[Patrick Turner]

flipper wrote:

Not a problem, per see. It's just that I never had occasion to observe
it for myself till this little guitar amp I built for the nephew. Did
a thanksgiving day field test at the family get together.

To keep costs down I'm using 12AT7s in the premp and first thing I
discovered on the test bench was being able to hear tapping in the
chassis in the speaker. Microphonics.

As luck would have it I just happened to have a low microphonic 6201
and, sure enough, no tapping sound with that one. Nice to see theory
play out in practice.

However, the source I normally use for tubes subs 12AT7WAs for 6201s.
Hmm. So I pulled out all my 12AT7s and a couple were WAs. One was
quiet (GE) but the Tungsol was microphonic as hell. And the 'plain'
Sylvania 12AT7 was as quiet as the GE WA.

One thing I learned, I can tell which will and which won't by simply
looking for the extra mica spacer. But the question is, how can I
order a 'quiet' one when 12AT7WAs don't necessarily have it but even
some 'plain' 12AT7s do?

Does anyone know if any of the 'current' production 12AT7s are low
microphonic? While the few I've got were distinguished by the extra
mica spacer when I look at photos at tube depot the Ei 12AT7 Elite
Gold claims low microphonics but doesn't have the extra spacer while
the 'cheap' Sino does (according to the pictures, the only one that
does). but make no microphonics claim.

It was kind of fun trying different dual triodes of the same basing.
While the microphonic 12AT7s were 'obvious' (if you tap the chassis),
they might be acceptable because it's not enough to oscillate but the
microphonic feedback screech with a 5965 ('computer' triode) was
stunning.

One must carefully sort through your 12AT7 to find ones that are
non microphonic. This means you should test them with a following high
gain amp and tap them lightly with a pencil.
Bad ones will make a ring waveform like a bell that may last
seconds and be high amplitude and somewhere between 500Hz and 5 kHz.
If used in a mic or phono preamp they are poison!

I have had one that was so bad that it oscillated with accoustic FB from
the speaker at low levels
even when used in the second gain stage in a phono preamp, see The
Rocket,
http://www.turneraudio.com.au/preamp...hono-2005.html

Tube damper rings did nothing to stop the
ring at about 1.5kHz that beset the music.

Even if there is no oscilation, a tube could be on the verge of doing
so,
and when a speaker response check is done with a slightly microphonic
tube present,
then there may be a peaked response at one or more F due to accoustic
resonance of
the tube structure.
My guess is that this is one explanation of why some tubes sound
different
although all the THD indicates blame-free 0.01% or less.

If one was extraordinarily lucky, a peice of music in the key of C
should sound richer if the tube Fo was a C note, but awful
if the Fo was at some non-musical tone which is 99% more likely.

So there isn't anything wrong with a 12AT7 as long as you pick one that
isn't microphonic, OR NOISY.

Theoretically, a 12AT7 should make an ideal phono imput tube, and
if 3 are paralleled, the Gm total of the 6 resultant triodes
is 24mA/V, some 15 times more than a 1/2 12AX7.
Therefore the noise is reduced by a factor of sq.root of 1/15 or from
about
2uV of the 12AX7 to about 0.55uV, which would in theory
allow MC direct connection without a step up tranny.
But you may find porcine flight is more likely to be possible
before such theoretical noise reductions are seen to be actually real.
I have found all tube inputs for MC to be too noisy
and A FET is the only way to go for low noise.

And the fet has no microphony. But it is much more sensitive to
the slightest magnetic field change so
the case for a phono amp with a fet should be heavy iron.

In the old days of the tube era the guys who serviced the tube gear used
in studios
placed the quietest and least micro tubes at the front ends and the
worst
went into the line level amps and power amps.

Fighting noise, distortion and microphony made work for a lot of ppl.
If you don't work at it, the gear won't perform optimally,
and after you have worked to optimise, the optimisation
does last forever.
Now there's no less work; the complexity is an order of magnitude
greater,
and there are fights over digital ****, as they churn out the
music all processed and packaged like pork sausages....



Patrick Turner.

