Filament DC offset for hum

[Rhodesplyr]

Recently, on another forum, someone posted a sketch of the circuit of 1940s guitar amp that appears to have been heavily modified at some point (including switching from cathode bias to back-bias), and one odd detail was that the center tap of the 6.3V winding is connected to a negative voltage node, the same one that supplies negative voltage to the output tube grids.

Someone else claimed that this was to reduce hum introduced by the filaments. I'm perfectly familiar with the idea of a positive DC filament offset to reduce hum, including the idea that the offset has to be high enough such that all portions of the heater waveform remain positive with respect to the cathode voltage.

I've never heard of a negative DC offset being used to reduce hum. Is the concept valid?

[loudthud]

Some Ampeg amps from the 60's use a negative bias on the heaters. I think the reasoning goes something like this: The heater wants to emit electrons from it's red glowing surface, although not as many as if it was white hot. The idea of positive bias on the heaters is that the cathode will repel those electrons. With negative bias on the heater the electrons will be drawn towards the cathode in relation to the voltage difference between the two. If the voltagee is more constant, the current is more constant so the influence of the line frequency is reduced.

[trobbins]

Biasing a heater in an indirectly heated valve to a DC bypassed voltage on either side of 0V typically causes an increase in the cathode-heater resistance. The minima in that resistance usually occurs around 0V, and the resistance typically plateaus to a higher level with bias beyond about 10-20V. I measured quite a few 12AX7 to check this, after I came across an old 1930's paper presenting measured results.

I noticed quite a variation in the measured levels, and the asymmetry. The general characteristic is not of a diode, as a few have analogised.

The aim is therefore to bias enough that the AC swing of the heater does not extend in to the lower resistance area, and to try different tubes. But note that the change is to heater cathode resistance only - not to the capacitance (which is likely to dominate hum for most situations).

Ciao, Tim

[Rhodesplyr]

I noticed quite a variation in the measured levels, and the asymmetry.

Question: I'm not sure I follow what you mean by asymmetry.

I recently repaired a 1953 RCA BN-2A mic preamp that has the 1620/6J7 filaments biased to around +12V. For a high gain circuit, it was surprisingly quiet with noise dominated by white noise. It had virtually no hum despite having AC heaters.

[R.G.]

Suppressing hum by doing [whatever] to heaters only works to the extent that the heaters were where the hum was coming from.

I had a Norton 850 Commando once. No matter what I did to the heaters, the mufflers were where the noise came from.

[trobbins]

My meaning of asymmetry is the variation in resistance with negative versus positive voltage bias. The following measurement plots show the likely variation that a 12AX7 could exhibit.

http://dalmura.com.au/projects/Heate...on%20plots.pdf

Extra comment related to the measurements is at http://www.diyaudio.com/forums/tubes...esistance.html

Only certain hum inducing mechanisms are related to the heater-cathode resistance of an input stage valve. Trying to improve this form of hum path would indicate that other typical hum issues have been addressed and shown to be at negligible levels.

[Rhodesplyr]

That's a fascinating graph. Thanks for posting.

Just to reiterate, my main point in posing this question was to verify that a negative voltage offset was a valid hum-reduction approach since I'd never heard of it. The rest about multiple noise sources goes without saying.

[catalin gramada]

I've never heard of a negative DC offset being used to reduce hum. Is the concept valid?

Hi
I used once for some 6SL7 and equivalents. It was very effective for some tubes and did almost nothing for other (depending of manufacturer). well the 6SL7 is a very strange tube for me, it develop a lot of EM couplings which are mainly responsable for noise. The worst was svetlana 6H9C(6N9S).I just cure completely the noise (including 50hz hum) by screening. Weird thing-the shield over bulb seems to acts like a sort of deflector and it affects the freq in hights. regards the heaters- yes negative dc reference works for me a lot (for this tube) but did not kill completely the 50hz noise alone but reduce subbstantial for some.
Catalin
later edit. from RCA manual:

[Merlinb]

Just to reiterate, my main point in posing this question was to verify that a negative voltage offset was a valid hum-reduction approach since I'd never heard of it.

