If it is directly heated then it shouldnt matter though i should think it shouldnt matter reguardless. Either the elctron is jumping off of the filament itself or a hot pieof metal which is being very closely heated by a heating filament. Either way you should be able to run one by itself.
Of course you could heat the other one up for a little extra glow.
I wonder if you could use the unused portion as a guide which would cancel distortion caused by vibrations in the tube?
Of course you could heat the other one up for a little extra glow.
I wonder if you could use the unused portion as a guide which would cancel distortion caused by vibrations in the tube?
There should be no problem at all with heating just one section of the 12AX7, the only thing that matters is the cathode temperature of the one that is being used. Just make sure that you have approximately 6.3V at the filament, dc will be quieter if this is being used in very low level applications due to reduced hum pickup.
As for the vibration cancelling idea, you can't assume that the two sections will be equally microphonic or have exactly the same resonances. If this is a concern I would isolate the socket from the chassis and use superflexible stranded wire to connect it to your passive components, alternately just build the whole thing on a mechanically isolated subchassis.
Tube dampers help to a very limited extent as do tube caps.
As for the vibration cancelling idea, you can't assume that the two sections will be equally microphonic or have exactly the same resonances. If this is a concern I would isolate the socket from the chassis and use superflexible stranded wire to connect it to your passive components, alternately just build the whole thing on a mechanically isolated subchassis.
Tube dampers help to a very limited extent as do tube caps.
Kevin, FWIW, I've found nearly a 6dB improvement in microphonics with a well-balanced phono input using two sections in the same envelope, at least at lowish frequencies (below 1kHz). The tradeoff is 3dB more hiss for the same tube count, but the effectiveness of this trick surprised the heck out of me.
cs said:
You should be fine. You'll need to halve your plate/cathode resistors (assuming using a resistor loaded common-cathode stage). You might have to try a few tubes to get ones which balance well.
You will double your gM by paralleling, which will lead to a reduction in noise. Gain will remain the same, due to the halved effective plate resistance.
Hi Sy,
I have tried the same approach (differential input) and also found it worked in an early phono stage I designed, while it was very effective against external noise pickup, it didn't seem help the issue of microphonics at all.
I ultimately ended up running two 12AX7 per channel with all sections in parallel with individual cathode resistors to assure current sharing between all of the sections. I think in total this front end was about 12dB quieter (measured) overall than a single section, but had the disadvantage of much higher miller capacitance. I was using a HO moving coil at the time and the higher capacitance wasn't an issue - with many mm and low capacitance cable it would not be an issue either.
The lower noise of course was due to the gaussian nature of the noise generated by the triodes, the lower effective source impedance, and lower resistances required for the plate load resistor and RIAA network that followed.
I have tried the same approach (differential input) and also found it worked in an early phono stage I designed, while it was very effective against external noise pickup, it didn't seem help the issue of microphonics at all.
I ultimately ended up running two 12AX7 per channel with all sections in parallel with individual cathode resistors to assure current sharing between all of the sections. I think in total this front end was about 12dB quieter (measured) overall than a single section, but had the disadvantage of much higher miller capacitance. I was using a HO moving coil at the time and the higher capacitance wasn't an issue - with many mm and low capacitance cable it would not be an issue either.
The lower noise of course was due to the gaussian nature of the noise generated by the triodes, the lower effective source impedance, and lower resistances required for the plate load resistor and RIAA network that followed.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.