Hi!
I would like to ask the tube Gurus:
Is it possible to build a preamp with one ECL82 per cahannel?
I have 2 Tungsram ECL82 Delta (golden pins) and would like to use them for an amp, but I wouldn't like to use output transformers. So I am thinking and looking for a preamp version with these tubes.
It would drive a SS-SE amp (Pass Zen).
Greets:
Tyimo
I would like to ask the tube Gurus:
Is it possible to build a preamp with one ECL82 per cahannel?
I have 2 Tungsram ECL82 Delta (golden pins) and would like to use them for an amp, but I wouldn't like to use output transformers. So I am thinking and looking for a preamp version with these tubes.
It would drive a SS-SE amp (Pass Zen).
Greets:
Tyimo
Yes, it is quite easy to do actually. For ideas look here (bottom of first section):
http://www.audiofanatic.it/Schemi/Tipo/Valvole/pre/Schemi_pre.html
The preamps presente there could be improved. In particular I don't like the idea of a mu follower with such disparate tubes, especially as the pentode part of the ECL82 triode strapped is a perfect low impedance triode to use as a follower on the output. I would also use a bootstrap cap from output to the triode plate circuit to give the triode a constant current plate load for better linearity, but then omit any cathode bypas cap as the gain which is already too large for a line stage, will be even higher. Underheating the tube somewhat will insure lower noise (although this will not be a large issue in this circuit), increased linearity and very long life.
http://www.audiofanatic.it/Schemi/Tipo/Valvole/pre/Schemi_pre.html
The preamps presente there could be improved. In particular I don't like the idea of a mu follower with such disparate tubes, especially as the pentode part of the ECL82 triode strapped is a perfect low impedance triode to use as a follower on the output. I would also use a bootstrap cap from output to the triode plate circuit to give the triode a constant current plate load for better linearity, but then omit any cathode bypas cap as the gain which is already too large for a line stage, will be even higher. Underheating the tube somewhat will insure lower noise (although this will not be a large issue in this circuit), increased linearity and very long life.
Note that bootstrapping the triode plate resistor from the pentode
cathode would increase output impedance to over 1k ohm (which
may be OK), because the pentode (presumably triode connected)
would no longer work as a CF, but as a unity-gain common-cathode
stage.
It would also maximise gain, which is more obviously not desirable.
I don't think linearity with just a plain plate resistor is a problem at
all in a line stage, esp. with the pentode section CF taking all the load.
I would use NFB from output to triode grid, plus another resistor from
input to triode grid, to get a buffered plate follower. It would be easy
to adjust gain, and distortion would be low (whether it "sounds"
right is another matter...). Only problem is that input impedance may
get a bit low, which dictates a low(er)-impedance volume pot...
Another idea is to make a cascaded CF where the triode section
drives the pentode, both CF-connected. But there is probably little
to be gained, if anything.
A third idea is to use the triode to buffer the pentode screen, for
improved PSRR. But this would require a higher voltage for the
buffer, maybe twice B+ (easy enough to derive with a couple of
caps and diodes).
Or forget about the triode and just use the pentode as common-
cathode line stage (or CF if you don't need any gain).
Oh, BTW, the triode could serve as a CF buffer for a Tape output,
and/or a buffer before the volume control. Then, a moderately sized
vol. pot could be used, but input impedance could be as high as you
like, since the vol. pot won't load the sources.
Well, I'm out of ideas now...
Morgan
cathode would increase output impedance to over 1k ohm (which
may be OK), because the pentode (presumably triode connected)
would no longer work as a CF, but as a unity-gain common-cathode
stage.
It would also maximise gain, which is more obviously not desirable.
I don't think linearity with just a plain plate resistor is a problem at
all in a line stage, esp. with the pentode section CF taking all the load.
I would use NFB from output to triode grid, plus another resistor from
input to triode grid, to get a buffered plate follower. It would be easy
to adjust gain, and distortion would be low (whether it "sounds"
right is another matter...). Only problem is that input impedance may
get a bit low, which dictates a low(er)-impedance volume pot...
Another idea is to make a cascaded CF where the triode section
drives the pentode, both CF-connected. But there is probably little
to be gained, if anything.
