Hello! I’m studying the behavior of op amps in simulation. I encountered strange behavior. My circuit is simulated normally with the value R4=1110, but if I change it to 1150 or value I get: doAnalyses: TRAN: Timestep too small; initial timepoint: trouble with xu1.x_u1:dvn-instance d.xu1.x_u1.d1
run simulation(s) aborted
If R=1111 — Simulation runing well.
I tried another amplifier and another spice model in a different circuit and I was not able to perform the simulation at all. Spice models from the site ti.com
Why are the input caps shorted? That’ll make trouble. The DC levels are completely off.
And please show the VSIN parameters.
Tnx for the answer. Funny but without shotring wires i have again:
doAnalyses: TRAN: Timestep too small; initial timepoint: trouble with xu2.x_u1:dvn-instance d.xu2.x_u1.d1
run simulation(s) aborted
VSIN parameters:
dc=0 ampl=0.1 f=1k td=0 theta=0 phase=180 ac=0.1 ph=180
dc=0 ampl=0.1 f=1k td=0 theta=0 phase=0 ac=0.1
Your VSIN parameters look off.
Try this:
dc=0 ampl=0.1 f=1k ac=1
for both and turn one of the VSIN sources upside down (same as 180 degrees phase).
And straighten out the wiring around the VSIN sources. It’s an eyesore and difficult to trace/follow.
Unfortunately, no changes.
doAnalyses: TRAN: Timestep too small; initial timepoint: trouble with xu1.x_u1:dvn-instance d.xu1.x_u1.d1
run simulation(s) aborted
Please open a new thread. You tacked this onto the end of a solved thread. I almost answered the wrong question.
I haven’t really used spice since college. (we don’t want to know how long ago that was) The original thread had a time step of something to the e-17 and was calling it microseconds. What happens if you increase the timestep?
“Timestep too small” is rarely the real problem. The message is a catch-all for everything ngspice can’t figure out.
Mostly, the problem is the model or the pin sequencing for the model.
Can you tell me where to download a really working spice model of any operational amplifier?
I tried 5 Op Amps from the ti.com. But “Timestep too small” again and again…
Hey, I’m the guy who asked the original question https://forum.kicad.info/t/timestep-too-small-error-in-simulation/51164, on top of which you intially posted yours.
The Op-amp model you got is probably fine. That said, I’ll attach the one I used successfully in my original question, maybe it’s of some help to you. But that wasn’t the problem in my case.
In my case, I had pin assignments wrong for some other component altogether (the NPN BJT, I think). Check that carefully for all of your components, even the tiny ones.
If it’s not clear what we mean by pin assignments, we’re talking about the following page in the component’s simulation properties. You can set up which pin numbers of the symbol correspond to which pin numbers of the simulation .lib file by modifying the table in the top half of the dialog box.
If you share your project files (including any libraries and stuff that you may have added), I’d be happy to check it out in KiCAD. But please check your pin assignments first.
LM741.lib (2.6 KB)
When I have trouble like this, I go back to the fundamental’s of (whatever it is).
I may be missing it but, I don’t see a TRAN statement in your schematic.
Below circuit (screenshot) works…
Thank you very much! I’ll check your model later. I don’t think it’s a problem with the model pins, but I’m sending you my sch-file and model. Please check it on your copy of Kikad. (I use the latest 8 version).
Pls dowload my files (new users can’t publish files) combo_panel.zip - Google Drive
By the way, i deleted TRAN statement from my schematic, but they are visible in the VSIN properties.
Will try draw new sch.
I’m pretty sure it is. I’ve done perhaps 20 circuits with opamps and simulation, and in 90% of the cases the problem was pin assignment in the model.
1 Like
I’m not sure that alone will work. You need a Directive; either from the Schematic and/or from the Directive panel…
It will.
OP, let’s go down to basics and get a few things fixed first:
For all parts on your schematic, tick “show pin numbers”.
For the VSIN sources, tick “show sim params”
Then we’re all on the same page.
After you’ve done that, show us the simulation properties of one of the opamps, specifically the “pin assignments” screen. Take a screenshot.
I’m not able to open your .sch file as it’s a newer version (I’m on 7), but it doesn’t matter. The .sch alone is useless.
I had a look at your files.
Is there a specific reason you chose the OPA1656 IC? It is a dual op-amp, but I’m not able to configure pins for two op-amps using KiCAD. This needs a more experienced set of eyes than mine.
Your numbering of U1 and U2 was also incorrect. It should be U1A for one op-amp, U1B for the other, and U1C for the power supply. They’re all part of the same physical IC.
I’d suggest sticking to a simpler single op-amp for now, unless you have some specific reason to use OPA1656. If your purpose is just to play around with op-amps in SPICE, that will work very well for you.
Good catch on the misnumbering of the amps, I missed that one.
Inserting an “Opamp_Dual” from the “Device” library and assigning a spice model to it is not hard. You only need to do it for one of the amps.
But you DO need to configure pin assignments for each of the amps.
Yes, but it worked with a short circuit on the input capacitors. I can’t specify a different pin configuration for side A and side B in this OA spice model. That is, I cannot indicate that there is an output at 5 and at 1 pin. Therefore, I use different copies of the components in the diagram. I think that now it is no longer very important which op-amp model to use, the main thing is to simulate it correctly. I’ll try it.
You have two unused opamps floating around somewhere in your circuit.
Did you run an ERC?
I’d never proceed to simulation until that’s either error-free, or at least qualified.
