diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Tubes / Valves (http://www.diyaudio.com/forums/tubes-valves/)
-   -   Cathode FB P.P. Strange behaviour. (http://www.diyaudio.com/forums/tubes-valves/122231-cathode-fb-p-p-strange-behaviour.html)

Larry Lomax 30th April 2008 10:50 AM

Cathode FB P.P. Strange behaviour.
 
Hello, I'm making some experiments with local F.B. on an 807 Push Pull amp. Let me introduce it:
Typical two 807 tube PP.
300V on screen grid.
700V on anode.
Fixed bias to obtain 25mA anode quiescent current.
10Kohm anode to anode load.
Resistive 8Ohm load on the secondary.
6SN7 driver after a phase splitter.
Everything works fine, as usual, 50W output.
After some experiments with ultralinear configuration, I changed to cathode feedback, quad/mcintosh style. Everything still worked fine, less power out, some around 35/40W, less distortion, wider band… as expected. The kink starts when output tubes are driven in to the clipping region and beyond. Well, for a start, peaks (triangular wave) are cut. All right, this is to be expected. Going further, rising driving voltage, output begins to square… this is to be expected too. But there is a point when with a little more increase in the drive anode voltage begins to fall, I mean, the almost square wave on the 807 anode lowers as if the anode power supply falls…
The power supply lower, off course, but jus some as 2V. I think this is not the cause of the lowering of about 100V on the anode AC. Is this strange?
Off course, who can be interested in this, who is so mad to drive so further the power stage. This is not all. When the drive is lowered to a point a bit less than clipping the power stage takes about a second to go back to undistorted triangle. And the concern is when full volume is reach and a peak in the music, like a deep bass, pushes the P.P. stage a little over clipping it will take a second to go back to normal functioning.
This doesn’t appear when the stage is in pure pentode operation, neither when it is in ultralinear connection, and neither when a lot of NFB is applied.
Power supply is steady at about 730/700V from 25mA quiescent current to full power a little before clipping. To clipping to the strange behaviour region power lower of about 2V. Anode power supply is from a 250W transformer followed by a choke input filter.
Driver supply is from another 330V transformer and supply the 807 screen grid too. But, this voltage is steady on 330V no visible change. Drivers voltage is filtered by a 2.7Kohm resistor and a 470uF capacitor and there is no evidence on drive voltage change reaching the kink over clipping.
No oscillations of any kind is present, neither motorboating, not before, not before clipping not after the kink.
The output transformer is huge and capable of handling up to 100W and over down to 12Hz. So I discarded any OPT saturation possibility.
Is this a “prerogative” of the cathode feedback output stages or I’m missing something stupid but I’m not capable to see?
Best regards.
Larry.

Miles Prower 30th April 2008 05:43 PM

^^^^

A schematic would certainly help here.

Quote:

When the drive is lowered to a point a bit less than clipping the power stage takes about a second to go back to undistorted triangle. And the concern is when full volume is reach and a peak in the music, like a deep bass, pushes the P.P. stage a little over clipping it will take a second to go back to normal functioning.
Sounds like normal overdrive "farting out".

jon_010101 30th April 2008 09:08 PM

With the cathode feedback, your driver distortion will be higher due to the higher necessary voltage swing. You may be seeing driver stage overload, rather than transformer saturation or power supply sag. In fact, the reduced power may also be a sign of this, as you should be able to get nearly full UL or Pentode power from CFB or McIntosh output stages, respectively. But both require a much higher drive voltage to counteract the negative feedback applied to the cathode.

Seeing the schematic would be helpful -- I'm also especially curious what transformer you are using, as it must be quite good!

Larry Lomax 5th May 2008 09:36 AM

Hello and thanks for answers.

Well... last night it cost me an 807 tube with a full anode red and the full valve turned out to blue glowing. Very pretty, but it seems the 807 didn't like it so much.

It appears to be something like a knee after clipping where the tube dissipate a lot of power to nearly destroy it, but I can’t figure out why. In the same circuit, same transformer without any cathode F.B. as the same to say pure class AB push pull, the most classical circuit, same quiescent current, about 20/30 mA, the stage enters clipping and going further power increase, and the wave on the load squares out more and more as usual.
With cathode F.B. it appears to be a real “farthing out” at large.
Driving is assured with more than good headroom by a 6SN7 which seems not to care about the load of the 807 grid, the 270Khom bias resistor and the 4.7Kohm stopping resistor in series with the grid. I attempt with different values just to see what happens, but, off course, the strange behaviour still persists.

I made the transformer, as I use to made for PP amps, this time I added a secondary to make tests connecting it in an ultra linear fashion. I really didn’t like it so much so I changed to cathode FB connection in 1:10. The transformer is made like the Williamson with many interleaving windings, split in two sections. The transformer is a pretty big one, with 400H open secondary impedance with a capacity of driving a load up to 75W at 12Hz, with an impedance of about 9500 Ohms anode to anode explicitly made for the 807 as datasheets require.
A note must be said, my tubes are NOS made in the USA by GE and are made in the 40es 50es but the characteristics are almost as those depicted in the datasheets.

Since 1974 I made a lot of amps based on 807’s, EL34’s, KT88, 300B’s, for radio and audio but I’m a first timer with UL’s and CF… and this is my first try…:whazzat:

Thanks for help and suggestions, cheers
Larry.


All times are GMT. The time now is 09:02 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2