Actions
Forum Thread Details
  • Replies 2 replies
  • Subscribers 462 subscribers
  • Views 309 views
  • Users 0 members are here
Related

Simulation software, close to real life?

Catwell
I recently built a circuit based on a schematic I drew in Multisim. To my dismay, it didn’t work properly. After closely inspecting the Multisim simulation, I noticed that it behaved exactly like the real life build. It’s amazing. I should have paid closer attention to the simulation. But I have a concern, how much weight should I give a simulation? Should I rely on it at all?
 
From your experience, how accurate have your simulation been to the final product?
 
Cabe

  • Hi Cabe,

     

    I use simulations very often, to design filters(Active, Passive, Digital, etc.). Any engineer who trust simulations completely is in for a big suprise. The worth of a simulation really stems from the accuracy of the model.

     

    A good example, is the humble op amp. 80% of the time the Ideal Op Amp model is accurate enough, however as we all know there's no such as an ideal op amp. Real op amps have finite gain, a finite input resistance, 3(sometimes 4) poles, and the list goes on. The ideal model doesn't take any of these into account, and like I mentioned 75-80% it is sufficient but you have to keep it's limitations in the back of your head at all times.

     

    If you simulate the circuit using Ideal Op Amps and find that your finished circuit doesn't work, try swapping out the ideal op amp model with a more precise model such as a single pole model. In filter design most of the time the single pole model is sufficient, but if the design is demanding you'll be better off using a more complex model, usually you'll find that the more complex model will simulate the error your observing in your real circuit.

     

    So the conclusion is, simulations can be very helpful and save you a lot of time, however always keep in mind the assumptions and simplifications used by the model, and MAKE SURE, that they hold for your current design.

     

    I personally use LTSpice, which is a free simulator from Linear Technology, I've learned how to work with it and I must say it's pretty powerful, but you have access to Multisim so take advantage of it but always remember the assumptions made in your model.

     

    hth,

     

    Jorge Garcia

     

    P.S I'm very broke, so you'll find that I tend to gravitate towards low-cost or open-source solutions, EAGLE, gerbv, LTSpice, Excel, etc.)

  • in reply to Jorge_Garcia

    Jorge,

     

    I agree, you can never truly trust a simulation. What I have found to be the most useful, is choosing parts that have an unique "Spice" model defined for them. This, from my experience, has not failed to predict what will happen in my real world circuit. If you are meticulous, you could get the model for each component. I would trust when is discovered via a simulation under those circumstances 99% of the time. Over all, I would say it is a good place to start in a design.

     

    I haven't used LTSpice all that much. I will definitely give it a try, with your recommendation. The following video shows it simulating a H-Bridge. It is impressive. I will have to compare it to Multisim in a future post for sure.

    This video is unavailable.

     

    Anyone know how easily LTSpice can work with Eagle?

     

    Cabe