A third idea is to use the triode to buffer the pentode screen, for
improved PSRR. But this would require a higher voltage for the
buffer, maybe twice B+ (easy enough to derive with a couple of
caps and diodes).
Or forget about the triode and just use the pentode as common-
cathode line stage (or CF if you don't need any gain).
Oh, BTW, the triode could serve as a CF buffer for a Tape output,
and/or a buffer before the volume control. Then, a moderately sized
vol. pot could be used, but input impedance could be as high as you
like, since the vol. pot won't load the sources.
Well, I'm out of ideas now...
Morgan
Morgan L said:
No, the idea was to use the triode connected pentode as a second stage follower ONLY so output would be from it's cathode, anode and g2 would be strapped to B+.
True. Indeed the gain is already too high, the triode in the ECL82 is fairly high mu. CCS plate loads do increase linearity but only really if you use as little further loading as possible, hence the follower. Any remaining excess gain would have to be dealt with by feedback as you propose (CF cathode to input triode grid, with series resistance to the said grid, with consequences as you mention).
Actually, that's probably the best idea, making the triode strapped pentode into a regular single common cathode stage. The mu would be low and Rp as well, just what one would want in a line stage. Personally, I find that for a line stage a 47k volume pot (or 50k) is perfectly acceptable but would not go below it. Using the triode as a buffer (CF) would alow one to use higher values. I don't have the ECL82 datasheet at hand to check capacitances but I doubt miller would be a problem with a 47k volume pot given the low gain. That would leave the triode as a tape out buffer, or for something altogether different (second stage of a phono amp... not the best choice but still fine).
ilimzn wrote:
'No, the idea was to use the triode connected pentode as a second stage follower ONLY so output would be from it's cathode, anode and g2 would be strapped to B+.'
Yes, that was clear to me. But earlier you said:
'I would also use a bootstrap cap from output to the triode plate circuit to give the triode a constant current plate load for better linearity, but then omit any cathode bypas cap as the gain which is already too large for a line stage, will be even higher.'
That was what my remark was about. Output Z can be measured by
sending a pulse into the output (CF cathode). A plain CF will "eat"
this pulse, whereas using it to also bootstrap the VA stage will relay
(much of) the pulse also to the pentode grid, which, in making
the pentode section (almost) a common cathode stage, will increase
output Z (though not dramatically). But enough about that - it
doesn't matter much, and we seem to be gravitating towards the
same suggestions on how to use the tube anyway. I totally agree
that 50k pot, plain triode-ct VA stage, slightly starved heaters, and
the possiblity of using triode as buffer, perhaps optionally
(say switchable), are good ideas that can serve as the basis of a
pretty good line stage implementation of this tube. Then there's
the op points...
Tyimo also asked: would it operate with under 200V?
For a line stage, 200V is OK, with about 100V across the tubes and
about the same across the load resistors (or CCS). Much under -
don't know, but preferably not. The pentode section will like a bit
of current, but we must avoid getting anywhere near risking grid
current, so a healthy Vak is good. Suitable currents are about
20-30mA for the pentode and 1.5-2mA for the triode, at said vol-
tages - IMO.
Morgan
'No, the idea was to use the triode connected pentode as a second stage follower ONLY so output would be from it's cathode, anode and g2 would be strapped to B+.'
Yes, that was clear to me. But earlier you said:
'I would also use a bootstrap cap from output to the triode plate circuit to give the triode a constant current plate load for better linearity, but then omit any cathode bypas cap as the gain which is already too large for a line stage, will be even higher.'
That was what my remark was about. Output Z can be measured by
sending a pulse into the output (CF cathode). A plain CF will "eat"
this pulse, whereas using it to also bootstrap the VA stage will relay
(much of) the pulse also to the pentode grid, which, in making
the pentode section (almost) a common cathode stage, will increase
output Z (though not dramatically). But enough about that - it
doesn't matter much, and we seem to be gravitating towards the
same suggestions on how to use the tube anyway. I totally agree
that 50k pot, plain triode-ct VA stage, slightly starved heaters, and
the possiblity of using triode as buffer, perhaps optionally
(say switchable), are good ideas that can serve as the basis of a
pretty good line stage implementation of this tube. Then there's
the op points...
