it emits a different number of electrons to the other end.

You can easily demonstrate this to yourself by looking at 300B or 845 curves. Hold anode voltage current constant, but chnage grid voltage by half heater voltage either side of your operating point, and you will see the problem.
 
diyAudio Senior Member
Joined 2002
HEATERS.

Brett,

The active regged split rail heater setup you mentioned for DHTs was also written up by John Camille in one of the early Sound Practices (SP5?) with apparently very good results in a 211 SET.

Interesting and flattering at the same time.
Little did I know that someone else had tried to tackle this already.

Gabe,

OK, EC8010 says, and I have read this from others, that DC on the filament of a DH tube (from triodes to pentodes, etc) causes uneven wear on the filament.

This particular phenomenon is called "notching".
It has been discussed here already on the forum a few months ago.
The term was brought back from memory by co-member D'Haen.

When fed by DC as EC8010 explains you stand a good chance that the cathode isn't heated evenly creating cold spots AKA "islands".
This in turn has an impact on the performance of the tube.

Cheers,;)
 
Ex-Moderator
Joined 2003
Even heating.

Hello Frank,

I'm not suggesting that the cathode is heated unevenly (although I would welcome discussion on "notching"), but I am saying that the DC voltage from one part of the cathode to the grid is different to another. AC heating averages these differences at a 50 or 60Hz rate, ensuring even wear.

Of course, these arguments only apply to directly heated valves.
 
Questions:

fdegrove, I remember the comment about notching. What is not clear is how that occurs. Is it a physical phenomenon that occurs due to the grid, inasmuch as the grid suppresses flow of current in certain areas of the filament? If that is so... then it would stand to reason that this would occur with AC as well as DC. The grid doesn't change location with AC does it?

Otherwise I would suspect more even heat distribution with DC versus AC, due to harmonics on the part of AC. Assuming clean DC.

EC8010, Is one end of the filament further away from the grid than the other? Or are you saying that one end of the grid is at 1/2 the DC voltage plus the bias voltage further away from the grid than the other, +/-? In other words, one end of the filament of a 300b is at 52.5 volts, while the other is at 47.5 volts, given a 50 volts bias? (I will have to check that out for myself!)

This is interesting and explains why in my battery tube radios when connected one way the tube doesn't conduct, since the filament voltage brings it into cut off.

But, given the age of my 201a tubes, and the fact that so many others do it, I do not see a problem with using DC on a 300b in the long run. I think it is a phenomenon as real but not as significant as is cathode stripping. It may only affect the tubes after 7-10 years of use! So... what of it?

My humble opinion,
Gabe
 
diyAudio Senior Member
Joined 2002
RE:Even heating.

Hi,

I'm not suggesting that the cathode is heated unevenly (although I would welcome discussion on "notching"), but I am saying that the DC voltage from one part of the cathode to the grid is different to another.

I wasn't trying to put words into your mouth...I take full responsibility for this claim I made.;)

A local search on the word "notching' will yield about 4 or 5 threads on it.

The reason I brought this up is that it's not commonly know amongst designers, but important nonetheless.

Too bad that file I intended to send you seems too large but it was discussed in there as well...in Dutch.

Another place holding useful info is the 1957 Sylvania Tube Manual and Tubecad.

Cheers,;)
 
Ex-Moderator
Joined 2003
Gabevee,

yes, I'm saying that one end of your 300B filament is at 47.5V, and the other at 52.5V, causing unequal emission. It's probably gilding the lily, but isn't that what we are about? And 300B's are expensive, so it's our duty to ensure longevity...

Frank,

time for me to learn some more...
 
diyAudio Senior Member
Joined 2002
NOTCHING.

Hi Gabe,

fdegrove, I remember the comment about notching. What is not clear is how that occurs. Is it a physical phenomenon that occurs due to the grid, inasmuch as the grid suppresses flow of current in certain areas of the filament? If that is so... then it would stand to reason that this would occur with AC as well as DC. The grid doesn't change location with AC does it?

From what I understand of it: somehow particles present in the heater material clogg together at one end either causing a heater to cathode short or, when it gets too heavy it actually breaks it.
This is a phenomenon that is well known in the lightbulb industry and between you and me I wouldn't loose sleep over it for audio use unless you want to run the same tubes for fifty years.
One way to prevent it is using AC for the heaters which, due to its pulsating action tends to clean this up.
Another solution, but no cure all, is to switch polarity of the DC heater occasionally.

What's more important is to heat the cathode evenly so electron emission is not concentrated on a particular end of the cathode.

BTW, the controlgrid doesn't really come into play here.

Otherwise I would suspect more even heat distribution with DC versus AC, due to harmonics on the part of AC. Assuming clean DC.

On the contrary, the alternating nature of the AC pulses heat the cathode more evenly, with DC applied you create an imbalance which is why I suggested the split -/+ DC arrangement using a virtual CT.

Other nasty problems can develop when using standby arrangements without really understanding how tubes react to them, IMO more tubes get poisoned by these arrangements than saved by them.

Cathode stripping, cathode poisoning, sleepy cathodes are all terms that come to mind.

Cheers,;)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.