Filament "polarity"?

[Earl]

I have read numerous times from cognizant amp builders/repair-persons...that when you wire the filament string, you keep the "A" filament wire connected to the same pin on every pre-amp socket, and the "B" filament wire connected to the same other filament pin on the socket.

IOW, "A" wire to all pin 9's, and "B" wire to all pin 3/4's.

My question is - will it make a difference in the amount of filament hum not conforming to this scheme? If so, can you explain the principal involved?

Thanks!

[Marc]

Originally Posted by Earl

I have read numerous times from cognizant amp builders/repair-persons...that when you wire the filament string, you keep the "A" filament wire connected to the same pin on every pre-amp socket, and the "B" filament wire connected to the same other filament pin on the socket.

IOW, "A" wire to all pin 9's, and "B" wire to all pin 3/4's.

My question is - will it make a difference in the amount of filament hum not conforming to this scheme? If so, can you explain the principal involved?

Thanks!

Earl,
I doubt seriously that this pin positioning would make any difference with AC filaments. Aesthetically it is very important to people who like their amps neat and orderly.... sounds like Hi Fi whoo-ee to me.


If it is truely important, I wonder how Fender kept all that green filament wire from getting confused over the years......

Marc

[Enzo]

Pins 4/5 by the way.

That was a favorite of Gerald Webber, wiring 12AX7s in "humbucking" mode.

I can't imagine it making any difference. The idea is apparently that somehow the hum from the heater AC would "cancel." Except that would mean that both sides of the tube were in the same phase relationship at the same signal level with the same number of stages after them. Not likely. How hum from the heater in the input stage tube would cancel against hum from the phase splitter tube is beyond me.

[MWJB]

Hi Enzo,

In actual fact Gerald says that it makes NO difference with preamp tubes. He only recommends observing polarity with respect to power tubes.

Sometimes it makes an audible difference, other times not - but for the time it takes to do you, may as well observe best practice (that's me saying that, not him :-)).

Marc - Fender couldn't keep from keeping that filament wire from getting confused over the years - many, many Fenders do not have the power tubes wired in 'correct' polarity.

[Enzo]

OK, been a while since I read Webber, but I could have sworn he referred to the 12AX7 wired for 6v as "humbucking." Since to him the current went up one side and back down the other. He saw that as some sort of cancellation.

[MWJB]

Sure, I seem to remember a similar description, but my interpretation was that 12AX7 were internally wired as humbucking, rather than being relevant to filament wiring to the tube socket (no reference to hand at present, may have dreamed it)?

[Earl]

Thanks for the comments, guys.

The only things I've ever come across, in regard to filament hum, is the 'standard' rules that applied to the filament circuit, rather than individual tube connections. The recent item concerning 'polarity' had me wondering.

My UL Twin has a "Hum Balance" pot, that you all know what it does...and adjusting it will eliminate (to my ears) almost all filament related hum. I doubt that amp I have has all the filaments wired the same.

[Enzo]

Yes, the 12AX7 internal "humbucking." DOn't now how that would work unless the two triodes met all the criteria I mentioned before as to level, phase relation, gain, etc.

[MWJB]

"My UL Twin has a "Hum Balance" pot, that you all know what it does...and adjusting it will eliminate (to my ears) almost all filament related hum. I doubt that amp I have has all the filaments wired the same."

Hi Earl,

I'd venture to say that the presence of your hum balance pot has more to do with general build quality, than with output tube filament polarity, specifically. Plenty of 50's, 60's & early to mid 70's Fenders (pre hum balance) are not wired with "correct" polarity at the power tubes, but still don't exhibit problematic filament hum.

After around '76 noticably less care seemed to have been taken with regard to wiring. Comparing the cost of a pot to that of a pair of 100ohm resistors, you can be sure that Fender only resorted to fitting the hum balance because they felt that they had to.

Generally, you only have a filament polarity problem if you can hear one & if rewiring cures it. As with other amp building "best practices" it's perhaps worth being aware of and incorporating (a small scale builder has the luxury of time to do this, less so on a production line), so that should you have a problem on powering up, you know what possible causes you have already eliminated.

