darligton Leslie

[sakis]

Things I have seen wrong with the Leslie `Sharma 1500

1 input: even though there is a very large capacitor shown (high frequency filter) in the schematic from input to ground I have never seen it installed. Filter like that could keep the input free from very high frequency signals (not audible)
• I choose to keep the same input cap IE 47mfd /63V
• Wired in series 2K resistor
• Add 200pf capacitor to ground (In parallel with existing 10K )

2 Vas amplifier: I found that miller cap on Vas amplifier was too optimistic at 68 pf and also wrong choice of part since its well known that simple 63 V ceramics are not so happy there … also most of those amplifiers I have seen designed around Darlington’s usually feature local miller or so called stabilization caps directly in the output transistors in the range of 470 or 560 pf
• I choose to go with 100 pf 3KV blue multilayer ceramic ( better safe than sorry )

3 Bias : … as seen from pictures and schematics bias is provided by a string of two 1N400X diodes and have no thermal compensation …since outputs are PTC this will mean that if the amplifier is stressed bias will increase together with the temperature of the output stage … this is what we call thermal runaway and if stressed a lot obviously will destroy the output section totally.
Since designing a real and complete VBE multiplier with adjustable bias and with thermal compensation was kinda complicated
• I choose instead to attach the diode string to the ballast resistors so if amplifier is stressed then ballast resistors will also get warm then bias automatically will be decreased .This will provide short of thermal compensation to the amplifier and hopefully not too much so we move then to a crossover distortion .

4 Outputs ..didn’t want to run around looking for parts from Mars so
• I choose to go with MJ 11015-16 ONSEMI well known pairs used in many Marshall and some Fender I believe.
• Added 1mfd of decoupling onatop of the reservoir capacitors
• Ultrasonic cleaning of the fuse holders they were actually black from oxides
• Obviously replaced all axial electrolytic
• Wash the boards
• And finally removed the ferrite from the inductor in the output ( simply I don’t like it )
• Add 3.3 R dumper resistor


Finally I would like to listen to some opinions around the bias thingy


Kind regards sakis

[CBS240]

Hi

SMD COG type ceramic chip capacitors are very good for compensation. Also they are easy to insert into any PCB by scratching a break in the trace at the most optimum location, that being right next to the high impedance node it is feeding back to. 100pf >200V is available, dirt cheap.

As for the bias arrangement, don't know what pics your referring too but Darlington output EF stages would require 4 diode drops to bias out of class B, and also the PTC is the product of the driver and output because they are on the same die and thus the same temperature...(PTCdriver X PTCoutput) does not = (PTCdiode + PTCdiode). This makes thermal runaway even worse, but if it were biased in class B with only two diodes it would not be such a problem....but then it would sound like crap, IMO. A Vbe multiplier using a TO-126 Darlington with a collector resistor to vary TC to closely match the outputs should do it. Or you could go all out and build a HEC module and insert it into the circuit.......Here is the paper for that. Use the 'amplified diode' circuit as the output stage within the GNF loop as if it were the Darlington followers and see if you get an improvement.

It may be nesacarry to add local compensation like Bob did in is vertical HEC amp, at least for the error amplifier.

[sakis]

the point on this was not to alter the amplifier but mostly provide some aditional safety
warming up the diode string ( since attached in the balast resistors ) will drope idle ... this will do for me

[ashok]

Appears to be an amp but I don't recall coming across this one before.
Was this one discussed on this forum ? A search didn't show anything.
Cheers.