Reply
 
Thread Tools Display Modes
  #1   Report Post  
 
Posts: n/a
Default Output transformer question

Why is it that running an output transformer into an open load at high
power levels will destroy it, where as running it into a short circuit
load won't ? Many thanks in Advance.
  #2   Report Post  
Phil Allison
 
Posts: n/a
Default




Why is it that running an output transformer into an open load at high
power levels will destroy it, where as running it into a short circuit
load won't ? Many thanks in Advance.




** In a phrase, such operation is capable of generating undamped, high
voltage transients in the primary that can cause insulation break down.

Shorting the secondary damps everything - especially the poor tubes.




........... Phil






  #3   Report Post  
Krzysiek Słychań
 
Posts: n/a
Default

You can use MOV's across the two halves of primary and all surges would be
reduced.


  #4   Report Post  
Phil Allison
 
Posts: n/a
Default


"Krzysiek Słychań"
You can use MOV's across the two halves of primary and all surges would be
reduced.



** MOVs have high and non linear capacitance.

The usual fix is connecting several fast, high voltage diodes in series
from plates to ground.





............ Phil



  #5   Report Post  
Mike Diack
 
Posts: n/a
Default

"Phil Allison" wrote in
:


"Krzysiek Słychań"
You can use MOV's across the two halves of primary and all surges
would be reduced.



** MOVs have high and non linear capacitance.

The usual fix is connecting several fast, high voltage diodes in
series
from plates to ground.


I notice spark gaps on some of the larger modulation transformers I have
kicking around. Cheap and effective.
M


  #6   Report Post  
DougC
 
Posts: n/a
Default

TubeGarden wrote:
Hi RATs!

To prevent damage to OT if speaker comes unhooked while playing loud music,
surround speaker wires with concertina wire

Happy Ears!
Al


Alan J. Marcy
Phoenix, AZ

PWC/mystic/Earhead

What about a much-higher-than-speaker-impedance resistor? 10X? 20X?
  #7   Report Post  
Tom Schlangen
 
Posts: n/a
Default


Hi Alan,

To prevent damage to OT if speaker comes unhooked
while playing loud music, surround speaker wires
with concertina wire


Yeah ;-)

For all those not lucky enough having a stash of NOS
concertina wire, a suitable wattage rating 150 ohms
resistor across the scondary will do. This was
and is good engineering practice with tube amps.

Astonishingly enough, such a "PT protection device"
worth a few cents is missing on most current
production tubed power amps, even very expensive ones.

Maybe one could make a fortune selling generic
wirewounds w/ gold plated leads for $499 incl.
VAT and shipping worldwide, claiming that
special concertina wire was used to make them.
Don't forget to add a "flow direction indicator"
mark on them and a nice leaflet with mumbo jumbo
install instruction, hehehe ...

Tom

--
When in doubt, use brute force.
- Ken Thompson
  #8   Report Post  
JamesG
 
Posts: n/a
Default

That's what I do for my guitar amps. I place a 100 ohm resister across the
output. Cheap insurance.

What about a much-higher-than-speaker-impedance resistor? 10X? 20X?



  #9   Report Post  
Sander deWaal
 
Posts: n/a
Default

"JamesG" top-posted:

That's what I do for my guitar amps. I place a 100 ohm resister across the
output. Cheap insurance.


What about a much-higher-than-speaker-impedance resistor? 10X? 20X?


That's a wise thing to do with any tube amp.
I use 68 ohms/2 watt carbon composite.

Someone who shall remain unnamed for the sake of common decency,
posted another good advice: in each tube anode leg 3 or 4 1N4007
diodes in series with the anode to ground.

--
Sander deWaal
"SOA of a KT88? Sufficient."
  #11   Report Post  
Tim Williams
 
Posts: n/a
Default

"Jon Yaeger" wrote in message
...
What about the tried and tested 10 ohm 1/2W cathode resistor? It's
protected my gear through some wild oscillations . . .


Fine for overcurrent protection, but not on a relatively unstable NFB amp
where the tubes aren't shorting, but oscillating (in fact, cathode current
will drop in class C, i.e., overdriven conditions, due to additional grid
bias).

For SE, you'll need a large amount of zeners, for about 2 x B+ rating. A
MOSFET might also be useful, since those have built-in zeners. Just leave
the gate tied to source with a resistor (and make sure it's an enhancement
mode!).

