Looks to be a good starting point.
2 minor problems with the sheet that I have noticed.
1. the number of intervals is set at 2 so the ODE doesn't give an accurate output, I've changed that to 2000 but you could leave the parqameter out it will default to 1000,
2. the ODE now gives a very fast response & I guess that this will be correct. So I have modified the end time to 20us so the graphs are showing the overall curves.
Whether the result is representative or not, I can't say.
Hope it helps
Thank you for your response. I think this has got me going on the right track. Do you know if Mathcad will allow you to carry units through the Odesolver or is it better just to leave it as is?
Addison Mauldin wrote:
Do you know if Mathcad will allow you to carry units through the Odesolver?
In Mathcad Prime! See please http://communities.ptc.com/message/167830#167830
The ODE solvers don't handle units. (That's one of Prime's selling features--it eventually will, we're told.)
The best way to do this (IMHO) is to divide each term by "UnitsOf(what the units should be)" and leave the units on the values. This will ensure that the problem scales correctly, if the length of the rails changes from 10 meters to 50 feet, the value will still report correctly to the problem. You can then multiply the solution by "UnitsOf(the correct unit)" to restore units to the answer.
Addison Mauldin wrote:
Thank you for the example Valery!!
Sorry for the Russian comments.
On the Mathcad Server
you can see English comments too.
See the video too - http://communities.ptc.com/videos/1699
Finally, something that Prime 2.0 can do better than MC15 ...
Ode with units is possible - just needs a little thought.
By the way I know the initial voltages are not zero....there has to be some initial voltage applied for the thing to work. I just don't know the values yet...
You have a number of issues here. The main one is that Odesolve doesn't like true subscripts. Better to set up the problem as a matrix equation (dY/dt = AY+B where Y is a vector of all your V's and I's, A is a matirx containing all the constants multiplying the V's and I's and B is a vector of any remaining constants). Then use one of the separate solvers like rkadapt, say.
I don't think your deltaz should be a function of time; surely it's just lt/N?
Zload should probably have units of ohms to make the last expression consistent.