Yes, it is just as valid the positive elevation. The idea is not to cut-off leakage current between heater and cathode but to saturate it (i.e. it becomes unvarying). Whether it is positive or negative doesn't matter, although it is true that the leakage is usually less when the heater is positive. See: Cooper, C. E. (1944). Valve Hum. Electronic Engineering, (July), pp72-5.

[trem]

Yes, it is just as valid the positive elevation. The idea is not to cut-off leakage current between heater and cathode but to saturate it (i.e. it becomes unvarying). Whether it is positive or negative doesn't matter, although it is true that the leakage is usually less when the heater is positive. See: Cooper, C. E. (1944). Valve Hum. Electronic Engineering, (July), pp72-5.

Once again, everything I Thought I knew is wrong.
I have that book (most of it is way passed my security clearance) and I will look that up.

I thought the problem with (A.C.) heater noise was....as the 6.3 VAC supply swings negative, the cathode starts looking more positive, and so the (instantaneous) negative heater would want to give up electrons to the Cathode.
If the heater were say 10VDC+, its electrons would not be attracted to the (more negative) Cathode of most preamp or power tubes.

But I must be wrong, because if it is OK for the heater to be more negative than the cathode, my understanding is not correct at all.?

[trobbins]

Perhaps the simplest way to think of it is a noise being injected in to the cathode of the input stage. That noise effectively is amplified, because it appears across the cathode - grid.

In an indirectly heated cathode, heater noise source in question is isolated from cathode by a resistance and a capacitance, so the noise couples in via two paths. DC biasing the heater voltage just changes the resistance path (by making the resistance higher), but doesn't change the noise leakage through capacitance.

Heater noise can also couple in by the wiring being close to the grid wiring for example. Heater noise can also include rectification hash from B+, and not just be 50/60Hz sine wave.

[Tooboob]

Great discussion ("I have this article from 1944", "I have it too" - what are the odds?) Great stuff to know.

[stevehoover]

I built a 63 Bassman Normal channel clone that no matter what I did, had a higher than normal 120hz buzz.
I de-elevated the heaters by connecting a tap off the bias supply to the center tap of the 6.3 heater supply.
It is now a very quite running amp.
Older texts that I found describing this technique mentioned that you need to be conscious of heater max voltage ratings. Also: it is Not recommended for circuits with cathode followers due to the elevated heater voltages found in those circuits. The technical explanation of how this works is above my pay grade, but it was very effective in my case.

[nashvillebill]

The Gibson GA-19RVT schematic even shows the heater center tapped connected to the cathode of the 6V6 (that's the way my 1963 Epiphone EA-15RV came wired, though not shown on its schematic). I did not realize however that it could be equally effective to make the heaters negative as well as positive! Learn something every day...de-elevating the heaters!

[uneumann]

I built a 63 Bassman Normal channel clone that no matter what I did, had a higher than normal 120hz buzz.
I de-elevated the heaters by connecting a tap off the bias supply to the center tap of the 6.3 heater supply.
It is now a very quite running amp.
Older texts that I found describing this technique mentioned that you need to be conscious of heater max voltage ratings. Also: it is Not recommended for circuits with cathode followers due to the elevated heater voltages found in those circuits. The technical explanation of how this works is above my pay grade, but it was very effective in my case.

The Bassman and other amps with a bias supply winding can easily generate a lift voltage for the filaments (or other purposes). I'm not sure where I've seen this idea, but this is an example from my own Bassman. It only requires 1 diode and 1 cap.



There are two single wave rectifier diodes and filter caps from the same filament winding. The lower diode and cap produces the normal negative voltage bias supply, the upper diode/cap creates a positive voltage supply for filament lift. That filter cap doesn't have to be that large since lift doesn't require much current. I used the 100uf cap since it was what I had at the time.

Since the Bassman does have a cathode follower, it's less stressful to that tube to use a positive lift.