Tim

--
"I've got more trophies than Wayne Gretsky and the Pope combined!"
- Homer Simpson
Website @ http://webpages.charter.net/dawill/tmoranwms


  #12   Report Post  
Phil Allison
 
Posts: n/a
Default


"Sander deWaal"


Someone who shall remain unnamed for the sake of common decency,
posted another good advice: in each tube anode leg 3 or 4 1N4007
diodes in series with the anode to ground.



** Sure wasn't me - I said to use FAST diodes.

BY227s or BY228s maybe.





............ Phil




  #13   Report Post  
Fabio Berutti
 
Posts: n/a
Default

A very interesting thread. AFAIK when a tube amp is powered with the
speakers unplugged, the output tubes are in fact presented an infinite load.
Consequently, the primary voltage is forced to swing up to very dangerous
levels and ultimately the insulating enamel will crack.
If the output is shorted, the opposite will occur: an enormous current surge
that will ultimately trash the tubes.
I use to provide a 10R, 1/4W cathode resistor, but it only protects from
over-currents (shorted output). HF oscillations can be ruled out connecting
a R-C filter across the tranny's connections to the anodes of the PP pair,
but still the case when a speaker is unplugged is not covered.
I suppose that a good resistor in parallel with the speaker is the easiest
and cheapest solution, which means it is the safest (all safety devices
should be intrinsically simple and failure safe). Now the question is, will
it affect the amp's performance? Would it be a nice idea to use an R-C
bridge tuned at some frequency? I don't think so.
I suppose a plain 9W, 150R wirewound resistor is fine for any amp up to some
100W (about 5W would be dissipated through the R with 8 Ohms boxes). A
bigger one (say 25W thick film type) would be safer, but more difficult to
install.

I'm running to "retrofit" all my units with this "safety device".

Ciao

Fabio


ha scritto nel messaggio
...
Why is it that running an output transformer into an open load at high
power levels will destroy it, where as running it into a short circuit
load won't ? Many thanks in Advance.



  #14   Report Post  
Phil Allison
 
Posts: n/a
Default


"Fabio Berutti"

A very interesting thread.



** Bet you say that to * all * the cute plates of spaghetti you see
........


AFAIK when a tube amp is powered with the
speakers unplugged, the output tubes are in fact presented an infinite
load.



** Not quite infinite - but very high indeed.

The voltage gain from grid to plate goes ballistic.


Consequently, the primary voltage is forced to swing up to very dangerous
levels and ultimately the insulating enamel will crack.



** In simple theory, it cannot swing to more than twice the DC supply.
However, when you factor in the effects of OT leakage inductance, the
various stray OT capacitances and the OP tubes acting like fast switches
when driven hard into a notionally infinite Z ..... bad things do
happen.




................ Phil








  #15   Report Post  
Tim Williams
 
Posts: n/a
Default

"Fabio Berutti" wrote in message
...
A very interesting thread. AFAIK when a tube amp is powered with the
speakers unplugged, the output tubes are in fact presented an infinite

load.
Consequently, the primary voltage is forced to swing up to very dangerous
levels and ultimately the insulating enamel will crack.


Of course, this will never happen with triodes when not overdriven - low
plate resistance restrains it. Pentodes are different and may oscillate
given the standard 20dB or so NFB, but a well-built amplifier which does not
oscillate ought to work fine in any situation...

Tim

--
"I've got more trophies than Wayne Gretsky and the Pope combined!"
- Homer Simpson
Website @ http://webpages.charter.net/dawill/tmoranwms




  #16   Report Post  
John Byrns
 
Posts: n/a
Default

In article , "Phil Allison"
wrote:


** In simple theory, it cannot swing to more than twice the DC supply.
However, when you factor in the effects of OT leakage inductance, the
various stray OT capacitances and the OP tubes acting like fast switches
when driven hard into a notionally infinite Z ..... bad things do
happen.


Your theory is too "simple", real theory says it can swing to more than
twice the DC supply, even without any leakage inductance, me thonks you
need a theory check.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/
  #17   Report Post  
Phil Allison
 
Posts: n/a
Default


"John Byrns"
"Phil Allison"


** In simple theory, it cannot swing to more than twice the DC supply.
However, when you factor in the effects of OT leakage inductance, the
various stray OT capacitances and the OP tubes acting like fast switches
when driven hard into a notionally infinite Z ..... bad things do
happen.


Your theory is too "simple", real theory says it can swing to more than
twice the DC supply, even without any leakage inductance, me thonks you
need a theory check.



