Tube failure mode: gassy?
 

OK, I'm trying to understand the failure of a 807 beam tetrode in a
homebrew push-pull audio amp.

Tubes are GE, about 45 years old, bought as new old stock several weeks
ago.

Being run with 400V on the plate and 255V on the screens.

Both behaved very nicely for a few weeks in my amp, and plate current
at idle followed the curves in the books just fine. They were
well-balanced.

But tonight under some listening stress the B+ fuse blew.

Turning the bias (supplied from a bias supply through a 100K resistor)
down to -40V, the good tube is drawing a small amount of plate current
while the other draws 40mA after warmup and current goes up and up and
up until it hits 100mA or so and I turn the amp off before blowing
another fuse. Check the grid current and as the bad tube warms up grid
current is increasing and increasing as well, to the point where the
-40V is being entirely dropped in the 100K resistor (I guess that means
about half a mA of grid current).

The tube with high current always had some violet-blue spots that
danced on the top but they were on the glass and mica so I figured that
was normal but now I don't know. There was also this strange orange
spot that was on a mica insulator there too.

Is this classic "gassy" tube failure? Towards the end it seemed to be
slightly more blueish on the top.

I've burnt up/blown up a lot of 6146 and other variant beam tetrodes
over the years but now I'm trying to get some understanding of what's
happening :-).

Will a certain chunk of NOS 807's simply be this way from age?

Tim.

Tube failure mode: gassy?
 

wrote:

Is this classic "gassy" tube failure? Towards the end it seemed to be
slightly more blueish on the top.

What condition was the getter in ?

Graham

Tube failure mode: gassy?
 

Pooh Bear wrote:
wrote:

Is this classic "gassy" tube failure? Towards the end it seemed to be
slightly more blueish on the top.

What condition was the getter in ?

Dark metallic. Not whitish, not even on the edges.

Just warmed it up again, and looked for interelement shorts. No shorts,
just lots of grid and plate current when negative grid bias and B+ is
applied, it seems to "run away" in fact (after 5-10 seconds of bias/B+
the currents start taking off steeply). No oscillations/parasitics on
the scope.

I actually have another GE 807 that's been in my junk box for many
years (presumed "junk"), and it has a similar getter and similar excess
currents (actually worse).

One thing I want to eliminate as a possibility is that something in my
homebrew amp did this tube in after just a few dozen hours of service:
390V on the plate, 255V regulated on the screen, about -23V on the grid
to give 38mA idle current, all following the classical AB1 parameters
in the 40's vintage tube manuals. This is MUCH relaxed compared to
typical ham transmitter usage I've done with 807/6L6GC/6146's in the
past (which is 700 to 750V B+, 80 or 90mA plate, class C so substantial
grid current, etc.) but that was ICAS.

Tim.

Tube failure mode: gassy?
 

wrote:

Pooh Bear wrote:
wrote:

Is this classic "gassy" tube failure? Towards the end it seemed to be
slightly more blueish on the top.

What condition was the getter in ?

Dark metallic. Not whitish, not even on the edges.

Just warmed it up again, and looked for interelement shorts. No shorts,
just lots of grid and plate current when negative grid bias and B+ is
applied, it seems to "run away" in fact (after 5-10 seconds of bias/B+
the currents start taking off steeply). No oscillations/parasitics on
the scope.

I actually have another GE 807 that's been in my junk box for many
years (presumed "junk"), and it has a similar getter and similar excess
currents (actually worse).

One thing I want to eliminate as a possibility is that something in my
homebrew amp did this tube in after just a few dozen hours of service:
390V on the plate, 255V regulated on the screen, about -23V on the grid
to give 38mA idle current, all following the classical AB1 parameters
in the 40's vintage tube manuals. This is MUCH relaxed compared to
typical ham transmitter usage I've done with 807/6L6GC/6146's in the
past (which is 700 to 750V B+, 80 or 90mA plate, class C so substantial
grid current, etc.) but that was ICAS.

If it's 'friend' hasn't expired I can't really see how your amp can be at fault.

Graham

Tube failure mode: gassy?
 

Tube failure mode: gassy?
 

You should also check to see if the coupling cap from the driving stage
is leaky.

Tube failure mode: gassy?
 

flipper wrote:

On 17 Jun 2006 05:42:10 -0700, wrote:

One thing I want to eliminate as a possibility is that something in my
homebrew amp did this tube in after just a few dozen hours of service:
390V on the plate, 255V regulated on the screen, about -23V on the grid
to give 38mA idle current, all following the classical AB1 parameters
in the 40's vintage tube manuals. This is MUCH relaxed compared to
typical ham transmitter usage I've done with 807/6L6GC/6146's in the
past (which is 700 to 750V B+, 80 or 90mA plate, class C so substantial
grid current, etc.) but that was ICAS.

