Linear power supply

[bountyhunter]

Better designs? Any examples?

Yes, here's a general design that can be easily sized up.

 

[wakoko79]

Thanks =), I'll try to understand this first though.

 

[wakoko79]

Yes, here's a general design that can be easily sized up.

@bountyhunter
As I can see in the circuit, U1B which is the voltage error amp also acts as a comparator the only difference I see here with my circuit is that it has a 0.033 capacitor there (I presume that is nano) as compensation. So, would it work in my original circuit? I'll try it with my circuit.

And the current error amp is the current limit right? I don't understand why the compensation caps for the U1A and U1B are different though.

 

[wakoko79]

*BUMP.

Please answer my last message sirs. T_T

 

[ronv]

@bountyhunter

As I can see in the circuit, U1B which is the voltage error amp also acts as a comparator the only difference I see here with my circuit is that it has a 0.033 capacitor there (I presume that is nano) as compensation.

Until you get the right size filter caps not much will help.
http://waynestegall.com/audio/ripple.htm
The IC is an amplifier not a comparitor.
The values appear to be Ufd.

And the current error amp is the current limit right? I don't understand why the compensation caps for the U1A and U1B are different though.

They are slightly different but not so much. The resistors are also different. There is no lead compensation, only lag.

 

[wakoko79]

@bountyhunter


Until you get the right size filter caps not much will help.
http://waynestegall.com/audio/ripple.htm
The IC is an amplifier not a comparitor.
The values appear to be Ufd.




They are slightly different but not so much. The resistors are also different. There is no lead compensation, only lag.

The lessons on my university about linear supplies are meager at best. They were just 3-5 slides (since SMPS is the focus, but it is quite difficult to design one). So I really don't know much about this.
I just realized that it is not a comparator, it's a integrator/LPF. I used a comparator before because with layman's intuition, I thought that an op amp reacting faster to output changes might be better. But now I think its they other way around. I don't know exactly why though. Any links/books/explanation please?

The other thing is that I am having a hard time finding an op amp that has a high enough maximum differential input, maximum input voltage, and maximum supply voltage. I scoured through digikey (I have no other choice but digikey) but finding an op amp with high max supply voltage alone is tad difficult already. Do you recommend that I just put an LM78xx for the op-amp's supply or would a zener do the job?

I'll also change the power section so that the pass transistors are at the top instead of the bottom (just like the normal linear regulators have).
With the old design, I researched about current hogging and thermal runaway when BJT's are just paralleled together without equalizing resistors.

I was convinced that thermal runaway happens because a slightly bigger collector current of one BJT will lead to a slightly smaller Vbe which will slightly increase collector current since the base current slightly increased as well(due to smaller Vbe). This ends up with that BJT hogging all the current that should have been shared to all parallel BJTs, thus the thermal runaway. So to counter this, a small resistor is placed at the emitter to counter the decreased Vbe (higher current > smaller Vbe > higher voltage drop to Re > stops further collector current increase)
But thinking back, I realized that all of the base and emitter of my pass transistors are tied together. Wouldn't that fact stop thermal runaway? I think it won't stop thermal runaway, so must mean my reasoning of how thermal runaway works is wrong.

 

[ronv]

The lessons on my university about linear supplies are meager at best. They were just 3-5 slides (since SMPS is the focus, but it is quite difficult to design one). So I really don't know much about this.
I just realized that it is not a comparator, it's a integrator/LPF. I used a comparator before because with layman's intuition, I thought that an op amp reacting faster to output changes might be better. But now I think its they other way around. I don't know exactly why though. Any links/books/explanation please?

They are kind of 2 different parts. The compariaor is optimized for switching while the op amp is for linear operation.

The other thing is that I am having a hard time finding an op amp that has a high enough maximum differential input, maximum input voltage, and maximum supply voltage. I scoured through digikey (I have no other choice but digikey) but finding an op amp with high max supply voltage alone is tad difficult already. Do you recommend that I just put an LM78xx for the op-amp's supply or would a zener do the job?

It depends on which circuit you are using now. In your original circuit the voltage from the op amp needs to go close to 30 volts. In this case you can use the LT1013 as the supply voltage goes to 44 volts so you could run it off the filtered 24 volts at 34 volts.

If you use bounty hunters you can use a low voltage op amp and a zener regulator.

I'll also change the power section so that the pass transistors are at the top instead of the bottom (just like the normal linear regulators have).
With the old design, I researched about current hogging and thermal runaway when BJT's are just paralleled together without equalizing resistors.

I was convinced that thermal runaway happens because a slightly bigger collector current of one BJT will lead to a slightly smaller Vbe which will slightly increase collector current since the base current slightly increased as well(due to smaller Vbe). This ends up with that BJT hogging all the current that should have been shared to all parallel BJTs, thus the thermal runaway. So to counter this, a small resistor is placed at the emitter to counter the decreased Vbe (higher current > smaller Vbe > higher voltage drop to Re > stops further collector current increase)
But thinking back, I realized that all of the base and emitter of my pass transistors are tied together. Wouldn't that fact stop thermal runaway? I think it won't stop thermal runaway, so must mean my reasoning of how thermal runaway works is wrong.

You need the small resistors in the emitter to make up for variations in base emitter voltage (mainly). They will balance out the load and add stability.