[dai h.]

IIRC there was a good thread on this on the old ampage (so prob'ly in the archives).

what comes to mind as re: heater hum, AFAIK, is that

a) tube heaters can be constructed differently, so there might be differences in hum from that resulting in hum differences between tubes

b) generally heater wiring should be routed at a right angle to the other wires (relative orientation of the fields as with orientation of transformers mattering and affecting hum) and down against the chassis (slight bit of shielding due to capacitive coupling--maybe a bit less buzz = higher harmonics of the 50/60Hz hum)

c) if there is one, the heater center tap should be in the right place in the grounding scheme (wired with the high current stuff since high return currents through ground ( = return path for the AC current), for ex. 12AX7 = 0.3A, EL34 =1.5A, etc.) for lowest hum (also assume the same w/fake fil.cnt. tap)

[Steve Conner]

I agree with Enzo: Hum can't possibly cancel between different tube sections.

It maybe theoretically could in power stages, but hum induced through the heater is not an issue in power stages, because the gain from the cathode to the speaker is so low, and a power tube cathode is always grounded with respect to AC signals.

Some 12AX7s are "humbucking" in that they have a spiral wound filament that aims to cancel out magnetically induced hum. This works on a per-tube-section basis irrespective of filament polarity or 6V/12V configuration. All NOS 12AX7s and EF86s that I know of had the spiral filament, butI'm not sure about new production tubes.

You can also get electrostatic induced hum, from the filament voltage capacitively coupling to other tube electrodes, or the filament actually emitting electrons itself, as opposed to the filament current inductively coupling into other wiring.

This electrostatic and filament emission hum is what the humbucking pot nulls out, and what you lose when wiring 12A*7s for 12V heater supply, because the hum cancellation needs the voltage across *each* filament to be balanced with respect to ground, and this is impossible when they're in series.

Floating the whole heater system a few tens of volts positive cures the filament emission problem but won't affect the electrostatic one.

[xFallen]

Originally Posted by dai h.

IIRC there was a good thread on this on the old ampage (so prob'ly in the archives).

what comes to mind as re: heater hum, AFAIK, is that

a) tube heaters can be constructed differently, so there might be differences in hum from that resulting in hum differences between tubes

b) generally heater wiring should be routed at a right angle to the other wires (relative orientation of the fields as with orientation of transformers mattering and affecting hum) and down against the chassis (slight bit of shielding due to capacitive coupling--maybe a bit less buzz = higher harmonics of the 50/60Hz hum)

c) if there is one, the heater center tap should be in the right place in the grounding scheme (wired with the high current stuff since high return currents through ground ( = return path for the AC current), for ex. 12AX7 = 0.3A, EL34 =1.5A, etc.) for lowest hum (also assume the same w/fake fil.cnt. tap)

I have read somewhere (I cannot locate the reference at the moment) that heater wires should be up and away from the chassis and other wires (as you mention orthogonal to signal and other wires where possible) the goal being to reduce coupling AC to the chassis and thereby keep the amp quieter.

I pose this as a point of discussion only as I am interested. I have no opinion one way or the other and cannot detect a difference in the two recent chassis I have wired. It makes sense to me that it at least could make some difference.

I ultimately settled on wiring my 5E3 and 5F2a heaters with wires twisted together above everything else away from chasssis or other wires. My Champ project used the flat against the chassis approach and the wires are not twisted. There is no serious hum in any of them but the Champ probably has a tad more than either the 5E3 or nearly identical 5F2a.


Barry

[MWJB]

Hi Barry,

My results have been the same as yours, no real audible difference between flat to the chassis or up & over the tube sockets...Victoria (and some other Fender tweed style builders) go the "up & over route" on their tweed repros.

It comes down to personal preference. Flat to the chassis does make tweaks & mods slightly more accessible with less chance of burnt filament wire insulation.

[Jag]

Hi Earl! Are you the same Earl from over at the FDP?

It doesn't make a difference on the preamp tubes. My own expierience is that it does make a difference on the output tubes.

