Elevating heater via 100K resistors...

[Cliff Schecht]

So I think the OP's question was misconstrued from the start. From what I got Bob understands that the heater needs some sort of centertap (be it physical or artificial) and was asking about a 100k voltage divider off of the B+ to elevate the heaters to a common-mode DC level. In this case you bypass the bottom resistor in the voltage divider with a bypass cap (10uF maybe?) and enjoy the added silence of adding DC to your heaters.

Although IMO the 100k voltage dividers make the cathode-heater much too high, I think you're better off when the DC is close voltage-wise between the heater and cathode, this way you get the least potential between the two and less coupling..

[diagrammatiks]

You can use up to about 100vdc for the dc elevation.

Just need to figure out the correct value for the voltage divider.

I think you need to keep the grounded leg of the divider at 200k or less to minimize loading on the b+ rail.

[Cliff Schecht]

You want the value to be larger than 200k, going lower will start loading down the B+.

[Fischerman]

You want the value to be larger than 200k, going lower will start loading down the B+.

Isn't that what the 150k 'FET simulator' resistor does as well? Just sayin'.

FWIW, I did the elevated heater trick (by making an extra B+ node) on two high gain amps a few years ago and used the power tube cathode in my Rocket-inspired amp. I can't remember the resistor values in those two high gain amps but I remember the voltage was in the 40vdc ballpark.

My first post here in forever and a day. I guess the itch never really goes away.

[FUCHSAUDIO]

There are choices, and each has it's advantages and can in some cases solve some problems with heater noise or tube failures etc.

Two resistors can simply balance a heater line without a center tap. If you don't have one, or it's not accurately centered on the winding, this can be a good solution. A hum balance control can do the same thing, but with a greater ability to null out the hum to the finest degree. Lifting the filaments is an old audio trick used in recording consoles and stereo stuff. It raises the filaments above ground on DC. The values of the two resistors can be adjusted to raise from little to many volts. A 400-V B+ feeding 1-M over a 10-K raises you 1% of the incoming voltage. This can be adjusted with a pot for lowest noise too. The filtering is not difficult, one small cap does it. It loads nothing on the supply. It should be done early in the supply so there is no time constant making the voltage climb slowly (like running it on an input tube B+ filter cap junction)

A simple non loading network can be connected at the choke output, and can raise the filaments for two purposes: One purpose is to reduce heater cathode differential (in a cathode follower or phase inverter this can be a problem). The second purpose is noise reduction by reducing the difference between the input tube cathodes and the heater line where the hum might be amplified in the lowest level stages.

Technically, the heater cathode differential should not be an issue with good tubes, but today's 12AX7's (usually the Russians are worst) will not tolerate the RCA manual rating of over 100-volts heater to cathode. The problem with "fixing that" with an elevated heater is you now reduce the h/c differential on the higher cathode tubes, but the input tubes now see a higher differential.

IN any case, any of these can and do provide solutions to noise on the heater lines (unless you just use DC like real men do...) lol.

[azatplayer]

Im in the midst of a JTM50 build and am looking at elevating the heaters.
Im using separate filters on all stages, all 32uf.
With the screen cap im using a dual 32uf can, one side unused.
COuld i place this divider across the can from the screen node side?
Looking at Andys pic here, if i dropped a 1meg across the can from screen to unused half, then 10k back to the cans ground lug, would this be a good spot to tap from? Regards grounding, im using local filters so star grounds on a path, the ground for the screen cap be ok for this idea? Or should it have a separate cap altogether?
Quiet is good

[martin manning]

Sounds fine to me. As Andy says, that will elevate the heater string to about 1% of B+. This amp has a cathode follower, so you might want to raise the reference voltage up to more like 10% of B+ by using a 100K in place of the 10k. That will reduce the heater-cathode voltage on the CF to less stressful levels.

[azatplayer]

Thanks Martin, i was thinking of getting that voltage up further.
Im running around 450V i think with this PT and GZ34 from memory, used the pT prior to this build. Might be a little high for this amp so will tune it as i go.
Was looking at maybe 60-80V a a reference, more concerned with grounding as im going with this thing. Tho as anal as you can be, theyre a pretty forgiving circuit. Done some lazy grounding on these before with no ill effects.
Still good practice is good practice.

