F4 Beast Builders

[Ian Macmillan]

You were right - it is in the owners manual See my last post.

 

[lhquam]

frankly , have no idea how to simulate power of

that's the thing I didn't even think of

I also developed a simulation of my amplifier that shows a similar, but not identical behavior. The simulated power supplies are SPICE pulse voltage sources of the form:

PULSE({Vinitial} {Von} {Tdelay} {Trise} {Tfall} {Ton} {Tper})

In the simulation below, Tdelay=20sec, Ton=60sec, Tper=120sec. Trise and Tfall were 1.6mSec, and not important. The circuit contains a diode so that when AC power is shut off, the rail voltage decay will be determined by the current thru the circuit, not back into the transformer simulated by the voltage source. The 3.5R resistor was empirically from the measured behavior of my amplifier, shown in post #1408.

 

[Zen Mod]

@ Ian

it teaches us that we can trust in Papa's words , as long math is not involved

in other circumstances , his math is impeccable

 

[lhquam]

I've not considered power off behaviour either, how serious an issue in practice do you think this is likely to be? If you have an amplifier already built then isn't it worth trying with a cheap loudspeaker?

Has there been any discussion of the shutdown circuit? I don't recall reading anything mentioning SCRs but I may have missed it.

When I turn off power to my amplifier, there is a very minor "click" in the speakers. Noticeable, but not in any way damaging to the speakers.

 

[Zen Mod]

me thinks we all are way past point of overengineering

 

[lhquam]

Does anyone know whether the production XA25 is totally free from power-on and power-off "thumps" or "clicks".

 

[Zen Mod]

well , it is production

and I know that every PL/FW product you can use even with Lowthers

edit : I believe Denis can confirm my strong hunch/experience

 

[lhquam]

Some speculation:
If the XA25V2 sense resistors have been reduced in resistance that might allow the rail voltages to be reduced and the OS idle current to be increased for increased class-A wattage and no change in overall power dissipation.

 

[Zen Mod]

I chose 35V rails and 0R1

do we know where origigi is now ?

 

[Ian Macmillan]

My guess is 28V rails and 2A bias. But that is all it is - a guess based on the spec.

 

[RobLK]

I went back to a simple biasing circuit I had from 2016 and tried it again. It seems to work in ltspice.

 

[lhquam]

I do not understand how changing the cascode base voltage would significantly affect the current thru the cascode transistor. If there were resistors from the cascode emitters to a reference voltage, such as ground, it might work.

 

[RobLK]

I tried it with the opto's disconncted and the bias is at 4.6 amps and with opto's bias is at 2.4 amps.

 

[lhquam]

What were to voltages on the cascode bases with and without the optos connected?

 

[RobLK]

only about 12v before and 11.5v after. Edited after value.

 

[RobLK]

I was using 0.4 ohm sense resistors above but I meant to use 0.5 ohm resistors to get a 2 amp bias current. With 0.5 resistors it shows 1.97 amps.

 

[lhquam]

 

[RobLK]

 

[RobLK]

It looks like it's changing the output bias by changing the resistance of cascode transistors.

 

[Ian Macmillan]

I don't think so. Changing the voltage at the base of the cascode transistors will not significantly alter the current through them. The input JFETs will see a different voltage at the cascode emitters but this does not change their behaviour much since they are effectively a current source into the cascode voltage source.