BUF634A: IV conversion of current DAC using composite amplifier

Well, you are adding the delay of the BUF634A inside the loop which is a manageable stability issue, but maybe there is an op amp with adequate output current to begin with -what do you think you need in terms of output current or min loading? 

What DAC are you using - most are complementary current output where using both outputs will improve HD2 if you care. 

Its not about current but I am looking at slew and settling time, In a delta sigma DAC , you have very fast changes on current (steps) so looking at a composite opamp that will track those

So the BUF634 is a unity gain voltage buffer - it cannot settle any faster than the core op amp - normally, composites with it are used for higher output current. You need something else I think. Only desired specs and source info can move this forward. 

In the many delta sigma DAC designs I have done, here is an example using the THS4561 FDA implimented with a MFB wrapped around it - not necessarily what you want, but the point is there are a lot of options, 

Hello Michael

   I am using a delta sigma DAC es9039q2m with a 2.4ma differential current output .

I am already using the OPA1612 for IV stage , but I am trying to make a better IV stage

    In order to do that , I am planing to use a FET opamp opa1656 (that has excellent noise for a FET)  but I want to beef up significantly the slew rate (at least 100V/us bandwith (at least 100Mhz)  and settling time  (300ns or better at 0.01%)  at  . So I was thinking to use a composite amplifier . (of course THD+N is very important)

   Any suggestions ?

Well not enough to go on yet, but couple of points, 

1. What maximum step size

2. Settling is a pulse response thing, HD is a sinusoidal test - doing both? 

3. For best settling, stay out of slew limiting - whether a part slew limits is a small signal response shape issue with desired output step size. By way of preview, the peak dV/dT on a 2nd order response shape is approx 2.85*(Vstep) *F-3dB (and this is including any peaking effects, applies roughly from 0.7<Q<4).

Tried to dispel some of the slew rate confusion in lower speed parts here - need to finish the higher speed slew rate article in the next couple of days, - so off to that, good luck with your questions. This one deals with transitioning from a largely slew limited step to a partially slew limited step for 1st order responses (but does add the effect of a real higher pole as well).. If you allow the edge to slew limit, you have transiently opened the loop requiring it to close again to find a final value - extending to the settling an uncertain amount. Stay linear if you can is the moral to this story.  

https://www.planetanalog.com/separating-linear-from-slew-limited-performance-in-amplifiers-part-1-of-2-the-signal-sped-up-insight-14/