Forums » CircuitLab Support Forum

Hey guys,

I'm a mechanical engineer and am still trying to learn the basics of circuits. For my project I have to simulate an automobile suspension using op-amps. I know the transfer function and have basically reconfigured the way mrobbins has his set up. Please help. I am working with this right now: https://www.circuitlab.com/circuit/hxp5z4/summer-setup/ based off of this: https://www.circuitlab.com/circuit/6tk255/project-setup/ Any bit of help is appreciated. Thanks

1) OA4 inputs are the wrong way round;

2) Use of "s" in a Laplace block is not recommended:

https://www.circuitlab.com/docs/circuit-elements/#laplace

The denominator should be the same or higher order than the numerator for best convergence.

Also, you may find that 1/s converges better as 1/(s+1p) where 1p = 1e-12.

See:

https://www.circuitlab.com/circuit/f73b76/modulated-sources-01/

and:

https://www.circuitlab.com/forums/support/topic/96jc4424/cls-laplace-h_s-block/

4) Use higher value resistors (and therefore smaller caps) in opamp circuits (1k - 100k or so). They do not drive low resistance loads well.

You might like to have a read of this too:

http://www.edn.com/design/analog/4376400/A-virtual-analog-computer-for-your-desktop

and:

http://www.edn.com/electronics-blogs/mechatronics-in-design/4369067/Circuits-without-wires-4369067

:)

Thank you for such a quick response, signality, and providing some very helpful knowledge! After making a few changes to my circuit I have modified my integrating op-amp by adding the proposed 1p resistor in series with the capacitor. Unfortunately this fix isn't as easy as I thought. Also communicating via email with my professor, and after he had a look at my circuit, he says that he doesn't see a double integrator and summation in my circuit, although I believe I do. He also says that I might have the transfer function modeled incorrect, but I believe it is T(s) = (Bs+K)/(Ms^2+Bs+k). Any thoughts on this? Again, this is where I stand https://www.circuitlab.com/circuit/xjx8x9/progress3/

Obvious errors:

1) OA4 is still has the input connections in the worng polarity.

2) In a 1st order RC circuit - i.e. an integrator, low paas filter, high pass filter - if you increase R then you decrease C. For a given time constant,

T = RC

Other stuff:

"I have modified my integrating op-amp by adding the proposed 1p resistor in series with the capacitor."

That's not what I was talking about. The small constant only applies to the use of the 1/s expression in a Laplace Block. It does not apply to lumped element circuits using opamps.

If you model:

T(s) = (Bs+K)/(Ms^2+Bs+k)

in a Laplace block and it does not give the same response as:

https://www.circuitlab.com/circuit/xjx8x9/progress3/

then there's either a mistake in the former or the latter.

If you assume that the:

https://www.circuitlab.com/circuit/xjx8x9/progress3/

expresson is modelled correctly and of course CL is simulaing correctly, then:

https://www.circuitlab.com/circuit/xjx8x9/progress3/

must be wrong.

Hint:

In:

T(s) = (Bs+K)/(Ms^2+Bs+k)

Bs+K is in the forward path;

1Ms^2+Bs+k is in the feedback loop.

I have been working on this all day and it has been driving me nuts. For one, I have simplified my model to represent T(s) = (Ms^2+Bs+k)/(Bs+k) instead of (Bs+k)/(Ms^2+Bs+k). Nonetheless you can see here that two feedback loops come from the change in the road's height to characterize the damping and spring force: https://www.circuitlab.com/circuit/y2s75n/model/

However, when I try substituting all the blocks as op-amps, I am constantly running into a problem when I get to the integrator. You can see here that without the integrator substitution, it the results are the same: https://www.circuitlab.com/circuit/78t6t2/attempt/

Now when I try to substitute the op-amp integrator I get a different result: https://www.circuitlab.com/circuit/nk9anq/model_attempt1/

Basically the same problem as before. I know there is a difference between ideal op-amp integrators and practical op-amp integrators (which are essentially low pass filters), but even when I use a low pass filter it still won't integrate the input as I would like. What am I missing?

Two problems in attempt1.

1) In you integrator OA6,

RC = 10k*100nf = 1e4*1e-7 = 1e-3

Hence 1/RC = 1k

You need 1/RC =1 if your integrator is to have the same scaling as 1/s.

2) You have forgotten that an opamp integrator inverts; i.e. it implements:

Vout(s) = -1/RC*Vin(s)*1/s

:)

As soon as I checked this post I JUST thought of that (Tau=RC)!!! Thanks :D

Signality, I have set my integrator to have a Time constant =1 Now what i don't understand is why the position results aren't the same when i substitute the op-amps in for the Laplace blocks:

https://www.circuitlab.com/circuit/m7ju8k/model_attempt1/

Concentrate.

You keep making the same mistakes.

1) Opamp +/- input polarity.

2) The basic opamp implementation of an integrator inverts. 1/s implementations do not.

3) Resistor values too low.

Also, please try to avoid editing circuits that have already been posted. Changes affect all instances so can render references in earlier postings confusing to others or just plain meaningless.

Take a copy and edit it then link to that one.

Thanks.