Plate-Grid vs. Plate-Cathode local NFB - Is it the same?

[Miles Prower]

I noticed Gary Pimm's Tabor amp uses a 6AU6 LTP DC-coupled to 1624 outputs. Isn't the 1624 very similar to an 807, and therefore to a 6L6GB? (Not 6L6GC, but that would be compatible.) If so, then that's another example of a simple local shunt feedback PP pentode design.

That's a very common misconception: the 1624 definitely is not an 807 with a DH cathode. The 1624 was intended for use in mobile transcievers, hence the DH cathode that could be turned off while listening. It's an RF type intended for the cut-off classes. Look at the specs: at V_PK= 400V and V_GK= 0V, the plate current is ~130mA for the 1624. This allows you to put it into Class B/C with less grid bias. Under the same conditions, the 807 pulls 195mA of plate current, as you would expect from an audio type, which is what the 807 started out as: the RF equivalent of the 6L6. The higher plate current makes Class A possible. Keeping PP 1624 finals in Class *₁ will result in ridiculous voltages -- 600V -- for about 10W of output. You can get that from 6AQ5s running at half the voltage.

This would account for the decisions made for the Pimm design, as that one is operating the finals at V_GK > 0

 

[rayma]

It uses -crossed- shunt feedbacks (a P-P design) to apply POSITIVE feedbacks to the driver loads via bootstrapping.

Also see the Luxman MB3045, with added CFs. https://elektrotanya.com/PREVIEWS/6...reo_differential_tube_amplifier_sch.pdf_1.png
Or the Mac MC275 amp. http://www.kbapps.com/audio/schematics/tubeamps/mcintosh/images/mc275.gif

 

[smoking-amp]

"Also see the Luxman MB3045, with added CFs. https://elektrotanya.com/PREVIEWS/63..._sch.pdf_1.png
Or the Mac MC275 amp. http://www.kbapps.com/audio/schemati...ages/mc275.gif"

Yes, a strong resemblance. With the 50% CFB of the Mac, it probably is not pushing the bootstrapped driver load R into the negative resistance region since bootstrap and cathode signals are the same level. (driver load R behaving like a CCS then)

Goldenbeer's design however is pushing the bootstrap level well beyond just tracking the grid drive level, pushing the driver load line to slant upwards at the HV end (opposite to normal load lines). So negative resistance occurs for the driver load. And that puts plentiful positive feedback into the driver for increased gain. (for the Mac, the load resistor has got nearly constant voltage across it, acting like a CCS, except for small distortion effects)

But then the Goldenbeer design might oscillate from the Pos Fdbk without an over-riding N Fdbk loop around it.

The Goldenbeer approach is interesting because it can remove the usual compressive effect from triodes, or go even further to produce expansive driver behavior (in class A) for compensating a compressive class A output stage. All those (usual) triodes with tilting over plate curves at HV can be fixed up with this. Cheap 6AS7, for instance, can be turned into 300B triodes in effect with this driver compensation effect (3rd harmonic compensation). However, won't work for SE amplifiers unless the OT has a P-P primary to give inverted phase for the P Fdbk.

A pentode driver with (fake/psuedo) neg. R loading might work for SE output directly, using output plate feedback to the driver screen grid instead. Something to explore.

..

 

[Wavebourn]

Thank you kind sirs!

Is "Schade" feedback any kind of NFB around a single stage, whether series or parallel applied?

Series, transformer driven - it was an example in his article. But the same cat can be skinned by multiple ways.

 

[centraltexastubeguy]

Type 1614 is the one which is electrically identical to the 807. My data sheet refers one to the 807 for curves. They are mechanically different.
The confusion regarding the 1624 may come from the statement in the data sheet which says it and the 807 are mechanically similar. The 1624 data sheet refers one to the 1619 for curves. As stated elsewhere in this thread, the 1624 and 807 are very different animals.
Rene