** And this "real theory" is to be seen where ??

Coming out your backside all that smoke ?

Considering a PP stage operating with secondary unloaded, plate voltages on
one side have to go negative for the voltage on the other side to exceed
twice the supply.





............... Phil




  #18   Report Post  
John Byrns
 
Posts: n/a
Default

In article , "Phil Allison"
wrote:

"John Byrns"
"Phil Allison"

** In simple theory, it cannot swing to more than twice the DC supply.
However, when you factor in the effects of OT leakage inductance, the
various stray OT capacitances and the OP tubes acting like fast switches
when driven hard into a notionally infinite Z ..... bad things do
happen.


Your theory is too "simple", real theory says it can swing to more than
twice the DC supply, even without any leakage inductance, me thonks you
need a theory check.


** And this "real theory" is to be seen where ??

Coming out your backside all that smoke ?

Considering a PP stage operating with secondary unloaded, plate voltages on
one side have to go negative for the voltage on the other side to exceed
twice the supply.


Exactly, there is nothing to prevent that from happening.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/
  #19   Report Post  
Phil Allison
 
Posts: n/a
Default


"John Byrns"
"Phil Allison"

** In simple theory, it cannot swing to more than twice the DC
supply.
However, when you factor in the effects of OT leakage inductance, the
various stray OT capacitances and the OP tubes acting like fast
switches
when driven hard into a notionally infinite Z ..... bad things
do
happen.

Your theory is too "simple", real theory says it can swing to more than
twice the DC supply, even without any leakage inductance, me thonks you
need a theory check.


** And this "real theory" is to be seen where ??

Coming out your backside all that smoke ?

Considering a PP stage operating with secondary unloaded, plate voltages
on
one side have to go negative for the voltage on the other side to exceed
twice the supply.


Exactly, there is nothing to prevent that from happening.

Regards,



** Fool's logic - there first needs to be something that makes it happen.

Either quit this pathetic cat and mouse stuff and post your info - or
else shut the **** up.




............ Phil


  #20   Report Post  
John Byrns
 
Posts: n/a
Default

In article , "Phil Allison"
wrote:

"John Byrns"
"Phil Allison"

** In simple theory, it cannot swing to more than twice the DC
supply.
However, when you factor in the effects of OT leakage inductance, the
various stray OT capacitances and the OP tubes acting like fast
switches
when driven hard into a notionally infinite Z ..... bad things
do
happen.

Your theory is too "simple", real theory says it can swing to more than
twice the DC supply, even without any leakage inductance, me thonks you
need a theory check.

** And this "real theory" is to be seen where ??

Coming out your backside all that smoke ?

Considering a PP stage operating with secondary unloaded, plate voltages
on
one side have to go negative for the voltage on the other side to exceed
twice the supply.


Exactly, there is nothing to prevent that from happening.


** Fool's logic - there first needs to be something that makes it happen.

Either quit this pathetic cat and mouse stuff and post your info - or
else shut the **** up.


There is something that makes it happen, did you ever hear of inductance?


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/


  #21   Report Post  
Phil Allison
 
Posts: n/a
Default

John Byrns
"Phil Allison"

Exactly, there is nothing to prevent that from happening.


** Fool's logic - there first needs to be something that makes it
happen.

Either quit this pathetic cat and mouse stuff and post your info - or
else shut the **** up.


There is something that makes it happen, did you ever hear of inductance?



** Still no info - just more asinine John Byrns cat and mouse stuff.




............ Phil




  #22   Report Post  
Patrick Turner.
 
Posts: n/a
Default

"Phil Allison" wrote in message ...
"John Byrns"
"Phil Allison"


** In simple theory, it cannot swing to more than twice the DC supply.
However, when you factor in the effects of OT leakage inductance, the
various stray OT capacitances and the OP tubes acting like fast switches
when driven hard into a notionally infinite Z ..... bad things do
happen.


Your theory is too "simple", real theory says it can swing to more than
twice the DC supply, even without any leakage inductance, me thonks you
need a theory check.



** And this "real theory" is to be seen where ??


Unfortunately, my postings don't seem to be appearing on the group
other than on the list on my PC, and the google record
shows that I have no contributions so far to this thread.
So my local server must have problems again with my newsgroup
access, so I will now post a total of the contributions
via the google account, which is slow and awkward to use,
in the hope that you all can see some explanation of the
voltage swing phenomena with unloaded PP pentode amps.