Basically, MWJB summed it up best in his first post. I too have had instances where it made no difference, however, most of the time it does make a difference IME. Sometimes the difference is not that great and I think it's one of the less critical areas regarding hum.

For those who really want to do everything they can, it certainly doesn't hurt.

[MWJB]

Hi Enzo,

Gerald's contention is this...

12AX7 have a centre tapped filament and are usually wired 6.3v humbucking. The statement was made in reference to 6EU7/12AX7 compatability, as 6EU7 cannot be wired humbucking.

[dai h.]

Originally Posted by xFallen

I have read somewhere (I cannot locate the reference at the moment) that heater wires should be up and away from the chassis and other wires (as you mention orthogonal to signal and other wires where possible) the goal being to reduce coupling AC to the chassis and thereby keep the amp quieter.

I pose this as a point of discussion only as I am interested. I have no opinion one way or the other and cannot detect a difference in the two recent chassis I have wired. It makes sense to me that it at least could make some difference.

I ultimately settled on wiring my 5E3 and 5F2a heaters with wires twisted together above everything else away from chasssis or other wires. My Champ project used the flat against the chassis approach and the wires are not twisted. There is no serious hum in any of them but the Champ probably has a tad more than either the 5E3 or nearly identical 5F2a.


Barry

keep in mind that when (as in Fenders) the heater wiring is done up and away, the heaters are still against the "chassis", that is the "chassis" in the form of the aluminum sheet on the inside of the cabinet cover.

(as you mention orthogonal to signal and other wires where possible) the goal being to reduce coupling AC to the chassis and thereby keep the amp quieter.

as for this point I think that is more about the relative angles/positions of the heater wires relative to the other ones. I think a good example of a clear rationalized approach (where you can see this in practice) is the Soldano SLO100 where the heaters are not twisted but paralleled strictly and are at 90 degrees relative to the other wires on the socket (which also seems to illustrate how twisting isn't required). The Reeves amps also but slightly different. re: AC coupling to the chassis, wouldn't that be favorable (as above) in reducing noise by a bit of capacitive coupling to reduce the buzz (so therefore they should be against the chassis)?

[DirtyGeorge]

Originally Posted by dai h.

IIRC there was a good thread on this on the old ampage (so prob'ly in the archives).

what comes to mind as re: heater hum, AFAIK, is that

a) tube heaters can be constructed differently, so there might be differences in hum from that resulting in hum differences between tubes

b) generally heater wiring should be routed at a right angle to the other wires (relative orientation of the fields as with orientation of transformers mattering and affecting hum) and down against the chassis (slight bit of shielding due to capacitive coupling--maybe a bit less buzz = higher harmonics of the 50/60Hz hum)

c) if there is one, the heater center tap should be in the right place in the grounding scheme (wired with the high current stuff since high return currents through ground ( = return path for the AC current), for ex. 12AX7 = 0.3A, EL34 =1.5A, etc.) for lowest hum (also assume the same w/fake fil.cnt. tap)

Why is c) necessary? It should not matter where you ground the center tap (assuming you are grounding and not connecting to a +Vdc bias). The center tap should have negligible current since all the filament current flows between the legs of the winding and none through the ground connection.

[dai h.]

it does matter where you put it, IME and the amount of currents (making the difference) was my rationale (I could be wrong). If it doesn't matter, you should be able to connect it to, say the input jack ground, but I think that would give you a major buzz (which means it does matter where it's placed). The fundamental reason is current flow in grounds.

[DirtyGeorge]

Originally Posted by dai h.

it does matter where you put it, IME and the amount of currents (making the difference) was my rationale (I could be wrong). If it doesn't matter, you should be able to connect it to, say the input jack ground, but I think that would give you a major buzz (which means it does matter where it's placed). The fundamental reason is current flow in grounds.

But the filament current flow does not pass through the center tap to ground. There should not be any appreciable current flow in the center tap ground.

DG

[dai h.]

wouldn't it be the same as a center tapped B+ winding (where it does matter where you put the center tap)?

[DirtyGeorge]

Originally Posted by dai h.

wouldn't it be the same as a center tapped B+ winding (where it does matter where you put the center tap)?

