Why "current starved" GC gain/input stages?

[Tom Schlangen]

Dear RATs,

some weeks ago I started a thread Re: Error in schematic?
or something like that.

Some answers had been very interesting, some had been below
the belt ... anyway, as I wrote I duplicated/breadboarded
the circuit:

http://www.ndh.net/home/schlangen/roehren/el34set/bench02.jpg

since the author didn't/couldn't respond to my
questions due to an overflowing mailbox above quota.

Meantimes, I did a lot of measuring and calculating on that
circuit, and yes, there is at least one error in the schematic,
but that isn't the topic of this article and left to the
reader to find out himself, if interested to do so ;-)

Also, I did a lot of reading RDH4 and calculating thru quite
some grounded cathode combined gain/driver stages to be found
on the web in quite similar GC circuits to the
one I breadboarded. I duplicated lots of them using
6SL7 or 12AX7 for roughly comparable gain, too, and
calculated and measured and compared thru them.

There is something I am wondering about, and I couldn't
work out the reason for so far or find it in RDH4:

Quite some GC gain/driver stages, mostly newer designs, use
very low anode current ratings in actual designs; e.g 6SL7
(one system) below 1mA, also for ECC83/12AX7, below 1mA
per system, too.

Looking at datasheets, these low current operation points
are often way below suggested values for these tubes, and
datasheet nomograms show less-than-ideal curves for such
low anode currents: These curves sometimes warp like
snakes (ahem) in low current ranges.

I can't measure distortions (I don't have the measuring
gear to do so), but the result of such low current
operation points should result in quite higher
distortion values than possible/optimal, am I right?

I suspect one reason being that many "newer" designs go
for as-high-as-possible gain values in such stages, with
very high value anode resistors to get most gain (thus
resulting in low plate current), and the authors don't
thinking about negative side effects, like higher distortion
due to nonlinearities.

Is there a reason for excessive low plate currents
in GC driver stages I overlooked?

Tom

--
Okay, maybe i am paranoid. But that doesn't mean
they are not out to get me. - unknown

[Choky]

"Tom Schlangen" > wrote in message
...
..
>
> Is there a reason for excessive low plate currents
> in GC driver stages I overlooked?
>
> Tom
>
> --

WTF is GC?
besides-main reason is lack of proper knowledge;
seems that these days everybody knows how to make best tube gadget.....
just think with your own head,and do not take any stage as good,even if is
made by "respected" maker.
hehe-I like to repeat long ago hear4ed lesson- "driver must dissipate at
least 1/10 of output stage dissipation".
off course that it is generalization (hehe,even rude one) but it is also
interesting one (old good food for thoughts ;).
I'm sure that Pat or some other clever rodent will giv'ya relation between
rise time,slew rate,upper F3,and other stuff (I don't remember that,just
because I always choose mucho current through my toobs-so I don't need rule
about minimum needed current),so ya'll see what is
important.

--
Choky
Prodanovic Aleksandar
YU

[Patrick Turner]

Robert Casey wrote:

> Patrick Turner wrote:
>
> >I normally run 1/2 a 12AX7 with 0.6 mA of plate current at idle.
> >I find the tube has lowest noise at that current, ie lowest flicker
> >noise
> >and grid input noise, and an example is at
> >http://www.turneraudio.com.au/htmlwebpgs02/schem10tubepreamp.htm
> >
> >6SL7 tubes also run at only a mA quite well.
> >
> >The loads for such tubes is usually 200k, so a 0.5 mA rrms current
> >change gives
> >100 vrms output, if the supply voltage and Ea is high enough.
> >No worries, and its done like this all the time in old gear.
> >
> >
> >
> >>
> >>
> This was also done in tube AM radios using the classic "AA5" design.
> The audio
> driver triode is the 12AV6, which is the same as half a 12AX7. The
> plate load
> is usually 470K and the B+ around 95V. They run at around 0.1ma plate
> current.
> See http://www.geocities.com/wa2ise/aa5diaa.jpg

Notice the way the 6AV6 triode is baised via a 4.7 Mohm resistor.
The tiny DC flow in 4.7M creates enough bias voltage to bais the tube.
This was very common practice for this sort of tube, because the cathode
MUST
be grounded, because of the detector diodes, and you can't have a normally
bypassed
Rk.