Here is how I have been replying to you all :-


wrote:

Why is it that running an output transformer into an open load at high
power levels will destroy it, where as running it into a short circuit
load won't ? Many thanks in Advance.


An OPT may not necessarily degrade, or necessarily short out
somewhere because of a nil load situation with a large signal.

If the tubes are triode, the plate resistance ( Ra ) of the triodes
provides
enough dynamic resistance between the ends of the primaries
and ground, so if you crank up the input without a load on the sec,
the voltage clips like as if there was a load present.
The positive going anode voltage in an AB amp tends to cause an increase
in the
tube current, even though the grid voltage is going negative, but the grid
swing voltage
is limited because triode gain is low, so you don't get the enormous over
voltage
swings with a triode amp.

But if the tubes are pentode of beam tetrode then the Ra is at least ten
times higher,
and without a load the signal voltage at the anodes does not stop
where you would expect clipping to occur, ie, nearly the same peak swing
as the
B+ voltage. But instead the voltage keeps increasing, perhaps to
3 times the normal maximum swing with a load of a few k a-a.
That is the observed phenomenon.
Just why this occurs is a little mysterious, when you consider what is
going on
with input grid v and plate v.
In a pentode/tetrode without a load, the gain is very high, nearly
equal to the U of the tube, which is about 130 for EL34, so
for a peak swing of -390 v, only +3 applied to the grid is required.
But we see perhaps a peak swing of -1,170 v, so one primary end connected
to the tube turning on
can swing 780 v below 0V.
Obviously the current in the tube has cut off.
The other tube is subject to a positive swing of +1,170 v and its grid is
going negative,
and since with a -v3 swing of only -3v, the anode current is not cut off,
and
it continues to cut off as the -ve grid swing continues.
Once the peak anode swing has reached the same value as the B+ voltage,
and one side of the tranny falls below 0V, the rising voltage side of the
tranny continues
to be controlled by the current turn off of the tube, and the stored
magnetic energy
in the inductance of the tranny releases enough energy to send the voltage
very high
indeed until some limiting factor kicks in, perhaps is arcing in the tube,

or across the insulation between P and S, or because the grid
swing cuts off all the current in the OPT, or between anode pin3 of the
octal socket and the earthy
heater pin2.
Once an arc gets going, there can be a catastrphic release of energy from
the
B+ supply caps, and the arc becomes supported by a rush if DC current and
associated
RF signals, and you have a stuffed OPT.

UL amps lay between triode and pentode amps, and will display some
"overshoot"
without a load.

OPTs should be wound to withstand such high voltages without any bother at
all,
and leaving 3,000 v DC applied to the P for a few minutes should
proove there is not problem with insulation.
If you have 0.7 mm of insulation thickness between P&S, then that's good
for
about 3,000v, and with the enamel of the wire, and varnish present,
the likelyhood of arcing is low.
With amps using a couple of thousand volts for the B+, the OPT
should be treated as a true HV device, with ceramic connectors,
and the whole thing encased in a sealed can containing suitable oil.

However, I have serviced many amps with only 300 v B+ rails,
and they arced. Cracks in insulation occur over the years, varnishing is
imperfectly done,
and water has been allowed to get in, and conductive tracks occur, and a
leakage starts, and stops with decent buzz and a blown supply fuse, or
maybe a fused OPT.
Often this is on the external leads into the tranny, and quite a few
trannies
are salvagable by picking out all the burnt insulation which is carbon,
which is conductive,
with an awl, and simply re-varnishing, and perhaps waxing.
If the tranny has sustained heat damage from being left unattended
with a saturated tube until the tube expired, 1/2 the primary may have
damaged insulation,
and active protection against tube DC ever rising above twice the idle
value should be
incorporated in every amp, and respond to each individual tube.
Older tube amps were often murdered this way, because the chemistry of the
varnish
wasn't too good, and the wire used in the OPTs designed by bean counters
was so thin that excess current soon had it cooking at 300C perhaps.


If I have doubts about an old tranny, I will heat it up to about 90C,
then place it in my vacuum chamber for 10 minutes, and the water
evaporates, and is sucked out of the
core wind up.
Varnish is allowed to flow back into the chamber and the vacuum kept as
hard as my reverse
connected compressor allows, and when the tranny is submerged in varnish,
the
air pressure is allowed back, and it forces the varnish into cracks
wherever it can go.
Baking at 120C for 4 hours cures the varnish. Polyurethane is OK for an
existing
tranny which has been varnished befow.
Waxed trannies can be boiled in oil in a vat for a couple of hours at
about 110C,
and allowed to cool to 60C before removing the tranny.
The moisture should be expelled by then.

