Filament AC or DC?

[Aiace]

I'm wondering what is the best solution for an indirect tube (in particular for 6AS7).
What's the pro and cons form both the solutions?

Thank you in advance for sharing your experience and opinion.
Aiace

 

[SGregory]

Look at the last post in the 6V6 linestage. A properly implemented DC heater will reduce background noise. You can see the 60Hz and subsequent harmonics associated with the use of AC on indirectly heated tubes. Noise is noise and the less of it the better.

Here:http://www.diyaudio.com/forums/tubes-valves/102352-6v6-line-preamp-23.html#post2836779

 

[Shoog]

AC on a 6AS7 is fine and doesn't generate any noise to speak of.
I have a PP 6AS7 amp and it is whisper quiet. I have always been of the opinion that there is some strong evidence that a poorly executed DC filament supply degrades the sound. It is quite a challenge to make a DC heater for a juice hungry tube like the 6AS7 - lots of extra heat in an already hot tube - will shorten amp life. Lots of extra components and big vulnerable caps to degrade in all that lovely extra heat.

If you are asking this question - you are unlikely to end up with the correct DC supply and so will end up with a worse sounding amp that cost you a packet more to build.

Don't do it !!

Shoog

 

[Aiace]

It seems that the opinions are diveded fifty-fifty

In the end, the concept is : a bad DC is worse than a good AC. This is sound like a Lapalice sentence to me

Is there any other experience out there ?

 

[GloBug]

It's good to hear the full story, helps to make an informed decision.

Both sound reasonable to me, like a lot things, not so black and white.

I would verify what Shoog said and go from there. Just weigh it out the benefits vs. efforts and compromised components for your particular project.

 

[Shoog]

It seems that the opinions are diveded fifty-fifty

In the end, the concept is : a bad DC is worse than a good AC. This is sound like a Lapalice sentence to me

Is there any other experience out there ?

That of course assumes that in all instances all advise is equally applicable. That most certainly is not the case here.

A few minutes searching Google will show that no one is using DC supplies on tubes like the 6AS7, which should tell you plenty about the utility of going down that path.

There are obvious benefits for using well designed DC heater supplies with all Direct heated tubes, there are no obvious benefits with Indirect heated tubes which have reasonable heater to cathode insulation. All that is needed to ensure that this is so is to raise the heater supply by about +20V above cathode potential. This can be derived with a simple resistor network from the B+ (which serves the dual purpose of been your B+ bleed).

For a full discussion of the issues;

http://www.freewebs.com/valvewizard/heater.html

Last time I looked into this in any depth there was a general consensus that a voltage regulated DC supply was inherently inferior to AC. The feeling from those who had tried it said that it was absolutely necessary to use a CCS DC supply - which is a little bit more complex to implement than a simple voltage regulator.

Shoog

 

[Yvesm]

My position is to go DC only when all other options was unsuccessfull.

With indirect heaters, remember to "bias" the heater above the cathode, this may save you for that expensive, inneficient and even noisy rectifier, even at very small level like in a phono preamp.

Yves.

 

[Aiace]

All that is needed to ensure that this is so is to raise the heater supply by about +20V above cathode potential.

remember to "bias" the heater above the cathode

This is not clear to me at all. Reading on the first few pages of Philips data sheets (I found them in this forum elsewhere) the filament must be at lower voltage than the catode. So, +20V sounds very, very, strange to me. Please explain me why.

it was absolutely necessary to use a CCS DC supply - which is a little bit more complex to implement than a simple voltage regulator.

Absolutely not more complex. You need 1 resitor less using a CCS than VCS using an LM1085.
But obviusly I agree with you: AC is easier then DC.

 

[suntechnik]

the filament must be at lower voltage than the catode

If a filament negatively biased relative to a cathode signal can possibly flow from a filament to a cathode and from a cathode to an anode make AC filament audible.

In case filament positively biased only leakages from a cathode to a filament would be possible. Actually positively biased AC filament common practice in old tube gear like EL84 based.

You can check maximum possible voltage filament-cathode in a tube specification.

IMHO solid state linear stabilized DСpreferable nowadays since it’s very simple and inexpensive using modern ICs.

 

[Aiace]

If a filament negatively biased relative to a cathode signal can possibly flow from a filament to a cathode and from a cathode to an anode make AC filament audible.

In case filament positively biased only leakages from a cathode to a filament would be possible. Actually positively biased AC filament common practice in old tube gear like EL84 based.

I admit this confuse me furthermore.....
As far as I figured out, it's the opposite.
Cathode emits electrons and they go stright to a positive voltage. So if filament is positive a few electrons go to the filament generating a leakage current modulated by tha AC of the filament.
If the filament is negative respect the catode no one electron can goes to a more negative potential.
The most of you say the opposite, so I think somenthing is running out of my mind,but what ????