But its real bean counter stuff.

Works OK though.

Patrick Turner

[Fred Nachbaur]

Patrick Turner wrote:
>
> Robert Casey wrote:
>
>
>>Patrick Turner wrote:
>>
>>
>>>I normally run 1/2 a 12AX7 with 0.6 mA of plate current at idle.
>>>I find the tube has lowest noise at that current, ie lowest flicker
>>>noise
>>>and grid input noise, and an example is at
>>>http://www.turneraudio.com.au/htmlwebpgs02/schem10tubepreamp.htm
>>>
>>>6SL7 tubes also run at only a mA quite well.
>>>
>>>The loads for such tubes is usually 200k, so a 0.5 mA rrms current
>>>change gives
>>>100 vrms output, if the supply voltage and Ea is high enough.
>>>No worries, and its done like this all the time in old gear.
>>>
>>>
>>>
>>>
>>>>
>>This was also done in tube AM radios using the classic "AA5" design.
>> The audio
>>driver triode is the 12AV6, which is the same as half a 12AX7. The
>>plate load
>>is usually 470K and the B+ around 95V. They run at around 0.1ma plate
>>current.
>>See http://www.geocities.com/wa2ise/aa5diaa.jpg
>
>
> Notice the way the 6AV6 triode is baised via a 4.7 Mohm resistor.
> The tiny DC flow in 4.7M creates enough bias voltage to bais the tube.
> This was very common practice for this sort of tube, because the cathode
> MUST
> be grounded, because of the detector diodes, and you can't have a normally
> bypassed
> Rk.
>
> But its real bean counter stuff.
>
> Works OK though.
>
> Patrick Turner

What's not obvious is that there is inherent DC negative feedback with
this biasing scheme, similar to using cathode resistor bias.

Let's say that as the tube ages, the plate current decreases. This will
mean less electrons are captured by the grid, which means that the grid
voltage becomes less negative, tending to increase the plate current.

As you say, it works just fine. But I've never seen it done in "hi-fi" -
wonder why?

Cheers,
Fred
--
+--------------------------------------------+
| Music: http://www3.telus.net/dogstarmusic/ |
| Projects: http://dogstar.dantimax.dk |
+--------------------------------------------+

[Patrick Turner]

Fred Nachbaur wrote:

> Patrick Turner wrote:
> >
> > Robert Casey wrote:
> >
> >
> >>Patrick Turner wrote:
> >>
> >>
> >>>I normally run 1/2 a 12AX7 with 0.6 mA of plate current at idle.
> >>>I find the tube has lowest noise at that current, ie lowest flicker
> >>>noise
> >>>and grid input noise, and an example is at
> >>>http://www.turneraudio.com.au/htmlwebpgs02/schem10tubepreamp.htm
> >>>
> >>>6SL7 tubes also run at only a mA quite well.
> >>>
> >>>The loads for such tubes is usually 200k, so a 0.5 mA rrms current
> >>>change gives
> >>>100 vrms output, if the supply voltage and Ea is high enough.
> >>>No worries, and its done like this all the time in old gear.
> >>>
> >>>
> >>>
> >>>
> >>>>
> >>This was also done in tube AM radios using the classic "AA5" design.
> >> The audio
> >>driver triode is the 12AV6, which is the same as half a 12AX7. The
> >>plate load
> >>is usually 470K and the B+ around 95V. They run at around 0.1ma plate
> >>current.
> >>See http://www.geocities.com/wa2ise/aa5diaa.jpg
> >
> >
> > Notice the way the 6AV6 triode is baised via a 4.7 Mohm resistor.
> > The tiny DC flow in 4.7M creates enough bias voltage to bais the tube.
> > This was very common practice for this sort of tube, because the cathode
> > MUST
> > be grounded, because of the detector diodes, and you can't have a normally
> > bypassed
> > Rk.
> >
> > But its real bean counter stuff.
> >
> > Works OK though.
> >
> > Patrick Turner
>
> What's not obvious is that there is inherent DC negative feedback with
> this biasing scheme, similar to using cathode resistor bias.
>
> Let's say that as the tube ages, the plate current decreases. This will
> mean less electrons are captured by the grid, which means that the grid
> voltage becomes less negative, tending to increase the plate current.
>
> As you say, it works just fine. But I've never seen it done in "hi-fi" -
> wonder why?

