New Posts  All Forums:Forum Nav:

12DQ6 SE Headamp?

post #1 of 10
Thread Starter 

Hi all,

 

I have a couple of 12DQ6s in my toolbox. I don't intend to use them as transmitting tubes because I have quite a few other tubes that will successfully fulfill that purpose. I am considering using these at low voltage to drive headphones through a transformer. Here is what I'm thinking...

 

- 12DQ6 in single-ended, parafeed mode.

- One stage of preamplification. NFB may be taken from output transformer to cathode of preamp triode. That will be if I need NFB. Preamp tube will be dual triode with 12v heater, not sure which yet. One triode per channel. Possibly 12SN7?

- Output transformer may be Edcor 15K - 600 ohm matching transformers (link below). With a 32-ohm load resistance, plate load resistance will look like 800 ohms.

 

http://www.edcorusa.com/p/158/xsm600-15k

 

- Supply voltage will be 30v. Filaments will be fed by constant current source transistors although no transistors will be in signal path.

 

The transformers are rated at 20-20KHz. I'll be using them in "parafeed", which means the transformer primary will be capacitively coupled to the tube plate and no current will flow through the primary winding. The 12DQ6 will have a resistor for DC current. If I idle the tubes at 15 mA each, that means I should be able to get 180 mW out of them into the 800 ohm transformer primary.

 

Questions, comments? I realize it's not an orthodox way to build a head amp but it's something I'm considering. I don't like buying high voltage parts and this seems like it would be something doable for a reasonable amount of change.

 

Ed

post #2 of 10

This sounds interesting, do you have a schematic? 

post #3 of 10
Thread Starter 

Not yet. Unfortunately, I was building a radio today with one of my 12DQ6s and found that it had a short between grid and cathode that only showed up after about a 10 minute warm up cycle. I had to toss the tube and choose another. I could look into using another tube. As for now, I won't be building this amp. I still wanted to pitch the idea because I know there's someone else out here who would.

 

Here's the problem...I know that a lot of "audiophile" headphone amplifiers can have outputs of up to 8w (some of the Schitt amps have this much) but I can't imagine that 1 watt could be very comfortable to listen to. I need to perform a measurement but I think 100 mW would be about as much as my ears could tolerate.

 

As for a schematic, take any generic SE tube amp with one triode stage as a preamplifier. That's pretty much what I'm thinking. The innovation is that I'm using a sweep tube with a low plate voltage, meaning it's not dangerous to touch the plate cap. The sweep tubes are built to have ultra-hot cathodes, letting them have a high plate current at low plate voltage. For a 6V6, you may get 35-40 mA at 150v with zero bias. A 6DQ6 would have considerably more with the same terminal voltages.

 

In building my radio receiver today, I found that the 12DQ6 had about 5.7 mA with zero bias at a plate voltage of 12v. At 30 volts, it might be a little higher. What you would need to do is to make the headphone driver of 32 ohms look something like 6400 ohms. That can be done using a 120v:6v power transformer (ratio of 20:1). The problem is that power transformers don't readily pass higher frequencies due to their higher parasitic capacitances. They would work but not well for audio.

 

Edcor has some coupling transformers that boast a ratio of 15K:600 at 20-20KHz frequency response. If you used a 32 ohm headphone drivers, the primary looks like a 512 ohm resistor. You want the output tube to idle at around 10 mA. Even better would be to use a parafeed where the transformer's primary sees no direct current.

 

So, with a plate voltage of 30, I think one of these tubes could successfully drive a transformer and therefore some headphones. It wouldn't be a great system but I think it could be a good system. The key would be the high plate current ability of a sweep tube.

 

Also, if you're one of those people who hates electrolytic caps in the signal path, you could rig up a negative power supply and feed the grids with a negative voltage. Class A amps have to have bypass capacitors with their cathode resistors unless you use negative bias on the grid. I don't find electrolytic caps to be objectionable but some of us do.

 

Ed

post #4 of 10

Surplus motor run caps should be cheap enough for any budget. I have used them with good success in a HV/low cost parafed amp. 

post #5 of 10
Thread Starter 
Quote:
Surplus motor run caps should be cheap enough for any budget. I have used them with good success in a HV/low cost parafed amp.

 

I've considered using them as power supply filter caps. Since they normally come in values like 8 to 10 uF, I'd have to get several of them to make one large bank. 60 uF is much more appropriate for a power supply blank.

 

If I built this I'd have to use a different tube. I don't have enough sweep tubes to use; I have a couple of 6DQ5s that will become transmitters. The idea behind using sweep tubes lies behind their massive, hot cathodes. At a plate voltage of 30, it will be able to pass enough current to put 100 mW into an audio transformer primary. Something like a 6V6, EL84, or other medium-power audio tube wouldn't be able to pass enough. I figure passing 20-25 milliamps would be enough.