(if I've counted right, we are still 50%-50% )

 

[Shoog]

I admit this confuse me furthermore.....
As far as I figured out, it's the opposite.
Cathode emits electrons and they go stright to a positive voltage. So if filament is positive a few electrons go to the filament generating a leakage current modulated by tha AC of the filament.
If the filament is negative respect the catode no one electron can goes to a more negative potential.
The most of you say the opposite, so I think somenthing is running out of my mind,but what ????


(if I've counted right, we are still 50%-50% )

It explains the principle in the Valve Wizard page I linked to.

It is not cathode to heater leakage which is of concern, it is heater to cathode leakage, and by raising the heater above the cathode it prevents leakage from the positive heater to the negative cathode.



Shoog

 

[Yvesm]

Consider the couple heater / cathode as a diode.
The heater is hotter than the cathode and is inside, the inner side of the cathode is not coated.
So we have a direct heated diode, the heater/cathode being the filament and the plate being this small pipe formerly named cathode !
This diode will conduct only if its plate (the former cathode) is more positive than its heater: the filament.

The maximum voltage between heater and cathode is seldomly less than says, 50 Volts AC.
Nevertheless check data sheet

Yves.

 

[dhaen]

"Quiet" AC

Better than DC: Spend some effort making the AC clean of dirt:

Use separate heater transformer
Filter (LC) on primary or secondary
Careful heater wiring placement

"Ground" the heater centre-tap (or resistive divider) in a carefully considered location. Precise position depends on your circuit and earthing scheme.

 

[Jebem]

This is sounding like rocket science, guys!
I enjoy participating in this forum, we can learn so much from other people.
And this one is a very interesting issue...

I think I know one or two things about this subject as well... mainly from my experience working with tubes at home and at work sionce the 70's... here it goes:

- Indirect heating means that the filament is not in the signal path in any way. Its only job is to raise the cathode temperature so it will start releasing electrons.

- Tube manufacturors do not recommend DC for the filaments, but AC. There is a good reason for this, as filament life will be shorter with DC (this is very well understood since Thomas Edison lamps).

- But we can look to the cathode - filament as a capacitor. And a capacitor will transfer energy.

- So, I like to use DC for my tube filaments, because I do not want to use AC wires near tube sockets, even on utput tubes due to the proximity to the control grid. The energy transfer from the AC filament to the cathode on one output tube is very small compared to the signal amplitude at the control grid, so it will not raise the noise level.

- However, if You leave the filament floating (AC or DC), meaning that you do not ground it at the transformer center tap or at one side of the filament supply, we will experience a lot of avalanche noise on the loudspeakers. So we accept electrons migrating from the cathode to the filament because those represents a tiny small current that will not affect the anode current in practical terms.

 

[Aiace]

ok, now it is more clear:
- Heater emits electrones to and they go to the less negative catode.
But... the heater-cathode is not a diode infact also the cathode emits electrones, so if heather is positive a current may flows.

My conclusion, in order ro reduce these currents, the heather and cathode should be connected together !! Is this right also for you or there is some else I missed?

- Tube manufacturors do not recommend DC for the filaments, but AC. There is a good reason for this, as filament life will be shorter with DC (this is very well understood since Thomas Edison lamps).

I think you are speaking aboaut the war betwen Edison and Tesla, but this wasn't related to the filament but only at their investments and poor elphants( Electrocuting an Elephant - YouTube).
And no, I've never read in the data sheets that the manifactures racommend AC. Where did you find this information ?

- But we can look to the cathode - filament as a capacitor. And a capacitor will transfer energy.

Right, infact data sheet report the value of capacitance between H-K (7pF for the 6AS7). So a leakage current (as little as you want) always occur if you use AC.

For the moment we can resume this:
1) DC is the best way to eliminate the Hum due the filament and/or the wires that connect the heater to the transformer.
2) DC with big currents it's a pain. A lot ot heat, big capacitors, noisy DC (due the residual AC on the DC, diode spikes and so on).
3) Also AC requires L-C filters to eliminate extra noise.
4) A dedicated trasformer for the filament is better. (Also in AC ???)
5) More info needed about the life of the tube in AC and DC

 

[nigelwright7557]

I think quite often hum is blamed on AC heaters when the hum is coming from other sources. I once placed a transformer near a valve and that caused hum pick up badly.
Layout of high impedance wires is very important.

 

[dsavitsk]

It depends on how low level the signal is. In power amps, AC us usually fine. In a phono stage, DC is likely necessary. For in between, generally AC for differential, DC for single ended. For a 6AS7? It's a noisy tube anyway -- what's a little more?

 

[Vinylsavor]

Hi!

DC heater in indirectly heated power amps are unnecessary. As has been mentioned before they are even a potential source for problems and even poorer sound if not executed very well.

Best regards

Thomas

 

[Dimitris AR]

I use only AC supply for the filaments in all the stages of an power amp. , as for the line preamp I use AC again but baised with say +50V and I think this reduces the secondary emission inside the tube , for the RIAA preamp I use only DC .

 

[2wo]

I go with AC as well...John