Possibly a silly idea

[Nick]

I was just thinking. In the thread about Simons phono stage, I went to check if the c3g was in triode mode, or pentode. Thinking if it was in triode, the other choice was to use a CCS on the anode, then the cathode resistor doesn't need bipassing.

But its pentode, so I thoght that was a shame. Then I had another thought, what happens if you stane a voltage reference on top of a pentode. In theory, it has a zero load (stay with me), so its more linear. But of course there would be no outout voltage as there is no load. But what if you sensed the current in the pentode, then you would have your signal again.

So what would happen if you stodd a triode on the top of a pentode, and took the triodes grid to a voltage reference, so it works as a cathode follower, then the voltage on the pentode is fixed, and it sees a zero (or low anyway, the rc of the cathode follower). then use a load resistor abode the triode, and take the signal from that point.

I may be going daft, I will try and simulate this tonight, and post a drawing.

As I say, just a silly idea I had over lunch.

[Lee S]

Hi Nick.

Isn't this what Rankin did in his Constant Current Output Stage Direct coupled 45 amplifier?

Used the triode section of a 6BM8 as an amplifier and the pentode section as the CCS for the cathode of the 45?

...or am I barking up the wrong tree?

Cheers,

[Nick]

No, its a bit stranger than that.

Let me post the drawing, it might make more sense then.

[Nick]

This was the idea. I guess it depends on the llinearty of the triode,

[Andrew]

Hi Nick,

I'm not getting all of this, so I'll ask a silly question if that's OK?

What's the relationship between the triode's load and the load the pentode sees? For a phono like Simon's the load needs to be fixed as the pentode load and RIAA Z are matched.

cheers,

-- Andrew

[Nick]

Not sure, I need to work out some maths. As far as I am thinking, the voltage on the pentode anode should be kept constant by the cathode follower action of the triode. So as the voltage is constant, its seeing a low load resistance, so should be at its most linear. The job of the triode and its load is to convert the change in current to a voltage that can be taken as the output.

I am open to peple saying its stupid, for example, one thing I considered, the triode may add more distortion than is removed by keeping the pentode loads constant.

I suspect, if we were in the SS world, a current mirror would be used in top of the pentode.

[Andrew]

Hi Nick,

Have you tried this in Spice, seems a good way to explore its workings, it won't tell us the whole story but....then suppose it depends on how long it will take someone to knock this up on a bench, less time if you have a bench HT supply.

I have a D3a pentode model, if its any help?

cheers,

-- Andrew

[Nick]

I have a D3a pentode model, if its any help?

Yes please, that would be a help. I only have it in triode.

[Andrew]

Here ya go, courtesy of SB, I beleive.

Nope its not working tried it as a .inc, as a .txt and a .gif. I can mail it if you prefer.

cheers,

-- Andrew

[Nick]

Cheers, got the email. Hopefully, in future .inc can be uploaded now.

[Nick]

Oh well, it seems to work. With a 10k top load, and 62R cathode resistor, gain is 90, and the pentode is seeing a load of 120R. Output impedance is controlled by the top resistor so is 10k.

According to spice, at 63v RMS out its at -39db 2nd harmonic.

Actually, what do you think, maybe thats not a bad result, a similar circuit with just the pentode, and a 5k load, gives a gain of only 43, and a 2nd harmonic at -43dB but thats only for a output of 31v RMS.

The silly circuit gives 2nd at -45dB for 31v RMS out, so it is slightly lower distortion.

Its just the right sort of voltage swing to drive a 300b though, and a 10k impedance isn't that bad. I must add it to the list of things to drive the 300b with and see what it sounds like. The one 150v VR tube can provide the screen supply and the grid for the top triode.

[Andrew]

Hi Nick,

Did you get chance compare the gain at both loads and against a standard bypassed cathode, or even fixed grid bias? 5k6 would be about right for Simon's phono. So many questions, would you be OK with sharing the sim? Might be easier.

It also occured to me that if Simon prefers fixed grid bias then both the S&B and the Llundahls seem to be OK with a bit of DC on the secondary. I've seen fixed grid bias done using a AA cell with the S&B and the TX isolates the CD fgrom the cart. Doesn't your phono put a bit of DC on the secondary?

It had already occured to me that Simon's lack of gain might be down to the unbypassed cathode resistor, which is why I suggested an LED. To be honest, I hadn't realised a bypassed cathode would make such a difference on a pentode, and can't find much in the modern books on this, I'll have to try the old books for an equation.

cheers,

-- Andrew

[Nick]

I have mailed you the sim I have been using, let me know anythink you find. I suspect it might turn out to be similar to a d3a with a CCS on its anode, but with the advantage of lower input cap due to the pentode.

I found the lundahls do not like any DC on their secondary at all. Thats why I had to change the biasing from T's original.

Of the top of my head, I would have thought the cathode degeneration would be exactly the same for a pentode and triode.

[Andrew]

I copied your circuit deleted the upper triode and set the B+ to 250v, output taken from the usual spot. Changed the load to 5k6 for both circuits.

Right dead quick shufty, what am I doing wrong?

I can't see the extra gain compared to the straight D3A with diodes in the cathode. I seem to get twice the gain for two 1N4002 in the cathode using virtually identical operating points.

I quick look at the FFT and distortion/noise do seem, however, to be lower, presumably more linear operation.

Confused......

-- Andrew

[Nick]

Yes, you are right, 5.6k load gives very slightly more gain with te extra triode (both cases unbipassed cathode). But I was using 10k. Try that. I know in your case you want a 5k6 output impedance, but this was more a experiment instead of something directly aimed at a particular use. I just thought of it whilst thinking about the phono, but it wasn't intended as a suggestion for it, just someting I hadn't seen done. (sure its in the old book though).

The idea was you can increase the load, and get more gain, but without having much effect on the pentode load line.

In the case of the 10k load the gain is higher and the distorton seems to be a lot lower than the single pentode