LT1166 Mosfet driver - once again

[the_manta]

Hi,
First of all, I’m usually building with tubes and do that quite a while now. I understand the concept of the usual solid state amplifiers. I have some books about that, too.
I recently got some LT1166. I want to experiment with hybrids a bit. The LT1166 looks interesting and even using Hexfets is no problem with it.
But I still don’t get what this device can do altogether.
First thing, nobody uses the input-Pin (Pin2) of the device. In every circuit with the LT1166, the signal is inserted at the constant current sources which deliver the Ub for the LT1166 and the Bias voltage for the Mosfets. I wanted to avoid that because I would need some tube-driver which can deliver quite some current (Gate capacity etc.).

Bob Cordell writes in his book: “Designing Audio Power Amplifiers” on page 544:



And on page 546:



Why does it then have an Input Pin ?
So it is exactly NOT used as in this basic schematic:



Instead they use it like this as explained above.



Here the upper Mosfet is controlled via the source current of the lower one. The error-correcting of the LT1166 is used here in extreme, isn’t it ?
Another Schematic can be found on John Broskie’s Website:
http://tubecad.com/2011/08/20/IRF%20...vo%20Large.png

diodes are used here to provide the bias. Why ? I thought the LT1166 would regulate the bias according to the source resistors on the Mosfets. Am I wrong here ?

Another thing is the offset control with the Integrator-OP on the bottom of the schematic. The Datasheet says that the LT1166 alredy has an offset control inside !

Datsheet: LT1166 - Power Output Stage Automatic Bias System - Linear Technology

Can someone help me on this ? This IC is still kind of a mystery to me.

[NMOS]

Originally Posted by the_manta

Hi,
First of all, I’m usually building with tubes and do that quite a while now. I understand the concept of the usual solid state amplifiers. I have some books about that, too.
I recently got some LT1166. I want to experiment with hybrids a bit. The LT1166 looks interesting and even using Hexfets is no problem with it.
But I still don’t get what this device can do altogether.
First thing, nobody uses the input-Pin (Pin2) of the device. In every circuit with the LT1166, the signal is inserted at the constant current sources which deliver the Ub for the LT1166 and the Bias voltage for the Mosfets. I wanted to avoid that because I would need some tube-driver which can deliver quite some current (Gate capacity etc.).

Bob Cordell writes in his book: “Designing Audio Power Amplifiers” on page 544:



And on page 546:



Why does it then have an Input Pin ?
So it is exactly NOT used as in this basic schematic:



Instead they use it like this as explained above.



Here the upper Mosfet is controlled via the source current of the lower one. The error-correcting of the LT1166 is used here in extreme, isn’t it ?
Another Schematic can be found on John Broskie’s Website:
http://tubecad.com/2011/08/20/IRF%20...vo%20Large.png

diodes are used here to provide the bias. Why ? I thought the LT1166 would regulate the bias according to the source resistors on the Mosfets. Am I wrong here ?

Another thing is the offset control with the Integrator-OP on the bottom of the schematic. The Datasheet says that the LT1166 alredy has an offset control inside !

Datsheet: LT1166 - Power Output Stage Automatic Bias System - Linear Technology

Can someone help me on this ? This IC is still kind of a mystery to me.

I prefer NMOS OUTPUT stage

IRFP240 and 9240 is not good solution because not really symmetrical

N-Chanell Mosfet are better solution

[nigelwright7557]

Originally Posted by NMOS

IRFP240 and 9240 is not good solution because not really symmetrical

They dont need to be symetrical, feedback compensates for any differences.

[djk]

"diodes are used here to provide the bias"

The diodes are symbols for constant current sources, you can use any constant current source design you desire, including a simple resistor (perhaps with a bootstrap cap).

"I wanted to avoid that because I would need some tube-driver which can deliver quite some current (Gate capacity etc.). "

Why? The FETs need a bunch of gate drive anyway.

[ingenieus]

I've had a go at simulating this circuit in LTspice a few weeks ago. It sits there doing endless iterations for quite a long time before coming up with the answer. It does work, though.

An internet search confirmed that others have had the same problem. Perhaps there is a problem in the model for the LT1166 included with LTspice. The likely answer seems that it is a complex chip to simulate. Anyway, my local suppliers do not stock it so there is not much point to investigating it further.

Is the LT1166 still available in other parts of the world?

[padamiecki]

Originally Posted by ingenieus

I've had a go at simulating...

Hi!
I simulated the circuit from app datasheet
and I do not like the distortion distribution
but Broskie likes it (lt1166)
and for me this is real indication
maybe will try this in practice

[Nico Ras]

Ingenieus, Arrow Altech is the local agents for Linear Technologies:
Montague Gardens Regional Sales Hub...Units J15 & J16
Centurion Business Park
Bosmansdam Road
Milnerton, 7441
Tel: (021) 555-1884 | Fax: (021) 555-1763

[ingenieus]

Originally Posted by Nico Ras

Ingenieus, Arrow Altech is the local agents for Linear Technologies:

Thanks.

[807]

Just for anybody who might be interrested. I just discovered that in LTspice it is very slow when the simulation option "Start external DC supply voltages at 0V" is off. Enable this and simulation is fast again.

[Edmond Stuart]

Originally Posted by the_manta

Hi,
[...]
I recently got some LT1166. I want to experiment with hybrids a bit. The LT1166 looks interesting and even using Hexfets is no problem with it.
But I still don’t get what this device can do altogether.
First thing, nobody uses the input-Pin (Pin2) of the device. In every circuit with the LT1166, the signal is inserted at the constant current sources which deliver the Ub for the LT1166 and the Bias voltage for the Mosfets. I wanted to avoid that because I would need some tube-driver which can deliver quite some current (Gate capacity etc.).
[...]

Hi,

The reason that nobody uses pin2 is that the error correcting part of the LT1166 is ill designed, totally!

@NMOS,
>I prefer NMOS OUTPUT stage
That's no answer to the question!

@ingenieus,
I also had convergence issues with the LT1166 model and was unable to fix it.

Cheers,
E.