I'm working on a design for a preamp with a phono stage. The phono stage is based on 2 EF40/86 and 2 5814A/12AU7 set up as amp -> RIAA -> amp -> cathode follower which then feeds into the main pre. The pre is all 12AU7s and 12AX7s. I'm running DC filaments across the board, so I settled on a parallel 12.6V to all tubes to reduce the current load on the regulators. Split supplies for the phono stage and the preamp stage. The issue is the two EF86 heaters at 6.3V each. My initial thought was to just run them in series, but after looking at some comments in other forum posts, there seems to be some concerns about heater mismatch and one tube pulling more current, so I thought I'd ask the hive mind.
Is there any way to balance the voltage across the two tubes? Virtual ground using 2 100R/5W power resistors? A current limiting resistor between the 2 filaments? Dropping resistors to turn the 12.6V to 6.3V (32R/2W)?
Any help would be appreciated!
Early schematic can be found here: https://www.diyaudio.com/community/threads/preamp-design-feedback.417521/
That thread apparently didn't get any traction. Probably the wrong forum section for that.....
Filament power supply below.
Thanks,
Chris
Is there any way to balance the voltage across the two tubes? Virtual ground using 2 100R/5W power resistors? A current limiting resistor between the 2 filaments? Dropping resistors to turn the 12.6V to 6.3V (32R/2W)?
Any help would be appreciated!
Early schematic can be found here: https://www.diyaudio.com/community/threads/preamp-design-feedback.417521/
That thread apparently didn't get any traction. Probably the wrong forum section for that.....
Filament power supply below.
Thanks,
Chris
Last edited:
Tubes in series have exactly the same current, so no problem there.
Running the 12a*7's at 12V is less current, same power as I understand it. So the 12V regulators don't need to provide the same current as 6V ones would. Though I certainly could be wrong....
C
C
A single 12AX7, for example, has two 6V filaments connected in series, with a center tap brought out (pin 9),
so they can be connected either in series or parallel. Each 6V filament takes 150mA, so in series the current
would be 150mA, and in parallel a total of 300mA per envelope.
so they can be connected either in series or parallel. Each 6V filament takes 150mA, so in series the current
would be 150mA, and in parallel a total of 300mA per envelope.
You don't need to balance the voltage across the tubes. There is no need. The fact that the individual tubes can be setup to have heaters in parallel or series should be a clue. It just doesn't matter.
In general 2 EF86 should be perfectly happy with series connection. Said that they are not designed for serial filament use thus 2 tubes from different vendors might be unbalanced enough to be of concern.
An alternative is to use PF86 ( serial filament variant used in TV set's) They draw 300mA, and has a voltage across
of 4.5V ( if i remember correct), thus 2 tubes in series builds to 9Volt, 3 V should be burned in a series resistor PF86 is significantly cheaper and more frequently sold.
An alternative is to use PF86 ( serial filament variant used in TV set's) They draw 300mA, and has a voltage across
of 4.5V ( if i remember correct), thus 2 tubes in series builds to 9Volt, 3 V should be burned in a series resistor PF86 is significantly cheaper and more frequently sold.
If you indeed observe a voltage difference between your series connected EF86 heaters, and if it is of any concern, i.e. if it is more than 5 %, you may parallel the higher voltage heater with a properly calculated resistor. NB: You need to measure and recalculate after a tube change.
Otherwise I second the PF86 recommendation that works well with DC heater supply.
Best regards!
Otherwise I second the PF86 recommendation that works well with DC heater supply.
Best regards!
Heaters are designed to generate a certain amount of (heating) energy into the cathode. Which results in a certain current draw at 6.3V or a certain voltage across the filament at 300 mA. It is just a matter of specification, either current or voltage. Both they have an U x I product which consumes the same amount of power.
With 2 6.3 V filaments it can be assumed safely that the currents are equal within production tolerances. If the tubes are from the same manufacturer and the same design. As correctly stated before by other members.
However, with different manufacturers who used a different design, the energy transfer to the cathode might be different and they might have needed to design the filament differently with a different energy and a different current.
Like others suggested, you simply can build the circuit and measure the voltages. If they are more than say 5% different, you can put a shunt resistor over the filament which drops the highest voltage. Because the difference will be small anyway the resistor is large and does not dissipate too much power. You would have to adapt the resistor for different tubes if you replace the tube.
Alternatively, petertub suggested to use a PF86 which needs 300 mA @ 4.5V. There also exists (existed?) the UF86 which is 100mA. The U-series was designed for radio sets which used fewer tubes than television. The current was lower so the voltage drop could be higher. The UF86 uses 100mA but at 12.6V, which would be a perfect match for you.
With 2 6.3 V filaments it can be assumed safely that the currents are equal within production tolerances. If the tubes are from the same manufacturer and the same design. As correctly stated before by other members.
