"Lost Steps"

Post Reply
User avatar
TReischl
Vectric Wizard
Posts: 4585
Joined: Thu Jan 18, 2007 6:04 pm
Model of CNC Machine: 8020 48X36X7 RP 2022 UCCNC Screenset
Location: Leland NC

"Lost Steps"

Post by TReischl »

Yesterday I went out to make my wife a small mold (about 2 X 2) for her pottery. A small cartoon dinosaur. I have been running a lot of these sorts of things lately and it has been pretty much cut and dried.

Not Dino! There they were, the dreaded "lost steps" !!!!!

First thing I did was make sure there were no foibles in the cnc file. That is easy with Mach 3, just take a good look at the cnc toolpath on the screen. Nope, no problem there.

Then I looked at my config screen. . .AHAH! I set up a rotary axis a while back and somehow there was data (like number of steps) from the rotary on my slaved axis. Fixed that. Ooops, still getting steps in same places. Grrrrrr. Learned in the process that a slaved axis uses the same steps as the master and it does not matter evidently if they are way out of whack. Interesting.

So then I moved on to the next simple thing to do (tearing the machine apart is not the simplest thing in my book), adjust the motor tuning down. No go, there were those lost steps. . . .

Ahhh, I found it! The belt on the drive was a bit loose, musta stretched over time. Fixed that, ran this huge test rapiding back and forth and then making sure it came back to zero. To qoute the Fonz: perfectomundo! That wrapped up my day yesterday.

So, this morning I go out to cut Dino. . . . you guessed it, there are those pesky lost steps again!

I have a machine built using CNC Router Parts components. It is fairly easy to pull a drive unit apart. Sure enough the set screw on the motor had worked loose. Fixed that.

LOST STEPS!!!!!!!!

Did a little more investigating. My machine has been running a lot for about five years. The pinion has worn enough that it bottoms out in the rack and allows the axis to be moved about 1/32 with power on the stepper. I now have some replacements on order.

This sort of thing is exactly why I built my own machine. I do not think a pinion gear wearing out after five years is a problem. I do a lot of 3D raster work and that axis sees a lot of travel. The pinion is aluminum and the rack is steel, so the pinion wears out which is a lot cheaper to replace than a rack.

Now, I "think" I know what the problem is. . . . .but after this episode. . . .those pesky steps just may rear their ugly little heads again.
"If you see a good fight, get in it." Dr. Vernon Johns

CarveOne
Vectric Wizard
Posts: 1479
Joined: Sun Nov 28, 2010 5:10 pm
Model of CNC Machine: Home built designs - 3'x4' and 5'x12'
Location: Eastern NC, USA

Re: "Lost Steps"

Post by CarveOne »

The two slaved step motors are running in opposite directions and will stay in sync only so long as they both get the same number of steps to each motor. Mach3 does not make any attempt to keep the two motors synchronized because there is no means for it to know that one motor has traveled a greater (or lesser) distance relative to the other motor. Slipping set screws is a problem with the r&p drive assemblies. My 5' x 12' home built machine sees very little use, and has the original version of the www.cncrouterparts.com drives. It has happened to me also.

Use red threadlocker on the screws and consider grinding a shallow flat on the motor shaft where the set screw leaves a mark on the motor shaft. I use a Dremel tool and a cut-off disk to grind the flats.

CarveOne
CarveOne
http://www.carveonecncwoodcraft.com

User avatar
TReischl
Vectric Wizard
Posts: 4585
Joined: Thu Jan 18, 2007 6:04 pm
Model of CNC Machine: 8020 48X36X7 RP 2022 UCCNC Screenset
Location: Leland NC

Re: "Lost Steps"

Post by TReischl »

That is exactly what I thought C1.

However, I set up the rotary on my machine quite a few months ago and have run a lot of jobs through with the crazy setting on the slaved axis. It was not just a little different but a whole lot. Correctly setup it is 1910 steps. It was set to over 5,000 on the slave.

I am going to be increasing the Z height on the machine by 6 inches in the next few weeks so I can do larger rotary work up to 8 inches in diameter. While I have the motors disconnected from the rack I am going to mess with the master/slave settings and see what really happens. I would think if the slave axis is set to over 5K steps then it should revolve at a very different rate than the master at 1910.

The motors I got from CNC Router Parts came with the flats ground on them. I am hesitant to put any type of loctite on them because they do need to come off once in a great while. Those set screws are very small and it is easy to crack them with too much torque. I figure if I have gone 5 years without it becoming an issue I can just check them about once a year and should be good to go.

I usually have more frequent problems with the Z axis ballscrew and loose set screws. If I were to build another machine I would take a serious look at using R&P on that axis.

