[Jon Yaeger]

A few years ago I designed a simple headphone amplifier for a neighbor,
using junk box parts that I'd salvaged from an old Wollensak recorder.

The recorder used a 7591A in SE mode with a small transformer. The spec
sheet boasted a wide bandwidth, which I didn't understand given the size of
the tranny.

Anyway, I put the 7591A to better use and put together a circuit using a
6SN7 and a 6V6 wired in triode mode.

The friend, who is also a musician, liked the performance of the amp.

I've got the parts to build another.

I'd like to know other recommendations for an output tube for this
application, e.g. 12B4. The ratio on the output transformer is about 19:1
and the primary resistance is 237 ohms.

I'll send a .pdf schematic to anyone who is interested.

Thanks for your input (or is that "output"?)

Jon

[Jon Yaeger]

in article , flipper at
wrote on 5/19/09 10:28 PM:

On Tue, 19 May 2009 21:17:04 -0400, Jon Yaeger
wrote:

A few years ago I designed a simple headphone amplifier for a neighbor,
using junk box parts that I'd salvaged from an old Wollensak recorder.

The recorder used a 7591A in SE mode with a small transformer. The spec
sheet boasted a wide bandwidth, which I didn't understand given the size of
the tranny.

Anyway, I put the 7591A to better use and put together a circuit using a
6SN7 and a 6V6 wired in triode mode.

The friend, who is also a musician, liked the performance of the amp.

I've got the parts to build another.

I'd like to know other recommendations for an output tube for this
application, e.g. 12B4. The ratio on the output transformer is about 19:1
and the primary resistance is 237 ohms.

I'll send a .pdf schematic to anyone who is interested.

Thanks for your input (or is that "output"?)

Jon

Are you looking for more power, or what? I ask because if the other
you made was 'good' then why change?

The 6L6 is sometimes used as a 'cheap replacement' but the gain is
less. But that wouldn't be as much of a problem in a scratch build.

If you want to spend a few bucks but still save vs the 7591, the 6GM5
is a 7591 in a different bottle and radiodaze sells them for 25 bucks,
a third the 7591. Well, for NOS that is. New manufacture JJ/Tesla
7591s go for 30 bucks a pair.

For triode I might try a 6EM7. Beefier than the 12B4 and you get an
'extra' gain triode.


Thanks for your suggestion.

Since it is a headphone amp, I'm not looking for more power. But I am
interested in more linearity, better sound, and using a "true" triode if it
has any real sonic benefits.

I'm familiar with the 6GM5. I guess could use a single Compactron like the
6T9 for reduced parts count, but I wonder if it would perform as well.

Jon

[Ian Iveson]

Jon Yaeger wrote:

A few years ago I designed a simple headphone amplifier for
a neighbor,
using junk box parts that I'd salvaged from an old
Wollensak recorder.

The recorder used a 7591A in SE mode with a small
transformer. The spec
sheet boasted a wide bandwidth, which I didn't understand
given the size of
the tranny.

Anyway, I put the 7591A to better use and put together a
circuit using a
6SN7 and a 6V6 wired in triode mode.

The friend, who is also a musician, liked the performance
of the amp.

I've got the parts to build another.

I'd like to know other recommendations for an output tube
for this
application, e.g. 12B4. The ratio on the output
transformer is about 19:1
and the primary resistance is 237 ohms.

I'll send a .pdf schematic to anyone who is interested.

Thanks for your input (or is that "output"?)

EL84 is an obvious suggestion.

6CH6 and other ~10W pentodes/beam tetrodes intended for TV
audio/video amps. Cheap and can make nice triodes. My
favourite coz I've got some.

Both halves of ECC88 or ECC82 or other small triodes with
low anode resistance. 12BH7 springs to mind but I haven't
the foggiest notion why.

All depending on what circuit you use and, crucially, the
primary inductance and saturation current of that
transformer. If you don't know you should measure, otherwise
you don't know what source impedance and load you should use
to get your required bandwidth and accommodate your required
drive voltage. Consequently you don't know where your
loadline might be for any valve, so how can you compare
linearity?

Also depends on what kind of output impedance you want. I
require a tenth of the headphone impedance, but many folk
are happy to drive with a high impedance.

You have so much power to play with, you can afford to lose
lots to manipulate the bandwidth and maybe output impedance
to where you want it.

