Using half of an opamp

[captntasty]

If I am using only half of a dual opamp should I ground the pins not used (5, 6, 7) or just leave them with no connection?

[R.G.]

 Tie the output pin to the (-) input, then tie the (+) input to Vbias, or to the output of the other half.

[oliphaunt]

Does it really matter?

[doitle]

You want to increase the loop gain to prevent the op amp section from going unstable with the open loop gain it normally would have. I think... RG will hopefully correct what I'm saying as I'm still in the middle of a controls theory course right now and dont understand it all 100%...

What I'm trying to say is that if you leave an op amp in an open loop configuration it likely has a gain of 100,000 or some other astronomically high amount. By adding a feedback loop you cut the gain down to say 100 or 10 and even though you aren't using the op amp section it's better to have it sitting doing nothing at a low gain than a high gain...

Again speculation on my part...

[R.G.]

Defeating oscillation in opamps has come a long way.

Most opamps these days are internally compensated to a gain of one. Some are only stable down to a gain of five or ten, and very, very few are uncompensated.

Here's something to carry to your controls theory class. Stability happens when the total forward gain is reduced to unity and below before the phase of the output reinforces the input - that is, gets to 180 degrees in an inverting amp. The number of degrees between 180 phase shift and the phase shift at unity gain is the *phase margin*. The point of single pole compensation, when you can use it, is that it uses a single dominant pole to reduce the gain to below unity before the second and following poles can add enough phase shift to make the system unstable.

Your feeling about loop gain and compensation is backwards, I think. Amps with little feedback need little compensation. Amps which feed back a lot need a lot of compensation. So you have to compensate unity gain feedback solutions more heavily than amps which work at higher gain. Amps which are compensated to be stable at unity gain are also stable with the same compensation at any higher gain. The response may not be as fast (i.e., bandwidth will be lower) for a unity-gain-compensated amp used at higher gains than you could get if you decompensated it for the higher gain, but it will be stable, and you don't have to be a controls wizard to get the compensation right in the face of part and temperature variation.

My controls class was fun... until I got the flu the week before the final! 

[captntasty]

Thanks R.G.  I'm going to tie the out put to the neg input and and the pos input to the output of the first half...  just out of curiousity, would it make any difference if I tied to Vbias?

For reference this is for building the 49er without the boost section. If anybody is interested I will post my pcb as long as it is alright with Paul of gaussmarkov (it is his board with the boost section removed).  I may even change it for a single opamp.