Headphone Tube Amp D3a

[eeyore]

Hi guys, just wanted to run past my design for a tube headphone amp. It is a combination of Andrea's design (here) and the design here. See attached schematic.

Would appreciate and comments/feedback. It has been designed so that a 5842/C3g can be used instead, as they have a similar operating point, and all one needs to do is to use a switch (and different sockets obviously) to switch our R3.

There isn't anything ground breaking here. Just a combination of a few people's design, and my first attempt in designing and understanding tube circuits. Any insights, comments and feedback welcomed!

[mach1]

Connect the earth end of the 0D3 to the top of the cathode resistors instead of earth and reduce their parallel value to 47 ohms to get the same bias point.

[eeyore]

I read that on the elf website, but trying to understand the advantages of that beyond removal of the cathode resistor. Does it take the cathode resistor out of the signal path?

[eeyore]

Epc no elf. Darn iPhone spell check.

[dsavitsk]

Originally Posted by eeyore

Epc no elf. Darn iPhone spell check.

ecp

The idea of trying the regulator to the cathode is that it simplifies the signal current loop and keeps the cathode bypass cap out.

I would also actually recommend using a solid state shunt element instead of the OD3 -- the Zout of the OD3 is ~300R which I found detrimental to the circuit. A MOSFET can be less than an ohm. And, you can use the OD3 to bias the FET if you want to keep the neat glow.

[regal]

Headphones especially Denons and Grados can pick up the slightest hum so you may not like the AC heat, I would at a miniumum elevate it +20V's. I would also caution 120hz buzz with sensitive phones with the small filtering you show, but it may be OK. You may want to try the Salas HV Shunt regulator. I just finished a WE417A SET headamp powered by the SSHV and it is incredible, others have similiar reports with the C3g + SSHV.

Did you see ecp.cc ?
With your schematic and the one above the CCS feeding two tubes will present sort of a push-pull situation rather than shunt + SET (which is what I think you want?) I always wanted to build this and see its behavior with a scope but never had the time, I'm just afraid the two tubes will play tug of war without the 2H cancellation.

[regal]

dsavitsk posted at the same time, One other thing is oscillation with the 5842/C3g/6C45p. With the 5842 you want to try a grid stopper on all four grid inputs ( I used 10k) and a cathode stopper (5R) and a plate stopper (1R). Cut the socket pins down and solder the stoppers directly on the pins.

[Stixx]

others have similiar reports with the C3g + SSHV.

Just a combination of a few people's design

That seems to refer to me...

Like the use of two halfwave rectifiers in the psu and Os-con in the cathode... a good sounding alternative to LED's...

I looks fine as it is, others have already given good advice (and I second regal on the SSHV regulator... very musical!)

[eeyore]

Would like to keep the entire design simple by using just. A VR tube. But will look into the MOSFET regulator option proposed.

Regal, am interested in how the two tubes form a push pull pair. As far as I can see the bias resistor as it is shown now will force 18mA through the D3a. The balance will be through the 0D3. Unless the bias point of the D3a changes, which I assume over time as the tube age rather the over a very short period, the currents should remain unchanged. However, if the cathode resistor is to be common to both Tubes I can see how that will cause the two tubes to become a 'push pull' where the bias of the D3a will change as the two tubes are tied together.

[eeyore]

Regal,

Noted about lifting the heater voltage to 20~25V. Will check the tables to see the maximum limits. But you still think that I need more filtering? LCLC followed by a CCS and a VR? I thought that'd be plenty? No prob to add another LC stage.