 

The other option would be to use grid #1 as the screen. It would increase acceleration of the current in the tube and therefore allow for a decent current at 30v plate. The second grid would have to contain the audio signal. Because these grids usually have a lower level of control over the electron stream, I would likely have to amplify the audio signal much more. I could make a tube "op amp", or a differential pair with a lot of feedback and control the gain to something like 30. A headphone amplifier doesn't need a whole lot of gain.

 

Ed

post #6 of 10

If only using 20-25mA, the 1/2 W version of the transformer may be better, instead of the 2.5W.  It has a lower DCR.

 

I'm planning a build around this, http://www.edcorusa.com/p/885/pcw10k-7k_300-32 with 6922 tubes at around 15mA per triode, HV.  Staring with a common cathode SE parafeed, and then morphing to a differential "supersymmetry" like design with a long tail pair.  The differential is going to be a bit tricky, to match gain on the positive and negative inputs.  I might look at the 1/2W version of the transformer you mentioned, using a cathode follower.  I don't have sweep tubes, unfortunately.


Edited by holland - 7/18/13 at 10:36am
post #7 of 10
Speaking of cathode followers you could always look at using a CCDA stage, I'm having a lot of fun playing with one at the moment for a project that I'm going to show off over the next few days.
post #8 of 10
Thread Starter 

The problem with using a tube amplifier and transformers is that you have to have a separate tap for each individual headphone impedance. The plate of any given tube wants to see a certain ohm value for the best power out. Even though we're talking 100-200 milliwatts here, we're also talking 30 volts at the plate.

 

I could also capacitor-couple the headphones to a cathode follower amplifier. That would mean the final tube would try to keep the same voltage across any load it sees, until the load impedance drops and the final amplifier can't keep up. There's a fatal flaw with that layout though...a class A amp has to have an idle current at least as high as any current it intends to supply. So, for 300 mW into 32 ohms I need to run the final amplifier at about 140 mA. This is next to impossible to do with a tube at low voltage. At high voltage, the cost of the transformer would drive me to make a 50w speaker amp instead of a 1/2 watt headphone amp.

 

So, there are two options...use a transformer and drive the headphones with low voltage, or use tubes for the voltage amplification and use a powerful MOSFET to drive the phones. My current headphone amplifier uses IRF610s in class A source follower mode, loaded by a current source, and driven by a 12AU7, all at 12 volts. It sounds great to me. Someone with a super-discerning ear would probably find something wrong with it. I don't have patience for such people.

 

My in-process headphone amp is something like the Gilmore Dynamic amp, where I have four pairs of transistors in a massive V-8 power amplifier stage with an op-amp controlling the front end. It's not super-powerful but I think it'll be better at driving low impedances than the amp I currently have. The more I think about it I don't think it's a good idea to build a tube amp. I can still draft up a schematic in case someone else wants to build and tweak my design.

 

Ed

post #9 of 10
Thread Starter 

The problem with using a tube amplifier and transformers is that you have to have a separate tap for each individual headphone impedance. The plate of any given tube wants to see a certain ohm value for the best power out. Even though we're talking 100-200 milliwatts here, we're also talking 30 volts at the plate.

 

I could also capacitor-couple the headphones to a cathode follower amplifier. That would mean the final tube would try to keep the same voltage across any load it sees, until the load impedance drops and the final amplifier can't keep up. There's a fatal flaw with that layout though...a class A amp has to have an idle current at least as high as any current it intends to supply. So, for 300 mW into 32 ohms I need to run the final amplifier at about 140 mA. This is next to impossible to do with a tube at low voltage. At high voltage, the cost of the transformer would drive me to make a 50w speaker amp instead of a 1/2 watt headphone amp.

 

So, there are two options...use a transformer and drive the headphones with low voltage, or use tubes for the voltage amplification and use a powerful MOSFET to drive the phones. My current headphone amplifier uses IRF610s in class A source follower mode, loaded by a current source, and driven by a 12AU7, all at 12 volts. It sounds great to me. Someone with a super-discerning ear would probably find something wrong with it. I don't have patience for such people.

 

My in-process headphone amp is something like the Gilmore Dynamic amp, where I have four pairs of transistors in a massive V-8 power amplifier stage with an op-amp controlling the front end. It's not super-powerful but I think it'll be better at driving low impedances than the amp I currently have. The more I think about it I don't think it's a good idea to build a tube amp. I can still draft up a schematic in case someone else wants to build and tweak my design.

 

Ed

post #10 of 10
Quote:
Originally Posted by Bow to Ed View Post

The problem with using a tube amplifier and transformers is that you have to have a separate tap for each individual headphone impedance. 

 

This is not strictly true. Yes, if you want the best efficiency, but efficiency is not an absolute requirement, particularly if the amplifier is overpowered and won't blow up due to the mismatch.

 

w

New Posts  All Forums:Forum Nav:
  Return Home