Ian

[Patrick Turner]

Jon Yaeger wrote:

A few years ago I designed a simple headphone amplifier for a neighbor,
using junk box parts that I'd salvaged from an old Wollensak recorder.

The recorder used a 7591A in SE mode with a small transformer. The spec
sheet boasted a wide bandwidth, which I didn't understand given the size of
the tranny.

Smaller trannies have wider BW than larger ones, if the interleaving
pattern is the same.


Anyway, I put the 7591A to better use and put together a circuit using a
6SN7 and a 6V6 wired in triode mode.

The friend, who is also a musician, liked the performance of the amp.

I've got the parts to build another.

I'd like to know other recommendations for an output tube for this
application, e.g. 12B4. The ratio on the output transformer is about 19:1
and the primary resistance is 237 ohms.

If you had a 5k load on the output tube, the match at the sec would be
5,000 / 261 = approx 20 ohms.

But with headphones the load match isn't very important as long as the
phones impedance is above 8 ohms.
The power/volume levels are low, and voltage levels tiny, and the OPT
will work with many types of OP tube. So the trick is to minimize the
hum and hiss, and the sound should be fine. Perhaps you need to place a
resistance divider across the sec, say 22 ohms and 4.7 ohms, and connect
the phones across the 4.7 ohms which will improve the SNR by 12dB and
allow higher signal levels to be used and a higher gain setting which
isn't so touchy at the bottom of the range.

Patrick Turner.





I'll send a .pdf schematic to anyone who is interested.

Thanks for your input (or is that "output"?)

Jon

[Jon Yaeger]

in article , Ian Iveson at
wrote on 5/19/09 11:26 PM:

Jon Yaeger wrote:

A few years ago I designed a simple headphone amplifier for
a neighbor,
using junk box parts that I'd salvaged from an old
Wollensak recorder.

The recorder used a 7591A in SE mode with a small
transformer. The spec
sheet boasted a wide bandwidth, which I didn't understand
given the size of
the tranny.

Anyway, I put the 7591A to better use and put together a
circuit using a
6SN7 and a 6V6 wired in triode mode.

The friend, who is also a musician, liked the performance
of the amp.

I've got the parts to build another.

I'd like to know other recommendations for an output tube
for this
application, e.g. 12B4. The ratio on the output
transformer is about 19:1
and the primary resistance is 237 ohms.

I'll send a .pdf schematic to anyone who is interested.

Thanks for your input (or is that "output"?)

EL84 is an obvious suggestion.

6CH6 and other ~10W pentodes/beam tetrodes intended for TV
audio/video amps. Cheap and can make nice triodes. My
favourite coz I've got some.

Both halves of ECC88 or ECC82 or other small triodes with
low anode resistance. 12BH7 springs to mind but I haven't
the foggiest notion why.

All depending on what circuit you use and, crucially, the
primary inductance and saturation current of that
transformer. If you don't know you should measure, otherwise
you don't know what source impedance and load you should use
to get your required bandwidth and accommodate your required
drive voltage. Consequently you don't know where your
loadline might be for any valve, so how can you compare
linearity?

Also depends on what kind of output impedance you want. I
require a tenth of the headphone impedance, but many folk
are happy to drive with a high impedance.

You have so much power to play with, you can afford to lose
lots to manipulate the bandwidth and maybe output impedance
to where you want it.

Ian


Ian,

I've got a lot of EL84s vs. 6V6s. I was also thinking about a 6FQ7 vs. the
6SN7.

My previous design used diodes and a large choke in a choke-input PS.

More questions:

1. Any sonic benefit to using a tube rectifier?

2. Any benefit to using soft recovery or HEXFREDs if solid state?

3. Is DC filament voltage a good idea?

Thanks in advance,

Jon

[Jon Yaeger]

in article , Patrick Turner at
wrote on 5/20/09 6:06 AM:



Jon Yaeger wrote:

A few years ago I designed a simple headphone amplifier for a neighbor,
using junk box parts that I'd salvaged from an old Wollensak recorder.

The recorder used a 7591A in SE mode with a small transformer. The spec
sheet boasted a wide bandwidth, which I didn't understand given the size of
the tranny.

Smaller trannies have wider BW than larger ones, if the interleaving
pattern is the same.


Anyway, I put the 7591A to better use and put together a circuit using a
6SN7 and a 6V6 wired in triode mode.

