diytube.com

I'm putting my current monster amps on hold for a while since the tubes alone are going to cost more than a college student's car:)

I have an octet of 6CB5As, a NOS sleeve of 6GF7A tubes, and some various 9 pins.

I can order 1.7K @ 100W Edcor output transformers, and I basically have the rest.

Does this look OK?



Thanks!

Blair

150Vdc on the screens, why?

ViynlSavor has an SE 6CB5A in triode mode @~425Vdc.

http://vinylsavor.blogspot.com/2012/06/ ... rcuit.html

I saw his blog. He says roughly ~20-25W class A in PP. I'm more interested in maximum output power. I wonder how many watts you can squeeze out of a pair in triode at 350V B+.

I've seen other references to this tube being used with between 100-150V on the screens, so that "why". That the only rhyme and reason:)

I'd even go with a pair per channel if it gave me ~ 50W in pentode.


Thanks!

The 6CB5A is a 23W tube. I suspect it would do a little less than an EL34 - about 12W for a pair of them strapped as triodes.

Right,

Which is why I elected pentode mode. I see on various data sheet 23W and 26W. Either way, George "Tubelab" has pushed a pair in screen drive to over 100W per pair without red plating them, so I suspect they will do a little more than their advertised plate rating because they were designed for 100% duty cycle.

As I said, I'd like to see ~50W out of them per pair if possible. It's an amp I want to build to hold me over a bit. I'm going to try to build it into a full chassis with no tubes exposed for little hands.

Thanks!

Blair

Will a 6GF7A have enough voltage swing to push a pair of 6CB5A tubes to full output? With other tubes, I just reference the data sheets to see what I need and go 20-30% higher than that.

How do you determine drive voltage needs on a sweep tube?

Thanks!

Blair

Does this chassis layout look OK? The 60W Edcors will fit in the upper right.

I'd like to keep circuitry as simple as possible which is why I want to use the 6GF7A as a single tube driver. The only concern is the ability of the 6GF7A to drive the 6CB5s to full output.



Thanks!

Blair

When working with sweeps, is it all trial and error, or is there a way to calculate the voltage swing needed to get full output? The 6GF7A I am proposing to use has a 64 gain section and a 5.4 gain section. I am thinking about using it like a 7247 driver.

Think I need an additional gain stage up front like in my drawing?

Blair

I never learned transconductance, they didn't teach it at ET 'A' school.

But the 6CB5 has a transconductance of 8800umohs.
The 6L6 has a transconductance of 5300umohs.
And a 7591 has a transconductance of 10,200umohs.
So the 6CB5 has a gain about inbetween a 6L6 and a 7591.

Of course the second section with a gain of 5.4 will be in a cathodyne circuit, so it won't have any gain.
With NFB, the first section's gain might be more like 10 to 20, so total gain will be 10 to 20.
If you do a LTP instead of cathodyne, and use 2 6CB5s you can get plenty of gain; 40 x 10 x 2 = lots. Leave off the cathode bypass capacitor on the first stage and you should be golden. Something like 25 X 10 X 2 = 500 - losses here and there.
A 6CG7 will work well to. It'll give you a gain of 10 X 10 to 20 = total gain of 100 to 200.

Make sure the transformers are on the bottom, so if it tips no one will get squiched.

Thanks!

All transformers will be inside the case like shown with the power transformer. The top slides off.

I think adding a 12AX7 in the first stage will give me plenty of overhead. Is the feedback point acceptable even if I add the 12AX7 up front?

Thanks!

Blair

A 12AX7 with local feedback would be good. You can lower the gain and even adjust it, to a point, with R41.
Then leave global feedback on the 6CB5.

Is the "local" feedback from the 680K back into the grid? I'm not familiar with "local" feedback.

Then just leave the global feedback around the rest of the circuit like shown?

Thanks!

Blair

Local feedback is negative feedback usually around one tube. Global is usually the entire circuit, but it has to be inverting to be negative. The 680K and 68K make a 10 to 1 or so voltage divider, so a portion of the anode signal comes back, but inverting, so it's negative and really kills the gain of the 12AX7.

Triodes usually don't need feedback because they are linear. The 12AX7 isn't very linear.

An unbypassed cathode resistor on a 12AX7 is a type of local feedback, but isn't much. If you put the signal from the anode back into the grid, you can have a lot of negative feedback and fix the nonlinearily of that tube. I'd rather use a linear tube to start with: 6N1P, 6AQ7, 6SL7, 6SN7, 6FQ7, 6CG7, 12BH7, 6DJ8, 6BZ7, 5751.

SRPP would be good for the first stage. Ready Ed's post #14 viewtopic.php?f=4&t=5951

I've got a few 6CG7s. I was just looking for a way to use the 6GF7As for an obscure tube compliment.

Think using three 6CG7 triode sections per channel in a similar topology to my schematic would be better?

As DeathRex noted:

6CB5 transconductance = 8800 (some datasheets have 5000-6000) micromhos
6L6 transconductance = 5300 micromhos.
7591 transconductance = 10,200 micromhos.

However, voltage gain (the amplification factor of a tube) is the mu:

6CB5 = 40
6L6 = 8
7591 = 17

On paper, the 6CB5 is easier to drive than the 6L6 and 7591, but Vinylsavor and Tubelab George do not not mention it. The mu of the 6CB5 is close to the zero-bias 838 tube.

http://tubedata.tigahost.com/tubedata/s ... 6/6BC5.pdf
http://tubedata.tigahost.com/tubedata/s ... /6/6L6.pdf
http://tubedata.tigahost.com/tubedata/s ... 7/7591.pdf