[dave slagle]

A local guy wants to do a c3M pentode feeding a LCR and then use a triode c3M for further gain.

My thinking is use a bifilar 1:1 to load the pentode and then feed a 1500 ohm LCR. The 1500 ohm resistor could then be placed across the IT secondary and the source impedance would then be 1500 ohms. The goal of the IT is to get the coupling cap out of the picture.

I played with spice a bit using a d3a model from steve bench and everything seems to behave as expected.

I was surprised at the difference in gain as you adjusted the load. taking the LCR impedance from 1K5 to 7K5 gave an extra 15dB of gain out of the d3a. Using a 600R LCR lost another 7dB of gain.

Since realistically we are swinging much less than a volt, i have to wonder how high you could go with the load to maximize gain. Of course he higher load will require more inductance out of the IT and that will come back an bite you at some point.

any thouhts?


dave

[IslandPink]

Whacky Stuff !
On stage 2 , why a C3m for gain ? Surely a C3g or a D3a has more triode gain , more suitable for a phono ?
MJ

[Andrew]

Hi Dave,

C3g is a really nice valve but...

One reason why Thorsten choose the E810F to drive the 600R LCR was that it has such a high gm. A gm of 50, is considered, I beleive, the test point.

The books say you can approximate the gain from a pentode as the load multiplied by the gm. So 600R * 50 would give a gain of 30, as you loose ten through the RIAA, then you get a gain of 3 out the end of the RIAA.

When we worked on the 1k5 I had setteled on it because it was low enough to be considered a low Z but would allow 2.5 times the load of the 600R, thus widening the range of pentodes we can pick to lower gm candidates D3A, 12GM7 etc. Also it had nice cap values....

Obviously we have to load at 1k5 otherwise the RIAA is screwed, so this gives us 1.5 * gm. For an E810F the 1k5 would give us a gain of 7.5 after RIAA, the second stage would only need a gain of about 13.

A C3g would be very nice, very quite, but has a gm = 15, or so, 15 * 1k5 = 22 or 2.2 after the RIAA, near enough to 2, which means the second stage has to find 50 to get typical MM levels at 40db. I would say that is getting quite hard to do in one stage, the gain of 33 the 600R needed can be done with a D3a triode. Of course, many system have more gain that is needed, especially if the preamp is active, so this may not be a problem.

For a C3g I'd be tempted to try around 3-5k, I recently did a matched 5k6 RC-based RIAA for the 'medium' gm pentodes, 5k6 has some since standard cap values. The whole thing was started by James as he really likes the C3g, I choose 5k6 becuase it matches C3g very nicely and a friend of ours has decided to trust me enough to build this.

Now, the next question is....how much does low Z matter?

One last question, I can't visualize your circuit, sorry, but it might be worth mentioning that pentodes and inductive loads can give odd results. Can you post the schem?

I am finally closing in on the PSU design for the 1k5 but still waiting for PSU iron

cheers,

-- Andrew

[dave slagle]

hey guys,

thanks for the input. I didn't pick the c3m and it is not etched in stone. Look at it this way, if you like the concept of a pentode input for a phono, this one might get built on this side of the pond.

[Andrew]

Hi Dave,

Just done a double take...have you got a D3a pentode model then? I have a Steve's triode connected D3A.

Anyway, I lurve the idea of a pentode or or cascode for an input stage

There lots of reasons why this should work really well, even better than a triode, and with high gm pentode/cascode you can have low Z as well.

As you know I intend to buid #4, its only a 10-15nF cap, after all, and the good thing about that choice is that you can use the cap to put the next valve into grid leak bias thus avoiding the cathode resistor/cap combo on V2 and getting a lower Zout to drive the line into the bargain.

cheers,

-- Andrew

[nickg]

Well, I own a example of the first, big cap into 600R, I would be happy to try replacing with the third option Dave if you want a test subject.

[dave slagle]

hey andrew,

i'm glad you are going for it and i agree that on paper the whole thing looks pretty appealing.

below is the tetrode model from steve.

[dave slagle]

[Andrew]

I could very well be wrong but the gm on the model looks a bit low to me.

According to the data sheet, the gm of the D3a is 35 so 1k5 * 35 should give an approximate gain of 50, I get more like 20.

So, Dave, you may find things are better when you build than you expect, gainwise.

cheers,

-- Andrew

[IslandPink]

Andrew -
Depends where you're running the D3a and what the model assumes . You'll only get S=35 at max conditions remember .

Mark

[Andrew]

Hi Mark,

Being new to pentodes, I could have easily screwed up, I used the Vg2, Ig2, Ia and Va values straight off the data sheet this should according to my calcs roughly give me.

B+ = 220v
Ra = 1k5
Rg2 = 12k
Vg2 = 160v
I g2 = 5mA
Rk = 47R
Vk = 1.25v
Va = 190v
Ia = 21mA

These are my calc'ed values, the model has this running hotter, nore like 28mA-30mA

Any thoughts?

cheers,

-- Andrew

[James D]

Chaps,

Telefunken curves give Gm=33mA/V at Andews op point. It is on the flatter point of the curve so the value is probably correct to 10-15%.

James

[dave slagle]

hey guys.

I plotted the curves in spice at the tfk operating plate and screen voltages and then dropped the tfk curves over them in photoshop.

dave

[Andrew]

Hmmm mutter...mutter..mutter...back to the drawing board on the phono then, I'm bit short on the old gain front.....mutter....mutter...I suppose I could use the dreaded ECC83 in the Aikido, I was hoping to get away with a 6072A.

But then there's also a CCS or choke loaded D3A.....options....

[Andrew]

Hey Dave,

Just a thought...why not dc-couple the LCR and use an output TX to drive the load?

The downside, of course, is the second stage would need a large resistor and a fairly big cathode bypass cap.

If you used another valve as a CCS or solid state CCS you could avoid the cathode cap, but would either need to cap couple the output or run in para-feed.

Just throwing some more alternatives into your mix.

Homing in on the first pass of my LCR PSU now, might be done by summer

all the best,

-- Andrew