Re: Long winded answer for SE fixed bias amp questions


"Phil S." <psymonds_no_spam@xxxxxxxxxxx> wrote in message
news:yOKdnZ1-6aNnZo3bnZ2dnUVZ_hSdnZ2d@xxxxxxxxxxxxxx

"J.P." <jpasano@xxxxxxxxxxx> wrote in message
news:1lr01356q1i4ksgt01tifsq71181mpp008@xxxxxxxxxx
On Sun, 1 Apr 2007 17:01:22 -0500, "Lord DSP" <Lord_DSP@xxxxxxxxxxx>
wrote:


"Phil S." <psymonds_no_spam@xxxxxxxxxxx> wrote in message
news:fpmdnTtgUsjLJJLbnZ2dnUVZ_hisnZ2d@xxxxxxxxxxxxxx
: "Mike Schway" <mschway@xxxxxxx> wrote in message
: news:mschway-9A960A.15354031032007@xxxxxxxxxxxxxxxxxxxxxxxxxxx
: > In article <rfqdnfYgdfi1RZPbnZ2dnUVZ_rKvnZ2d@xxxxxxxxxxx>,
: > "Phil S." <psymonds_no_spam@xxxxxxxxxxx> wrote:
: >
: >> I'm using the Vibro Champ front end, two 12X7's and that part is
built.
: >> For the back end, I'm building for a fixed bias 6AQ5. I've got what
I
: >> think
: >> is a proper layout for the bias supply and the power stage. Where
I''m
: >> getting cold feet is the proper way to link the two parts (bias
supply
: >> and
: >> preamp input).
: >>
: >> I've got output from the preamp 12AX7 plate into a 100K plate load
: >> resistor
: >> and into a .02 coupling cap. (We just ignore the tremolo circuit
here.)
: >>
: >> The other end of the coupling cap goes into the Grid of the power
tube?
: >>
: >> Also hanging from the power tube Grid is the 220K that conneccts to
the
: >> bias
: >> supply?
: >>
: >> Do I also need a grid leak resistor, one end to the Grid the other
end to
: >> Ground? Or is this only for the cathode biased circuit? It is on
this
: >> point that I'm unclear. Since there is no PI in a SE amp, I'm having
: >> some
: >> difficulty understanding what gets left out.
: >>
: >
: > A grid leak resistor in this application is not really a grid leak
: > resistor, think of it as part of the bias supply.
: >
: > It's basically equivalent to the ground side trim resistor on many
bias
: > pots At the right value, it enables you to adjust the correct bias
: > voltage with the pot. If it's too small a value, though, it'll
prevent
: > you from being able to dial in the correct level. (the cathode's
: > grounded, right?)
: >
: > You might not even need the 220K resistor (which used to be connected
to
: > ground)...just add in the bias feed to the downstream side of the
: > coupling cap before the 1K5 grid snubber resistor. In a push/pull amp
: > the pair of 220K bias feed resistors are there to isolate one
: > alternation from the other.
: >
: > Good luck.
: >
: > --Mike
:
: Mike,
: Thanks very much (you, too, Steve) for the information. I have a
lingering
: question, raised by your discussion above involving the 1K5 grid
snubber. I
: see amps with and without the grid snubber. I had not planned on using
one,
: but, as long as the topic is in play, would you mind elaborating a bit
on
: reasons for including it, values typically used, and opinions on whether
it
: affects tone, etc.?
:
: Not too technical here, just looking for some basics. Thanks.
: Phil


Before I start, I want to establish a few conventions for
the following dissertation
Whenever I refer to overdrive, I will be speaking
of *output tube clipping type overdrive* and not preamp
clipping type.

When I refer to any DC voltage values and
currents, they will be arbitrary values chosen for ease of
understanding and might well be a far cry from actual
operating values.

And when I refer to "coupling caps"
I will always be refering to the signal coupling caps that
couple the AC signal to the grids of the output tubes
When I say "primary side of the cap" I am refering to the
side of the coupling cap that is connected to the phase splitter.
When I refer to the "secondary side of the coupling cap,
I am refering to the side that is connected to the grids of
the output tubes



If overdriving the output hard enough to
to cause forward biasing of the output grids, then a
"snubber"resistor in series with the grid
will somewhat lessen the amount of electrons that will leak
into the coupling caps due to "grid rectification" effects.
Whatever electrons accumulate on the cap from this
rectification process are *additive* to the bias from
the bias supply and will cause biasing a effect that
increases the bias, This shifts the bias in the 'cold direction"
This "bias jacking" effect can be seen as crossover distortion
during clipping of the output section on some amps.