[Andy Evans]

Two things I've observed abut microphonics:
1. There seems to be microphonics from the natural construction of the
tube, and another kind from a fault condition where it has been dropped
etc.
2. The natural kind of microphonics seems to be less obtrusive in push
pull (balanced)stages than SE. I use DHTs and don't have much of a
problem since my system is balanced.

[Patrick Turner]

Andy Evans wrote:

Two things I've observed abut microphonics:
1. There seems to be microphonics from the natural construction of the
tube, and another kind from a fault condition where it has been dropped
etc.
2. The natural kind of microphonics seems to be less obtrusive in push
pull (balanced)stages than SE. I use DHTs and don't have much of a
problem since my system is balanced.

OK, so you have two DHT triodes in an LTP balanced circuit
and one tube is noisy, and the other ain't.

The LTP would simply halve the noise so you'd get
two phases of the noise at each output.

Seems to me the balanced circuit hasn't much to offer on noise reduction
but
does offer much lower THD/IMD and gives CMR.

Patrick Turner.

[Sander deWaal]

Patrick Turner said:


Tube damper rings did nothing to stop the
ring at about 1.5kHz that beset the music.


Tube dampers do nothing, period.


--
- Ever seen someone with 5.1 ears? So, what does that tell you? -

[Andy Evans]

OK, so you have two DHT triodes in an LTP balanced circuit and one tube
is noisy, and the other ain't. The LTP would simply halve the noise so
you'd get
two phases of the noise at each output.

I'm talking two good tubes (LTP with CCS underneath) , but both prone
to misrophonics in a SE topology.


Seems to me the balanced circuit hasn't much to offer on noise
reduction
but does offer much lower THD/IMD and gives CMR.

It's not really noise reduction in the conventional sense - not sure
what it is, but it does seem to work in practice - the tubes behave
themselves and don't scream and howl.

[Patrick Turner]

Sander deWaal wrote:

Patrick Turner said:

Tube damper rings did nothing to stop the
ring at about 1.5kHz that beset the music.

Tube dampers do nothing, period.

--
- Ever seen someone with 5.1 ears? So, what does that tell you? -

They have pro-logic surround sound brain.....

Patrick Turner.

[Patrick Turner]

flipper wrote:

On Sun, 26 Nov 2006 10:37:47 GMT, Patrick Turner
wrote:



flipper wrote:

Not a problem, per see. It's just that I never had occasion to observe
it for myself till this little guitar amp I built for the nephew. Did
a thanksgiving day field test at the family get together.

To keep costs down I'm using 12AT7s in the premp and first thing I
discovered on the test bench was being able to hear tapping in the
chassis in the speaker. Microphonics.

As luck would have it I just happened to have a low microphonic 6201
and, sure enough, no tapping sound with that one. Nice to see theory
play out in practice.

However, the source I normally use for tubes subs 12AT7WAs for 6201s.
Hmm. So I pulled out all my 12AT7s and a couple were WAs. One was
quiet (GE) but the Tungsol was microphonic as hell. And the 'plain'
Sylvania 12AT7 was as quiet as the GE WA.

One thing I learned, I can tell which will and which won't by simply
looking for the extra mica spacer. But the question is, how can I
order a 'quiet' one when 12AT7WAs don't necessarily have it but even
some 'plain' 12AT7s do?

Does anyone know if any of the 'current' production 12AT7s are low
microphonic? While the few I've got were distinguished by the extra
mica spacer when I look at photos at tube depot the Ei 12AT7 Elite
Gold claims low microphonics but doesn't have the extra spacer while
the 'cheap' Sino does (according to the pictures, the only one that
does). but make no microphonics claim.

It was kind of fun trying different dual triodes of the same basing.
While the microphonic 12AT7s were 'obvious' (if you tap the chassis),
they might be acceptable because it's not enough to oscillate but the
microphonic feedback screech with a 5965 ('computer' triode) was
stunning.

One must carefully sort through your 12AT7 to find ones that are
non microphonic. This means you should test them with a following high
gain amp and tap them lightly with a pencil.
Bad ones will make a ring waveform like a bell that may last
seconds and be high amplitude and somewhere between 500Hz and 5 kHz.
If used in a mic or phono preamp they are poison!

Yes, already tested and done, not that I had all that many to test.

