better 6au6 alternative

[nhuwar]

Im using a schematic that uses 2 6au6's for driving my final tubes but I want to use something better. I'm just not sure want to use so if anyone has any suggestions i'm listening cost no object.


Thanks

Nick

 

[wa2ise]

It's hard to say unless we know which characteristic of the tube in its circuit you want improvement in.

 

[nhuwar]

Well the circuit is a altec lansing 260a that I'm kind of coping but I was told by a member that I should use something other then a 6au6. It's running at 180vdc b+ 120vdc g2 and I think it's got -25vdc with respect to cathode. Push pull stage 1 to 1 impedance to 2 813's in push pull.
I'm not exactly sure what charicteristic to improve though but the amp as designed only go to 15khz.

Nick

 

[ray_moth]

If the person who advised you to change to a different tube didn't give any reason, then it's hard to know what's supposed to be wrong with the 6AU6.

What is it that you would like to improve? If it's noise or microphonics, then just changing to a diffrent brand of 6AU6 may solve the problem. There's quite a bit of variability between different examples: for instance, I found GE 6AU6s to be highly microphonic, whereas RCA ones were quiet.

 

[Miles Prower]

Im using a schematic that uses 2 6au6's for driving my final tubes but I want to use something better.

I'd give that a real good re-think. The 6AU6 is already pretty darn good. Granted, it ain't exotic or particularly "sexy", but there's a good reason for that: the 6AU6 does just about anything quite well: audio, RF, DC (error amp or CCS). That's why you see them used so frequently.

 

[ray_moth]

Glad to hear you say that, Miles. I'm thinking of using 6AU6s in more applications. I've already tried them as a preamp in triode mode, with the plate and g3 grounded and g2 acting as the anode (v. small signal, or they overload), also in pentode mode as a CCS in the tail of a 6SL7 LTP splitter.

I really want to use a 6AU6 in pentode-mode (instead of an EF86) as the first stage of a Mullard-style PP amp, diract coupled to a 6SN7 LTP splitter (with another 6AU6 as CCS in the tail), which will drive EL34s in pentode-mode PP. You need a lot of voltage gain, to be able to apply generous NFB for pentode-mode OP tubes, and the 6AU6 certainly gives that.

I'm inclined to put as much voltage gain as possible in the first stage, for which I think the 6AU6 is a good candidate. I can then use a lower gain, more rugged tube (e.g. 6SN7) as the splitter/driver. I don't consider a 12AX7 (or 6SL7) to be a good choice here, as the splitter/driver, especially since the EL34s will be fixed-biased so their grid resistors need to be lower than in Mullard's cathode-biased design.

I'm not sure yet what the optimum operating point is, for linear operation of a pentode-mode 6AU6, as a common cathode AF voltage amp . Any ideas about that?

 

[Miles Prower]

I'm not sure yet what the optimum operating point is, for linear operation of a pentode-mode 6AU6, as a common cathode AF voltage amp . Any ideas about that?

Yep.

An externally hosted image should be here but it was not working when we last tested it.

Here's a loadline for pentode mode operation. Now, that 30K loadline (blue) looks pretty darn good. Of course, this is an estimate based on the second harmonic only. However, if it produces very little second, it ought to be good for higher order harmonics as well. (There is a graphical method to estimate higher order harmonics, but that's so involved that by the time you work that out, you could have built a prototype and o'scoped the design.)

However, we have another problem. That 30K line crosses the "knee" of the plate curves at -1.0Vdc, and not the zero volt line. If you can be certain that the input voltage will never rise higher than +/-1.5V, you're OK. If it doesn't, your operating point goes off the chart, and into a very non-linear, not to mention possible screen grid poofage, territory.

It would be nice if that -1.0Vdc grid line were the 0V grid line. So what you can do is transpose it by reducing the nominal screen voltage. If the grid lines are dropped by one volt, then what is now the plate current at Vgk= -3.0Vdc should go to Vgk= -4.0Vdc, so read off the characteristic the resulting plate current. Next, go to the forward transfer curves which plot plate current against screen voltage and find the curve where Vgk= -3.0Vdc, and the plate current is equal to the new value you want at that grid voltage. Then interpolate what the corresponding screen voltage should be. It's not necessary to be uber-accurate about this since VTs are rather more "forgiving" than are transistors of all sorts. Furthermore, these curves represent a non-existant "ideal" VT, and are the average results of measuring the actual characteristics of a certain sample set of production VTs. Here, I came up with a Vsgk= 116Vdc, and called the new screen voltage 120Vdc since that's close enough. Next, look up the grid bias at the new screen voltage that gives the design nominal Q-Point plate current. In this situation, that comes to -1.6Vdc instead of the old value of -2.5Vdc. You can also read off the new screen current, if you need to know that for selecting a series dropping resistor (definitely not recommended: use a voltage divider instead to stabilize the screen voltage and improve linearity). Reducing the nominal Vsgk by 30V will put the plate characteristic in a more favorable relationship to the design load.

It does not necessarily follow that this will give the best distortion behaviour, but it certainly gets onto the in-field, if not actually on home plate (Baseball reference there).

As with any other circuit: design, test, optimize before you build permanently.