Need help rigging scissor like mechanism

[x2ego]

Been trying to figure this out for a while and my head is spinning with all the tuts and suggestions I've looked at for rigging this.  I model, yes.  I rig, no.  Can someone please help me make this work?  I tried bones and IK solvers, I tried constraints, although I can't remember how to make those work either, I tried a lot of stuff.  The pivots all align properly and I've placed dummies on every pivot in case they are needed.  In the pic, the right most link drives it all.  Force is applied from right to left and the axis is constrained by the slot it is riding in.  From there, the force is transferred to the second link which is anchored on it's left pivot.  As such, the second link travels counter clockwise around it's left fixed pivot driving the next two links.  Those two links are constrained by the slot at the top, resulting in movement confined to the horizontal.  I have the mechanism working by doing keyframe animation moving each individual link frame by frame (yeah, it was a huge pain), but I would like it to be rigged properly. 

[leeminardi]

Essentially what you have is a 6-bar linkage that is the combination of two 4-bar linkages.  The bottom portion of the mechanism can be considered as a piston pushing a crank.  The applied force is the piston action and the second link is the crank.  The pivot for the second link should be a its lower left joint.  The pivot for the first link should be at its upper left.  A dummy should be placed at the bottom right corner of the first link and linked to the first link.  The first link should be linked to the second link.  Once you have that select the second link and apply an HI IK solver clicking on the dummy for the end of the chain.  At this point you should be able to move the IK Goal to the left and right and the first and second link should appropriately follow.

The next thing to focus on is the upper portion of the mechanism. Here we can consider the second link as a crank and the horizontal rod as the piston confined to horizontal motion. The cranks is driving the piston.  It is similar to the tutorial I posted here.

I am traveling until next Wednesday and away from my computer otherwise I would provide more detail and a picture or 2.  Hopefully this is enough to get you started.  

[x2ego]

Hello Lee,
thank you for taking the time to wrestle with this.  After I posted this, I was catching on a little bit more and actually examined and dissected this video: https://www.youtube.com/watch?v=kqt-H1MC0Uw

I thought this might help me with the linear and rotational aspect of the project.  It's a very simple solution and I had some fun quickly reproducing in it Max.  I used to be a bit more versed with this stuff, but have been away from rigging for a long time.

Let me take a shot at what you describe and get as far as I can.  Maybe I'll surprise you and have it running by then.  It helps to have another set of eyes on a problem for a fresh perspective.  Let me hop to it and thank you for guiding me along a different path.

Eric

[leeminardi]

The tutorial you reference provides an approximate solution.   The connecting rod and piston do not stay together as they should.  It is a quick and dirty approach for when close enough is good enough.  The author and several comments reflect this if you dig down enough.

My piston tutorial provides an exact solution.  It requires a little math but hopefully I have done the hard part.

Let me know if you have any questions.  I have my Chromebook with me.  It's good for email and internet browsing  but unfortunately not Max.

[x2ego]

Lee,

I followed your instructions and the first two links are moving as they should.  Next, as in your text, I tried to focus on the upper portion by applying the simple piston method, but it gave me unexpected results with the dummy jumping far to the left when I tried the same inherited movement values.  At first I thought that maybe the pivots weren't lined up, but they were good.  Thinking I set up the links between the dummies and the model links incorrectly, I deleted the dummies and repeated the procedure.  Same result, so somehow it's not working as expected.  Since the piston method is not the best solution, I decided not to waste too much time on it and report back here. 

Next I will try and follow your tutorial and see if I can apply it.  Half way there.

[leeminardi]

Good luck. If you cannot get it to work post your file by Wednesday and I will look at it on Thursday of that is not too late.

[x2ego]

Having some issues the values for the lengths of the links in my model.  Using the Max measuring tool and snapping between the pivot points on the ends, I made a note of the results for the parts corresponding to the crank (link 2 @ 34.294") and the rod (link 3 @ 40.259").  My model is oriented differently than you example, so my axis of rotation is x.  Dutifully following your example in the tutorial, I created the variable crank, rod, and alpha with their appropriate values.  Next I inserted your mathematical formula and hit evaluate.  The computer didn't blow up, my model didn't disintegrate, but the part of the model the expression was applied to move much more than it should have, albeit in the right direction.  My guess is that the values for the crank and rod are to blame.  I checked in the units section of Max and it's set to US Standard decimal inches.  Since the model was created using inches, I would have thought that the measurements of the rod and crank, also measured in inches would be correct, but something is amiss.  

I'll keep working on it.

[x2ego]

Another thought came to mind that maybe the reason it's also not working as expected is the fact that your equation was based on the orientation of your example.  My layout isn't the same, although the principle applies.  I don't have the mathematical acuity to make the necessary changes, but maybe you can put me on the right path once more.

[leeminardi]

1. Make sure both the display and system units are the same.  If they are not it will screw up the distances you use.

2. Check out my Part 2 slider-crank tutorial here.  The Part 1 tutorial assumes that the center line of the piston passes through the center of the crank.  Part 2 allows for an offset "H". Your linkage has an offset.  In addition, Part 2 links everything to a base dummy allowing the mechanism to be freely oriented in 3D space.  Your linkage, as mine, has the piston move in the x direction so you should be able to use my equation for the piston  but should change the sign from + to - at ... + sqrt(rod^2... to ...-sqrt(rod^2...  This should compensate for the piston being to the left of the crank.