I've got a 6EM7 that does the same thing and it's visual appearance is
perfect as well so I doubt it's your circuit.

Ironically, it turned out to be fortuitous in my case because I was
testing a circuit design intended to force PP current balance and that
tube was a great test subject. Dern thing works even though it would
run away on it's own.

A servo presumably ?

Graham

Tube failure mode: gassy?
 

wrote:
Pooh Bear wrote:

wrote:


Is this classic "gassy" tube failure? Towards the end it seemed to be
slightly more blueish on the top.

What condition was the getter in ?


Dark metallic. Not whitish, not even on the edges.

Just warmed it up again, and looked for interelement shorts. No shorts,
just lots of grid and plate current when negative grid bias and B+ is
applied, it seems to "run away" in fact (after 5-10 seconds of bias/B+
the currents start taking off steeply). No oscillations/parasitics on
the scope.

I actually have another GE 807 that's been in my junk box for many
years (presumed "junk"), and it has a similar getter and similar excess
currents (actually worse).

semi-dumb question: where are you measuring the bias voltage? (directly
at the tube, or on the other side of the grid resistor?)

I'm wondering if a bad coupling cap is causing the bias at the tube to
drift towards 0v or even positive.
--
Doug Smith W9WI
Pleasant View (Nashville), TN EM66
http://www.w9wi.com

Tube failure mode: gassy?
 

Doug Smith W9WI wrote:
semi-dumb question: where are you measuring the bias voltage? (directly
at the tube, or on the other side of the grid resistor?)

I'm wondering if a bad coupling cap is causing the bias at the tube to
drift towards 0v or even positive.

Well, I WANT the bias to be circa -20V or -25V. There's a negative
supply and some pots which form a stiff voltage divider and supplies
grid bias through a 100K resistor.

On the "bad" tube grid current is so high that it does indeed drift up
to 0V in a minute or so. Yeah, I know, there's not supposed to be grid
current until the grid goes positive. I suspect this is the root of the
problem.

The coupling capacitors are indeed good and moving the tube around it
follows the tube.

Tim.

Tube failure mode: gassy?
 

wrote:

OK, I'm trying to understand the failure of a 807 beam tetrode in a
homebrew push-pull audio amp.

Tubes are GE, about 45 years old, bought as new old stock several weeks
ago.

Being run with 400V on the plate and 255V on the screens.

Both behaved very nicely for a few weeks in my amp, and plate current
at idle followed the curves in the books just fine. They were
well-balanced.

But tonight under some listening stress the B+ fuse blew.

Turning the bias (supplied from a bias supply through a 100K resistor)
down to -40V, the good tube is drawing a small amount of plate current
while the other draws 40mA after warmup and current goes up and up and
up until it hits 100mA or so and I turn the amp off before blowing
another fuse. Check the grid current and as the bad tube warms up grid
current is increasing and increasing as well, to the point where the
-40V is being entirely dropped in the 100K resistor (I guess that means
about half a mA of grid current).

The tube with high current always had some violet-blue spots that
danced on the top but they were on the glass and mica so I figured that
was normal but now I don't know. There was also this strange orange
spot that was on a mica insulator there too.

Is this classic "gassy" tube failure? Towards the end it seemed to be
slightly more blueish on the top.

I've burnt up/blown up a lot of 6146 and other variant beam tetrodes
over the years but now I'm trying to get some understanding of what's
happening :-).

Will a certain chunk of NOS 807's simply be this way from age?

Tim.

I have very little direct tube experience, but from what I've read it
sounds like the grid is emitting. Apparently this happens when some of
the oxide coating from the cathode gets onto the grid, which gets hot
and emits electrons, which tends to pull the grid more positive. The
current flow on the grid heats it up more, which makes things worse, and
it runs away.

In all the old tube books this is a result of mistreating the tube --
running so much cathode current that the oxide vaporizes a bit and
condenses on the grid (although it may also happen from positive ion
bombardment -- hmm). But who knows what 50 years of sitting in a drawer
(or getting carted around) will do?

'Regular' grid current happens when the grid goes positive and collects
electrons, tending to pull the grid more negative -- this is why you can
self-bias a class C amplifier or oscillator with a resistor to ground.

I have no idea of how to actually test this theory, or what other
mechanisms would cause this problem. The best I can do is mention that
while 807's are normally $18 at Antique Radio Supply they're on sale
right now at $10, and 1625's are still a better buy at $5.50.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html