However, with different manufacturers who used a different design, the energy transfer to the cathode might be different and they might have needed to design the filament differently with a different energy and a different current.
Like others suggested, you simply can build the circuit and measure the voltages. If they are more than say 5% different, you can put a shunt resistor over the filament which drops the highest voltage. Because the difference will be small anyway the resistor is large and does not dissipate too much power. You would have to adapt the resistor for different tubes if you replace the tube.
Alternatively, petertub suggested to use a PF86 which needs 300 mA @ 4.5V. There also exists (existed?) the UF86 which is 100mA. The U-series was designed for radio sets which used fewer tubes than television. The current was lower so the voltage drop could be higher. The UF86 uses 100mA but at 12.6V, which would be a perfect match for you.
A lot of great comments here. Thank you all.
I'm not going to change tubes at this time since I have a bunch in the bin that I'd like to use.
As I see it, I have a couple of options:
1. Let'er rip with the series filaments and hope for the best. I'm getting less comfortable with this choice.
2. Run everything off 6.3VAC. The tranny has 2-6V/4A windings, so power isn't an issue, but I am worried about noise. Especially with the pentodes.
3. Run everything off 6.3VDC. The total draw would be 2.2A. Even if I split this into two rails using the two windings, 1.1A through a TO-220 regulator is a bit high for my liking without good heatsinking.
5. Get Edcor to wind me a custom tranny with 12, 6 & 5V windings. This may be simplest since they usually wind to order.
4. Create a separate 6.3VDC rail. I was putting the two 6V windings in series to generate the 12VAC input. I think I can split off 6V from one of the windings and use that to generate a 6.3VDC rail for the two pentodes. And then I can split the 12A*7's between the two 12.6VDC rails with each sourcing less than 500mW. Something like this (without all the LM317 trash):
I've never seen anyone do this though. Is this a possibility, or a total fail? I do have a separate regulator circuit that is generating 5VDC for an Arduino Nano to control the input switching. I can feed that 6.3V just as well. Of course I was getting that off a separate 5V tranny winding. And I don't think i can generate 6.3VDC off a 5VAC winding consistently. Though I may be wrong there as well.
Thanks for looking,
Chris
I'm not going to change tubes at this time since I have a bunch in the bin that I'd like to use.
As I see it, I have a couple of options:
1. Let'er rip with the series filaments and hope for the best. I'm getting less comfortable with this choice.
2. Run everything off 6.3VAC. The tranny has 2-6V/4A windings, so power isn't an issue, but I am worried about noise. Especially with the pentodes.
3. Run everything off 6.3VDC. The total draw would be 2.2A. Even if I split this into two rails using the two windings, 1.1A through a TO-220 regulator is a bit high for my liking without good heatsinking.
5. Get Edcor to wind me a custom tranny with 12, 6 & 5V windings. This may be simplest since they usually wind to order.
4. Create a separate 6.3VDC rail. I was putting the two 6V windings in series to generate the 12VAC input. I think I can split off 6V from one of the windings and use that to generate a 6.3VDC rail for the two pentodes. And then I can split the 12A*7's between the two 12.6VDC rails with each sourcing less than 500mW. Something like this (without all the LM317 trash):
I've never seen anyone do this though. Is this a possibility, or a total fail? I do have a separate regulator circuit that is generating 5VDC for an Arduino Nano to control the input switching. I can feed that 6.3V just as well. Of course I was getting that off a separate 5V tranny winding. And I don't think i can generate 6.3VDC off a 5VAC winding consistently. Though I may be wrong there as well.
Thanks for looking,
Chris
Last edited:
That's technically true, yes. Series operation (of heaters designed for parallel operation) will exaggerate any differences in their heater power. But for only two heaters it is a pretty minor concern. Three might be pushing your luck.
Since you have a bunch of tubes with 2 filaments, that when serial connected provide a centerpoint (pin 9) you could use those filaments as "shunts" for your serial connected heaters. This would level out small differences between filaments and the, in my view, quite small, but nevertheless real, possibility of a run away ef86 heater.
Offcourse, operating with 1 of the ef86 pulled, or should one of the heaters fail, things would go seriously out of balance, so it would be a good thing to have shutdown, or at least some form of alert, should that happen.
Offcourse, operating with 1 of the ef86 pulled, or should one of the heaters fail, things would go seriously out of balance, so it would be a good thing to have shutdown, or at least some form of alert, should that happen.
Last edited:
and don't forget, the All American Five AM radio had no transformer, all the tubes' heaters were wired in series to get to 110v-ish.
it's just a hunk o tungsten stuffed with energy to heat up the cathode.
it's just a hunk o tungsten stuffed with energy to heat up the cathode.