Ok, let's break it down a bit.
Picture it in terms of the coupling cap. On the preamp side,
(primary side) the DC voltage on the cap is sitting up at, lets
say, 200 volts. On the output tube grid side, (secondary side)
let's say -40 volts.

So, this gives us 240 volts across the cap at quiescent.

Now, let's hit it with a 50 volt AC signal that is heading
in the positive direction.
The grid side at -40 volts is (signalwise) going to follow the
other side at 200 volts. So, let's track what happens
at several points in time.

Initially before the signal starts to go positive,
the grid side of the cap is at it's quiescent value of -40 volts
At this point in time let's say that we have 4 ma worth
of electrons flowing out of the grid towards the negative
bias supply.

Now lets go to when signal has bumped up the primary side
side of the cap to 420 volts.
Ok, since the secondary side of the cap
simply follows the primary side, then our voltage
at the secondary will be -20 volts and we should have
*halved* the grid current to -2ma. Also our secondary voltage
should likewise go up to -20 volts
But wait a minute, that
didn't happen. Instead we have something like -30 volts
and -3ma
( a gross exageration BTW, for illustrative purposes. In
reality the actual values are so small as to possibly be
considered negligible)

So instead of the voltage across our coupling cap
staying at 240 volts, we just gained 10 volts and it's
now 250 volts !

So, why the non linear tracking ? Well, in essence,
because the impedance of the grid in this region
*isn't linear.* Doh!
It is in fact, a negative impedance. IOW, as the signal is
going up, the input impedance of the tube is going down.

So already we have a bit of a shift in the average DC
voltage across the cap, but it gets worse, way worse.
When the signal gets high enough so that the grids approach
zero volts, they shift to being forward biased and then really
start to source electrons big time. IOW, the grid impedance
goes way, way down at this point.
Of course impedance is the inverse of conductance,
so perhaps a better war of stating it would be to say that the
grid *conductance* goes up. So if the grid conductance goes
up and it starts to conduct more, where does it conduct from.
Yup, the negative bias supply, but also the *coupling cap*.

So now, we're actually starting to *source* electrons out
of the grid into the coupling cap, which really makes the
secondary side of the coupling cap even more negative
(with respect to its primary side).
So we just that we picked
up another 20 volts on our coupler and now measure
an average DC value of 270 volts across it instead
of it's original quiescent (no signal) value of 200 volts.

Ok, all fine and dandy until the signal starts to come
back down. Remember that our tube was initially biased at
-40 volts. So as our signal starts to come down,
let's say the primary side of the cap comes down to
230 volts. Subtracting the 270 volts across the cap
from 230 volts gives us -40 volts yikes! We're already
hitting the quiescent bias value on the grid of the output
tube and our signal has another 30 volts to go before
before it even reaches zero
( which equates to 200 volts on the coupling cap primary)

So when the signal does get to zero and the coupling
cap primary voltage likewise gets back down to 200 volts,
the grid of our output tube will be down at 200 volts minus
270, volts, which comes to -70 volts !!!
( when it should be only at the quiescent value of -40 volts)

So as you can see, the output tube is going to get cut off
way before it's time. Which means that since it's now not
even operating for a complete half cycle, it's been jacked
into class "C" operation
You can see this effect on some amps as crossover
distortion when the outputs clip.

In fact, the reason I know this is just because of that.
After seeing this crossover distortion during output
clipping a few times, I just had to know what the cause
was. I tracked it down to the the output section coupling
caps and the aforementioned reasons.
In fact, when I experimented with direct coupling the
phase splitter with transistors instead of capacitors,
the crossover distortion dissappeared.

Ok, 1.5K will do little to suppress this. 10K and up starts
to suppress it a bit more. But then there's the question of
whether one even wants to suppress it at all.

Here's my take on it. This kind of crossover distortion
is pretty much masked (soundwise) by the fact that the
amp at this point is already kicking out it's full volume.

