Machsupport Forum

Mach Discussion => General Mach Discussion => Topic started by: rrsquez on November 19, 2012, 05:26:42 PM

Title: Z-Axis Repeatability: Small Changes
Post by: rrsquez on November 19, 2012, 05:26:42 PM
Hello All, I am using MACH 3 to route out my Printed Circuit Boards (PCBs) designs. The main problem that I am encountering is that the Z-Axis looses track of position. My plunge depth is only 0.01 mm, so I cannot tolerate very much error.

For regular DIP or 1st gen surface mount parts, this is not much of a problem. The problem is that I cannot do the latest generation of surface mount footprints without wiping out traces. This is because my "V" bit gets wider as it goes deeper. It goes deeper (hence wider) unintentionally.

The problem occurs for even a small design with just 1 fine pitch part (0.65 mm spacing). Starting off, my vertical repeatability was very good. However, after a couple of hundred lines of code I can tell that the cuts get progressively deeper (can tell by sound). Looking at the Z Axis values, all should be well. However, when I stop the code, and recheck my vertical axis, I see that it has drifted (deeper). It appears that MACH 3 has lost track of a few steps. This problem is not as bad for X and Y axis, but for Z axis, my tolerances are a lot smaller.

I have noticed that the loss of tracking is always unidirectional (not random). Steps are always lost to make the machine cut deeper than intended, never in the opposite direction.

I have read as many post about "Z-Axis Repeatability" as I could find. I have kept my lead screws clean and properly lubricated. I have already decreased my velocity and acceleration to as far down as I could manage. I have also slowed down my feed rate to less than 60 mm/min, my plunge rate is 1 mm/S and my lift bit rate to 750 mm/min. I have also limited the amount of clearance to 1 mm (was 4mm) to decrease the amount of Z-Axis travel. None of the above affected the amount of error I detected (no improvement). I feel that I must be doing something wrong.

I'm not sure how to manage the "backlash", yet. I'm not sure if that is the problem (or one of), but it is something that I have not tried yet.

I would greatly appreciate any advice that anyone can give me.

Thank you, Richard V
Title: Re: Z-Axis Repeatability: Small Changes
Post by: BobsShop on November 19, 2012, 06:16:42 PM
How far are you lifting your quill (z) after cutting.  A technique used on older manual mills and lathes is to back out several revolutions from the surface you are working on then cranking back in the same amount.  Using this method can compensate for any sloppiness in the the lead screw.  I use a small taig mill that doesn't have ball-screws.  When engraving metal, I generally back out .100 of an inch.  Yes, this does increase the time it takes to finish a job.

Although this small mill's repeatability is questioned my some, I find my cuts are consistent.  Have to be honest though.  I doubt the cuts are within .01 mm.  I will be interested in seeing how you address and solve this concern.

Bob@BobsShop
Title: Re: Z-Axis Repeatability: Small Changes
Post by: rrsquez on November 19, 2012, 08:21:07 PM
Hello Bobshop, thank you for your reply. I thought I already mentioned this in my post: I am only lifting my "quill" by 1 mm off of my PCB surface. Previously I had this set to 4mm. Reducing the amount of Z-Axis travel did not help my problem, though.

I think your solution implies that I am unable to set the proper distance from the surface. This is not the case. In fact, I get good results when I first start off. It is only after the machine has been running for a while that the problem shows up.

It appears that Mach 3 loses track of the Z-Axis motion after several cycles of up/down motion.

REPOST:
Hello All, I am using MACH 3 to route out my Printed Circuit Boards (PCBs) designs. The main problem that I am encountering is that the Z-Axis looses track of position. My plunge depth is only 0.01 mm, so I cannot tolerate very much error.

For regular DIP or 1st gen surface mount parts, this is not much of a problem. The problem is that I cannot do the latest generation of surface mount footprints without wiping out traces. This is because my "V" bit gets wider as it goes deeper. It goes deeper (hence wider) unintentionally.

The problem occurs for even a small design with just 1 fine pitch part (0.65 mm spacing). Starting off, my vertical repeatability was very good. However, after a couple of hundred lines of code I can tell that the cuts get progressively deeper (can tell by sound). Looking at the Z Axis values, all should be well. However, when I stop the code, and recheck my vertical axis, I see that it has drifted (deeper). It appears that MACH 3 has lost track of a few steps. This problem is not as bad for X and Y axis, but for Z axis, my tolerances are a lot smaller.

I have noticed that the loss of tracking is always unidirectional (not random). Steps are always lost to make the machine cut deeper than intended, never in the opposite direction.

