"screens are an impractical load for that configuration"
Yeah, probably need to put in some followers of some kind. However, if the UL taps turn out to be close to the correct voltage swings required for linearity, then the UL tap, LTP plate resistor, mod might be workable without the followers.
Don
Yeah, probably need to put in some followers of some kind. However, if the UL taps turn out to be close to the correct voltage swings required for linearity, then the UL tap, LTP plate resistor, mod might be workable without the followers.
Don
"first disclosure by Walker of his circuit involving the cathode in the output transformer. (QuadII)"
How about partial cathode feedback (CFB, Quad II) plus ultralinear (UL) plus Circlotron alltogether:
http://www.diyaudio.com/forums/showthread.php?threadid=19624&highlight=
Then we can add "More Than UltraLinear" output error feedback (above posts) or Hawksford style error correcting output:
http://www.diyaudio.com/forums/showthread.php?threadid=35001&highlight=
for "Beyond All Things Linear" performance.
By the way, Hawksford cites Llewellyn, 1941, patent # 2,245,598 as the origin of the error correction idea (done in tube technology).
http://patft.uspto.gov/netahtml/srchnum.htm
Don
How about partial cathode feedback (CFB, Quad II) plus ultralinear (UL) plus Circlotron alltogether:
http://www.diyaudio.com/forums/showthread.php?threadid=19624&highlight=
Then we can add "More Than UltraLinear" output error feedback (above posts) or Hawksford style error correcting output:
http://www.diyaudio.com/forums/showthread.php?threadid=35001&highlight=
for "Beyond All Things Linear" performance.
By the way, Hawksford cites Llewellyn, 1941, patent # 2,245,598 as the origin of the error correction idea (done in tube technology).
http://patft.uspto.gov/netahtml/srchnum.htm
Don
Perhaps the original 6L6 circuit I was rumbling on about with high thd in UL was the fact that the 6L6 family aren't kinkless tetrodes whereas the KT & EL family is.......perhaps this is the conflict with the distortion digression ?
I'm implying that the KT90 used in an UL circuit doesn't behave as a kinkless penny hence it's orig home in frame output circuits and not really in audio.
any digression ??
richj
I'm implying that the KT90 used in an UL circuit doesn't behave as a kinkless penny hence it's orig home in frame output circuits and not really in audio.
any digression ??
richj
Hmmm, I thought the 6L6 WAS the 1st of the Kinkless tetrodes?
Here's an update on "more than UltraLinear"
T2a and T2b can be wound on the same core, maybe a ferrite E core. "Foll." is the location for possible emitter followers if needed for handling screen currents. But, assuming the UL taps provide close to the correct screen voltages for linearity, the turns ratio on T2a and T2b could be set up so V3 and V4 do not need to provide much current. Zeners Z1 and Z2 provide for operation with screen voltages below B+.
I added the T2a and T2b xfmr. so as to get rid of the cross coupling in the earlier version. This way the circuit can operate in class aB without feedback having to cross thru the output xfmr. when one output tube is cut off. (Well, some will still, due to the other side of the LTP) Hopefully, providing better linearity and bandwidth in the feedback loop this way. (I suppose one could eliminate the cross xfmr. effects altogether by just AC grounding the common cathode point of the LTP) Again, the resistors R1 and N*R1 are set up for an anticipated gain of N for the output tubes V1 and V2. (for operation without T2a and T2b, the zeners get changed back to resistors and the LTP plates to V1/V2 screen connections get crossed, and emitter followers "foll." will probably be needed.)
Don
Here's an update on "more than UltraLinear"
T2a and T2b can be wound on the same core, maybe a ferrite E core. "Foll." is the location for possible emitter followers if needed for handling screen currents. But, assuming the UL taps provide close to the correct screen voltages for linearity, the turns ratio on T2a and T2b could be set up so V3 and V4 do not need to provide much current. Zeners Z1 and Z2 provide for operation with screen voltages below B+.
I added the T2a and T2b xfmr. so as to get rid of the cross coupling in the earlier version. This way the circuit can operate in class aB without feedback having to cross thru the output xfmr. when one output tube is cut off. (Well, some will still, due to the other side of the LTP) Hopefully, providing better linearity and bandwidth in the feedback loop this way. (I suppose one could eliminate the cross xfmr. effects altogether by just AC grounding the common cathode point of the LTP) Again, the resistors R1 and N*R1 are set up for an anticipated gain of N for the output tubes V1 and V2. (for operation without T2a and T2b, the zeners get changed back to resistors and the LTP plates to V1/V2 screen connections get crossed, and emitter followers "foll." will probably be needed.)
Don
Attachments
smoking-amp said:
Apologies......youre right
. I can only presume that the screen grid electrode /beam plate spacing is the critical parameter as when the tubes are operated as triodes then the thd is reduced.....because there is no internal feedback within the tube. I find a conflict in the Mullard EL34 20watt amp UL circuit....to what is being discussed.....
From a text written in 1960; Quote << With this particular screen grid ratio it has been found that improved linearity is obtained at power levels above 15W when resistors of 1 K are inserted in the screen grid connections......the consequent slight power reduction in the peak power is negligible ........It will be seen that excellent linearity is obtained up to 20V i.e 27W output ........under optimum conditions the UL design can deliver about 65% of the maximum o/p of the pentode circuit but with considerably less distortion ......at the same time the internal resistance is reduced to the level of a triode output stage >>. End quote...That last bit is the strongest arguement in favour of UL but that high value (1K) in the EL34 g2 to me rules out the idea of current sourcing.
richj
"That last bit is the strongest arguement in favour of UL but that high value (1K) in the EL34 g2 to me rules out the idea of current sourcing."
