[Don[_4_]]

Here's link to the next iteration of the schematic:
http://mysite.verizon.net/vze11z1rm/

Zooming in on the page helps the legibility.

I cleaned up the OPT, and changed the bias to a "failsafe" mode.
I'm working on the frequency shaping shelf networks etc.
Please comment.
-Don

[Patrick Turner]

On Sep 7, 11:59*am, Don wrote:
Here's link to the next iteration of the schematic:http://mysite.verizon.net/vze11z1rm/

Zooming in on the page helps the legibility.

I cleaned up the OPT, and changed the bias to a "failsafe" mode.
I'm working on the frequency shaping shelf networks etc.
Please comment.
-Don

I can see the schematic OK from http://mysite.verizon.net/vze11z1rm/


What's the big problem that cannot be easily fixed?

But regardless of any stated bothers, here's a list of what I'd so
with your schematic...

1, use 0.22uF input cap plus R2 = 100k to give a passive HPF with pole
at 7.2Hz.

2. V1 A and B grids should each have a series 4k7 grid stopper,
because high gm triodes can sometime oscillate at RF when grids are
tied together, 6DJ8 sure can, maybe 12AT7. But when in doubt, take the
precaution.

3. Why not bypass R3 cathode R ro 12AT7? 220uF plus 0.47 is about
right. Then Ra of the 12AT7 will remain low, and not have early pole
with Miller C into 6SN7A.

4. With GNFB connected, compensation 3k3 plus C8 120pF need very
careful choice using a radio tuning gang and 25k series pot. Place
0.22uF across 8 ohm output with no other load, use 5khz square wave,
and maybe there are RF oscillations. Usually they can be stopped with
the right value R&C, but there may be some ringing on the square wave,
and some reduction is usually possible with the C7 across FB resistor
and the zobel R&C across the OPT sec.

Its utterly impossible to calculate the time LF and HF constants for
unconditional stability especially without a load, and most tube amps
oscillate badly unless extreme HF and LF gain is curtailed by stepping
networks which effectively push phase shifts away from the AF band
while reducing open loop gain at the extreme F.

5. Instead of C1 = 0.068uF with 1M bias Rg, LF in-stability is stopped
dead when 0.47uF is used off anode followed by 1M in parallel with
0.047, then 220k bias Rg. Bass remains excellent, and recovery after
overload gets better.

So, I'd say you have a lotta things to work out by careful trial and
measurement in order to achieve unconditional stability regardless of
load, including no load at all, or using any value of C across any of
the outlets, all while maintaining the widest possible full power
bandwidth into pure resistance. I do this sorta stuff every other
week, and know by experience. But those who don't do it for a living
build good oscillators which perplex them severely. But all you need
is as much inventive and intuitive attention to detail and science as
a humble sod like I have and if I can do it then so can you.

The process of trimming the amp after initial wiring to stop it
oscillating is called critical damping, and after doing several amps
the average tech "gets it" but many won't, they expect to calculate
and apply results, and that's never worked for me. I just do it. Just
don't leave the amp oscillating at full power at 100kHz for very long.
It invites smoke out, and you can't get it back in.

6. 6SN7A&B cathodes have CCS SS tail, but this goes to 0V so Ek may
only be +8Vdc, and that's all that is across the CCS. I'd have a rail
of say -20V at least to which the bottom of CCS man be taken.

7. The 0.039u + 39 ohms zobel across 16 ohm coil becomes a mainly
resistance load at above 104kHz.
I'd be using R = 16 ohms, and C maybe up to 0.27uF, but sometimes such
Zobels make matters worse, even though intuition leads you to believe
that the output zobel MUST work, yet it don't. Think perhaps of using
2 x Zobels = 3k3 x 10W plus 1nF or 2.2nF across each 1/2 primary of
the OPT. Excessive HF gain in OP tube can occur as leakage L which is
a finite amount of inductance between load and anode can cause OPT
tubes to have large ringing at HF square waves yet you see nothing
much at V0. Quality of OPT can determine the amount of global NFB.
The well damped tube amp should withstand testing with a square wave
and will usually show most ringing when slight C load iis used on
output, and at low level when open loop gain is highest in class A
operation. The error signals at KT88 grids and anodes should not be
excessively ringey, or distorted as the amp vainly tries to correct
its open loop shortcomings, at least until over 1/2 full PO, ie, -3dB
below clipping with NFB connected.

Most of my schematics at my wesbite have all sorts of compo methods to
enable the best sound from tubes. But what I have at my site is valid
only for the schematic and the OPT I have used, and what works best in
your amp will be different because you have a different OPT with
different parasitic L and C elements.
So be prepared to be frustrated by the inexplicable, the irrational,
but never give up asking questions.

Patrick Turner..

[Arny Krueger[_4_]]

"Don" wrote in message
...
Here's link to the next iteration of the schematic:
http://mysite.verizon.net/vze11z1rm/

Zooming in on the page helps the legibility.

I cleaned up the OPT, and changed the bias to a "failsafe" mode.
I'm working on the frequency shaping shelf networks etc.
Please comment.
-Don

My first question when I saw this chematic is why not connect the two
sections of the first dual triode in cascode?

http://www.freewebs.com/valvewizard/cascode.html

[John Byrns]

In article ,
"Arny Krueger" wrote:

"Don" wrote in message
...
Here's link to the next iteration of the schematic:
http://mysite.verizon.net/vze11z1rm/

Zooming in on the page helps the legibility.

I cleaned up the OPT, and changed the bias to a "failsafe" mode.
I'm working on the frequency shaping shelf networks etc.
Please comment.
-Don

My first question when I saw this chematic is why not connect the two
sections of the first dual triode in cascode?

http://www.freewebs.com/valvewizard/cascode.html

Perhaps the cascode wasn't used because it had more gain than was needed.

Or one could even go so far as using a pentode instead of the 12AT7.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Don Pearce[_3_]]

On Wed, 07 Sep 2011 13:33:42 -0500, John Byrns
wrote:

In article ,
"Arny Krueger" wrote:

"Don" wrote in message
...
Here's link to the next iteration of the schematic:
http://mysite.verizon.net/vze11z1rm/

Zooming in on the page helps the legibility.

I cleaned up the OPT, and changed the bias to a "failsafe" mode.
I'm working on the frequency shaping shelf networks etc.
Please comment.
-Don

My first question when I saw this chematic is why not connect the two
sections of the first dual triode in cascode?

http://www.freewebs.com/valvewizard/cascode.html

Perhaps the cascode wasn't used because it had more gain than was needed.

Or one could even go so far as using a pentode instead of the 12AT7.

Or make the input stage a Blumlein pair instead of simply parallelling
them. That would, I think, be a better use for a pair of triodes.

d