Thinking of building a low gain preamp

[Palustris]

The 6W6 is interesting for several applications. Here are the plate curves; I don't remember where I found them.

 

[BinaryMike]

Ah, yes. Thanks. That's probably from Vademecum section 216 at tubedata.org.

 

[nerdorama]

Frank's tube data has it: http://www.mif.pg.gda.pl/homepages/frank/sheets/093/6/6W6GT.pdf

It's my first stop for tube data sheets. I keep the link to his search page readily available: http://www.mif.pg.gda.pl/homepages/frank/vs.html

Anyone ever use, or consider, the Russian directly heated 4P1L? I have a handful. Triode connected they get a lot of good press either for small SE amps and in line stages. I'm thinking of using them as PP drivers of interstage transformers.
John

 

[thorpej]

Another random thought for the low impedance anode follower might be a 6AS7 or 6080. Its a dual direct heat triode meant as a voltage regulator. They're also relatively cheap. 280 ohms of plate resistance also makes it much lower than a 300B.

Spent some time playing with this today. Using a IXYS 10M45 as a plate load, set for ~20mA, it would be able to easily drive a solid state power amp (usually around 47K input impedance, IIRC). Simulates with a voltage gain of about 1.7 (4.6dB). Simulation shows that a 120Vac power transformer would be sufficient, and at this operating point, both the 10M45 and the 6AS7 dissipate less than 2W each.

 

[gadget73]

That would probably be sufficient. Only thing is that its phase inverting, though nothing says you couldn't put another stage in front to correct for that or add more gain if you really wanted it.

The tube ought to last just this side of forever under those conditions too.

 

[thorpej]

That would probably be sufficient. Only thing is that its phase inverting, though nothing says you couldn't put another stage in front to correct for that or add more gain if you really wanted it.

The tube ought to last just this side of forever under those conditions too.

Eh, I don't let myself get too worked up about phase inversion. So long as you invert every signal an equal number of times between the source and the speakers, it doesn't really matter (ok, maybe it might if you happen to have a multi-room system with overlapping "audible" zones... but then speaker orientation makes a difference, too!)

I figure the original phase is essentially random anyway, so ...

 

[gadget73]

yeah, its just something that occasionally comes up. One of my preamps inverts. At least one of my amps inverts too. I don't tend to use them with each other, not for any specific reason, but it just doesn't seem to be the combo I usually grab.

 

[thorpej]

Happy New Year, AK peeps! Since finally getting at least part of my garage workspace set up after a major home remodel, I've been trying to blast through projects in the hopper, including this one. I had some 2 channel CCS plate load boards made, and assembled one of them yesterday and got it all dialed in. I have a stash of 6080s to play with, and I hope to start experimenting with it soon -- waiting on 4.7uF film caps that I ordered yesterday (but will alas arrive while I'm out of town on a business trip).

(Sigh, my hand-drilled heat sink pin holes were not quite aligned correctly for the sink on the left, hence the dab of black liquid insulating material as insurance against the sink, which is not insulated from the 10M45, from shorting to the resistor solder pad... I'll have to be more careful next time.)

 

[thorpej]

So, I breadboarded the circuit from post #44, but with the slight difference that the CCS load is dialed in at 40mA. The results are spot on with simulation, with gain of 1.75. As you can see from the spectrum analysis peak-hold plots, distortion is very low ... practically in the noise floor, about 60dB down with 1V peak input. Noise is a bit more of a factor with lower input levels, although it's still not too bad (~50dB down with a 100mV peak input signal). This is with regulated DC heaters. Removing the cathode bypass capacitor doesn't seem to improve distortion or noise, and seems to lead to a bit of high-frequency roll-off.

I'll rewire the socket shortly to see how well matched the other triode in this particular 6080 is.

 

[thorpej]

Gain of the second triode in this bottle is a teensy bit higher... 1.88 (vs 1.75). It's not really that far off, but perhaps something that should be addressed with a couple of small trimpots to balance out the sides at build time, allowing them to be re-adjusted whenever the tube is replaced. Performance of the second triode seems otherwise equal.

 

[PakProtector]

The 12B4 is a very fine tube to use for this app. I would urge you to reconsider, just because it isn't in production any more. Cascode CCS with mu-follower output works *REALLY* well too. They deliver really low output impedance, and isolate the triode from the PS. A single L-C supply is more than quiet enough, even with just +critical inductance, and a few uF of capacitance. heat sinking the upper MOSFET is easy... A pair of 6W4 damper diodes will be *MORE* than adequate.

Best o' luck with your project.
cheers,
Douglas

 

[thorpej]

I may reconfigure this as a cascode as well, just to fiddle around. The disadvantage of using 2 6AS7s is that you need 5A of steady-state heater current (and nearly 7 for the cold-start surge).

 

[thorpej]

It occurs to me, after reviewing Millett's lo-mu preamp schematic, that you could get the desired gain from a 6AS7 by operating it as the input stage of a cascode. The CCS tubes in the lo-mu preamp could be converted to cascode upper stages with 430R anode loads, yielding 3X gain and 430R output Z if 6AS7 gm is close to the datasheet value of 7000. Another very nice way to accomplish this might be with PNP transistors in a folded cascode arrangement. The transistors are sonically transparent when used in this way.

The rated gm values for 6AS7 / 6080 seem to be all over the map. The Svetlana 6AS7 data sheet says 5500, the Philips 6080 data sheet says 6500, and the GE 6AS7GA data sheet says 7000.

 

[djnagle]

I really like what my Millet low mu did for my system. If you are thinking about it, the parts list lists a transformer with tthe filement section tthat is too small.

 

[kward]

Hmm djnagle, looks vaguely familiar...

 

[thorpej]

The results are spot on with simulation, with gain of 1.75. As you can see from the spectrum analysis peak-hold plots, distortion is very low ...

Had a few spare minutes today to pencil out a power supply for this. Because the voltages are so low, even very good electrolytics are cheap, so I splurged on a 2-choke filter; the Triode Electronics Dyanco replacement chokes are relatively inexpensive and physically small. It simulates flat-as-a-pancake at the final filter cap.