PSpice Convergence Problem

[Bob Penoyer]

I have a circuit comprising one OP-77/AD op-amp that is basically a
Sallen-Key filter with some other resistors and capacitors around it.
All capacitors use a tolerance (10%) and all resistors use both a
tolerance (1%) and a tempco (100 ppm/C).

I've had convergence problems during worst-case simulations at some
temperatures. The OP-77/AD seems to be at the heart of the problem but
I'm stuck with it because I'm analyzing an existing circuit.

All of the power supplies (+/-15V for the op-amp, and +5 for another
item) have both 5% voltage tolerances and 200 mVpp noise at 20 kHz
bandwidth.

I've made these modifications to the default simulation settings:
VNTOL = 1E-4
ABSTOL = 1E-10
GMIN = 1E-9
ITL1 = 5000
ITL2 = 2000
ITL4 = 2000

Is there something else I should do to enhance convergence?

What about the solver? Is the default best? What are solver 0 and
solver 1?

Is selecting GMIN stepping helpful?

Thanks,

Bob Penoyer

 

[Jim Thompson]

On Wed, 14 Sep 2005 21:04:29 -0700, Bob Penoyer
<bob@NOSPAMbobpenoyer.com> wrote:

I have a circuit comprising one OP-77/AD op-amp that is basically a
Sallen-Key filter with some other resistors and capacitors around it.
All capacitors use a tolerance (10%) and all resistors use both a
tolerance (1%) and a tempco (100 ppm/C).

I've had convergence problems during worst-case simulations at some
temperatures. The OP-77/AD seems to be at the heart of the problem but
I'm stuck with it because I'm analyzing an existing circuit.

All of the power supplies (+/-15V for the op-amp, and +5 for another
item) have both 5% voltage tolerances and 200 mVpp noise at 20 kHz
bandwidth.

I've made these modifications to the default simulation settings:
VNTOL = 1E-4
ABSTOL = 1E-10
GMIN = 1E-9
ITL1 = 5000
ITL2 = 2000
ITL4 = 2000

Is there something else I should do to enhance convergence?

What about the solver? Is the default best? What are solver 0 and
solver 1?

SOLVER=0 is the old original solver before it was "new and improved".
I have found the setting SOLVER=0 will often work on difficult
circuits.

Is selecting GMIN stepping helpful?

Sometimes ;-)

Thanks,

Bob Penoyer

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| E-mail Address at Website Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.