high votage output stage

[latala]

During my years in audio I required a higher voltage output stage than could be provided by the then available output transistors
so I developed the attached output stage and during the last thirty years i have used it several times with success as i have never seen it used else where i would apreciate your comments and thoughts re this attached sketch
Please remember that this was developed in the days of the 2n3055/2n2955 the modern devices of today were but a dream
and i suppose could be used with the modern devices for the killowatt amps etc
all comments are welcome regards Trev

[jaycee]

Looks similar to the cascode output stage used in the Leach SuperAmp (aka the Double-Barelled Amplifier)

http://users.ece.gatech.edu/~mleach/superamp/

[darkfenriz]

Hi latala
This circuit is very clever, on paper it looks like doubling the maximum Vce of output transistors. Are there any benefits campairing to vintage series connection (as in quoted Leach)? Definitely on scheamtic it looks more clever and simpler, but what more?

best regards
Adam

[latala]

i havnt seen the super leach but if you study this you will find its not a cascode at all but a pair of series connected complimentary transistors with drive taken from both sides of the driver
regards Trev the 1975 ref is from an old sketch pad thaty i found in the garage
believe me if you saw my garage you could believe i havent been there since then
there are disadvantages such as twice the normal sat voltage per side but thats a small price to pay
the seies conncted resistor are to balance out the supply voltage across the series connected devices
i have a few more quirky designs such as these is there any intersest in these
i would like to see them in the worlld rather than just fade away
regards trev

[Samuel Jayaraj]

latala, how stable is the output stage that you have shown? If output devices are 3055/2955 or equivalent types, then what is the driver transistors minimum specs.

When are your other quirky sketches being posted?

[latala]

the out put stage was as stable as the conventional darlington pair type and provided that the vas stage is correctly compensated no problems at all were encountered
from memory the circuit was a conventional dif pair feeding a common emiter vas amp with a pair of transistors connected a a constant current load bias being set via a rubber diode
i have always tended to run the diff front pair at about 0.5 ma per device with 5 ma on the vas stage
compensation around the vas stage being a nominal 100 pf
but here the compensation cap is from the collector of the vas amp to the base of the feedback leg of the diff pair
this i have found always sonded nicer smoother
also the drive from the vas stage to the driver transistors was via a typical 220 resistor
this improved recovery from over drive conditions
as i said at the begining this schematic is more of use to the builder who happens to have the transistors in the junk box etc
i will be pleased to discuss or supply a more complete schematic if required
i dont know much about pcb programs can you people out there read qroute 3 schematics

regards trev

[CBS240]

Originally posted by latala
i have always tended to run the diff front pair at about 0.5 ma per device with 5 ma on the vas stage

The bandwidth gain product is lower at smaller currents in most small signal devices. How does this affect the operation of the diff pair? I am fond of SOT-23 transistors and lower currents is helpful here.

[powerbecker]

Hello Trev,

I have strong doubts that this series connection will divide their Uce equal! The both transistors and Re s had to be very equal and this over a broad range of current and voltage.
Functional You switched two high current sources together, with
a load of 5k.....this arrangement has much more amplification than a VAS.
Regards

Heinz

[Eva]

That arrangement would need at least a balancing buffer.

[anatech]

Hi Trev,
That's very similar as seen in the Bryston 4B. It's been around for a long, long time. They used 2N3773 and 2N6609 as outputs, MJE15028 and compliment as the driver.

Apparently it's stable, and it works. The biggest difference is the polarity of the driver transistor.

-Chris