I have read as many post about "Z-Axis Repeatability" as I could find. I have kept my lead screws clean and properly lubricated. I have already decreased my velocity and acceleration to as far down as I could manage. I have also slowed down my feed rate to less than 60 mm/min, my plunge rate is 1 mm/S and my lift bit rate to 750 mm/min. I have also limited the amount of clearance to 1 mm (was 4mm) to decrease the amount of Z-Axis travel. None of the above affected the amount of error I detected (no improvement). I feel that I must be doing something wrong.

I'm not sure how to manage the "backlash", yet. I'm not sure if that is the problem (or one of), but it is something that I have not tried yet.

I would greatly appreciate any advice that anyone can give me.

Thank you, Richard V

Title: Re: Z-Axis Repeatability: Small Changes
Post by: BR549 on November 19, 2012, 09:48:21 PM
You are losing steps on the lifting of the Z. That indicates your Z motor is NOT strong enough to do what you want to do. I could be from wrong driver settings,  too low of voltage for motors , feedscrew too fine causing motor to have to go TOO fast(out of torque range), Machine binding, Motor too small for Z loads. Acceleration setting too fast for motors, ACtive high/low setting for drives could be wrong.

Those topics have been covered about a zillion times please do a search.

Just some thoughts, (;-) TP

 
Title: Re: Z-Axis Repeatability: Small Changes
Post by: DHay13 on November 19, 2012, 10:03:55 PM
I just had something similar happen to me yesterday but I thought I might have caused it somehow. I was doing a 'test cut' on a piece of MDF with a simple text engrave. I set the Z depth to -.1 and it worked great. I then tried another one with -.1 again but it was much deeper this time. I zeroed my Z the same as the first piece. It looked like my bit might have dropped down slightly. So I re-adjusted it and re-set my 0. After re-starting my program it was about .2 above the material but the DRO said -.1. I ended up zeroing the Z with the bit about .05 above the surface and running the program again. The cut was perfect.

I haven't looked at it again to figure it out but that's what I experienced.
Title: Re: Z-Axis Repeatability: Small Changes
Post by: Tweakie.CNC on November 20, 2012, 03:33:55 AM
Hi Richard,

Make sure that your V bit is actually being held securely in your collet (and that the collet is the correct size for the tool) - It is often the simplest things that cause us the problems.  ;)

Tweakie.
Title: Re: Z-Axis Repeatability: Small Changes
Post by: RICH on November 22, 2012, 07:26:34 AM
Ricahard,
I agree with Terry, reply #3. Trying to maintain axis position / repeatabliy to .01MM ( .0004") requires excellent hardware.
You can have that much backlash even with good components and a properly configured /adjusted machine.
You don't mention what your calculated resolution is and please note that chances are your machine will not achieve calculated resolution
due to many factors.

RICH
Title: Re: Z-Axis Repeatability: Small Changes
Post by: DHay13 on November 22, 2012, 11:43:45 AM
Not to hijack the OP's thread but it appears we are having a very similar issue. I cut a piece this morning that required multiple toolpaths. The problem was that I somehow forgot to set the toolpath for my outer perimeter cut >:(. So I jumped over to my other PC and created that as a standalone operation. I then imported it to Mach3 and hit 'Go to Z'. The X and Y went to where they were supposed to but then the Z dropped down to about .1 below where my surface is (luckily it was in a pocketed area so no crash). So it should have read about -.1, but read .000. So my Z 0 dropped itself about .1 from the time I started to the time I finished. I didn't measure the depth cuts but everything looks perfect so I don't know what happened? Have to get ready to eat some turkey soon so I'll measure the depth later.
Title: Re: Z-Axis Repeatability: Small Changes
Post by: DHay13 on November 27, 2012, 11:01:35 PM
A little more info. Zeroed all 3 axis and hit 'Regen Toolpath' then lifted the Z and hit start. I lift the Z because otherwise it will drag across the material at 0. The bit moved to the entry point and started its helix. When the Z DRO got to '0' the tip was still about .1 above the workpiece. I then stopped Mach3 and shut off the spindle and re-zeroed the Z after setting it back down on the material. It worked fine after this. The same exact thing happened a couple of times so it is repeatable. Just don't know why.
Title: Z-Axis looses steps
Post by: rrsquez on October 29, 2015, 11:00:12 PM
Hello All. It's been a while now, but I'm back trying to make PCBs with my machine. I've just read my old post and I am having the exact same problem.
I know how to tie down a collet,
I know how to zero my Z axis,
I know the resolution isn't perfect; I get -0.094 rather than -0.01 mm for my Z axis.

I believe the problem that Mach3 looses track of the Z axis. In fact, it looses track in the same direction; Z always goes deeper. I notice this because it bores down into my PCB surface, wiping out the copper that I needed to stay behind. MACH 3 never looses track of X & Y. If it were a "motor power" as someone suggested, it takes more power to dig deeper into my PCB material. Maybe it's when I lift the Z axis?

