Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Tubes / Valves All about our sweet vacuum tubes :) Threads about Musical Instrument Amps of all kinds should be in the Instruments & Amps forum

Local feedback between grid-cathode
Local feedback between grid-cathode
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 22nd July 2019, 12:42 AM   #31
smoking-amp is offline smoking-amp  United States
diyAudio Member
 
smoking-amp's Avatar
 
Join Date: Dec 2001
Location: Hickory, NC
Plate to grid N Fdbk receives another hidden benefit beyond what the usual N Fdbk and loop gain formula would indicate. This comes from the sensitivity of the N Fdbk injection point tracking (non-linearly) the non-linearity of the distorting device in such a way that further correction is achieved for free. When the device current goes up, its gm increases, causing the output Rp to decrease but also the input sensitivity to correction to increase similarly.

Output stage CFB also achieves this extra benefit.

For a TWO stage (driver and output stages) P-P design, with N Fdbk around them both ("local" ie, not passing thru the OT) a similar benefit can be obtained by "crossed" N Fdbks back to the driver grid1's. (or even to the driver grid2's if precautions are included for grid 2 impedance variation)

An example of this is ThorstenL's amplifier design (below) (Which incidentally lowers the input impedance to the driver stage grids, similar to shunt Schade. So a current driver pre stage is needed to drive it. )

Push pull from nose to tail...distortion signature

Crossed N Fdbks have often been touted as cancelling odd harmonics better
than the usual non crossed N Fbks (back to driver cathodes). This can be true IF the scaling of the driver is correctly done. The driver must be operating over a similar portion of a similar non-linear device curve as the output device is. (usually not designed specifically to do this, but could be done) The driver needs to have similar curvature features as the output device obviously.

If this scheme could be done perfectly, (matched similar driver and output tubes and matching non-linear regions of operation) then the output stage distortion could be completely corrected with a -finite- amount of N Fdbk. Practical considerations mean this only partially approaches that ideal, but can exceed the N Fdbk - loop gain formal expectations. A conventional "linear" global N Fdbk around all can iron out the remaining residual distortion.
Attached Images
File Type: jpg ThorstenL_dc shuntfb.jpg (62.6 KB, 142 views)

Last edited by smoking-amp; 22nd July 2019 at 12:56 AM.
  Reply With Quote
Old 22nd July 2019, 02:22 AM   #32
smoking-amp is offline smoking-amp  United States
diyAudio Member
 
smoking-amp's Avatar
 
Join Date: Dec 2001
Location: Hickory, NC
addendum:

Making the driver stage distort similarly (and with crossed N Fdbks) like the output stage also injects further distortion into the forward path. Not good.

But if the input signals to the driver stage were to be inserted at the more linear driver screen grids, then the N Fdbks to grid 1 would still get the enhanced correction for fixing the output stage (tracking non-linearity) while the input signal would not be further distorted by the drivers.

This leads to an amplifier that would likely need pentodes (or similar, cascode, Mu follower ...) every where for gain (well, except not necessarily for the outputs), which would be rather unusual.

I haven't ever considered this topology before, although I see someone used a screen driven driver for an 845 amplifier:

Try A Screen Driven Driver Stage Article By Rikard Berglund From Sound Practices Issue 8, Winter 1994/1995

Makes me think: 6GF5 drivers, 6GE5 outputs. Matched non-linearities. (should work for most any TV Sweep tube outputs, they all are just scaled up versions of 6DQ6, 6GE5/6JN6 anyway. )

Last edited by smoking-amp; 22nd July 2019 at 02:42 AM.
  Reply With Quote
Old 22nd July 2019, 02:46 AM   #33
gabdx is offline gabdx  Canada
diyAudio Member
 
Join Date: Jan 2008
magic, looks like a pentode on your schematic, so you don't have an UL transformer?

With UL there is absolutely no reason to connect the grid of the output to its anode with feedback.

PRR said it too, if you want to waste some gain and linearize the output with local feedback, if UL is not enough or not a possibility, you use cathode feedback through a dedicated winding or through the 16 ohm tap if you are in SET. In push pull you need dedicated CFB windings.

Usually GNF is the best solution, but the CFB helps the output drive Z, you reduce GNF by the same ratio...

Scenario, the amp works with 15 db feedback, you install a CFB transformer with 4 db feedback, you reduce GNF by 4 db.
__________________
amplificateurslegrand.ca

Last edited by gabdx; 22nd July 2019 at 02:59 AM.
  Reply With Quote
Old 22nd July 2019, 02:56 AM   #34
smoking-amp is offline smoking-amp  United States
diyAudio Member
 
smoking-amp's Avatar
 
Join Date: Dec 2001
Location: Hickory, NC
If you look at the curves SpreadSpectrum posted for KT88 shunt Schade (post # 27) versus some UL curves, there is no comparison for linearity. Shunt Schade beats UL hands down, as well as typical $$$ triodes.