The friend, who is also a musician, liked the performance of the amp.

I've got the parts to build another.

I'd like to know other recommendations for an output tube for this
application, e.g. 12B4. The ratio on the output transformer is about 19:1
and the primary resistance is 237 ohms.

If you had a 5k load on the output tube, the match at the sec would be
5,000 / 261 = approx 20 ohms.

But with headphones the load match isn't very important as long as the
phones impedance is above 8 ohms.
The power/volume levels are low, and voltage levels tiny, and the OPT
will work with many types of OP tube. So the trick is to minimize the
hum and hiss, and the sound should be fine. Perhaps you need to place a
resistance divider across the sec, say 22 ohms and 4.7 ohms, and connect
the phones across the 4.7 ohms which will improve the SNR by 12dB and
allow higher signal levels to be used and a higher gain setting which
isn't so touchy at the bottom of the range.

Patrick Turner.





I'll send a .pdf schematic to anyone who is interested.

Thanks for your input (or is that "output"?)

Jon


Thanks, Patrick, for your always helpful advice!

[Don Pearce[_3_]]

On Wed, 20 May 2009 10:06:33 GMT, Patrick Turner
wrote:

If you had a 5k load on the output tube, the match at the sec would be
5,000 / 261 = approx 20 ohms.

But with headphones the load match isn't very important as long as the
phones impedance is above 8 ohms.
The power/volume levels are low, and voltage levels tiny, and the OPT
will work with many types of OP tube. So the trick is to minimize the
hum and hiss, and the sound should be fine. Perhaps you need to place a
resistance divider across the sec, say 22 ohms and 4.7 ohms, and connect
the phones across the 4.7 ohms which will improve the SNR by 12dB and
allow higher signal levels to be used and a higher gain setting which
isn't so touchy at the bottom of the range.

The best way to set these resistor values is so that the volume in the
headphones is the same as that from the speakers, so there are no ill
effects from either plugging in or unplugging the phones. This also
helps avoid the problem of playing phones at too high a level.

d

[[email protected]]

On May 20, 11:43 am, Jon Yaeger wrote:
in article , Ian Iveson at
wrote on 5/19/09 11:26 PM:



Jon Yaeger wrote:

A few years ago I designed a simple headphone amplifier for
a neighbor,
using junk box parts that I'd salvaged from an old
Wollensak recorder.

The recorder used a 7591A in SE mode with a small
transformer. The spec
sheet boasted a wide bandwidth, which I didn't understand
given the size of
the tranny.

Anyway, I put the 7591A to better use and put together a
circuit using a
6SN7 and a 6V6 wired in triode mode.

The friend, who is also a musician, liked the performance
of the amp.

I've got the parts to build another.

I'd like to know other recommendations for an output tube
for this
application, e.g. 12B4. The ratio on the output
transformer is about 19:1
and the primary resistance is 237 ohms.

I'll send a .pdf schematic to anyone who is interested.

Thanks for your input (or is that "output"?)

EL84 is an obvious suggestion.

6CH6 and other ~10W pentodes/beam tetrodes intended for TV
audio/video amps. Cheap and can make nice triodes. My
favourite coz I've got some.

Both halves of ECC88 or ECC82 or other small triodes with
low anode resistance. 12BH7 springs to mind but I haven't
the foggiest notion why.

All depending on what circuit you use and, crucially, the
primary inductance and saturation current of that
transformer. If you don't know you should measure, otherwise
you don't know what source impedance and load you should use
to get your required bandwidth and accommodate your required
drive voltage. Consequently you don't know where your
loadline might be for any valve, so how can you compare
linearity?

Also depends on what kind of output impedance you want. I
require a tenth of the headphone impedance, but many folk
are happy to drive with a high impedance.

You have so much power to play with, you can afford to lose
lots to manipulate the bandwidth and maybe output impedance
to where you want it.

Ian

Ian,

I've got a lot of EL84s vs. 6V6s. I was also thinking about a 6FQ7 vs. the
6SN7.

My previous design used diodes and a large choke in a choke-input PS.

More questions:

1. Any sonic benefit to using a tube rectifier?

2. Any benefit to using soft recovery or HEXFREDs if solid state?

3. Is DC filament voltage a good idea?

You are asking the kind of questions that just get people hot and
bothered. You shoule not even ask but try each one yourself and see if
there is any difference. Actually, for your results to be valid you
have to get OTHERS' response in a way they don't know which is which.

