SPICE Model for AD8610 does not work

I'm using the AD8610 in a simple buffer application. AD8627 and ADA4622 are working as expected but in the prototype i am bound to the AD8610. See the part of the simulation below.

Followup: I tried also to eliminate R1, R2, C1 and C2, same result

AGND is just a virtual GND for the single supplied circuit and should have a voltage of 2.5V. The right side is a square voltage generator with the AGND as offset. The pp-voltage is 400mV and the AD8610 acts as buffer. Here is what the waves should look like (simulated with AD8627):

v(generator) and v(generator_out) are congruent... v(agnd) is 2.5V

Now the waves with original AD8610 spice model:

v(generator_out) is wrong.

I compared both models and found a way to make the circuit run, but I have no idea what the changed parameter does actually or where I can find it in the datasheet. Hope anyone can help me.

Left: AD8627 / right: AD8610

If I change V3 to 5 or something the circuit works correct. What is V2 and V3 for? Are the values correct?

Thanks in advance and best regards

Fabian

Parents
  • 0
    •  Analog Employees 
    on Jun 5, 2018 1:50 PM

    Hi Fabian,

    I tried to emulate your circuit and I noticed that the voltage reference LT6656-5 is affecting the LTspice simulation with AD8610 causing slow convergence or at times non-convergence and incorrect output voltage. Even if you change V2 and V3 to 5 in the AD8610 spice model as what you mentioned on your first post, this will only solve the problem on certain values of supply voltage for example it will work on 12V but if you try 7 or 8V, still the error is present. I tried replacing the LT6656-5 with an ideal voltage source and output voltage is correct for AD8610 for single supply voltages from 0 to 26V. Please see attached image.

    You could use AD8627 or ADA4622 for now since its already working fine with LT6656-5 in this circuit without problem.

Reply
  • 0
    •  Analog Employees 
    on Jun 5, 2018 1:50 PM

    Hi Fabian,

    I tried to emulate your circuit and I noticed that the voltage reference LT6656-5 is affecting the LTspice simulation with AD8610 causing slow convergence or at times non-convergence and incorrect output voltage. Even if you change V2 and V3 to 5 in the AD8610 spice model as what you mentioned on your first post, this will only solve the problem on certain values of supply voltage for example it will work on 12V but if you try 7 or 8V, still the error is present. I tried replacing the LT6656-5 with an ideal voltage source and output voltage is correct for AD8610 for single supply voltages from 0 to 26V. Please see attached image.

    You could use AD8627 or ADA4622 for now since its already working fine with LT6656-5 in this circuit without problem.

Children
No Data