Methods to prevent such voltage excursions in pentode amps are rarely
used. But to me its an essential prevention measure, and series diodes
should be connected
from each anode circuit to ground to prevent the anode voltage going
negative below 0V.
Also it is essential to connect R+C networks across the primary, or each
1/2 primary,
to provide a load at HF, because inductive speaker impedance rises with F,
so that at
100 kHz, there is no load on the amp, and output tube gain has risen, with
a certainty
of RF oscillation, and high HF voltages.

See the schematic at

http://www.turneraudio.com.au/htmlwe...00ulabmono.htm

There are 3 x IN4007 in series with 1M R across each to equalise the
voltage
across the diodes, and this gives a 3,000 PIV rating, so the diodes won't
crap out.
They'll only conduct when the anodes go more than about 1.8 v negative.

The owner with this pair of amps will never blow his OPTs.


Patrick Turner.




Mike Diack wrote:

"Phil Allison" wrote in
:


"Krzysiek Słychań"
You can use MOV's across the two halves of primary and all surges
would be reduced.



** MOVs have high and non linear capacitance.

The usual fix is connecting several fast, high voltage diodes in
series
from plates to ground.


I notice spark gaps on some of the larger modulation transformers I have
kicking around. Cheap and effective.
M


Arcing damage can occur in low voltage trannies where the B+ is only 300v.
I seen it happen.

A spark gap is an entirely impractical solution, and the diodes
are better.
Phil says high speed diodes would be used, but I found
slow "garden variety" IN4007 had enough bandwith to stop
excess voltages for the BW involved.

Patrick Turner.


Jon Yaeger wrote:

in article , Sander deWaal at
wrote on 10/8/04 5:15 PM:

"JamesG" top-posted:

That's what I do for my guitar amps. I place a 100 ohm resister across the
output. Cheap insurance.


What about a much-higher-than-speaker-impedance resistor? 10X? 20X?


That's a wise thing to do with any tube amp.
I use 68 ohms/2 watt carbon composite.

Someone who shall remain unnamed for the sake of common decency,
posted another good advice: in each tube anode leg 3 or 4 1N4007
diodes in series with the anode to ground.


What about the tried and tested 10 ohm 1/2W cathode resistor? It's
protected my gear through some wild oscillations . . .


10ohms for Rk won't protect the wild swings of OPT voltages as I pointed out
in a previous post.
The 100 ohms does little better, since its a load of over 10 times the value
normally used.

ONLY the diodes will work as well as they do.

Patrick Turner.

Phil Allison wrote:

"Sander deWaal"


Someone who shall remain unnamed for the sake of common decency,
posted another good advice: in each tube anode leg 3 or 4 1N4007
diodes in series with the anode to ground.


** Sure wasn't me - I said to use FAST diodes.

BY227s or BY228s maybe.


I don't care if someome hinted I was commonly indecent,
since it was I who said IN4007 would be OK.

The shunt capacitance of the OPT is usually enough to
cause the response of the output stage to be -3 dB at say 26kHz if
the tubes were EL34, in pentode with Ra = 12k, and Csh looking into the
primary end
= 500 pF.

So at 260 kHz, the response is down 20 dB, and excessive voltage peaking
seems unlikely,
because 500pF imposes a load = 1.2 kohms.

But use fast diodes if you must.
I have had no troubles with IN4007.

Patrick Turner.



........... Phil


Fabio Berutti wrote:

A very interesting thread. AFAIK when a tube amp is powered with the
speakers unplugged, the output tubes are in fact presented an infinite load.
Consequently, the primary voltage is forced to swing up to very dangerous
levels and ultimately the insulating enamel will crack.


The load isn't infinite, but say you have 40 mA of idle current in the tube
turning off, ie, grid going -ve, then the reduction of current in that half of
the OPT winding
causes the high voltage we see.
At the LF, the load is the Lp of the OPT, and at HF, the load
is the shunt C, and maximum gain
is where the Lp and Csh have the least effect which is at MF, between 400Hz and
2 kHz.


If the output is shorted, the opposite will occur: an enormous current surge
that will ultimately trash the tubes.