My recommendations are no, no and no to each question you have
asked, except DC heaters to the earlier stages
would be good if not too much bother. But I could be wrong. You won't
know unless you try every permutation yourself.

Actually, I think solid state headphone amps are actually better
except for Stax style headphones with high voltage bias.

[Patrick Turner]

Jon Yaeger wrote:

in article , Ian Iveson at
wrote on 5/19/09 11:26 PM:

Jon Yaeger wrote:

A few years ago I designed a simple headphone amplifier for
a neighbor,
using junk box parts that I'd salvaged from an old
Wollensak recorder.

The recorder used a 7591A in SE mode with a small
transformer. The spec
sheet boasted a wide bandwidth, which I didn't understand
given the size of
the tranny.

Anyway, I put the 7591A to better use and put together a
circuit using a
6SN7 and a 6V6 wired in triode mode.

The friend, who is also a musician, liked the performance
of the amp.

I've got the parts to build another.

I'd like to know other recommendations for an output tube
for this
application, e.g. 12B4. The ratio on the output
transformer is about 19:1
and the primary resistance is 237 ohms.

I'll send a .pdf schematic to anyone who is interested.

Thanks for your input (or is that "output"?)

EL84 is an obvious suggestion.

6CH6 and other ~10W pentodes/beam tetrodes intended for TV
audio/video amps. Cheap and can make nice triodes. My
favourite coz I've got some.

Both halves of ECC88 or ECC82 or other small triodes with
low anode resistance. 12BH7 springs to mind but I haven't
the foggiest notion why.

All depending on what circuit you use and, crucially, the
primary inductance and saturation current of that
transformer. If you don't know you should measure, otherwise
you don't know what source impedance and load you should use
to get your required bandwidth and accommodate your required
drive voltage. Consequently you don't know where your
loadline might be for any valve, so how can you compare
linearity?

Also depends on what kind of output impedance you want. I
require a tenth of the headphone impedance, but many folk
are happy to drive with a high impedance.

You have so much power to play with, you can afford to lose
lots to manipulate the bandwidth and maybe output impedance
to where you want it.

Ian

Ian,

I've got a lot of EL84s vs. 6V6s. I was also thinking about a 6FQ7 vs. the
6SN7.

My previous design used diodes and a large choke in a choke-input PS.

More questions:

1. Any sonic benefit to using a tube rectifier?

Not that I am aware of.

2. Any benefit to using soft recovery or HEXFREDs if solid state?

Not that I am aware of.

3. Is DC filament voltage a good idea?

In input and driver stages it is a good idea.

if the design is good, you will not see any rectifier switching products
in the output.
If there is no hum or such switching spikes, the rectifiers and
smoothing circuits are well designed.
Exotic diodes in audio amps don't cure problems if the placement of
parts and screening and earth pathing ain't done right.

Patrick Turner.

Thanks in advance,

Jon

[Patrick Turner]

Don Pearce wrote:

On Wed, 20 May 2009 10:06:33 GMT, Patrick Turner
wrote:

If you had a 5k load on the output tube, the match at the sec would be
5,000 / 261 = approx 20 ohms.

But with headphones the load match isn't very important as long as the
phones impedance is above 8 ohms.
The power/volume levels are low, and voltage levels tiny, and the OPT
will work with many types of OP tube. So the trick is to minimize the
hum and hiss, and the sound should be fine. Perhaps you need to place a
resistance divider across the sec, say 22 ohms and 4.7 ohms, and connect
the phones across the 4.7 ohms which will improve the SNR by 12dB and
allow higher signal levels to be used and a higher gain setting which
isn't so touchy at the bottom of the range.

The best way to set these resistor values is so that the volume in the
headphones is the same as that from the speakers, so there are no ill
effects from either plugging in or unplugging the phones. This also
helps avoid the problem of playing phones at too high a level.

I think this design was for a dedicated headphone amp only. Speakers
won't be used so they won't be switched off when phones are plugged in.

Magnetic coupling between an OPT and a nearby mains PT is deadly for a
headphone amp SNR.

If you don't have potted trannies, a remote PSU is a good idea.

The Ming Da Chinese phone amp sufferes from magnetic coupling because
they didn't have potted trannies and they have the PT far too close to
the OPT.

Patrick Turner.

d