why pentode has a sweeter sound than the beam tetrode?
Alfred
Can anyone tell me
1)why the pentode has a sweeter sound than the beam tetrode?
2)what is the reason for European EL34 g3 not to internally wired with
K(cathode)?can I assume that
g3 can also be used at higher or lower voltage
than the cathode if I don't wired the g3 & K(which is pin 1&pin 8)together
externally?
I mean g3 & K must they work at same voltage level?
3)and how can a power tube hold it's biasing tightly?
Thanks for anyone who can answer my questions
Alfred
Ian Iveson
1. History. I don't think this is always true, but in circuits with
intentional harmonics like guitar amps, they do have that reputation, if by
sweet you mean second and third harmonics. Some see the beam tetrode as a
compromise to get round contemporary patents, others as a triumphant
advance. I'm European so I'm kinda biased.
2. That would be the 6CA7? The EL34 has an internal connection I believe.
All proper EL34s are European I think. China might not agree. I believe
that a Musicman design used 6CA7s driven from their cathodes by transistors,
cascode style. The grids are grounded and I think g3 is grounded too,
bypassing the transistor. Apart from that I have never seen k and g3 not
connected.
3. Why not? Once a decent valve has bedded into its groove, it rarely needs
adjustment for fixed bias. With cathode bias it will always be a bit
elastic, unless you use a perfect constant current source. But absolutely
tight is not the same as tightly bound to the changing characteristic of the
valve. The perfect system would adjust the valve to its best operating
point over the life of the valve. Now there's a challenge.
There are some gurus kicking around here. But will they bite?
regards, Ian
"Alfred" <mkg...@hotmail.com> wrote in message
news:9763dg$2oj$1...@coco.singnet.com.sg...
Punkerdubh
- My data book says that the EL34 and the 6CA7 use the same
basing (8EP).
- I've seen schematics for 6CA7 based Traynor guitar amps that
tie g3 to the bias supply (about -50V):
http://www.lynx.bc.ca/~jc/721005_YBA3.gif
I'm told that this increases reliability because the suppressor
suppresses better in this configuration. I don't know if it's true,
but that's the folklore that I heard.
- After the first 50 hours of use, tubes are pretty stable but if they
get a good hard physical shock that mucks about with the internal
geometry the parameters can shift a bit.
- Punkerdubh (not a guru either)
Sheldon D. Stokes
<ianives...@virgin.net> wrote:
>But absolutely
>tight is not the same as tightly bound to the changing characteristic of the
>valve. The perfect system would adjust the valve to its best operating
>point over the life of the valve. Now there's a challenge.
>
>There are some gurus kicking around here. But will they bite?
>
Not that I'm a guru, but:
IF you refine your goal slightly to: "The perfect system would hold
the valve's operating point fixed regardless of it's aging
characteristics" IF the operating point is selected as the best
point, then that's not too hard a task.
Imagine a servo loop with a very long time constant, say 10 seconds or
so. The input of the servo loop is the voltage across a small cathode
resistor, and the output adjusts the grid voltage. This can be easily
accomplished using a single dual op-amp and a small handful of passive
parts.
The advantages are: Perfect balance between output tubes, consistant
bias through the life of the tube. And adding a comparitor circuit
and a couple LED's could allert you that the servo had hit a limit; IE
your power tube is old and tired.
Toroid output transformers for tube amps are a really good idea, given
their wide bandwidth and tight coupling between primary and secondary.
But unless you gap them (which hoses the tight coupling), they are
very intolerant of DC on the core. A simple servo loop like this
could remove all but traces of a DC offset on the core.
Sheldon
Ian Iveson
can use a constant current source at the cathode, but that has repercussions
because it must be bypassed by a cap, and the combination is likely to fall
short of the goal of infinite DC resistance and zero AC resistance at all
frequencies.
But neither of these schemes allows for changes in characteristics, beyond
maintaining the same DC current. My contention is that, if the
characteristics change enough to require a change of bias voltage, then the
ideal operating point may move, and not necessarily in the exact direction
that the servo or CCS takes it by way of compensation. Obviously an
automatic bias control saves the valve from running away, and keeps it
fairly close to optimum, but it is something of a stab in the dark.
Thinking about it now, I suppose that the only parameter remaining, after
adjusting the cathode current and the grid voltage, is Va. As the
characteristic curves flatten over time, perhaps a compensating increase in
Va would be in order?
I suspect this is an academic point compared to all the other ageing
effects, really.
regards, Ian
"Sheldon D. Stokes" <sto...@spinn.net> wrote in message
news:1mtd9tgpefqhu29sg...@4ax.com...
Patrick Turner
Your message of at least twopence worth should surely be raised to two
bob , no ?
I am one who backslided into calling pentodes pentoads . If you look in
the Radiotron Designer's Handbook you will see
the attrocius THD analysis for a 6F6 when driven hard single ended . My
own THD measurements agrees with the "bible"
every time I've measured a beam tetrode or a pentode . I've never been
able to distinguish which is best , or worse ,
when how does it sound is all the concern .
Last week I had to raise 3 valve radios from the dead , the oldest was
an Australian Tasma from about 1935 with an all
pentode circuit . One had an EL33 which is a pentode which is very
similar in characteristics to an EL84 .
The other had a 6V6 which is a little beam tetrode of similar power to
the EL84 or EL33 . These valves were used for
many years in countless radios and TV sound amps , nearly always with no
NFB and a very sensitive speaker that measured
poorly . Despite an attrocius damping factor and no bass and a bit of
hum these old bangers are remarkably
easy to listen to .
We get a Goon Show episode every saturday here and I reckon its as good
as when I heard it in the
fifties . My own kitchen radio is an AM thing I dreamed up 3 years ago
with a better reflex bass and dome tweeter .
