For 300V operation there are some HV transistors which were designed for fluorescent tube drivers that might work in a low power Class A output stage. They were available in complementary pairs, but you would need to check SOA for class A operation. Low/medium power MOSFETs might be a better choice as the HV would mean the 3...5V gate drive is not as significant. With 150V for a long tail pair the long tail resistor should be big enough not to require a CCS.
Try a 6L6 or similar.
More seriously, crack out the product selector at RS or Element 14 and see what's available. RS list over 200 NPN transistors with >400V Vce
Class-A & Current-Dumping & Class-S
The Linn Valhalla seems a JLH inspired design. Linn implied the Lingo was class A whilst Valhalla AB. As far as I know Valhalla is class A. It allows an all NPN output ( BUX84 ). The bottom is a current mirror of sorts. The PSU seems to be a voltage multiplier when 115V. I calculate 11 mA used as A current of the outputs. The VAS is another BUX84 driven with emitters joined BC327. This is driven by an op amp. The feedback seems to be both AC and DC of the simplest type. If I'm not wrong the output stage is switched off via the current mirror when not in use ( Q11,22 via flip flop ). This circuit was close to being 100% reliable if the 3 x 47 uF 250( 200 ) V caps inspected yearly. 220 uF 250V is the ideal upgrade. The Valhalla looks to be an ideal music amplifier if modified. It's distortion is better that 0.05%. If anyone cares to look at it I never completely understood it.
http://www.hifi-manuals.com/400422/L/Linn-LIG/Valhalla____________-D-EN1-950-LIG.gif
I did try EL84. It wasn't ideal. The Sugden A21 is interesting.
The Linn Valhalla seems a JLH inspired design. Linn implied the Lingo was class A whilst Valhalla AB. As far as I know Valhalla is class A. It allows an all NPN output ( BUX84 ). The bottom is a current mirror of sorts. The PSU seems to be a voltage multiplier when 115V. I calculate 11 mA used as A current of the outputs. The VAS is another BUX84 driven with emitters joined BC327. This is driven by an op amp. The feedback seems to be both AC and DC of the simplest type. If I'm not wrong the output stage is switched off via the current mirror when not in use ( Q11,22 via flip flop ). This circuit was close to being 100% reliable if the 3 x 47 uF 250( 200 ) V caps inspected yearly. 220 uF 250V is the ideal upgrade. The Valhalla looks to be an ideal music amplifier if modified. It's distortion is better that 0.05%. If anyone cares to look at it I never completely understood it.
http://www.hifi-manuals.com/400422/L/Linn-LIG/Valhalla____________-D-EN1-950-LIG.gif
I did try EL84. It wasn't ideal. The Sugden A21 is interesting.
My brother was always intrigued by the Sugden A21. Now I see why. This is so logical. Douglas Self implies it has very poor symmetry. That really isn't important as that defect only makes for the nicest distortion type. The middle transistor is a buffer. The first a voltage amp. Note, despite the AC coupling between stages the global feedback via R8,9 C16 is far from being complex.
Richard Allan Class A amplifiers
Richard Allan Class A amplifiers
IMO that SRPP OPS have no voltage gain or is very low ,
since OPS top part(VT3&VT4) acts as bootstrapped emitter follower with no voltage gain very close to 1 ,
and in the same time bottom OPS part(VT5&VT6) is collector follower but work with significant local negative fdbk derived via divider R16/ R17 ( Shade-fdbk) enhancing transfer characteristic closer to some linear triode ,
for sure that OPS have no symmetric AC output wave but acts more as some OTL SET tube amp or so .
IMO that R20(1R5) is inserted there from reason that lower collector follower OPS part to never work under 0 load value , in case of short output or some very low LS load , but is also there to ensure SRPP action .
Member
Joined 2009
Paid Member
The Sugden is more complex than it seems. As said local feedback will fix the gain. I will have to build it to be sure. To get a bit of output stage voltage gain would be neat. The PW Texan being one. If in Paul Kembles pages see B&O designs also. They are very simple.
The problem with all JLH's is heat. It would be possible make a larger standard JLH. JLH dealt with the DC coupling in a later version. If I did a bridge version I would try to get my inverting signal from the simplest circuit possible. A Concertina phase splitter should be fine.