I have had one that was so bad that it oscillated with accoustic FB from
the speaker at low levels
even when used in the second gain stage in a phono preamp, see The
Rocket,
http://www.turneraudio.com.au/preamp...hono-2005.html

Tube damper rings did nothing to stop the
ring at about 1.5kHz that beset the music.

Hehe. Yes, the 5965 was like that. Although I was able to get it to
stop squealing by placing the amp on something 'soft' rather than the
hard table. Not a 'solution', just an observation.

During this rare and bad occurence of microphony I was able to pick up
the offending amp with both hands and the noise still happened.
Padding under the amp didn't work for me either.

Only tube replacement worked.

Actually, it was worse. The squealing was so bad I doubt one could
hear music through it and loud enough that I didn't even try.

Even if there is no oscilation, a tube could be on the verge of doing
so,
and when a speaker response check is done with a slightly microphonic
tube present,
then there may be a peaked response at one or more F due to accoustic
resonance of
the tube structure.
My guess is that this is one explanation of why some tubes sound
different
although all the THD indicates blame-free 0.01% or less.

Interesting thought.

Something must be responsible for the sound changes heard during tube
rolling exercizes.
I call it magic, but its something real alright.


If one was extraordinarily lucky, a peice of music in the key of C
should sound richer if the tube Fo was a C note, but awful
if the Fo was at some non-musical tone which is 99% more likely.

So there isn't anything wrong with a 12AT7 as long as you pick one that
isn't microphonic, OR NOISY.

That is about it with ANY preamp tube


I seem to be fairly lucky in the noise department, at least as far as
the tubes go.

I do have a 'problem' with noise from the little notebook I'm using as
a sound source to test with. It's definitely power supply related in
the notebook as it vanishes when it's on battery only but I can't make
complete sense of it because it doesn't show up on my other amps, such
as my original 'Plain Jane' 6V6 SE, regardless of the power source.

That needs a bit of clarification. On the guitar amp I split the two
preamp stages and added a switch selectable 'CD level' input. The two
stereo channels are resistor mixed into mono. There was still too much
gain for 'CD level' direct so that's knocked down before going into
the final preamp stage before the phase splitter. So the 'CD Level'
input consists of two series 33ks (the stereo mix) into a 15k to
ground (for the level drop). Then into the switch to the master volume
preceding the last preamp stage.

Ok, so the point is it's not going into the first preamp stage with
gobs of gain so it shouldn't be a case of just more amplification
making the noise louder.

Could be grounding, I suppose, but it's the same for all the amps. I
do have less B+ filtering on the guitar amp, though (well, on the
power tubes. I split the preamp stage off on a separate filter chain
to isolate it from power section B+ droop under load)). The only other
significant thing I can think of, off hand, is Plain Jane has no metal
chassis whereas the guitar amp does. Makes me wonder if all that
wonderful 'shielding' is acting to induce ground noise into the
circuit.

It's not really a problem for the intended purpose as the guitar ain't
got no power supply but I'd like to find it for my own edification.

Btw, if you've ever heard 'PC noise' you'd immediately recognize it as
it mimics PC activity. Hard drive running, mouse movement, etc. So I
know it's coming from there.

Locating noise sources and eliminating the problem is a challenge some
days...
and nights...


Theoretically, a 12AT7 should make an ideal phono imput tube, and
if 3 are paralleled, the Gm total of the 6 resultant triodes
is 24mA/V, some 15 times more than a 1/2 12AX7.
Therefore the noise is reduced by a factor of sq.root of 1/15 or from
about
2uV of the 12AX7 to about 0.55uV, which would in theory
allow MC direct connection without a step up tranny.
But you may find porcine flight is more likely to be possible
before such theoretical noise reductions are seen to be actually real.
I have found all tube inputs for MC to be too noisy
and A FET is the only way to go for low noise.

And the fet has no microphony. But it is much more sensitive to
the slightest magnetic field change so
the case for a phono amp with a fet should be heavy iron.

In the old days of the tube era the guys who serviced the tube gear used
in studios
placed the quietest and least micro tubes at the front ends and the
worst
went into the line level amps and power amps.

That, of course, is what I did. They all work just fine for the phase
splitter.

The SNR at power amps will be better.

Patrick Turner.

