Best input/driver stage for 807 SE?
Henry 007
6SL7 and 6SN7 available, but I'm not sure what the best options are. Is SRPP
good sounding? I am keen on accuracy and low distortion, rather than an
overly warm sound. I want this amp to have slam!
The project is coming along nicely, have got some NOS 1945 807s, some 1960s
600V oil caps (2x8uF, 2x40uF) and a huge power transformer dated 1959!
Outputs will be 125ESE, and the chassis is an old PA amp. Should be fun to
make!
~h~
Patrick Turner
Henry 007 wrote:
I guess I could say the slam and accuracy depends on the maximum power you want,
and what mode of operation you will be using.
If you had a single 807 in tetrode, expect about 9 watts and somewhat high thd
and very high Rout, and to get the Rout down to about 1 ohm,
about 20 dB of NFB will be required, so probably the 6SN7 halves could be
cascaded
to provide enough gain to allow the loss in gain of 20 Db when NFB is applied.
6SN7 is a medium u tube, with good linearity, and the drive voltage for an 807
in tetrode is low, so linearity will depend on the NFB applied to fix the THD
of the output tube, which will have at least 10 times the THD as the driver
stage.
The 6SL7 is a high u tube, more suited to use as an input tube where
higher gain is required if the drive to the output tube is higher as it would be
if the 807 was set up in Triode.
In this case, one could be flexible.
The 807 in triode needs an average of 2 to 3 times the drive voltage as tetrode
mode,
to make a maximum of 7 watts, but an average usable power of 4 to 5 watts.
The input capacitance of the 807 in triode is higher than in tetrode,
so there is only one choice, 6SN7, and it could be used
with each half cascaded. The amount of loop FB does not need to be as high as
the tetrode tube, since Rout is already low due to the triode connection.
I find about 12 dB of NFB is OK for most class A triode amps.
If there is no intention to use any global NFB, or any other FB loop,
then for triode, you'd only need 1/2 a 6SN7 to make about 28 vrms drive voltage,
and then only 1.8 vrms is required at its grid for full power.
The other 1/2 of the 6SN7 can be a preamp stage with an unbypassed
cathode R so the gain is about 10, so a 0.18 vrms input signal will produce full
power.
Even if 12 dB of global NFB is used in the 807 triode amp,
the input voltage with NFB will be about 0.72 vrms.
Most CD players produce plenty of voltage, and so do most phono amps.
If a pair of 807 are paralleled in triode, the power output will be greater,
depending on the 807 load, but one thing is for sure, less
input voltage will be needed for the same sound level, perhaps 0.5 vrms of
signal,
and you should find this very suitable with most sources.
Using a paralleled 6SN7 as a driver for the two 807 outputs would be better,
and then you could have a choice of which input tube to use, either
another 6SN7, or a 6SL7, which will reduce the input voltage
required to between 0.35 and 0.25 vrms.
I use a 13E1 in SE amps of my own
These need about 55 vrms at the output tube grid for 22 watts of output.
I use an ECC32 which is slightly higher gain than a 6SN7,
and a 12SL7 input tube, and I have about 16 dB od global NFB.
Sensitivity is about 0.5 vrms.
The schematic and notes are at
http://turneraudio.com.au/htmlwebpgs02/schemseul22w.htm
Food for thought.
Patrick Turner.
Ronald Pit
I didn't build it , but I found this :
http://members.home.nl/triode.dick/tekeningen/se807.gif
Hope it helps ....
Ronald
"Henry 007" <future_unity@say_no_to_spam_hotmail.com> schreef in bericht
news:CAkG9.20534$_h6.3...@news.xtra.co.nz...
Patrick Turner
Ronald Pit wrote:
> Hi Henry ,
>
> I didn't build it , but I found this :
> http://members.home.nl/triode.dick/tekeningen/se807.gif
>
> Hope it helps ....
>
> Ronald
It is a nice circuit in the above URL.
I probably wouldn't bother with the SRPP 6SN7 driver.
A normal plate loaded 6SN7 would easily make enough drive voltage for
an SE 807 in beam tetrode. ( not pentode, since an 807 is NOT a pentode )
Any 2H distortion in the driver tube tends to cancel the 2H in the output tube.
Its not a full cancelation, since the output tube might have
8% of 2H at some load, and the driver will only have 1% of 2H,
so to get good cancelation you would have to design the driver stage "badly"
on purpose, by using a low load R which is too low.
I find this consideration not worth doing, and I just use an R
loaded driver stage, and settle for minimal 2H cancelation.
In the above circuit, the open loop gain,
assuming a 5 k to 8 ohms OPT is approx 10 x 17 x 30 divided by 25 = 204.
B, the fraction fed back = 1.5 / 82 = 0.0182.
Therefore the gain with FB is 204 / ( 1 + [ 204 x 0.0182 ] ) = 43.3.
The gain reduction is 4.71 times, which is an appied 13.5 dB of global NFB.
This is about enough FB to reduce the Rout to about the same as a triode amp
with no NFB.
Probably however, a pair of 807 in triode in parallel
would produce slightly more power, but have a far more blameless
THD profile, as each tube would see a 10k load, and could be biased
at a lower tube current.
I suspect the load tolerance would be a lot better, as even with no global FB,
the Rout at the sec is 1.28 ohms, plus winding resistance of about 0.6 ohms =
1.88 ohms,
and the applied NFB loop would be about 6 dB, since the triode gain is
less than half the tetrode's, so Rout would end up at approx 1 ohm.
Maybe using Rfb of 47 k intsead of 82 k will improve the Rout further.
In this case the SRPP could be retained, as it is lower Rout
than a normal R loaded stage with an unbypased cathode R.
The power supply uses 33uF, 10H, 33uF, for the main anode supply.
I would use 2 x 47 uF for C1, in series, and then about 3 H,
and then 2 x 470 uF in series for the B+ supply.
This will keep the charge currents in the tube rectifier low,
but reduce ripple voltage at C2 from 0.065 vrms to 0.046 vrms,
and I suggest the larger caps are cheaper than the larger choke,
and that the 235 uF anchor for the OPT will work better than the 33 uF at LF,
where at 30 Hz, the reactance of 33 uF is 160 ohms, and capable of
creating some IMD.
Patrick Turner.