Gradually losing z-position during 3D relief carving

Hello all,

I use a diy router table to do wood relief carvings with 1/4" tapered ball nose bits and a 110V Chinese spindle with er11 collets. I have a consistent problem where the z-position slowly moves upward to the point where the bit is no longer cutting. For example, I would zero the bit and position it at z1.5 coordinate. The bit is exactly touching the workpiece. Then after running for a few hours, the bit is barely cutting. I position the bit at z1.5, and there is a 0.020-0.030" gap between the bit tip and the workpiece. So its like the bit is slowly slipping upward to the point where it no longer is cutting (I ran roughing passes before this with a 1/8" ball nose). This issue has consistently occurred about 6 times now as I run over the same workpiece, each time using a different clamping method, collet nut, collet, and bit.

The part that makes no sense to me, is that there seems to be no cutting force on the bit at all. At most, the 0.75mm radius tapered ball nose from spetool is only taking off 0.02" DOC at 10% stepover (200ipm, 100ipm in z, 12000 RPM). The bit feels like it is tight. So how could the bit be physically moved upward with such little cutting force? I rationalized that the repeated up and down motion of the bit touching the workpiece gradually makes the bit slip upward, but it still doesn't make sense to me.

The other thing is, I tried different clamping methods. I always snap the collet in the nut first. At first, I stuck the bit all the way up in the collet where it was sticking out 1" in the back. I realized this would reduce the clamping force. So next, I pushed the bit all the way but not past the end of the collet, and backed it up so there was a 1/8" gap at the end of the collet. I believe this is the correct way to install the bit in er style collets. After I did this, I reran and the bit actually cut deeper for a time, until it eventually shifted upward to the point of noncutting after the 2-1/2 hour mark. So this tells me the problem has something to do with the clamping force being insufficient, which causes the bit to gradually slip. After this, I tried swapping collets and trying out different clamping forces with the wrenches, resulting in the same thing: The bit z position gradually shifts upward until it is not cutting.

It makes sense to rule out machine mechanical issues, because the problem is so consistent. It never slips downward, for example. It doesn't seem to be a g-code or mach 3 issue, because that one time I changed the clamping method to have the bit not stick out the collet, it improved but did not eliminate the issue. So to me, it seems like the clamping force is simply insufficient to hold the bit even if I crank down on the wrenches with medium or high force on different brand new collets ( I tried chinese collets and technik collets er11).

So I am at a loss. Does anyone have an idea for how I can increase the clamping force, if that is even the problem? Or have some other idea for the possible cause of this problem? If you need more information, please let me know. Thanks.

Do you use any weight compensation for the z-axis like springs or gas struts?

Hi david, thanks for the reply.

No, I don't have any weight compensation of any kind on the z axis, it didn't really cross my mind when I built it.

I was trying to troubleshoot earlier, and this might be more in line with your comment, but I think it has to do with somehow missing steps on the z-axis, not the bit slipping. While the bit slipping seemed to explain the problem, I simply don't observe it slipping, and don't see how it could slip according to my original description.

I'm not really familiar with weight compensation mechanisms on the z axis. Is that something that could prevent the z-axis from losing steps in only the + direction? Perhaps my machine is losing steps on the downward motion which is assisted by the axis weight?

It might help to see a mention of the feed rate in the negative Z direction.The situation described could possibly be caused by the machine missing steps during the more aggressive plunge cuts of the 3D surface.A modest size machine with ER11 collets won't have the same capacity for cutting as an industrial machine and the speed and feed rates may have to be reduced from the published recommendations for heavy duty machinery.

Running Mach3? Known issue.
You could be missing one step when the direction changes. Your controller or pulse engine can't handle a simultaneous step and change in direction. The lost Step pulses add up.

Try flipping the Active State for the Step pulse on the Z axis in Mach3. ie, if it was Active HI, make it Active LO.

Has been found to work in other cases.