Fighting noise, distortion and microphony made work for a lot of ppl.
If you don't work at it, the gear won't perform optimally,
and after you have worked to optimise, the optimisation
does last forever.
Now there's no less work; the complexity is an order of magnitude
greater,
and there are fights over digital ****, as they churn out the
music all processed and packaged like pork sausages....



Patrick Turner.

[Patrick Turner]

Andy Evans wrote:

OK, so you have two DHT triodes in an LTP balanced circuit and one tube
is noisy, and the other ain't. The LTP would simply halve the noise so
you'd get
two phases of the noise at each output.

I'm talking two good tubes (LTP with CCS underneath) , but both prone
to misrophonics in a SE topology.

Have a schematic?

How can SE topology be balanced?



Seems to me the balanced circuit hasn't much to offer on noise
reduction
but does offer much lower THD/IMD and gives CMR.

It's not really noise reduction in the conventional sense - not sure
what it is, but it does seem to work in practice - the tubes behave
themselves and don't scream and howl.

Depends on how they are used.
One can lead them to screaming and howling very easily.
Full moons and vampires are not even required.

Patrick Turner.

[Andy Evans]

wrote:
Two things I've observed abut microphonics:
1. There seems to be microphonics from the natural construction of the
tube, and another kind from a fault condition where it has been dropped
etc.
2. The natural kind of microphonics seems to be less obtrusive in push
pull (balanced)stages than SE. I use DHTs and don't have much of a
problem since my system is balanced.


I don't follow the reasoning there. You hoping both halves inject the
same noise in phase?

I'm not an engineer - I'm basically a hobbyist. I can only report than
I've built about 5 balanced line stages with no real microphonics
problems, and one SE one with large microphonics problems, plus that
two colleagues had the same experience. We're talking exactly the same
valves here, so the valve isn't the variable in this case. I'd be
pleased if anybody could explain.

[Patrick Turner]

Andy Evans wrote:

wrote:
Two things I've observed abut microphonics:
1. There seems to be microphonics from the natural construction of the
tube, and another kind from a fault condition where it has been dropped
etc.
2. The natural kind of microphonics seems to be less obtrusive in push
pull (balanced)stages than SE. I use DHTs and don't have much of a
problem since my system is balanced.

I don't follow the reasoning there. You hoping both halves inject the
same noise in phase?

I'm not an engineer - I'm basically a hobbyist. I can only report than
I've built about 5 balanced line stages with no real microphonics
problems, and one SE one with large microphonics problems, plus that
two colleagues had the same experience. We're talking exactly the same
valves here, so the valve isn't the variable in this case. I'd be
pleased if anybody could explain.

Where is the schematic of what you have built?

Microphony is yet another thing that delivers a spurious
add on signal into the signal path and I can't ( yet ) see
that a balanced circuit would give any protection against
microphony effects unless the resulting spuriae signal
appeared in common mode across the balanced output.

Patrick Turner.

[Patrick Turner]

Locating noise sources and eliminating the problem is a challenge some
days...
and nights...

Yeah. I did improve the guitar input a bit by slightly rerouting the
ground, but it wasn't really a problem to begin with. But it surprised
me as I was investigating the notebook noise. I mean, the guitar input
phone jack is screwed to the case, and that is a metal to metal
ground, yet I could touch the tip of the notebook jack to the sleeve
on the guitar phone jack and hear the dern music (not much but no
question it was there), yet not hear it if I touched it to the chassis
right there at the same place the jack is grounded to it. But it
measures '0' ohms, as near as my meter can read it. There must have
been 'some' fractional ohms somewhere creating an incredibly small
signal but with all that gain... sheesh.

Guitar amps are the land of the primitive.

Lotsa gain and high noise, but guys like to play with some noise
as it gives a certain edge...

If they were to have a real quiet input with a j-fet then
this noise would be astonishingly lower, and then
they'd not know what to think about it....

I had built a separate little two transistor microphone preamp to use
with it but, hell, the dynamic mic worked just fine straight into the
hi gain guitar input.

I'm beginning to think the notebook noise into the 'CD line' input is
simply a matter of gobs more gain than on Plain Jane so it's now
audible. For one, there's no noise from connecting just the grounds
between the two so it's definitely on the signal itself.

