6sn7 operating points

[amperex]

I believe the Jeff Korneff preamps uses that configuration. I heard those preamps- very nice. I built a simple 6P5GT RC coupled preamp & it sounds very good. I would need the Korneff preamps next to this new DIY 6P5GT with an A/B test to test for best sonics. Anyways, very close to one another.

The 6P5GT I like are all round plate including Tung-Sol, RCA Cunningham & Sylvania in that order. They are all very close to each other in performance.

[PRR]

>> the 6B4G's need about 60v grid

> At that voltage, you'd probably want to use ...

I think Richard was getting the 60V from 6BX7 with transformer coupling, a fairly reasonable proposition.

Then his question is how to drive 6V into 6BX7 with 6SN7. At 6V signal and 250V supply, 6/250= 3%, we are not really near the 10%-20% range where we have to optimize. With tight layout and no NFB, 6SN7 with 220K and 90V at the plate will work for 6V to 20KHz.

For maximum output we want the plate load somewhere between the saturated Rp and the next-stage Rg. Say Rp=6K, Rg=1Meg, calculate 77K, use 47K. Bias the plate at 0.5 to 0.7 of B+, say 120V to 170V. Expect something like 20% of B+ as peak voltage: about 50V peak. As you point out, an R-C driver is not quite enough to slap a 2A3/6B4 working on the same supply rail (it is marginal with cathode-biased 2A3/6B4). That's probably why the transformer-coupled 6BX7 is in there: it could easily slap 160V peak from 250V rail.

For maximum gain from the 6SN7 stage, use a very high plate resistor, almost as big as the next-stage Rg. Say 470K. Then run as much current as possible while still making your output swing. That can sometimes be 50V or less on the plate. But at this impedance, stray capacitance droops the top of the audio band, and this low plate voltage may not be stable with age or tube-swap. And many classic triodes "bend" around 50V. 100K-220K and 70V-120V is a better bet. Gain will not be much less than the max-gain case (and gain was not demanded by Richard), things can drift all over and still leave 6V headroom.

It is sorta a case of: hard to go wrong. Dang near any bias point will work. I'd be more concerned about the parameters of the coupling transformer: 6BX7 is beefy, but a poor choice of transformer will strain it, and to get real beef in 6BX7 you need good DC current which strains the iron. That core and its windings becomes the pivot of the whole amp.

[AudioGeek]

i looked at the standard charts and drawing a load line from 260v... 30K load line (3*Rp)... looks pretty linear around 4v bias 150v plate and 6mA.

Remember i just need 6v out of it.

Hey, if it sounds bad i can always change it.... since RC coupling to next stage.

btw, should I put a cathode bypass cap on the 6SN7?

and finally, the 6BX7 will be driving a Electra-print phase splitting IT. gotta email jack and see what they are rated for. at least 25mA i believe... and thats what I am going to run.

Thanks!

[PRR]

> 30K load line (3*Rp)...

Gain will be low, and distortion not as low as it could be.

But you may already have all the gain you need, and the first stage's distortion is sure to be very small compared to the driver and output stages.

My main objection to 30K 6mA is that the first stage needs the best-filtered power. Given a choice between a 1mA stage and a 6mA stage, the 1mA stage needs 1/6th the power filter capacitance for the same buzz. True, the cost is not large either way, not compared to the overall amp. And you are not building for Profit (commercial designs often starved stages to shave pennies in power filtering).

> should I put a cathode bypass cap on the 6SN7?

Bah, don't think too much about small details. Big details that will derail the project if wrong, that's where you should do the hard thinking. But a cathode cap is a $0.25 problem: if thinking-time is money and you think about it for a few minutes, you are wasting money. Buy the cap, leave space, then try it both ways. The main difference will be ~5dB of gain: this alone (how it matches your signal source strength and listening levels) may be the deciding factor. If the stage were working harder, it would affect THD and bandwidth, but this stage works so E-Z that it hardly matters. (Obviously if you were going to make a million units, knowing in advance if you needed a million $0.15 caps might justify a LOT of thought.)

Yeah, if you decide you like the gain with a $0.25 cap, you might wonder if a $25 virgin-silver cap would sound better. But if you like it best without the cap, that question never comes up.

[AudioGeek]

supply is 5u4g to 15uf to 10H (100ohm) to 150uf to output tubes... input is tapped of at the 150uf to a 5K resistor to a 47uf cap.

output is PP and 120mA total. input runs at 26mA total.

ripple to input stage is less than 1mV

[AudioGeek]

i have 6 47/47 500v elna cerafine caps (before they went expensive)...

so I could always add an extra cap for filtering for the input stage as it needs it more than the output.

The output is not as important due to the PP topology.