Shunt Schade and CFB have the advantage of tracking non-linearity in the N Fdbk path. (series Schade too, need an interstage xfmr though)

Triodes have 3/2 power law at the plate versus 4/2 power law at grid 1. Same thing for UL, 3/2 power law at the screen grid, 4/2 power law at grid 1. Tracking errors.

Look at the 6HB6 curves below. 1st pic is g2/g1 triode (like a real triode or UL mode), and the 2nd pic is shunt Schade. Same tube.
Attached Images
File Type: jpg 6HB6_Triode.jpg (41.3 KB, 136 views)
File Type: jpg 6HB6_Schade.jpg (45.0 KB, 81 views)

Last edited by smoking-amp; 22nd July 2019 at 03:07 AM.
  Reply With Quote
Old 22nd July 2019, 03:20 AM   #35
Wavebourn is offline Wavebourn  United States
Anatoliy Lisovskiy
diyAudio Member
 
Wavebourn's Avatar
 
Join Date: Sep 2006
Location: Pleasant Hill, CA
Local feedback between grid-cathode
Quote:
Originally Posted by gabdx View Post
With UL there is absolutely no reason to connect the grid of the output to its anode with feedback.
.
Vice verse.
If input goes to control grid, feedback to screen grid, "correction" is much worse that if both input and feedback go to the same grid. Transfer functions when controlling by different grids are different, they do not match.
__________________
!!!Warning!!!
Single Ended Class-A-Holism is addictive!!!
  Reply With Quote
Old 22nd July 2019, 05:07 AM   #36
SpreadSpectrum is online now SpreadSpectrum  United States
diyAudio Member
 
SpreadSpectrum's Avatar
 
Join Date: Nov 2007
Yep, it is best to correct the nonlinearity in the electrode from which it originated. The result is far less distortion. That's why UL curves don't look anywhere near as good as plate-grid feedback curves!
__________________
My Blog: http://tubeswithatwist.blogspot.com/
  Reply With Quote
Old 22nd July 2019, 10:01 AM   #37
LKA is offline LKA  Europe
diyAudio Member
 
LKA's Avatar
 
Join Date: Dec 2014
I've tried to compare CFB and SCHADE OPS in simulator. Everything is on the picture.
Attached Images
File Type: png OPS-KT88-CFBvsSCHADE.png (99.2 KB, 138 views)
  Reply With Quote
Old 22nd July 2019, 11:12 AM   #38
savu is offline savu  Ireland
diyAudio Member
 
savu's Avatar
 
Join Date: Feb 2007
Location: Dublin, Ireland
Send a message via MSN to savu
Hi LKA,

Have you tried 100% CFB and triode connected KT88 or 6C33C? Drive voltage would be much higher but you get full benefits of the much lower output impedance.

Silviu
__________________
1'st rule of chess:
The only way to get smarter is by playing a smarter opponent.
  Reply With Quote
Old 22nd July 2019, 11:59 AM   #39
LKA is offline LKA  Europe
diyAudio Member
 
LKA's Avatar
 
Join Date: Dec 2014
Hi, no, I've never tried.
CFB+TRIODE, output impedance is the lowest but distortion not.
Attached Images
File Type: png PP CFB OPS KT88.png (163.3 KB, 92 views)
  Reply With Quote
Old 22nd July 2019, 02:42 PM   #40
SpreadSpectrum is online now SpreadSpectrum  United States
diyAudio Member
 
SpreadSpectrum's Avatar
 
Join Date: Nov 2007
I built a 100% CFB (follower) SE amp on the bench with and Edcor output transformer and got a 0.6 Ohm output impedance, real-world measurement. Of course, I used a Mosfet as the output device but I don't think a tube would have made that much higher. Once you make a follower amp with an output transformer, the copper losses in the transformer usually dominate the resulting output impedance. The output transformer copper losses account for 0.53 Ohms of that 0.6 Ohms.

I've published driver designs on my blog that can drive follower output stages to clipping with very low distortion.
__________________
My Blog: http://tubeswithatwist.blogspot.com/

Last edited by SpreadSpectrum; 22nd July 2019 at 02:46 PM.
  Reply With Quote

Reply


Local feedback between grid-cathodeHide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

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

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
LTP feedback return. Cathode or grid? Elerion Tubes / Valves 8 15th May 2019 11:31 AM
Control Grid Feedback (as opposed to Cathode FB) and Miller Capacitance ballpencil Tubes / Valves 10 18th November 2015 12:07 AM
Partial feedback (anode grid) with cathode follower swordsman Tubes / Valves 6 9th July 2011 06:29 AM
Neg-Feedback into Grid of gain stage instead of cathode? Alastair E Tubes / Valves 3 25th January 2011 10:19 AM
300B SE cathode feedback to grid? alan_982002 Tubes / Valves 3 29th March 2010 01:04 PM


New To Site? Need Help?

All times are GMT. The time now is 05:59 PM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Copyright ©1999-2019 diyAudio
Wiki