Indeed, ultimately.
A KT88 can generate an instantaneous short term turn on current
of 1 amp, so if the turn ration is 12:1 for the 1/2 primary to the output sec,
then you get a max current of 12 amps at the output.
This might immediately fuse a transistor, but not a tube, you need to exceed the
max ratings for
a lot longer before the tube ****es itself.


I use to provide a 10R, 1/4W cathode resistor, but it only protects from
over-currents (shorted output). HF oscillations can be ruled out connecting
a R-C filter across the tranny's connections to the anodes of the PP pair,
but still the case when a speaker is unplugged is not covered.
I suppose that a good resistor in parallel with the speaker is the easiest
and cheapest solution, which means it is the safest (all safety devices
should be intrinsically simple and failure safe).


Forget "good" resistors connected across the load, or permanently
connected internally in the amp. Bad ones are also useless.
The only solution is careful design and protection circuits, excellent
insulation, and thoughtfully designed NFB loops.


Now the question is, will
it affect the amp's performance? Would it be a nice idea to use an R-C
bridge tuned at some frequency? I don't think so.


?,
You're right.


I suppose a plain 9W, 150R wirewound resistor is fine for any amp up to some
100W (about 5W would be dissipated through the R with 8 Ohms boxes). A
bigger one (say 25W thick film type) would be safer, but more difficult to
install.

I'm running to "retrofit" all my units with this "safety device".


150 ohms across the output will do little.

Diodes are the way to go, and no, they will not make any sonic difference to the
sound,
and the extra distortion under conditions right up to clipping
is maybe 0.001%.

6 diodes are cheaper than an OPT.

Patrick Turner.


Phil Allison wrote:

"Fabio Berutti"

A very interesting thread.


** Bet you say that to * all * the cute plates of spaghetti you see
.......

AFAIK when a tube amp is powered with the
speakers unplugged, the output tubes are in fact presented an infinite
load.


** Not quite infinite - but very high indeed.

The voltage gain from grid to plate goes ballistic.

Consequently, the primary voltage is forced to swing up to very dangerous
levels and ultimately the insulating enamel will crack.


** In simple theory, it cannot swing to more than twice the DC supply.


Oh but it can.

Think about the *pentode* output tubes sitting there with 40 mA of idle
current.
One tube turns on, grid goes +ve, and pulls its end of the OPT down to near 0V,
and the grid loses control.
Remember, with no load, the tubes are acting as nearly pure voltage field
effect devices,
with SFA current change.
The other grid turns the other tube off, and when the "on" tube loses contact
with the idle primary current, it becomes disconnected from the amp, and its if
someone
wrenched the tube from its socket.
Meanwhile, the tube with the -ve going grid v and current cut off going on
still has control of its end of the OPT, and current
is cut off futher as the grid signal goes more -ve, so the unloaded end of the
OPT
simply climbs to whatever possible voltage it can.
The grid voltage at idle may be -20v, and to get a 400v peak swing +ve, only
maybe
3.3v grid v is needed because without a load, gain = Gm x Ra = U.
As the grid swings more -ve, the Gm of the pentode reduces, Ra even rises
somewhat,
so to get another +400v anode change, you may need -6v at the grid, and
then a further +400v may need 11v grid change, and then the grid hits grid
current,
and further control of the tube is stopped.
The fixed voltage of the screen prevents the electrostatic feedback effect that
occurs with triodes,
which dictates that a rising anode voltage tends to increase the tube current,
ie, Ra is
10 times lower than a pentode.
Not all triodes are like this; some transmit triodes, ( 833 ? ) have plate
curves
very similar to pentodes, and have enormous gain without a load.
Square waves with no load produce "interesting" voltages.....



However, when you factor in the effects of OT leakage inductance, the
various stray OT capacitances and the OP tubes acting like fast switches
when driven hard into a notionally infinite Z ..... bad things do
happen.


Well yes.
When the OPT is connected to the high pentode ra-a
impedance, the open loop bandwidth is typically from say 150 Hz to 5 kHz,
but at 1 kHz, the the load is indeed a very high value, and the peak unloaded
voltages
can strain hell out of the insulation.

Pentode voltage swings need to be clamped with diodes.
The OPT needs to be designed to take 3,000vrms a-a applied to the
primary with a DC voltage present = 500v, or in the case
of transmitter tubes, a lot more than this capability.
Its actually easy to achieve all this with HV OPTs, if you use oil bath
insulation in sealed
pots, with attention to moisture and pollution prevention.
It used to be routine procedure to thoroughly encase all HV trannies,
and use oil insulant, ( some of which was extremely toxic, like the crap
in HV oiler capacitors .)
Thousands of amp and modulator trannies were done this way,
eg, the trannies at commercial AM transmit stations....