I do think this may be a speed related issue traveling in Z axis. I was wondering if anyone has figured out a good way to deal with this problem? I wish I could add an encoder to my Z axis, but does Mach 3 support it now? In the older version it doesn't. I look forward to any good responses to this post.

Thank you, Richard V
Title: Re: Z-Axis Repeatability: Small Changes
Post by: Tweakie.CNC on October 30, 2015, 02:34:29 AM
Hi Richard V,

Try halving the Velocity and acceleration you have set for your Z axis in Motor Tuning (save the changes).

If this cures the problem then you need to find the optimum reliable settings for this axis.
If it doesn't cure the problem then I suggest you thoroughly check the mechanicals - such things as loose set screw in couplings, binding or tightness of slides or screw etc.etc.

Tweakie.
Title: Re: Z-Axis Repeatability: Small Changes
Post by: rrsquez on November 28, 2015, 10:46:47 PM
Hello Tweakie. What you suggested is a good idea: halving the speed, and checking the mechanicals. I took apart my machine and thoroughly cleaned the moving parts. I also tuned the motor to give me a slow response to G00 commands (rapid move). I also made sure my feed rates were set low (like 20 mm/min). I even tried using painfully slow speeds, but that didn't help.

Despite all of this, I still keep losing counts. It's always in one direction; my cut depth is set to z = -0.01 mm. I can hear that something is going wrong as the milling takes place. At first it is good, but after repeated up and down commands, it keeps cutting deeper. If I pause the run, and re-zero the Z axis, I see that my Z axis is off by 0.23 mm (roughly). So instead of cutting to z=-0.01 mm, I'm going to z = -0.23 mm (or so). If I were to let it go, it would probably get worse.

In reviewing Part 3 of the Tutorial on YouTube, I saw a configuration for extending the pulse duration. The video explains that that older controllers might have problems with reversing the direction a lot, as I do in Z axis (making a PCB). The video explains that 99% of users don't need anything other than a "0" entry. Does anyone think this will help?

Thank you, Richard V
Title: Re: Z-Axis Repeatability: Small Changes
Post by: Tweakie.CNC on November 29, 2015, 01:26:02 AM
Quote
In reviewing Part 3 of the Tutorial on YouTube, I saw a configuration for extending the pulse duration. The video explains that that older controllers might have problems with reversing the direction a lot, as I do in Z axis (making a PCB). The video explains that 99% of users don't need anything other than a "0" entry. Does anyone think this will help?

I think you need to try changing the value and see if it improves the situation.
(Checking my set-up I use 2 for step pulse width).

Is your Z axis stepper motor getting hot (approx. 60 deg.C) or is it running cold ??

Tweakie.
Title: Re: Z-Axis Repeatability: Small Changes
Post by: joeaverage on November 29, 2015, 01:30:17 AM
Hi rrsquez,
what is the resolution of your setup ie steps per mm. What you have described sounds like the machine is driving down by say 100 steps
but only retracting 99 steps. After a number of moves the depth of cut will increase.
I cut PCB's as well using PCB-gcode plugin for EAGLE, are you doing similar? I had to do quite a bit of work to get the bed of the machine
flat with respect to the z axis and can maintain cut depths within 8um ie 0.008mm and was quite pleased to achieve that. I would have thought
cut depths of 10um ie .01mm quite a challenge, certainly not one I can match.
For most stuff I do I use a 0.4mm two flute endmill, I find the clean cut by comparison to a V bit preferable. For really fine quad flatpacks I have
to use a V bit tho. Doing this I can achieve trace spacings/widths of 0.2mm, I had hoped to get to 0.1mm but found the results not reliable enuf
to warrant populating the board.

Craig
Title: Re: Z-Axis Repeatability: Small Changes
Post by: RICH on November 29, 2015, 07:24:43 AM
Hmm....

0.01mm ( 0.0004") depth of cuts which requires a vey accurate SYSTEM which includes tuning and setup.
 
What kind of machine are you using?
Post your xml file.

Quote
If it were a "motor power" as someone suggested, it takes more power to dig deeper into my PCB material. Maybe it's when I lift the Z axis?

As Terry replied it could be the motor. For very small moves the motor  needs to be capable of delivering high acceleration and thus at 0.01mm
move it's almost instanteous.Thus you can loose steps which could be random or consistant.

You need to meticulously tune the steppers and then find your max
vel and accel that delivers repeatable results. Very seldom will the machine be able to move at it's resolution. Practicaly speaking "zero"
backlash is not to be had.

It gets very difficult to find those very small things that affect accuracy since many things can be the cause of it.
You should be able to say, with out doubt / absolutely, that it is not a speed problem.

First tune the steppers, do tests, to verify movements, etc. to know exactly what each component of the system is contributing or limited to.

BTW,
Resolution is 1/steps per unit...so for say 23000 steps per inch
resolution is 0.000043".

RICH