Until I get the parts from CNC Router Parts I will be running all rotary jobs since that axis is not having a problem.

Mostly, I wrote this post because we see folks here that have similar issues and I thought it would be good for them to see what others do when it happens.
"If you see a good fight, get in it." Dr. Vernon Johns

User avatar
zeeway
Vectric Wizard
Posts: 3157
Joined: Thu Feb 11, 2010 9:24 pm
Model of CNC Machine: Self-built
Location: SC, USA

Re: "Lost Steps"

Post by zeeway »

I have two motors running my long axis, with one of these set as a slave. Each side also has a home limit switch - I think I remember correctly that they are Hall effect switches. When I "home all" on Mach 3, the main and the slave re-synch.

I also use CNC Router Parts rack and pinion drives on x and y...every so often I re-tension the belts. My last major issue came with the Z axis set screw let loose on a big piece of walnut...was not happy about that. I try to keep an eye on that now.

Angie

User avatar
martin54
Vectric Archimage
Posts: 7339
Joined: Fri Nov 09, 2012 2:12 pm
Model of CNC Machine: Gerber 48, Triac PC, Isel fixed gantry
Location: Kirkcaldy, Scotland

Re: "Lost Steps"

Post by martin54 »

I usually have more frequent problems with the Z axis ballscrew and loose set screws. If I were to build another machine I would take a serious look at using R&P on that axis

What sort of problems do you have with the z axis Ballscrew Ted, funny because personally I would prefer ballscrews to rack & pinion on a machine, especially the z axis. Personal choice I guess for a lot of people.

User avatar
TReischl
Vectric Wizard
Posts: 4585
Joined: Thu Jan 18, 2007 6:04 pm
Model of CNC Machine: 8020 48X36X7 RP 2022 UCCNC Screenset
Location: Leland NC

Re: "Lost Steps"

Post by TReischl »

martin54 wrote:I usually have more frequent problems with the Z axis ballscrew and loose set screws. If I were to build another machine I would take a serious look at using R&P on that axis

What sort of problems do you have with the z axis Ballscrew Ted, funny because personally I would prefer ballscrews to rack & pinion on a machine, especially the z axis. Personal choice I guess for a lot of people.
Hello Martin!

It is not so much that I have lots of problems with the Z axis ballscrew, other than the usual coupling set screw issues most people have from time to time. Like I said earlier I am hesitant to goop any of the loctite products in a small set screw hole cause it is so easy to crack or strip them out in aluminum.

Way, way back, before I am became a propeller head software writer I was a mechanical engineer. One of the things I designed was large flat sheet laser cutters for the steel rule die business. Back in those days gantries were pretty much unheard of and moving tables were typical. We built big honking 4 X 8 tables using 2 inch diameter ball screws tensioned between mounting blocks to help reduce the S curve whips inherent in spinning long cylindrical shafts. That, and a significant amount of power is required to just turn the screw due to their weight and inertia.

What I do not like about them these days is maintaining them. On my first machine I did a "ball screw" type design but the nuts were not the recirculating industrial type. They were delrin with a spring gizmo to reduce backlash. Replacing those is a rather tedious process if you ask me. The other thing I noted is that I could not achieve very high rapid speeds or cutting speeds. Those screws were double starts so getting the machine up to speed was problematic. Five start on an x or y axis results in some loss of accuracy. I never got that machine to rapid above 150 IPM and the cutting speed top end was about 90 IPM.

When I decided to build a new machine I took a serious look at R&P. Maintenance is easier in my opinion. To replace anything in the drive system all that is required is removing one bolt, loosen a turnbuckle and the whole assembly comes off the machine. There are no issues with creating vibration by spinning a long slender shaft at higher speeds. Since the rack is steel and the spur is aluminum the rack does not incur any significant amount of wear. Also, since the rack is upside down and outside the cutting area it does not accumulate debris. Here is a pic of what it looks like from the work area side:
M1.jpg
The drive system for the axis you can see on the right side of the pic is outside that aluminum extrusion.

I like that the spur gear is pulled into the rack by a stiff spring to eliminate backlash. On my last machine those little delrin nuts had been cross slit with a spring wrapped around to provide anti backlash. But I think the inertia of the gantry could easily open that spring causing a bit of overtravel.

Is R&P perfect? Oh heck no. On large hi speed industrial machines it has been replaced by either helical drive R&P or linear drive motors.

On my current machine I can achieve rapids of 800 IPM and easily cut at 250 IPM. I posted a video a while back of the machine cutting at 800 IPM. Here is a link:

https://www.youtube.com/watch?v=Pt8n43_YRjI


There is no way those little plastic nuts would have handled those g forces.

