Measuring Gantry Rigidity

Questions/answers/discussion about initial setup of your CNC Shark

Moderators: al wolford, sbk, Bob, Kayvon

Post Reply
JayMcClellan
Posts: 18
Joined: Mon Feb 25, 2013 8:53 pm

Measuring Gantry Rigidity

Post by JayMcClellan »

I'm about to make some modifications to my shark that should make the gantry much more rigid, and I wanted to get some quantitative measurements of the rigidity before and after the modification. I came up with this procedure for measuring the deflection of the router bit when a known force is applied in different directions, and I decided to post the procedure here along with my initial results in case anyone else wants to make comparable measurements on their system. I arbitrarily chose a force of 5 pounds applied 1 inch below the collet, which is more force than one might expect in practice but it's within the range of possibility when cutting aggressively, and it gives an easily measurable deflection on my shark (alas). I plan to repeat this measurement after my modifications, and then periodically thereafter to help detect any changes in the system such as looseness in the linear bearings for example.

Equipment
1/4" diameter steel rod about 2" long, with a hole drilled near one end and a loop of wire through the hole
Dial indicator and a way to mount it to measure deflection of the bit (I clamped down a steel plate and then used a magnetic base to hold the indicator)
A push/pull scale capable of measuring 5 pounds

You can buy a commercially-made push/pull scale (Amazon has some for under 20 bucks) but I made my own with an aluminum rod passed through a spring and secured at one end. Holding the free end of the spring, it can pull on the hook or push on the end of the rod. I calibrated it by marking how far the spring stretched with a 5-pound weight hanging from the hook.
Image

Procedure
1. Move the router to the center of travel in X, Y and Z. It doesn't have to be exactly centered but it should be close. Leave the controller powered on so that the servos will hold their position, but make sure the router is switched off. Since the actual height of the router can be changed by sliding it in the clamp, I also recorded the height of the collet above the bottom plane of the Z stage, which was 0.75 inches in my case.

2. Install the rod in the collet, with the hole positioned 1 inch below the collet.

3. Mount a dial indicator parallel to the X axis and touching the rod at the hole position. The indicator should have a flat tip so that minor movement of the rod perpendicular to the indicator won't affect the reading.
Image

4. Zero the dial indicator, then measure the deflection while applying a 5-pound force in the +X direction. In the photos I don't quite have the spring all the way to the mark, as it's tricky doing this with one hand while photographing it with the other. Also record the deflection after the force is removed, as it may not return to zero.
Image

Repeat, applying the force in the -X direction.
Image

5. Position the dial indicator parallel to the Y axis, and repeat the measurements by applying force in the +Y and -Y directions.

6. Remove the rod and position the dial indicator vertically below the collet. If there's insufficient room, you can remove the collet nut or place the indicator against the router to the right (+X) or left (-X) of the collet.
Image

Measure the vertical deflection while applying a 5-pound force downward (-Z) in line with the spindle axis, and also record the deflection after the force is removed. I had a hard time holding and photographing this so I cheated and rested my hand on top of the router but don't do that!
Image

Repeat, applying the force upward (+Z) along the spindle axis. I tied a piece of string across the top of the router in order to pull it upward with the scale.

7. For each of the 3 axes, add the absolute values of the deflections with force applied in each direction, and then divide by 2 to get the average. Do the same for the deflections after the forces were removed, to get a measurement of the "play" in the system. A large amount of play probably indicates that something is loose.

Results
Here are the results of my initial measurements on my stock Shark HD 2.0:

X deflection 0.022", play 0.0025"
Y deflection 0.014", play 0.0015"
Z deflection 0.010", play 0.0015"

As a rough rule of thumb, we can expect the positional error due to flexing of the gantry to be within the deflection numbers (plus or minus) with up to 5 pounds of force applied to the bit. And we can expect the error to be within the play numbers (plus or minus) with no force applied to the bit. The total accuracy of the system may be worse due to other factors, but it's unlikely to be better. Whether these numbers are good or bad depends on what you want to do with the system, but I expect that these are typical numbers for this model of shark. The newer models should have less deflection since they have an aluminum back plate on the gantry, which will be my next project. The Y and Z numbers are closely related since they come largely from twisting the gantry due to the applied torque, but if one of them suddenly goes up then I think it might indicate something loose along the corresponding axis.

Post Reply