No, in the center tapped PT the current flows alternately through the center tap and each leg, steered by the diodes. In the center tapped AC filament circuit there are no diodes blocking current between the legs so no current flows through the center tap (it just flows from one leg to the other). Grounding the center tap only references the filament winding to ground such that one leg is +3v peak and the other is -3v peak. The polarity switching every half cycle.

DG

[dai h.]

so if my notion that it was high current returns making a difference is incorrect, what would explain a difference in getting less buzz when placing the heater center tap with the high current returns (B+ center tap, spk. jack return) and getting more when putting it towards the preamp? And why is the heater center tap a fat wire in my Marshall PTs? If there was no significant current flowing, couldn't they just use say 22AWG just for the fil. cnt. tap as with the B+ winding wires?

[DirtyGeorge]

dai h.
I can't explain why there would be a difference with grounding closer to the preamp. I always tie my filament center tap (usually virtual with two 100 ohm res) to a voltage divider for a DC offset. I have not experimented with grounding it.

As for the filament winding center tap being heavy gauge wire, I suppose the x-former manufacturers can't assume the user intention. You could use it for a full wave rectifier circuit and then there would be current flow through the center tap.

I guess one of us is going to have to measure the center tap and see if there is current flowing.

DG

[drewl]

I didn't measure, but when fixing an amp with a loose center tap, I remember a decent "snap" when I grounded it manually while it was on...seemed like a fair amount of current to me!

I know I've seen a few amps with dc preamp heaters that had the polarity's mixed....VHT comes to mind.

[gdgross]

Assuming a PP amp, I always wire my output tube heaters out of phase.

The thought being that any induced hum will be canceled in the OT, much like power supply ripple. I have no idea if it makes any difference in reality, but it's one of those things that is good practice and lets me sleep better at night.

[Enzo]

There are small currents from electrons flowing off the filament and to the cathode within the tube. Those 60Hz currents WILL share the ground path. That is the reason elevating the heaters to DC higher than the cathode reduces hum it stops that current.

Gerald's contention is this...

12AX7 have a centre tapped filament and are usually wired 6.3v humbucking. The statement was made in reference to 6EU7/12AX7 compatability, as 6EU7 cannot be wired humbucking.

Yes, without belaboring it, he says that, but the only way that would actually cancel hum would be if the two triodes are at the same signal level, in phase, and have the same gain structure after them until the signal paths are mixed.

Since the two heaters in a 12AX7 are joined at one end, and the two free ends are bonded to make this humbucker, can we really assume the two heaters in a 6EU7 are not configured exactly the same within the tube, other than the pins used? In the 12AX7 wired 6v, the heater current runs from pin 9 through the two filaments and out pins 4/5, and then reverses since it is AC. In the 6EU7, heater current enters pin 1, flows through the two filaments, then exits pin 2, and then reverse. Only difference is you can't split them into 12v on the 6EU7.

[dai h.]

There are small currents from electrons flowing off the filament and to the cathode within the tube. Those 60Hz currents WILL share the ground path. That is the reason elevating the heaters to DC higher than the cathode reduces hum it stops that current.

is that so called "heater cathode leakage"? So if the heater center tap needs to be connected to where the cathode is connected maybe that explains why the buzz was lower (and thus the correct place to put it)? (A return path for this small current loop.) Would that make sense as a correct explanation?

[dai h.]

Originally Posted by DirtyGeorge

Why is c) necessary? It should not matter where you ground the center tap (assuming you are grounding and not connecting to a +Vdc bias). The center tap should have negligible current since all the filament current flows between the legs of the winding and none through the ground connection.

okay I think I "see" this (mentally) better. So say you have no fil.cnt. tap, and a fake one w/two 100ohms, if the heater current for the tubes was going directly through the 100 ohm Rs, they would burn out. When operating normally, they do not, so they can't be a direct path for the heavy current. Say when you wanted to drop the voltage, and used an R to do it, you'd want it directly in the current path, and that would be in series with the outside htr. windings and not the cnt. tap.