One way to build the best of all worlds machine would be to take the ESP Audio El Cheapo 60 watts and take it up to 100 watts. At 100 watts you would have to accept it's a raw device for power. The PSU would need to be about 90 VDC. If the PSU halved ( 30- 0 -30 or 0-30 V rms ) Then a 25 watt design is available from the same parts. As 5 watts class A would be OK 0.7A standing current would do it with margin to spare. The bias could be the very crude diode type that should be about 20 mA class AB and 0.7A class A. The El Cheapo is enough like the JLH in the amount of open loop gain to be able to compete. I would place a thermal switch in circuit to change the class A bias if the amplifier got too hot. This would be due to the amplifier trying to offer more current which the JLH can't do. What you would be doing is the Yamaha idea more to the JLH way of doing things.
Rod Elliot seems to both love and hate El Cheapo. In class A most of it's problems go away. In AB it will go loud. In some ways this type of class A ( A + AB ) is a variation of class G problems without the complexity and some of the advatages. The simpler 40 watt El Cheapo might be OK. It's nearly the amp below.
Modifying and Restoring the Cortina 3070 Page-3
The problem with all JLH's is heat. It would be possible make a larger standard JLH. JLH dealt with the DC coupling in a later version. If I did a bridge version I would try to get my inverting signal from the simplest circuit possible. A Concertina phase splitter should be fine.
One way to build the best of all worlds machine would be to take the ESP Audio El Cheapo 60 watts and take it up to 100 watts. At 100 watts you would have to accept it's a raw device for power. The PSU would need to be about 90 VDC. If the PSU halved ( 30- 0 -30 or 0-30 V rms ) Then a 25 watt design is available from the same parts. As 5 watts class A would be OK 0.7A standing current would do it with margin to spare. The bias could be the very crude diode type that should be about 20 mA class AB and 0.7A class A. The El Cheapo is enough like the JLH in the amount of open loop gain to be able to compete. I would place a thermal switch in circuit to change the class A bias if the amplifier got too hot. This would be due to the amplifier trying to offer more current which the JLH can't do. What you would be doing is the Yamaha idea more to the JLH way of doing things.
Rod Elliot seems to both love and hate El Cheapo. In class A most of it's problems go away. In AB it will go loud. In some ways this type of class A ( A + AB ) is a variation of class G problems without the complexity and some of the advatages. The simpler 40 watt El Cheapo might be OK. It's nearly the amp below.
Modifying and Restoring the Cortina 3070 Page-3
If you want more power from the JLH design, look at the Class-A site: by doubling up the output transistors (see the "JLH for ELS" section) and raising the rails and standing current, you can easily get 40 watts and verging on 60. Of course very large heatsinks are a must and a monoblock approach would be better.
Regards
Mike
Regards
Mike
JLH shows a design for a pair of Class A amps driven by a phase splitter in his 1970 postscript to the original 1969 article (see figure 4)
Nige
I guess that you already know about ALTO- UK made amp , which is in IMO conceptually direct copy of JLH1969 A-class amp , but is so modified to work in much cooler a/B-class ,Audio Innovations Alto
I guess that you already know about ALTO- UK made amp , which is in IMO conceptually direct copy of JLH1969 A-class amp , but is so modified to work in much cooler a/B-class ,Audio Innovations Alto
Banet, Yes I did and no I didn't. I thought it more of a Leak type design. I never saw the biasing. I must look to find-out more.
The problem with more power is it seldom gets you what you want. 10 watts class A is enough. My 21 watt AB 5 watts A would be ideal. Not least that it was 40 watts into lets say 3 ohms. It's not the 10 watts that matters, it that it can't swing more current. Over-biased class AB can.
The RCA design of 1959 looks very interesting as the Eico Cortina. It would not quite match the JLH for distortion. It would be very easy to set up. Just play with biasing diodes and resistors until it settles at lets say 0.9 to 0.7 amps depending on heat sink.
http://www.tronola.com/RCA_13-10_circuit.gif
The problem with more power is it seldom gets you what you want. 10 watts class A is enough. My 21 watt AB 5 watts A would be ideal. Not least that it was 40 watts into lets say 3 ohms. It's not the 10 watts that matters, it that it can't swing more current. Over-biased class AB can.
The RCA design of 1959 looks very interesting as the Eico Cortina. It would not quite match the JLH for distortion. It would be very easy to set up. Just play with biasing diodes and resistors until it settles at lets say 0.9 to 0.7 amps depending on heat sink.
http://www.tronola.com/RCA_13-10_circuit.gif
That rca amp is funny. What are trying to accomplish with these diodes in output? A short-circuit protection ... of income, haha? The way I see it, these diodes would (to some extent) prevent the R13/R14 to blow in case of overload ... but instead it would to some extent guarantee that either a diode or at least one of output transistors would blow.
Either way, a semiconductor company would surely stay in business ...
Either way, a semiconductor company would surely stay in business ...