Think of the way car ignition systems have been done;
no problem with high voltages, maybe 20,000, until you touch a live lead
when farnakling with a car.
What really hurts is when you hit the raised bonnet with the back of your
skull,
and that's another dent to pay for.

With OPTs using +500v B+, and plain octal tubes,
normal good design and impregnation is OK; no need for the oil.
But wires to the winding eds should be sleeved where they enter the coils,
and kept apart where voltages of opposite polarity are present.
I have seen arcing occur through insulation between wires from each anode
being too close to each other.

The higher the F, the greater the likelyhood of an AC started arc
at HV.
The use of R+C across each 1/2 of the OPT does not have mucg effect at
audio F, because you don't want to waste power
in R strappeed across the OPT anywhere, including the secondary,
where an R+C is also ideally required.

The R+C networks are solely used to provide some sort of
*resistive* load to the transformer at F above 50 kHz, and *no lower*.
At HF, the effects of the many complex arrangements of parasitic
leakage L, shunt C, and winding self capacitance can result in
series or parallel resonant circuits whose behaviour cannot be entirely
accurately
predicted or modelled on a PC.
Engineers use to working it all out on paper or a PC throw up their hands in
despair
when dealing with OPTs.
I smile at such despair, and simply juggle the *critical damping*
resistance values empirically, until it measures right, with excessive FB
connected,
and is stable with any value of capacitance whatsoever connected across the
output terminals.
Then when the oscillations at HF have been stopped, the NFB is eased off
to the value it is to be under normal op conditions.

As long as no idiot audiophile or nutter goes in with a soldering iron removing

parts he thinks are unnecessary, all will stay stable for the next 50 years.
The non-idiot audiophiles will respect what the technical minded
engineer has prepared for him to enjoy the music.

There is no excuse for building an amp which becomes a bloomin
oscillator under any load condition whatsoever.
Plenty of tube amps will oscillate themselves to death
just by connecting 0.22 uF across the output.
It shows the designer didn't do his homework.


Patrick Turner.






Coming out your backside all that smoke ?

Considering a PP stage operating with secondary unloaded, plate voltages on
one side have to go negative for the voltage on the other side to exceed
twice the supply.





.............. Phil

  #23   Report Post  
Phil Allison
 
Posts: n/a
Default


"Patrick Turner."
"Phil Allison"

** Sure wasn't me - I said to use FAST diodes.

BY227s or BY228s maybe.


I don't care if someone hinted I was commonly indecent,
since it was I who said IN4007 would be OK.



** 1N4007s will conduct fast enough stop the overvoltage problem - but
not switch off fast enough and so fail short due to rapid overheating if
subjected to constant switching operation at high frequencies.

This is *no theory* but proven by tests in a Marshall 100watt ( 4 x EL34)
guitar amp being driven hard into a simulated speaker load at 5 kHz.

BY227 / 228s ran quite cool in the same tests.





............. Phil


  #24   Report Post  
Fabio Berutti
 
Posts: n/a
Default


"Phil Allison" ha scritto nel messaggio
...

"Fabio Berutti"

A very interesting thread.



** Bet you say that to * all * the cute plates of spaghetti you see



Only to the ones prepared by Mom when I manage to get back home, I'm a
terrific cook...



Reply
Thread Tools
Display Modes

Posting Rules

Smilies are On
[IMG] code is On
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Question - Old Jensen Audio Output Transformer for Alnico Speaker Ether Vacuum Tubes 6 April 3rd 12 09:01 PM
Altec 15356a Line Transformer Servin Pro Audio 2 October 12th 04 05:14 PM
Question for Output Transformer Gurus frank Vacuum Tubes 9 January 14th 04 05:08 PM
Dumb output transformer question Ed Vacuum Tubes 2 November 24th 03 10:02 AM
Output Transformer Rewind Question Steve O'Neill Vacuum Tubes 2 October 1st 03 06:59 AM


All times are GMT +1. The time now is 04:02 PM.

Powered by: vBulletin
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Copyright ©2004-2024 AudioBanter.com.
The comments are property of their posters.
 

About Us

"It's about Audio and hi-fi"