LM4780+ 2 Way Plus Servo

[jackinnj]

Originally Posted by peranders

Did you have the same circuit as with the one you have drawn? Did you connect the input signal to the surrogate opamp?

Yes -- it's merely the servo circuit from one of the national semi datasheets connected via a 205k resistor to the inverting input of the LM4780.

I'll breadboard it this week and run it through the network analyzer.

[klewis]

Jack,

I think you need to use the LM411 or LM412 for the servo. Low initial offset and low offset drift. Offset is also adjustable. Or the LM412, same low offset and low drift not adjustable. Look at the national datasheets - compare the spec to the LM4562. The non-Jfet type opamps have much more offset and offset drift with temperature, both of which you're trying to eliminate by using a servo.

Self's book has a chapter on servos.

Ken

[peranders]

Originally Posted by jackinnj

Here's a bode plot of the servo itself -- the response is down 56dB @100Hz if I've drawn everything correctly!

I put in the values into LTSPice and the -3dB is at 0.015 Hz!! I did also include the control object and applied the signal at the normal input.

[jackinnj]

Originally Posted by klewis

Jack,

I think you need to use the LM411 or LM412 for the servo. Low initial offset and low offset drift. Offset is also adjustable. Or the LM412, same low offset and low drift not adjustable. Look at the national datasheets - compare the spec to the LM4562. The non-Jfet type opamps have much more offset and offset drift with temperature, both of which you're trying to eliminate by using a servo.

Self's book has a chapter on servos.

Ken

You mean, I think, LF411 and LF412. Seems also that it's a good idea to use polypropylene for the integrator caps, and low leakage diodes. I thnk I have some texas instruments quad jfet opamps somewhere in the lab.

I wonder, also, whether to make the LM4780's inverting, and use the inverting servo.

[klewis]

Originally Posted by jackinnj

You mean, I think, LF411 and LF412. Seems also that it's a good idea to use polypropylene for the integrator caps, and low leakage diodes. I thnk I have some texas instruments quad jfet opamps somewhere in the lab.

I wonder, also, whether to make the LM4780's inverting, and use the inverting servo.

Jack, you're right, "LF" though when googling this morning pre coffee, LM still got me to the correct data sheet...

Ken

[peranders]

Originally Posted by jackinnj

You mean, I think, LF411 and LF412. Seems also that it's a good idea to use polypropylene for the integrator caps, and low leakage diodes. I thnk I have some texas instruments quad jfet opamps somewhere in the lab.

I wonder, also, whether to make the LM4780's inverting, and use the inverting servo.

Polyprop caps will not add much here. Polyester will do fine unless you really like big polyprops

[jackinnj]

Skipping the buffer, inverting and employing the LF412 and low leakage diodes:

[klewis]

Jack,

With the DC servos, you can direct couple the amp, so, if any of the capacitors are DC blocking and not for band pass purposes, they could be eliminated.

Ken

[Redshift187]

Also, you can use a JFET like a PN4117 or 2N4117 for a low leakage clamp instead of diodes, like I've done with my servo design.

[klewis]

Jack,

No need to invert the servo. Here's the schematic from Doug Self's book. I've tried it, it works. The values shown give unity gain at 1 Hz. The two grounds shown should be connected to signal ground. Ignore the op amp designation, just grabbed it quickly to make the schematic.

Ken