If I were to build a super accurate machine I would probably use ballscrews and sacrifice speed for accuracy. But those would have to be ballscrews with recirculating bearings.

Well, that is my thinking on the subject, of course, it is nothing more than my opinion. :D
"If you see a good fight, get in it." Dr. Vernon Johns

User avatar
martin54
Vectric Archimage
Posts: 7339
Joined: Fri Nov 09, 2012 2:12 pm
Model of CNC Machine: Gerber 48, Triac PC, Isel fixed gantry
Location: Kirkcaldy, Scotland

Re: "Lost Steps"

Post by martin54 »

Ted, things have moved on a bit since you built that machine lol, sounds like you were using ACME leadscrews rather than ballscrews if you were using delrin nuts. I wouldn't build a machine using them either lol.
Speeds can be increased on long lengths by using rotating ballnuts rather than rotating the shaft & from what I have read ordinary shaft seals do a pretty good job of keeping swarf out.

It's not for everyone though you are right, I don't know anywhere near enough to be able to say that one is better than the other, both have pro's & con's & there are so many variables to take into account with a machine build.

Other thing I was going to mention was have you considered a moveable table rather than a machine with a much higher gantry? Have seen a few designs where the table can be raised & lowered with relative ease when increased z height is needed for rotary work. Or put the rotary axis off the end of the main table at a lower level.

User avatar
TReischl
Vectric Wizard
Posts: 4585
Joined: Thu Jan 18, 2007 6:04 pm
Model of CNC Machine: 8020 48X36X7 RP 2022 UCCNC Screenset
Location: Leland NC

Re: "Lost Steps"

Post by TReischl »

Now you are starting trouble Martin! :idea: ..... :D

Moving tables, grrrrrrr, and double grrrrrrrr!

Those were the first tables I designed those many years ago. WOW! Do they take up floor space! A four foot table needs at least 8 feet of space to move.

Ours worked pretty good but we were using Thompson shafting, linear bearings, the whole works.

One of my last designs was a rotating ball nut on a moving table. We could achieve decent speeds with it and very little vibration. Laser cutters are very sensitive to vibration, makes for a very lousy cut.

Yup, you are right about the acme screws. But hey, I was doing it on the cheap.

This morning I put in the new gantry riser blocks. Added 6 inches to the Z clearance. So now the rotary axis will sit in a sort of "well" at the back of the machine and not be in the way so much. I still have a 2' X 4' work area on the front of the machine. Will be adding a torsion box to take up the additional 6 inches. Sounds kinda crazy, add 6 inches to the gantry, add 6 inches to the table- - -huh? what? I need the extra height so I can turn 8 inches on the rotary and have clearance for longer tool lengths. I could not put a cutter in the machine hanging out of the collet by 2 inches and still clear anything more than 4 inches in diameter. Now I will be able to. I suppose I could have just put risers under the base rails but I liked the idea of putting in new extrusions better. Less work, that is for sure.

Interesting side note, as I was adjusting the accels/decels the other day I was noticing that the g forces were like .09. Talking to a friend who still works for the laser guys, he was saying they are now getting 2 g's during accel. WOW! I looked it up to make sure he had his info right, and he did. It is something to see a big honking gantry start moving at 2000 IPM and not fly off the machine when it stops.

So, I need to make a torsion box and get everything squared up again.
"If you see a good fight, get in it." Dr. Vernon Johns

User avatar
FixitMike
Vectric Wizard
Posts: 2173
Joined: Sun Apr 17, 2011 5:21 am
Model of CNC Machine: Shark Pro Plus (retired)
Location: Burien, WA USA

Re: "Lost Steps"

Post by FixitMike »

TReischl wrote: So, I need to make a torsion box and get everything squared up again.
If you want a really stiff torsion box, make the internal spaces triangles rather than rectangles. Much, much, stiffer. If I remember right, you can use half the internal material and get six times the torsional stiffness.
Good judgement comes from experience.
Experience comes from bad judgement.

User avatar
martin54
Vectric Archimage
Posts: 7339
Joined: Fri Nov 09, 2012 2:12 pm
Model of CNC Machine: Gerber 48, Triac PC, Isel fixed gantry
Location: Kirkcaldy, Scotland

Re: "Lost Steps"

Post by martin54 »

Ted, probably my badly worded post but no I am not talking about a fixed gantry type machine where the table moves.

What I was trying to put across was a moving gantry machine but the table can be moved up & down increasing or decreasing the gantry height, not something you would want to do on to regular a basis though as I believe you really need to resurface the table to ensure it is still perfectly flat.

Same with the rotating ball nuts, they are housed in the gantry if I remember rightly & drive the gantry backwards & forwards along the fixed ballscrew.