Jane is only a couple of watts whereas the guitar amp is 20 and it's
got significantly more gain than needed for the 20 even with knocking
the input down 6 dB. And the CD being quiet would seem to confirm
that.

Theoretically, a 12AT7 should make an ideal phono imput tube, and
if 3 are paralleled, the Gm total of the 6 resultant triodes
is 24mA/V, some 15 times more than a 1/2 12AX7.
Therefore the noise is reduced by a factor of sq.root of 1/15 or from
about
2uV of the 12AX7 to about 0.55uV, which would in theory
allow MC direct connection without a step up tranny.
But you may find porcine flight is more likely to be possible
before such theoretical noise reductions are seen to be actually real.
I have found all tube inputs for MC to be too noisy
and A FET is the only way to go for low noise.

And the fet has no microphony. But it is much more sensitive to
the slightest magnetic field change so
the case for a phono amp with a fet should be heavy iron.

In the old days of the tube era the guys who serviced the tube gear used
in studios
placed the quietest and least micro tubes at the front ends and the
worst
went into the line level amps and power amps.

That, of course, is what I did. They all work just fine for the phase
splitter.

The SNR at power amps will be better.

Yup.

This was a fun project. Now all I need to do is write him an
instruction manual, especially how to bias the thing, and it'll be
done. Well, my part. The brother-in-law is going to do the cabinet
work and buy a proper guitar speaker.

Good luck.

Patrick Turner.

[Andy Evans]

I don't have a schematic I can post (and it's not gain responsible, I
don't think) but I can describe the stages:

Balanced:
Long tail pair with constant current sink under it (see Morgan Jones
p134). 300 ohm grid stoppers, 560k grid leak. Anodes connected push
pull to Lundahl LL1689, B+ (150vDC) to centre tap of primary. Balanced
outputs from secondary into a 27k shunt attenuator. That's it. I've
been using various DHTs like 30, 31, 12A. I get same results with 27k
anode resistors and 1k Russian teflon caps instead of the transformer.

SE:
Same power supply as above. 250k volume control on input, 300 ohm grid
stoppers. 27k anode resistor, can't remember cathode resistor but it
was bypassed with a 47k polypropylene cap. Output caps 1uF Russian
teflon, then 560k to earth on output.

I think that's approx correct. Andy

[Andy Evans]

SE:
Same power supply as above. 250k volume control on input, 300 ohm grid
stoppers. 27k anode resistor, can't remember cathode resistor but it
was bypassed with a 47k polypropylene cap. Output caps 1uF Russian
teflon, then 560k to earth on output.

What's the other half of it doing?

There isn't another half - that's the single ended version with the
microphonics.

[RickH]

flipper wrote:
On Tue, 28 Nov 2006 14:29:54 GMT, Patrick Turner
wrote:


Must be something funky in how they did the grounding because one side
is still earthed at the junction box, the safety is just redundant.
Not to mention they're on the same outlet so they're commoned 'right
there'. Maybe setting up some kind of ground loop with the switcher
spikes on the 'active' ground but the 'safety' being quiet, or some
such mystery. Doesn't make much sense because the amp is still
grounded so if the switcher is bouncing the 'active' ground then it
should be all over the place whether the switcher's safety is there or
not. Either that or they 'safety' grounded some part of the notebook
interior without properly referencing the internal supply back to the
brick. Or a plain ole ground loop inside the thing.

I found it; didn't say I had made sense of it.

snip

If your neutral wire (usually a white wire in USA) is earthed in the J
box, then you have an outlet that is a definite electrocution hazard,
and probably installed by a do-it-yourselfer pretending to be an
electrician. The safety ground should NEVER be used as a return path,
(even though neutral and safety are actually bonded in the main breaker
box or at the meter). In branch circuits only neutral and hot MUST
carry the load, and safety ground must NOT be carrying any load,
otherwise it (safety ground) will not be there in the event current
should begin to flow through a persons body. There are hundreds of
deaths a year due to improperly wired neutrals, especially in
situations where the victim may be barefooted on concrete (garages,
basements, etc.). If at any point in a branch breaker circuit your
neutral wire should become grounded, means you have essentially lost
your safety ground, this is a common blunder. I'd have an electrician
re-wire that outlet and also check out the others.