Tyimo also asked: would it operate with under 200V?
For a line stage, 200V is OK, with about 100V across the tubes and
about the same across the load resistors (or CCS). Much under -
don't know, but preferably not. The pentode section will like a bit
of current, but we must avoid getting anywhere near risking grid
current, so a healthy Vak is good. Suitable currents are about
20-30mA for the pentode and 1.5-2mA for the triode, at said vol-
tages - IMO.
Morgan
Hi,
one should consider that the pentode section of ECL82/6BM8/PCL82/16A8, when being triode strapped, is not a particulary linear one. Have a look at:
http://www.tubes.mynetcologne.de/roehren/daten/ecl82pentode_as_triode.pdf
This might be relativated by the small signal swing needed for a preamp output.
Talking about S/N ratio of a preamp. the 300mA series heater PCL82/16A8 variant (which probably is much cheaper to get in decent quality than ECL82/6BM8) with a DC 300mA CCS at heaters (LM317...) will eleminate any heater-induced hum in such a preamp.
Tom
one should consider that the pentode section of ECL82/6BM8/PCL82/16A8, when being triode strapped, is not a particulary linear one. Have a look at:
http://www.tubes.mynetcologne.de/roehren/daten/ecl82pentode_as_triode.pdf
This might be relativated by the small signal swing needed for a preamp output.
Talking about S/N ratio of a preamp. the 300mA series heater PCL82/16A8 variant (which probably is much cheaper to get in decent quality than ECL82/6BM8) with a DC 300mA CCS at heaters (LM317...) will eleminate any heater-induced hum in such a preamp.
Tom
Triode+Pentode in CF, 250V:
http://forum.electronicwerkstatt.de/phpBB/uploads/02355_1078860208_head04.gif
Or a mu follower with pentode section as upper tube. Proposed by elektor not so long time ago.
http://forum.electronicwerkstatt.de/phpBB/uploads/02355_1078860208_head04.gif
Or a mu follower with pentode section as upper tube. Proposed by elektor not so long time ago.
Tyimo,
What's the sensitivity of the power amp for full O/P? I'm thinking of a different approach.
If less than 2 V. RMS will drive the power amp into clipping, you don't need any gain at all and a buffered level control would be FINE. Wire the triode as a cathode follower with the pentode wired up as a constant current sink for the CF's load.
What's the sensitivity of the power amp for full O/P? I'm thinking of a different approach.
If less than 2 V. RMS will drive the power amp into clipping, you don't need any gain at all and a buffered level control would be FINE. Wire the triode as a cathode follower with the pentode wired up as a constant current sink for the CF's load.
Tubes4e4 said:
Just looked at it and you are right, of course. Seems like the only way to get decent linearity is to use a CCS plate load, which complicates things. OTOH, the gain would be low (mu is about 8 IIRC) and an unbypassed cathode resistor would make it lower still (at the expense of output impedance, though). For a couple of V RMS it may still have decent enough distorsion specs even without these extra measures.
Eli Duttman said:
Hey Eli!
I just sent you a reply to your EMail.
Actualy I'd probably recommend using the 1:2 at the input of the amp depending on your buffer's output impedance and whether it can drive a 1.5k ohm load.
I'm often reluctant to recommend using those 1:2 mic trannies except where one has control over what's driving it as their 1.5k ohm input impedance (when they're properly loaded) will cause some output stages to choke. You know, like most tube output stages.
se
Hi! I am back again.
The situation is changed, my new power amp need 15 V input and has 330K Zin. (it is a mosfet follower). Mostly I use it with CDplayer which has more than 2 V output.
So, I would like to ask the Tube Gurus to suggest me a good preamp with ECL82 tube.
Until now I found 2 design and I would like to hear (read) your oppinions.
Here is the first competitor:
The situation is changed, my new power amp need 15 V input and has 330K Zin. (it is a mosfet follower). Mostly I use it with CDplayer which has more than 2 V output.
So, I would like to ask the Tube Gurus to suggest me a good preamp with ECL82 tube.
Until now I found 2 design and I would like to hear (read)
your oppinions.Here is the first competitor:
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.