[FUCHSAUDIO]

In hindsight, I think a smaller cap (maybe 1-uf) is ok. If the top resistor is higher than 100-K, there would be a time constant. You don't need much filtering really, just a little bit. I've used 1-M over 10-K and a .1 cap in Marshalls that worked just ducky. I think the heaters cancel most of the hum on their own, and the elevation fixes that 120-Hz annoying hum that heater-cathode differential can bring. I had a JCM-800 with all new caps, tubes, and a dandy hum the customer said "it's always done that since I bought it in the 70's...). This circuit fixed it right up.....

[Gaz]

There are choices, and each has it's advantages and can in some cases solve some problems with heater noise or tube failures etc.

Two resistors can simply balance a heater line without a center tap. If you don't have one, or it's not accurately centered on the winding, this can be a good solution. A hum balance control can do the same thing, but with a greater ability to null out the hum to the finest degree. Lifting the filaments is an old audio trick used in recording consoles and stereo stuff. It raises the filaments above ground on DC. The values of the two resistors can be adjusted to raise from little to many volts. A 400-V B+ feeding 1-M over a 10-K raises you 1% of the incoming voltage. This can be adjusted with a pot for lowest noise too. The filtering is not difficult, one small cap does it. It loads nothing on the supply. It should be done early in the supply so there is no time constant making the voltage climb slowly (like running it on an input tube B+ filter cap junction)

A simple non loading network can be connected at the choke output, and can raise the filaments for two purposes: One purpose is to reduce heater cathode differential (in a cathode follower or phase inverter this can be a problem). The second purpose is noise reduction by reducing the difference between the input tube cathodes and the heater line where the hum might be amplified in the lowest level stages.

Technically, the heater cathode differential should not be an issue with good tubes, but today's 12AX7's (usually the Russians are worst) will not tolerate the RCA manual rating of over 100-volts heater to cathode. The problem with "fixing that" with an elevated heater is you now reduce the h/c differential on the higher cathode tubes, but the input tubes now see a higher differential.

IN any case, any of these can and do provide solutions to noise on the heater lines (unless you just use DC like real men do...) lol.

Unless I'm having a brain fart, in your schem with the 400vdc B+ example, 396 volts is getting dropped across the 1M resistor. I just wanted to mention that this is above the typical 350 volt rating of most 1/2W resistors. Even though the power dissipation is small, it may still exceed the voltage ratings of most common resistors.

[FUCHSAUDIO]

Usually, the voltage rating refers more to insulation breakdown (the coating on the part) than the part itself. I have used 1/2-Watts without issue, maybe a 1/4" watt on the bottom even....If a 1-Watt would help you sleep better at night, have at it....

[heisthl]

Unless I'm having a brain fart, in your schem with the 400vdc B+ example, 396 volts is getting dropped across the 1M resistor. I just wanted to mention that this is above the typical 350 volt rating of most 1/2W resistors. Even though the power dissipation is small, it may still exceed the voltage ratings of most common resistors.

Just series 2 470k as a sub for the 1M.

[Cliff Schecht]

It's funny this thread pops back up now. I was thinking about this the other night and came to the conclusion that adding DC to the heaters reduces capacitive coupling between the heater and cathode. Looks like I had the exact same thought half a year ago, forgot and re-remembered it .

[Colossal]

came to the conclusion that adding DC to the heaters reduces capacitive coupling between the heater and cathode.

I do this routinely on amps and especially those with a cathode follower. It a) makes for a quieter amp, in some cases graveyard quiet - no hum or buzz whatsoever b) is useful for where a cathode follower may be operating at higher than normal voltages and c) is quick and easy to do. It is especially easy on cathode biased amps where you've already got a 'built in' voltage source at the power amp cathode bypass cap. I usually tap off the screen voltage node, set up a divider of 180-220k as the dropping resistor, add a 250V electrolytic, and use 47k to ground. The resistors only need to be 1W or so, but I typically use 3W metal oxide to hedge against any failure.

[Gaz]

Usually, the voltage rating refers more to insulation breakdown (the coating on the part) than the part itself. I have used 1/2-Watts without issue, maybe a 1/4" watt on the bottom even....If a 1-Watt would help you sleep better at night, have at it....

Interesting, I'd never heard that before. I personally have used a 220k/56k divider later down in the B+ rail.

It should be done early in the supply so there is no time constant making the voltage climb slowly (like running it on an input tube B+ filter cap junction)

I was also wondering if you could explain this. I don't really understand what voltage you're referring to and why it would be a problem. Thanks!