2-D plot resolution

[pzimpelman]

I made a shear and moment diagram plot in mathcad and it should look like this:

However, in mathcad, mine is looking like this:

The problem, in particular is that on the moment diagram at about x= 1" there's a "knee" in the moment diagram that shouldn't be there.

Does anyone know the cause of the "knee?" Anyway to increase the resoltion of the plot?

[MikeArmstrong]

What happens when x = gap? The definition of v is the issue.

Please upload a worksheet.

[pzimpelman]

Thanks for your response

[RichardJ]

You don't show how x is defined, but looking at the graph I'll guess it's a range variable 0*in,1*in..15*in. So you get a point every 1 inch. Make the increment much smaller, say 0*in,0.01*in..15*in.

If that's not the problem, please post a worksheet.

[pzimpelman]

How exactly would you decrease the increment?

[MikeArmstrong]

See attached.

[RichardJ]

Don't rely on the quickplot feature of Mathcad. It gets lots of people into trouble. Use a range variable to define the x-axis range and increment.

[MikeArmstrong]

Maybe that is a flaw with the quickplot feature. I think the user should be able to change the quickplot scale, to allow as in this instance, the increment to be reduced.

[RichardJ]

It's because of the units. Quickplot seems to get confused by units. I personally never use it, because it's too unpredictable.

[MikeArmstrong]

Richard Jackson wrote:

It's because of the units. Quickplot seems to get confused by units. I personally never use it, because it's too unpredictable.

Poor behavior if you ask me

[Werner_E]

Richard Jackson wrote:

It's because of the units. Quickplot seems to get confused by units. I personally never use it, because it's too unpredictable.

The quickplot normally has a quite nice adaptive way to chose the number of points and for some nasty functions it was hard to duplicate the (good) quickplot results using range variables.

But when using units the quickplot feature usually fails badly.

We can see what happens when we format the plot to show the individual points while its unclear why this happens.

Even more confusing is the "solution". Curiously its not a modification on the abscissa but on the ordinate. Simply multiplying the quickplot variable by the unit makes the feature work again. But the calculation time for the second plot in this example is quite prohibiting.

So normally the advice not to rely on the quickplot feature but use your own range is sure a good one.

[RichardJ]

Interesting solution

But the calculation time for the second plot in this example is quite prohibiting.

That's easily fixed.

[Werner_E]

Richard Jackson wrote:

Interesting solution

Yes and the question is what x-values the function is fed with when we use the quickplot and why adding the unit in the function call changes the spacing on the abscissa. Mysterious!

But the calculation time for the second plot in this example is quite prohibiting.

That's easily fixed.

Ok, in case of this function its really easy 🙂

[RichardJ]

Ok, in case of this function its really easy 🙂

Tom posted a worksheet that showed how to convert an integral with a variable limit to an ODE, which makes it much faster. I can't locate the thread though, and (regrettably) I didn't keep a copy of the worksheet.

[Werner_E]

Richard Jackson wrote:

Ok, in case of this function its really easy 🙂

Tom posted a worksheet that showed how to convert an integral with a variable limit to an ODE, which makes it much faster. I can't locate the thread though, and (regrettably) I didn't keep a copy of the worksheet.

This one?

http://communities.ptc.com/message/78968#78968

[RichardJ]

That wasn't the one I was thinking of, but it will work. Mostly, anyway. The worksheet fails in MC15, so I'll have to figure out why.

Edit. Because the Mupad symbolics return a result that the numeric processor does not understand. So I'm not sure there's an easy fix.

[AlanStevens]

It's clearly a resolution problem. Presumably a low resolution has been set because the Moment integration takes too long otherwise. But the integration is simple to do analytically, so:

If you set n to 20 in the above you get the same corner cut-offs in M as in your original.

Alan

[pzimpelman]

Thanks!