[x2ego]

Just got done going through the second video.  Still trying to digest it.  In the meantime I'll attach a video of what I have and a Max file if and when you have time to peek at it.  Too bad your Chrome Book doesn't Max, lol. 

[leeminardi]

1. Forget about the orange link for now.

2.Place a "base" dummy at the pivot point of the crank.

3. Link all components to the dummy.  This will ensure the coordinate system for all components will have their reference point as the center (pivot point) of the cranks.

4 Place a dummy at the top end of the green link. It will act as the piston moving only left/right. This dummy will only be linked to base dummy (not to any links).   The expression controller described in my Part 2 video should be assigned to this dummy.  The green link should have a LookAt constraint assigned to it pointing to the dummy acting as the piston.

[x2ego]

A bit of frustration setting in.  When I put the base dummy at the pivot point of the crank and I link the crank to it, the crank assembly tilts?  First thing I did was check the pivot points - dead on x/y/z on both.  Next I checked RCS and both are the same orientation in the world view.  Out of frustration I applied xform to the crank with no change at all.  Still doing the tilt hop when linking to the dummy.  What basic step am I missing here?  Btw, google didn't bring up much in the way of help on the subject of objects tilting or moving when linked to a dummy.  I feel like a dummy right now.

[x2ego]

Sorry, I figured it out.  I had to unlink everything before linking to the base dummy.  Let me try this again.

[leeminardi]

I wish I could look at your file and help you out.  I hope you figure this out before I return from vacation this Wednesday.

It is best to have the pivots of objects orthogonal to the world coordinate system when possible.  Having the base dummy should simplify things but all relevant objects should be linked to it directly or via a parent/child hierarchy.

One other thing that may cause a problem for you (like you need something more) is that the equations I posted assume that when the crank is vertical the angle alpha is zero.  You should check this for your model.  If it isn't zero then you would need to add or subtract the angle from alpha in the expression.  For example, cos(radToDeg(alpha)) would become cos(radToDeg(alpha) + 12.4) or cos(radToDeg(alpha) - 12.4) if  12.4  was the angle (in degrees) shown when the crank is vertical.  

Good luck!

[x2ego]

Good points Lee, thank you.  My process is like an ant hill.  I gather non stop and put it all in a pile in my mind.  At first things make sense, then not so much, but it all goes in.  Eventually I get overloaded, things start to overlap, and a fog sets in.  At that point I have to step back and take a break from it all.  Once I do that and totally shut myself off from it, my mind starts to settle.  It may take a few days, or it might happen overnight, but often I will wake up and have an epiphany.  My subconscious keeps working on the problem and often comes up with the answer or method to overcome what my waking mind was incapable of doing.  Kinda happy it works that way, although when I'm in research mode, I'm a bit OCD and don't seem to know when to stop unless exhaustion gets the better of me, lol.  See what I mean, 6 am on a Sunday morning. 

[leeminardi]

Whatever process works for you. Go for it.

One more thing to keep in mind, I assume your action is parallel to the XZ plane, if so the Y coordinate of ALL moving objects should be the same. If they are not it could cause the tilting you reference.

[x2ego]

Just wanted to check in.  Will get back to the rig tonight.

[x2ego]

Tried several times to get your method to work on my model, but I think because the orientation is different and because I suspect the units somehow don't wanna play nice, I wasn't able to.  Not deterred, I decided to try and replicate your model with the same orientation and almost have it, but the vid below is my second attempt and both times, the piston is in the wrong place and out of sync.  I checked and double checked the equation - this is a direct copy/paste from the Max Expression Controller window - crank*sin(radToDeg(alpha))+sqrt(rod^2-(crank*cos(radToDeg(alpha)))^2), but I'm sure it's correct.  In your YT vid, you weren't specific about the placement of the three components, but I placed them just like you had in the side view.  I made sure the pivot points were all on the same horizontal plane.  I did notice in the wireframe view, that the piston pivot attached to the rod in an unexpected way after pressing "evaluate".  I would have expected the piston to attach to the end point of where I measured the rod.  It did not.  This again, might have something to do with the units???

[x2ego]

Here's a step by step wireframe that you might use to interpret the issue?

[leeminardi]

@x2ego one major problem that I see is the when the crank is at 3 o'clock (horizontal pointing to the right)  the piston should be at its furthest position to the right.  Likewise, when the crank is at 9 o'clock, the piston should be at its furthest to the left position.  The fact that the piston is not at its extreme positions for these two cases indicates that your angle alpha is not 0 when the crank is pointing straight up.  To fix this you can either:

1.  Rotate the crank so that it is vertical and with transform type-in on note the angle for Y. Include this phase angle with alpha in the expression (e.g., cos(radToDeg(alpha) become cos(radToDeg(alpha) + the phase angle in degrees)

2. With the crank vertical change rotate its pivot (affect pivot only) so that Z is pointing up.

With either method  verify that the piston position makes sense with the angle of the crank.

As I mentioned, I will be home on Thursday and can look at your Max file then.  Be sure to post the .max file (NOT rar) you want me to look at in a 2021 or earlier version.