Lesson Learned: Do Not Let the Steppers Skip when Zeroing

Feature Request
1

My experience may sound stupid to many, but I thought someone new to machining like me may save a cutter and some material after reading this.

I was cutting some 1/8 6061 Aluminium plate, and zeroed the machine as usual, but while I was zeroing the Z axis, I pressed 1mm instead of 0.01 once (perhaps skipped more 0.01 steps when I pressed the cutter against the plate), causing the motor to skip. Then the cutter went straight into the plate when cutting and snapped. I thought I had the wrong feedrate but it’s the zeroing that caused the problem. Although the cutter looks right on top of the plate, the program thought it’s way lower because the steppers on the Nomad do not have any feedback system.

So once you think you let the stepper motor skip, restart Carbide Motion and let it home again. Even you don’t break anything the zeroing will be off and the part won’t come out correct.

I set the topic to Feature Request because I’d like to have an option to home the machine again without restarting the program. I do not have the latest version so sorry if it has already been added.

I tried again with correct zeroing later, and the part comes out nicely. If anyone is interested in it I would love to share the settings and pictures of the finish. I used Inventor HSM as my CAM software for a while and can at least program some easier parts. It appears to be very similar to Fusion 360, so if anyone is interested I can explain the workflow too.

Thanks for reading. Hope this helps.

p.s. Is the correct term for a stepper motor not being able to perform a step due to lack of torque/blockage “skipping”? This is the best I can describe it.

2 Likes
2

In the ShapeOko forums, we usually termed them as missed / lost steps, but the machine skipping works too.

1 Like
3

In the ShapeOko forums, we usually termed them as missed / lost steps, but the machine skipping works too.

:+1:

I learned and teach it as dropped or missed steps. YMMV.

Then the cutter went straight into the plate when cutting and snapped

Even with experience mistakes can happen. Thanks for sharing this! Others can learn from our mistakes.

On this end: Been there! Done that!

Although the cutter looks right on top of the plate, the program thought it’s way lower because the steppers on the Nomad do not have any feedback system.

Yup, that’s a “Don’t do that!” kind of thing. :scream:

By-the-by, this can happen when one collides with a fixture of the stock too.

I once helped debug a problem where parts were coming out ever so slight off - but only for the machining done after a certain point of the job (part of a tool change). This was a high speed machine and it was something the human eye would not catch… the tool ever so slight grazed a fixture and one, yes, one step was lost.

Eventually, I tracked it down to an incorrectly modeled fixture. It was ever so slightly larger than the specification in the CAM package. The anti-collision software said everything was OK (which was correct) but in “real life” it wasn’t. Yeah, they probably should be passing so close to the fixture but because it was production shop the time saved eventually added up to measurable $$$.

Touch top Z0 requires a delicate hand. I teach the technique like this:

Use a bright light (trivial with today’s bright LED flashlights) shining across the stock. Manually rotate the end mill so that flutes “just so” and one can see a long shadow of the end mill and light underneath it. The long shadow does not touch the end mill. As one steps down, the shadow eventually touches the end mill and the light underneath the end mill disappears. Z0.

There is a way to avoid having to do this, or at least do it less frequently.

A) Strap on a spoiler board.

I use a sea-of-holes bed so I cut a custom spoiler for a job or a group of jobs.

B) Mill the spoiler board flat. Now it is flat and square to the machine.

C) Touch top Z0 on the top of the spoiler board.

D) Set Z0 of the CAM to the bottom of the part (NOT the top).

As long as the spoiler remains in good shape, one doesn’t have to zero Z again.

A variant of this technique is commonly used for wood.

A & B remain the same.

C) Put two pieces of writing paper on top of the milled spoiler board. Z0 to that.

The trick is to move down in Z until the paper can just barely move. Set Z0. Remove the paper.

D is the same.

When the part if milled, the Z0 is offset from the spoiler by about 0.006. When the part is done, there will be a tiny bit of wood left - onion skin. Remove the part and the onion skin simple breaks away with no damage to the finish of the part.

If one is using on square end end mills, the spoiler should NEVER be touched!

mark

2 Likes
5

Thanks for the tips Mark. I am using a spoiler board, but since I am using the carpet tape to fix the Al plate onto it, there is some distance between the bottom of the Al plate and the spoilerboard. Since my part includes milling part of the part to half its thickness, it might not be as accurate if I zero to the top of the spoilerboard. I used the paper method, and I got a nice 0.62-0.63 thickness on the area that should be 0.625.

The finishing on the sides seems a little rough. I am going 8000rpm, 12in/min and 0.01 stepdown, using Lakeshore Carbide 1/8’’ ZrN coated corner radius flat endmill.

I want to try using a multimeter, and connect the probes to the Al plate and the cutter to see if they touch by testing the resistance. The position that I get a reading should be where the cutter touches the plate. I wonder if anyone tried this before.

1 Like
6

I understand. The general cases I mentioned above do assume one has a means to clamp the stock without tape (e.g. step block and clamps, sea-of-holes bed).

I want to try using a multimeter, and connect the probes to the Al plate and the cutter to see if they touch by testing the resistance. The position that I get a reading should be where the cutter touches the plate. I wonder if anyone tried this before.

That is how most Z zero plates work. It is an excellent solution. Many CNC machines have automatic Z zero plate routes. Lay the Z zero plate (of a known thickness) on the top of the stock. The machine gently lowers the tool (of a known length due to automatic tool length determination) until the circuit closes. Z0 on the top of the stock is then simple math.

Here is an example:

mark

7

I come up with 11 IPM for finishing, 5 for fine finishing (keeping the same RPM and step down).

I’m tempted to say let’s run things faster to get a better finish - 10K RPM, 13 IPM.

mark

8

Thanks! I will try that next time.

9

Thanks for the explanation! Glad to know it’s an established method.

10

I keep the paper. If the paper is just barely scratched after machining the zero is essentially as good as it gets. If the paper is actually cut, the zero is a bit low. If the paper is untouched, the zero is a bit high.

Printing alignment marks / patterns on the paper is helpful. That makes getting the material close to square / aligned, even in the middle of the waste board, a breeze.

I don’t know for certain but the paper seems to add some friction helping to keep the material in place.

As an added bonus the part shadow left on the paper is just too cool!

11

As strange as it sounds, the paper is more abrasive than the material which is being cut, and will dull the endmill even more.

12

I keep the paper. If the paper is just barely scratched after machining the zero is essentially as good as it gets. If the paper is actually cut, the zero is a bit low. If the paper is untouched, the zero is a bit high.

Agreed.

Due to positioning errors (0.001-0.005" repeatability in CNC machines in the classes we’re talking about) the paper gets ripped up, sometime quickly.

The purpose of the shift - by the paper - is to prolong the use of the spoiler board. Leaving the paper helps but also loses the flatness of the machined spoiler - paper has variations often larger than the CNC machine does!

The onion skin of the wood serves the same purpose. One can almost see through it and one can see where it is thicker or missing.

As an added bonus the part shadow left on the paper is just too cool!

It is cool and I have left it in from time to time as a teaching tool.

I don’t know for certain but the paper seems to add some friction helping to keep the material in place.

Paper can be highly abrasive… and make more of a mess.

Not with my expensive end mills. YMMV. :grinning:

mark

13

I must be buying better laserjet paper than you. :wink:

14

Ditch the paper and use shim plastic. You get a known thickness that way. I also use it for zeroing the Z,you can deduct the thickness for an even closer tolerance.