Cheers
Roger

Quote:

---

Originally Posted by RCaffin

Running Mach3? Known issue.
You could be missing one step when the direction changes. Your controller or pulse engine can't handle a simultaneous step and change in direction. The lost Step pulses add up.
Try flipping the Active State for the Step pulse on the Z axis in Mach3. ie, if it was Active HI, make it Active LO.
Has been found to work in other cases.
Cheers
Roger

---

Only on the Step though.
If you do it on Dir the axis Direction reverses.

Question. Is the stepper strong enough or maybe too strong?.
I see no mention of the stepper/ driver/ voltage he is using.

Quote:

---

Originally Posted by davida1234

Do you use any weight compensation for the z-axis like springs or gas struts?

---

THIS ^^^

Hi there,

Thanks for the response. I mentioned in the OP I used 100ipm on the z, but I did not mention the motors. I am using nema 23 425oz-in at 36V and 4.2A on the z. I've jogged it at 200ipm with no stalling, so I figure 100ipm is safe for my machine. Also, I tried backing off the speed and acceleration by half and had the same issue.

I am barely making any cuts with this pass, its simply a finish pass over a roughed out 3d relief. So there's hardly any loading on the bit. So I think the motor is capable of moving the z at 100ipm without stalling, but possibly there is some issue with the lack of a counterbalance as suggested by others that could cause the missed steps. After all, the z does move up and down very rapidly with very different loading on the ups and downs.

Hi Rcaffin and dazp,

I haven't heard that one before, I will give it a try and update the situation, thanks.

Hello all,

My axes steps are all active low. I spend a while trying to test out the active low --> active high on the z step pulse idea. I tried just deactivating the "step low active" setting in mach 3, which caused the axis to not move. So I tried rewiring to an active high configuration on my stepper driver (DM542A) and breakout board (DB25-1205). However I could not get it to work in the active high (common cathode) configuration (Pul+ and dir+ connected to BOB pins, Pul- and dir- grounded). Apparently I saw some others who had the same issue with this BOB. I measured the voltage at the BOB pins and measure 2V, not 5V. I think its a design flaw in the no-name DM25-1205 BOB as others have suggested.

So I was forced to put the wiring configuration back to active low. If anyone has another idea for how I could switch to an active high step, I'm all ears, but currently I don't know how to do it with my BOB.

At this point it does make sense what Roger suggested, where steps are being lost due to the simultaneous direction change and step pulse. It seems others tried switching from active high to active low, and that fixed their issue. However I am already in active low. Besides, I don't understand why switching that would eliminate the problem. After all, why don't I see any step losses on my other axes which all run on active low? Is there some other way to stop losing steps due to the dir and pul interference besides switching to active high?

I should also mention, I use ethernet smoothstepper. I'm open to any more ideas and assistance as I run more tests myself, thank you all.

Update.

I seemed to have solved my problem. I heard form the folks over at warp9 and some other forums on here that some stepper driver models simply cannot handle fast direction changes like the operation I was doing. I upgraded my DM542A to a DM556T (stepperonline). Wow what a difference.

For anyone reading this with the same problem, or if you are looking to build your own machine, please do yourself a favor and just buy from stepperonline. I fell for the trap of buying cheap crap on ebay and the quality is severely worse than stepperonline's products. And their prices are so low now anyways. I had to upgrade my motors to stepperonline motors also in the past, and they are significantly better, aka they don't stall nearly as fast as cheap chinese steppers. My z-axis feels like a new machine, so quiet and smooth. I'm getting new drivers from stepperonline for all my axes. My z-position is perfect now thanks to the new driver.

Thanks everyone for the suggestions, they helped lead me to my solution. Hopefully this helps someone in the future.

Quote:

---

Originally Posted by Eric1

Hi there,
at 36V .

---

This is the issue.

I know the motors in question. Generally have an inductance value of 3.5mh.
Can't achieve velocity and torque is low on 36v.
you need at least 48v.
Best solution is to swap your drivers out for DM860 instead.
Run them at 60v and set driver current to the next level below 4.2A.
If you have no heat issues up the current to the next level up and so on.

Many people do this as have I. Tthe improvement is vast!!!!.