Mechanism: User-defined measure not available in Motion Study

[Pettersson]

Hi, everyone. Here's hoping someone recognizes this problem.

I have a simple assembly with an arm and a spring. I want to make a motion study measuring the force from the spring in the tangential direction that the arm is moving. To do this, I've set up some measures for the force of the spring, as well as some position measures that allow me to calculate the force component in the relevant direction. I've then set up a user-defined measure that looks like this:

Spring_load * X_coord / Spring_length

The Spring_loas measure is a Net Load measure for the spring. The X_coord and Spring_length measures are position measures.

So far so good. I can run a Dynamic Analysis and calculate my measure. I can pull up a graph and see how it behaves over the analysis time. But now I want to do an optimization study for this. To do this, I create a Motion Analysis. Now, in the area where I get to select my parameters, all the direct measures (Spring_load, X_coord and Spring_length) are there, but not my user-defined measure. What's even weirder, some troubleshooting shows me that if I remake my calculation to this:

X_coord / Spring_length

... I get to include it in my motion analysis. But as soon as I add the Spring_load measure in my calculation, it disappears from my motion study options. I'm able to use a derived measure based on position measures, but as soon as I add the spring force into the calculation, it can no longer be included in my analysis. Anyone have any idea of what's going on here?

[gfraulini]

Within the user defined measure, had you set the measure unit?
It could be given by that.
In any case I don't be surprised by this because, for example (if I well remember) in Mechanism you can't utilize user defined measure not even for servo motors.
Monday I'm going to take it a look. Years have passed since I done my last motion analysis feature...

[Pettersson]

Thanks for your answer. Yes, I have set a unit, as you can see in the screenshot. I've tried setting it to both Force and Dimensionless, but it doesn't help. And the problem isn't that it's user-defined, since it works just fine if I remove the Spring_load from the calculation. As long as I work with just position measures it's fine, but when I add the net load measure, it's not available anymore. I's really weird.

For the moment, I managed to make a workaround since a spring force is so easy to calculate. I basically made my own Relation Analysis calculating the spring force and used that in my Motion Analysis, but that's not that elegant, and only works with such a simple thing as a spring. For other types of studies, that obviously wouldn't be an option, so I'm still hoping to figure out the issue here.

Thanks again for your attention. I expected this thread to not get any replies, so I was very happy to read yours!

[gfraulini]

I've done some test and I'm not be able to utilize the "tangent_load" for optimization/feasibility study.
Even by specifying a relation with data taken from motion analysis

• max_sping_load
• max_x_coord
• min_spring_length

I can't utilize it for the optimization.
It appears you can use only dimensions/parameters/measures directly created by Creo

[Pettersson]

Thanks for having a look. As I wrote above, I manage to get access to the user-defined measure in my motion study as long as I don't include the spring_load measure in the calculations. It's really weird. The issue is with getting the user-defined measure into the motion analysis. I have no problems running the optimization study.

I can run studies for optimizing calculated parameters without problems. Uploading the model I used in the end. This doesn't use the spring in mechanism, though, but rather calculates the spring load through a relation. But I can run an optimization study trying to minimize the spring force difference (DF:Tangent_force_diff) while keeping the tangent force (F:Tangent_force) to (for example) 10 N. And the tangent force difference is calculated in a relation analysis feature.

[SweetPeasHub · Accepted solution]

Its just a very annoying bug with a simple work around. 

Do not copy the measure from another measure.  Do not edit the measure with additional variables after apply/OK the first time.

In one step create a user measure from scratch with Spring_load*X_coorddp/Spring_length.

This should work, let me know if it doesn't for you.

Why?  I think there are data type issues not properly allocating space for the data series when the types change due to an edit.  Something is missing in the definition that does not pass into the CAD side.

As a note, there are two other ways to measure the force you are wanting.

1. net load on the servo motor getting torque.  Then a measure to divide by 60mm.  (or you can use a separation measure to get the 60mm distance but watch this bug turning up again if you edit/change the 60 constant to the separation measure. (create a whole new measure instead)

2. add a dummy pin part where the spring connects.  Get the connection reaction for the pin joint referenced to the moving arm. I prefer this because it is more directly obtaining what you want and will not have the bug that user defined measures can have.

[SweetPeasHub]

something is up with this thread. it is not showing my answer and has widget errors. does anyone else see that?

[SweetPeasHub]

Hi SweetPeasHub,

Great job! Pettersson marked your answer to a question as an accepted solution!

Subject Re: Mechanism: User-defined measure not available in Motion Study
Date: 05-11-2019 04:23 AM
Its just a very annoying bug with a simple work around.

Do not copy the measure from another measure. Do not edit the measure with additional variables after apply/OK the first time.

In one step create a user measure from scratch with Spring_load*X_coorddp/Spring_length.

This should work, let me know if it doesn't for you.



Why? I think there are data type issues not properly allocating space for the data series when the types change due to an edit. Something is missing in the definition that does not pass into the CAD side.



As a note, there are two other ways to measure the force you are wanting.



1. net load on the servo motor getting torque. Then a measure to divide by 60mm. (or you can use a separation measure to get the 60mm distance but watch this bug turning up again if you edit/change the 60 constant to the separation measure. (create a whole new measure instead)



2. add a dummy pin part where the spring connects. Get the connection reaction for the pin joint referenced to the moving arm. I prefer this because it is more directly obtaining what you want and will not have the bug that user defined measures can have.

[Pettersson]

Yes, there's something weird. I received an e-mail with your answer and marked it as a solution there, but couldn't see the actual reply in the thread. You were on the money: deleting the measure and recreating it made it work for me. Seems like there's a bug with user-defined measures being redefined. Does anyone know if PTC has been made aware of this issue?

The solution with putting a dummy pin is a good one, too. I thought it annoying that I couldn't measure reaction forces at the mounting point of the spring, but I didn't think of making a small pin and attaching the spring to that. Thanks for that idea!

(It's rather ironic that a thread about a bug in Creo would suffer from a bug in the forum sotware.)