I am now working to get an old japanese Trio amp
to work. The four EL84 tubes for the two channels have the cathodes all tied
together,
and the summed current used for the heaters in the phono stage, where
the second half of the 12AX7 has its cathode grounded, and has a
3.3 M bias R. Caps are mainly ceramic, the RIAA filter network resides
on a single ceramic type of chip.

Where a corner should be cut, it should be cut,
so says Bean.

Meanwhile, at the end of the 9th stage of Le Tour, one Mr Beloki
got a flat tye on a sharp corner at some speed down hill, and he broke his leg,
elbow, and wrist, and was carried off the tarmac screaming in pain.
Lance, who was right behind him was forced into making a detour off the road,
across a field, then rejoined the race by cutting across the corner along a
foot track in the grass, and then continued
to firm his grip on the Yellow, showing his superb presence of mind..

There are corners, and corners.

Patrick Turner.





>
>
> Cheers,
> Fred
> --
> +--------------------------------------------+
> | Music: http://www3.telus.net/dogstarmusic/ |
> | Projects: http://dogstar.dantimax.dk |
> +--------------------------------------------+

[Patrick Turner]

Yves Monmagnon wrote:

> Hey Pat !
> Look at http://letour.francetv.fr/

I think I am terrible at trying to read Le French,
so I been getting my race info from BBC.
every 5 or 10 minutes, just hit 'reload' abnd the page gets updated with the
latest.

Its some relief from my round of emails.....

I used to go road racing here, and I know what those guys are going through
in the Tour, and it isn't a place for people who are not tough.
When I raced, in a local club here of about 150 riders,
there were always guys getting over broken collar bones
and arms, and every so often somebody died on their bike.

>
> French speaking ....
> No race to day !

Gee, they took it easy in stage 11, but then after two days in
mountains, the cream rose to the top, and the real men now can sit back a bit,
and not have worry about some nobody winning, because they are already so
far in front on time.

Patrick Turner.

>
>
> "Patrick Turner" > wrote in message
> ...
> > Meanwhile, at the end of the 9th stage of Le Tour, one Mr Beloki
> > got a flat tye on a sharp corner at some speed down hill, and he broke his
> leg,
> > elbow, and wrist, and was carried off the tarmac screaming in pain.
> > Lance, who was right behind him was forced into making a detour off the
> road,
> > across a field, then rejoined the race by cutting across the corner along
> a
> > foot track in the grass, and then continued
> > to firm his grip on the Yellow, showing his superb presence of mind..
> >
> > There are corners, and corners.
> >
> > Patrick Turner.

[Tom Schlangen]

Hi Patrick,

even after consulting RDH4 once again I am not fully
convinced about the "excessive" low current plate
curves not resulting in unlinearity = distortion.

But maybe even some ammount of distortion is wanted
in the driver stage to cancel out distortion of
the power stage. But ... did the author of such a
circuit really think of this as a wanted feature?
We don't know, but when looking at some circuits,
I doubt it that the author did think so far.

>> Is there a reason for excessive low plate currents
>> in GC driver stages I overlooked?
>
> Bean Counters.

Or, in the case of current, counters of electrons
passing by ;-)

Tom

--
fnord (now you see it, now you don't)
- R.A. Wilson / FZ