User avatar
TReischl
Vectric Wizard
Posts: 4585
Joined: Thu Jan 18, 2007 6:04 pm
Model of CNC Machine: 8020 48X36X7 RP 2022 UCCNC Screenset
Location: Leland NC

Re: "Lost Steps"

Post by TReischl »

FixitMike wrote:
TReischl wrote: So, I need to make a torsion box and get everything squared up again.
If you want a really stiff torsion box, make the internal spaces triangles rather than rectangles. Much, much, stiffer. If I remember right, you can use half the internal material and get six times the torsional stiffness.
Makes sense to me (appeals to my engineering side) but. . . . . . I don't wanna cut all those miters where it meets the side pieces.

Seriously, it will be bolted down to another torsion box that is covered by 1 1/2 thick material. The only reason I am doing a torsion design is for the support in the middle.

I am thinking it was Ian Kirby (the dude with the tennis shoes in some woodworking mag a while back) who did an article on torsion boxes. He had one made out of 1/4 ply mounted on the wall and he was sitting on it with no sag. An ol' Ian liked his pints. . . .or donuts.

BTW, the main table for my machine is a torsion box, works great.

Thanks for the idea.
"If you see a good fight, get in it." Dr. Vernon Johns

User avatar
TReischl
Vectric Wizard
Posts: 4585
Joined: Thu Jan 18, 2007 6:04 pm
Model of CNC Machine: 8020 48X36X7 RP 2022 UCCNC Screenset
Location: Leland NC

Re: "Lost Steps"

Post by TReischl »

martin54 wrote:Ted, probably my badly worded post but no I am not talking about a fixed gantry type machine where the table moves.

What I was trying to put across was a moving gantry machine but the table can be moved up & down increasing or decreasing the gantry height, not something you would want to do on to regular a basis though as I believe you really need to resurface the table to ensure it is still perfectly flat.

Same with the rotating ball nuts, they are housed in the gantry if I remember rightly & drive the gantry backwards & forwards along the fixed ballscrew.
I tell ya Martin, I am getting old and my reading comprehension has gone south! You were pretty clear, I just saw "where the table moves" and that did it, LOL.

The way this is going to work out is that I will be mounting a sub/spacer table to take up the six inches I have increased the height. But if I need to do something that does not fit under the spindle I can pop that table off and gain 6 inches of height. I just finished up getting the rotary mounted and I will be able to easily hang a tool out of the collet 3 inches and machine an 8 inch diameter with room to spare.

For those of you reading that about the tool hanging out 3 inches: Doing rotary work can quickly create these kind of problems, not much can be done about them except slow down the feed rate. And I do not like a tool hanging out that far either.

About those rotating ball nuts. . .what I did on that machine was put the nut in a housing, put a GTD pulley centered on it, then the motor was fixed to moving table (could have been a gantry). So the motor spun the housing with the nut in it and voila! Moving table without spinning a big ol' ball screw. There were a bunch of bearings and stuff too so the thing would not wear itself out.
"If you see a good fight, get in it." Dr. Vernon Johns

User avatar
TReischl
Vectric Wizard
Posts: 4585
Joined: Thu Jan 18, 2007 6:04 pm
Model of CNC Machine: 8020 48X36X7 RP 2022 UCCNC Screenset
Location: Leland NC

Re: "Lost Steps"

Post by TReischl »

I got this put back together again.

The replacement pinion fixed the slop in the axis.

Here is my rebuilt setup so I could swing 8 inches on the rotary axis:
20150809_144703.jpg
20150809_144717.jpg
The only thing left to do is mount the vise back on the table. I use it a lot.

The table is 2' X 4'. I have not done anything even close to that big in a long, long time.

I did a bunch of improvements while I was at it. The t slots on the base table were mounted with #8 pan head screws. That was on the recommendation of the seller. What a pain, the heads stuck up just enough so they would tend to catch but not enough to make me remove over 100 screws, c'sink, then rescrew. Since I was putting new t slot on the torsion box, I immediately counter sunk the holes. Ahhhh, life is good now.

I also squared the machine (.004 in 48 inches as close as I can tell), then machined the front and right side edges.

It is setup so I can remove the torsion box and the rotary and have the full table with tslots in it. I highly doubt I will be doing that anytime soon, if ever.

I used my typical technique for mounting tslot: Cut plywood boards, put t slot in between boards, screw it all down. That way I can change out spoiler boards without having to remove tslot. Works great. If I need to add a crosswise tslot, I just pull out the board, cut it where I want the tslot to go and trim off about .75 from one of the boards. Too easy.
"If you see a good fight, get in it." Dr. Vernon Johns

Post Reply