Not sure what you mean by current sourcing here. The 1K in the EL34 g2 would seem to me to indicate they effectively reduced the % screen tap somewhat by screen current induced voltage drop across the 1K. Perhaps just fine tuning the tap? There would be some non-linear effect from the 1K near plate saturation (low plate voltage) when screen current picks up faster than plate current, but that would seem to just make the plate saturation non-linearity worse, so probably not the reason for the 1K.
Don
Not sure what you mean by current sourcing here. The 1K in the EL34 g2 would seem to me to indicate they effectively reduced the % screen tap somewhat by screen current induced voltage drop across the 1K. Perhaps just fine tuning the tap? There would be some non-linear effect from the 1K near plate saturation (low plate voltage) when screen current picks up faster than plate current, but that would seem to just make the plate saturation non-linearity worse, so probably not the reason for the 1K.
Don
Even my thoughts about that high value 1K g2 resistor...different tubes i.e Kt88 recom typ 270R UL fixed bias and 100R cathode bias for roughly same a-a impedance..Interesting that swapping tubes in the same circuit config doesn't effect Pout....i.e a 6550A for a KT88. Most of us know this....In my appl notes for 6550A, the g2 res can be omitted......I don't think there is any fine UL tap tuning, except the reduction of the leakage inductance as the UL tap is moved towards the full primary.
The metastable space charge going on inside an o/p tube is complex enough. .....someone from the past had worked out what was going on in an UL comfiguration..slide rules (slow.. painful process).....not just trial and error with the grids shaped with a snips and pliers.....
I came across a salesman from a tube company who said < we busted an 88 open and measured everything....the interelectrode gaps etc.........and when they potted a clone if behaved completely different ? >
As I see it, it's back to the physics books....if there are any.Any clue of what's going on in the radiotron handbook ?
richj
The metastable space charge going on inside an o/p tube is complex enough. .....someone from the past had worked out what was going on in an UL comfiguration..slide rules (slow.. painful process).....not just trial and error with the grids shaped with a snips and pliers.....
I came across a salesman from a tube company who said < we busted an 88 open and measured everything....the interelectrode gaps etc.........and when they potted a clone if behaved completely different ? >
As I see it, it's back to the physics books....if there are any.Any clue of what's going on in the radiotron handbook ?
richj
Afaik, the only purpose for that nice screen resistor is to limit the screen current to a level that will not literally melt the screen.
As I mentioned before there is a certain California based (two of them actually) tube amp company that claims 75 watts from a pair of EL34s... which they get at the expense of the tube's life. Reads nice on the spec sheet tho'....
If you reduce the screen resistor to a manageable level, you find that the power drops back down to a more manageable 50-60 watts (max).
On the EL34 you can often peak through a slot into the innards - at least you can on some of the Russian types - you can watch the temperature/color of the screen grid under load.
The phrase "flash bulb" comes to mind.
Btw, a common cause of failure for the 6146 tube in RF service is excessive screen current (visible flash!).
As far as "omitting" the series resistor for the 6550A? I doubt that will work in practice... maybe if they wound the screen grid with 20ga.?
_-_-bear
As I mentioned before there is a certain California based (two of them actually) tube amp company that claims 75 watts from a pair of EL34s... which they get at the expense of the tube's life. Reads nice on the spec sheet tho'....
If you reduce the screen resistor to a manageable level, you find that the power drops back down to a more manageable 50-60 watts (max).
On the EL34 you can often peak through a slot into the innards - at least you can on some of the Russian types - you can watch the temperature/color of the screen grid under load.
The phrase "flash bulb" comes to mind.
Btw, a common cause of failure for the 6146 tube in RF service is excessive screen current (visible flash!).
As far as "omitting" the series resistor for the 6550A? I doubt that will work in practice... maybe if they wound the screen grid with 20ga.?
_-_-bear
...but that's not an UL circuit then...
what I meant is that attempting to "squeeze" the maximum power output from an amp in any mode with a pentode tube usually forces the screen to be run in excess of the power level that is *best* in terms of tube life/non-flaming, sparking.
...in specific, the current crop of import tubes won't handle as much *juice* as NOS 6CA7s or EL34s once did - so pushing the limits results in short tube life or failure, usually failures. So, a "manageable B+" voltage is indicated. ie. *less power output* rather than more.
but feel free to run ur tubes however you wish...
_-_-bear
what I meant is that attempting to "squeeze" the maximum power output from an amp in any mode with a pentode tube usually forces the screen to be run in excess of the power level that is *best* in terms of tube life/non-flaming, sparking.
...in specific, the current crop of import tubes won't handle as much *juice* as NOS 6CA7s or EL34s once did - so pushing the limits results in short tube life or failure, usually failures. So, a "manageable B+" voltage is indicated. ie. *less power output* rather than more.
but feel free to run ur tubes however you wish...
_-_-bear
Hi..there...
this snip makes nice bedtime reading.
http://www.webace.com.au/~electron/tubes/screens.htm
richj
this snip makes nice bedtime reading.
http://www.webace.com.au/~electron/tubes/screens.htm
richj
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.