However it does influence the tone of the overdrive.
It tends to make it a bit softer and and creamier and bit more
harmonically rich sounding.
The downside is that it makes the bass at
overdrive a little less solid and a bit flabby sounding by
comparison.

By comparison? with what? Well by comparison with
what the amp sounds like if you totally eliminate this
de-biasing (load line shifting) effect. When I direct coupled
the output section to the phase splitter with transistors instead
of capacitors, the amp became very tight sounding on the low
end when overdriven. Someone like Dick Dale would
probably go nutzoid over an amp like that. Whereas
someone like Carlos Santana might prefer cap coupling.

Hmm, was going to go on about grid blocking and Miller
effects, but I'll have to get back later on that as I'm late
to be somewhere. Gotta jet

Good example...J.P.

Sorry to disagree, but this guy is a jack ass. I wouldn't even see the post
except that it is quoted in the reply, as he is in my kill file. Which part
of "not too technical" did he not comprehend and it's off the topic, to
boot? What's this horseshit about "...output section to the phase splitter
with transistors..." ROTFLMAO!

I've already got the board built. It's a neat layout with power supply,
bias supply and preamp all on the board, each in it's own section. I've
still got to find an inexpensive chassis for it, but that's the easy part.

Phil, Lord DSP is my friend and colleague Dave Moore - a brilliant engineer
and what he wrote is perfectly revelant to guitar amps in general whether
he directly addressed your concerns or not.

Dave owns and likes and plays and has an appreciation for tube amps.
He thinks SS amps are simply poorly engineered for the musician.
He also thinks any type of electronic device, even digital is just more
paint on the pallette to use for music.

However I think part of his reaction to many in this group is due to their
prejudiced tube snobbery. But tubes are made of sand also - well the glass is.
Dave happens to like the advances in technology even though one so-called
`advancement` comes in the form of miniaturation and as a repair guy,
I hate that crap, but Dave gets around it by soldering SMD ICs to regular
sized DIP sockets for his experiements.

I'm sure Dave is aware that most mixing boards are now all full of SMDs
and are basically disposable which is another gripe of mine,
so naturally I'd like to see the continuance of tube technology for guitar
amps as well as the size of required high watt components to be able
to use in hand-wired circuits for ANY electronic device designed for making
music. After all, ANYONE can swap computer cards. In fact, if you buy any
of those newly manufactured 1/2 watt carbon comps (which perform just as
well as any NOS) to use in your projects, have you noticed that their leads
aren't as long as original Allen Bradleys or any other NOS ones and that's
another beef of mine.

Ok, I like what we call the `pure` tube tone without silicon devices, but many
good sounding tubes amps have silicon rectifiers in them. Should we convert
them all to use tube rectifiers including a tube to replace the rectifier diode in
the bias supply circuit?

The real issue is tone, ain't it? Great tone if we can get it, right? But tone is
subjective and art is anything you can get away with, right? So what the hell
is really wrong with sand?

.

Relevant Pages

  • Re: Long winded answer for SE fixed bias amp questions
    ... :>> The other end of the coupling cap goes into the Grid of the power tube? ... let's say -40 volts. ...
    (alt.guitar.amps)
  • Re: Long winded answer for SE fixed bias amp questions
    ... :>> The other end of the coupling cap goes into the Grid of the power tube? ... let's say -40 volts. ...
    (alt.guitar.amps)
  • Re: Long winded answer for SE fixed bias amp questions
    ... :>> Do I also need a grid leak resistor, one end to the Grid the other ... of *output tube clipping type overdrive* and not preamp ... When I say "primary side of the cap" I am refering to the ... let's say -40 volts. ...
    (alt.guitar.amps)
  • Re: Long winded answer for SE fixed bias amp questions
    ... :>> Do I also need a grid leak resistor, one end to the Grid the other ... of *output tube clipping type overdrive* and not preamp ... When I say "primary side of the cap" I am refering to the ... let's say -40 volts. ...
    (alt.guitar.amps)
  • Re: Regen Question
    ... plate of the tetrode that of a cascaded second stage), or grid voltage ... run at best gain and the amount of feedback is controlled via pot or throttle cap. ... 12BE6 tube (I actually held one end of a .01 cap and touched the OTHER ...
    (rec.radio.amateur.homebrew)