It has an EL34 strapped as a triode , driven by a 12AX7 , some NFB , and
the radio gets everything they broadcast as it's
got 10 Kc bandwidth . Now I don't really mind the simpler older pentode
/ tetrode concoctions but I know my kitchen radio IS JUST SO MUCH BETTER
.
As for your concerns about G3 , the suppressor grid , let me suggest you
read the RDH about valve function . In many valves
G3 has its own pin because designers might want to use G3 as a control
grid , especially in oscillators , in FM quadrature demodulators , etc ,
etc , etc .
Well Ian , did I bite ?
From Patrick Turner
website http://www.turneraudio.com.au
Ian Iveson
had been incorporated in the internal connections, like the EL84. Muxed ip,
obviously. Sorry. I could have looked at the EL34 amp just yards away from
me, so there is no excuse!
What brought on my confusion is that I have never seen the EL34 connected in
any way other than g3>k, unlike the 6CA7, where we now have two examples,
Traynor and Musicman, that use it connected differently.
So why have EL34s never been used like the 6CA7? Perhaps that valve had a
quirk that was overcome by the EL34?
cheers, Ian
"Punkerdubh" <punk...@newsguy.com> wrote in message
news:976pp...@enews2.newsguy.com...
Sheldon D. Stokes
>But neither of these schemes allows for changes in characteristics, beyond
>maintaining the same DC current. My contention is that, if the
>characteristics change enough to require a change of bias voltage, then the
>ideal operating point may move, and not necessarily in the exact direction
>that the servo or CCS takes it by way of compensation. Obviously an
>automatic bias control saves the valve from running away, and keeps it
>fairly close to optimum, but it is something of a stab in the dark.
That's true, but a change in the tube curves (not just a shift),
strikes me as only being possible with a change in
cathode/grid(s)/plate physical geometry. I could be wrong, but if I'm
right and the tube geometry is changing, I'd say you've got bigger
problems brewing than slightly sub-optimized bias settings.
The condition of the cathode would cause the curves to shift around,
but I'm not sure I see where a change in curves could happen that
wouldn't be compensated for by a bias voltage adjustment. It's very
possible that I'm missing something obvious here.
Sheldon
Ian Iveson
a beautiful Chinese plotter by Dieter Wolbart, who visits the group
occasionally.
Comparing a NOS Telefunken with a well used Mullard (a double-getter with
distorted anode from much overheating for years), there is no evidence of
degradation apart from a fuzziness on the part of the Mullard, due no doubt
to the shift in internal geometry. Looking at the construction, it is
difficult to see how the geometry can change in a way that coherently
changes the characteristics, because it is a set of concentric cylinders,
more or less. So for example a shift of the anode would result in the same
mean distance to the cathode. Now you know better than me but I would
assume that this distance is not linearly related to any of the parameters
of its characteristics, so they will change in a complex way due to the
eccentric distribution of distance between electrodes. The buckling of the
Mullard's anode seems to lead to a double trace on the plot.
What surprises me is that the plots of the two valves are nearly identical
through the useful range. The Mullard's pentode plot is squashed down at
the very top, so it saturates earlier. Unfortunately I can't see the scales
on the plots, and as far as I know the plotter may adjust to fit the screen.
But I assume it would do this in a linear way so it does not detract from my
point.
So based on little evidence this is my theory so far. Imagine (!) that the
grid lines on a pentode Va/Ia plot are leaf springs attached to a pivot at
the origin by means of a clamp that allows them to slide against each other.
Now embed the whole assembly in jelly. If you are still with me, take the
flat of your hand and push the top spring at 90 degrees to its direction.
So the lines squash down, fan-style almost, but with the upper lines
squashing quicker than those nearer the bottom.
If this is how they go, then saturation occurs earlier as the valve gets
older, and the non-linear region leading up to that point occurs at
decreasing current levels.
Now take either the servo bias or the constant current source at the cathode
method of autobias. They both adjust the operating point by maintaining a
constant DC current, and in both cases the effective grid voltage is
adjusted to suit. Now it seems to me that this is the very opposite of a
sensible response. This DC nfb means that the valve digs its heels in as
the non-linear region encroaches from the top left of the plot.
What would be better would be to slide the operating point down and to the
right. Now this is in a similar direction to where the load line would take
it if the bias was fixed, as far as I know, except I imagine the best
direction would be more directly downwards, current-wise.
This would require positive dc feedback and a positive adjustment to Va.
Doing that without changing the voltage across other valves in the circuit
is not a trivial problem it seems.
The downside of course would be rapid meltdown in some circumstances,
possibly when first attempting to set the bias!
The upside would be that the power, voltage or current output would drop to
match the linear capability of the valve as it ages.
After all this I guess that it would not be worth the hassle. But it is
more possible now than it was, because of the opportunity for an intelligent
servo and regulated supply, that could detect unsafe conditions.
That would be an interesting discussion topic I fancy: the scope for
intelligent ancillary circuits. I wonder what boards are available,
Sheldon? Microprocessor control would be a neat selling position too.
regards, Ian
"Sheldon D. Stokes" <sto...@spinn.net> wrote in message
news:npfg9t4nsq9911vkg...@4ax.com...
Alan Douglas
Ian Iveson wrote:
>What surprises me is that the plots of the two valves are nearly identical
>through the useful range. The Mullard's pentode plot is squashed down at
>the very top, so it saturates earlier.
Probably some parts of the cathode are running out of emission.
> Unfortunately I can't see the scales
>on the plots, and as far as I know the plotter may adjust to fit the screen.
>But I assume it would do this in a linear way so it does not detract from my
>point.
It doesn't adjust: it's a Chinese copy <g> of a Tek 570.
73, Alan