New Class A, Super-A, Non-Switching : need a revival ?

[tiefbassuebertr]

Originally Posted by StevenOH

The Sansui Tanaka patent is US4401951. It also has the discussion of a few other "nonswitching" patents.
The common mistake with this approach is to leave Power stage with the new type of distortions due to unpredictable nonlinearity of P. transistors currents at the small levels (and near the signal zero crossing!).

orig posted by Steph_tsF:
This the famous Nelson Pass patent, indeed. Thanks 1000 times for reminding us. It is panicking to see those japanese companies (Matsushita, JVC, Pioneer, Sansui) rushing around 1978-1979 for Non-Switching Class AB, while Nelson Pass had it done as soon as Feb 1975, with his US patent granted Nov 1976.

here the link therefore and some others in this case:
Patent US4401951
Patent US6429744
Patent US5019789
Patent US6297699

Steph_tsF: I also had the same thought many times.

[kenpeter]

This was the simplified version of Pass' circuit from his patent.

At first glance, looks very similar to mine. But Pass' load isn't
forced to use only the diode path, so not sure how he keeps
a reactive load from sometimes depleting the emitter reserve?

Whatever else: even in simplest form, its more parts than I'd
ever throw at it. Still want to work up a sim of this one to see
if it might actually perform better in some subtle way I did not
expect???

I think crossing here is still shaped by the log law of emitters?
I'm not entirely sure yet if his quasi-complimetary diodes are
the primary shapers? I do see he sticks yet another diode into
parallel with some of his emitter resistance in Fig 5. Maybe to
do some shaping? Would have been redundant in my scheme...

[steph_tsf]

let's try a summary about the class AB Non-Switching, or class B Non-Switching output stages for Audio Power Amplifiers. We are not including any class AA or class A+ here.

Patents :
Technics New Class A :
JVC Super-A : US4274059
Pioneer Non-Switching : US4215318 and US4254379
Sansui Non-Switching : US4401951
We need to add :
Hitachi Superlinear : US4345215
Pass Non-Switching: US3995228
Dodson Hyperbolic Class A :
Thornton Non-Switching : US4160216
Hawksford Non-Switching :
All patents are downloadable from Patent Searching and Inventing Resources

AES preprints and papers :
Technics New Class A :
JVC Super-A :
Pioneer Non-Switching : AES E-Library: Non-Switching Amplifier
Sansui Non-Switching : AES E-Library: New Biasing Circuit for Class-B Operation
Sansui Non-Switching : AES E-Library: New Biasing Circuit for Class B Operation
Hitachi Superlinear :
Pass Non-Switching:
Dodson Hyperbolic Class A : AES E-Library: Quasi-Class A/Improved Class AB Bias Loop
Thornton Non-Switching : AES E-Library: A Power Class A Technique with Open Loop Stability
Hawksford Non-Switching : AES E-Library: Error Correction and Non-Switching Power Amplifier Output Stages
All AES papers are downloadable (subscription needed) from AES E-Library

Commercial products :
Technics New Class A : SU-V5 SU-V7 SU-V8 SU-V9
JVC Super-A : AX1 AX2 AX3 AX4 AX7
Pioneer Non-Switching : SA7800 SA8800 SA9800 A5 A6 A8 A9 A60 A70 A80
Sansui Non-Switching : AU-D5
Hitachi Superlinear : HA-3700 / HA-4700 (BJTs) and HA-7500 MK2 (MOSFETs)
Pass Non-Switching:
Dodson Hyperbolic Class A :
Thornton Non-Switching :
Hawksford Non-Switching :
Many service manuals are downloadable from Free Manuals User / Service / Schematics for Download | HiFi Engine

Any omission ?
I have not included some techniques used inside opamps (Burr-Brown, Texas Instruments), barely applicable to Audio Power Amplifiers.

Cheers,
Steph

[event horizon]

Originally Posted by steph_tsf

Any omission ?

Cheers,
Steph

Yes, like i mentioned, the very early Technics "hi end" power amplifiers used a class B amplifier to drive a low voltage full current class A output stage Here is an example, the SE-A1, if i remember correctly these things cost many thousands of ££££. Looks like they call it Class A+.

They don't make them like they used to

[steph_tsf]

Originally Posted by StevenOH

I will show one "well forgotten" circuit from the '60s with the interesting properties. It is SRPP (in Cl. A bias), but is Non-Switching in Cl. AB bias. T11 will never switch off. The Non-switching condition for T12 is V(D5)≥0.5V + V(D6)max. D6 should be Schottky (to minimize power waste) with the appropriate current rating and in TO-220 or TO-3P case. Obviously, it is the voltage follower/power buffer and T1 provides the Error Correction function for Pushing and Pulling action of T11, T12. The bias with the shown values: T1, T12 ~0.4 mA, T11 ~10 mA, D6 is (may be) non conducting. D5 is generally not necessary (only for non-switching T12 and Cl. A bias) and can be shorted. C61 provides bootstrapping for R59.

Can somebody derive a stronger version, powered by +42V/-42V and able to deal with 5 amps ? How should we name this, if we succeed ?

[StevenOH]

Trio-Kenwood pat. US4342966.

[steph_tsf]

Originally Posted by event horizon

The very early Technics "hi end" power amplifiers used a class B amplifier to drive a low voltage full current class A output stage Here is an example, the SE-A1, if i remember correctly these things cost many thousands of ££££. Looks like they call it Class A+. They don't make them like they used to

I know. Class A+ and class AA are expensive solutions, with two separate amplifiers for each channel. Over time, the class A+ got replaced by the class AA, where the second amplifier is a total feedback current buffer. You may consider the class AA as an upgraded Quad 405, still with a balancing bridge, but this time it is wideband (four resistors) and there is a total local feedback around the current dumper, which is a dedicated amp with a big open loop gain, instead of a dumb darlington. I think I'll open a few separate threads about Hitachi Dynharmony class G, Technics and Pioneer class A+, Technics class AA, QSC floating supply bridge. This present thread is about fiddling with the Vbe multiplier. Okay, the Thornton scheme is not what we may call a Vbe multiplier fidle, but as the Technics New Class A was already there, I found a nice parallelism between Technics New Class A, and Thornton scheme. They both displace the problem, using fast diodes. All classifications have limitations.

Cheers,
Steph

[event horizon]

No problem Steph I just thought i'd mention it for completeness & the fact that it had a true class A output stage.

Besides sooner or later (depending on how long it takes me to get stuff together) i'll be building a few of similar design 250W+ RMS 8ohm & 500W+ RMS 4ohm. There's plenty to go about

[kenpeter]

Will this do?

[steph_tsf]

Originally Posted by StevenOH

Trio-Kenwood pat. US4342966.

Thanks for the info. Attached there is a .txt file containing the upgraded info.