Gib quality - performance gains

[flick]

I've completed a truing and re-finishing job on my gibs, and just wanted to share my experiences and observations.

The problem: my mill was horribly noisy when asked to cut anything more substantial than PVC. Experimenting in cold-rolled 1018, slotting with a 3/8" endmill, I found that I couldn't take much more than 0.025" DOC, 0.002" CLPT. I used good endmills: 2 flute kennametal, 4 flute hanita varimill. Varied my speed from 500 to 3000 RPM without much change. At 0.05" DOC it would quiet down at about 0.0005" CLPT. Not acceptable.

I won't go through my entire troubleshooting process, but I measured 2 variables that I thought were representative of how well my mill would/should perform: free movement at the end of the table, and dynamic torque on the ballscrew at low speed.

Free movement was measured with an Asimeto dial test indicator mounted on the base of the machine, touching the left-most end of the table on the back face, with the table in the middle of it's travel. Torque on the ball screw was measured with a Gray Tools 0-75 in.lbs torque wrench. I had to remove the X-axis motor, and make an adapter to drive the ballscrew with the torque wrench.

Initial readings were: 0.020" total free movement, 12 in.lbs torque.

I regret now not taking more torque measurements as I tightened up X axis gib. After I had made it as tight as I could with a screw driver, I was still getting 0.018" free movement. I then used a small ratchet, and got it down to about 0.015" free movement before I started losing steps. That was when I pulled the gib, and discovered that it was bowed, twisted, and poorly finished from the scraping job it had been subjected to. I hastily decided to refinish all my gibs, and took the rest of them out and started working on them before I took the time for any more measurements. As it turns out, I couldn't be happier with that hasty decision.

It was a lot of work. I'm estimating about 80-100 hrs. That involves a lot of false-starts and blind alleys with my scraping methods, and I could probably get back to this point now that I've had my scraping education in about half that time. But I'm not completely finished yet either.

My Process: I used the belt sander to make a scraper out of an old file - large (10"?) radius on the end, and about a 5 degree negative angle to the rake. I started by truing up the non-working sides against a surface plate with bluing. I ended up removing a lot (a LOT) of material from the X gib. The Y gib needed less attention, and the Z least of all. To get a good finish I had to sharpen the scraper after each pass. Then I wanted to make them perfectly match their mating tapers, so I used magic marker on the working side, inserted them into the dovetail, tightened the screws until I got an increased drag reading, then removed the gib, and scraped where the magic marker had rubbed off. I used magic marker instead of bluing, because I wanted something that would dry, and only come off with significant rubbing.

I ended up rushing the job towards the end. Of course I wanted 100% contact, but some paying work came my way, and I had to get the machine up and running quickly. I will put more work into the gibs soon, but these are the results so far:

0.004" free movement, 12 in.lbs drag
0.0015 free movement, 15 in.lbs drag

Baseline torque for the X and Y axes are each 10 in.lbs with the gibs installed loosely, with significant clearance. This is the amount of drag created just from the weight of the table on the ways and preload in the ballscrew assembly. The baseline for Z is 15 in.lbs in the up direction, due to the weight of the head. As many of you know, amount of torque required to move the head in the other direction with a loose gib is 0.

If you're wondering how the 3/8 endmills are cutting now, I haven't had any time to tune up the parameters, but initial results are making me happy: When slotting 1018 with a carbide 2 fluter I was able to take .125" DOC, 0.0025" CLPT at 2000 RPM, and NO SQUAWKING! I can't do much better than that with the Haas VF4 at my day job.

[instructor37]

[QUOTE=flick;756413]I've completed a truing and re-finishing job on my gibs, and just wanted to share my experiences and observations.

That was when I pulled the gib, and discovered that it was bowed, twisted, and poorly finished from the scraping job it had been subjected to. I hastily decided to refinish all my gibs, and took the rest of them out and started working on them before I took the time for any more measurements. As it turns out, I couldn't be happier with that hasty decision.

QUOTE]

I've read other forums on this subject, and our Patriot machines also had bowed gibs as well. It seems to be a standard of Chinese manufacture to leave the gib to a blacksmith, no matter how well the rest of the machine is built. I left the job up to the students and they did an excellent job of getting all the gibs flat and re-doing the locator indents for the adjuster screws. Once the gibs are flat and adjusted properly the machine performance is 100% better.

[BobWarfield]

Great post. I lapped my own gibs and it really improved performance. I like to run them very tight. Love the idea of being able to measure the ballscrew torque.

Best,

BW

[keen]

Hi Guys - Over the years I have noticed gibs on various of my machines are 'bent' - probably due to difficult to control stresses in the production grinding processes. The factory fixture would usually hold the gib in place during the grinding or milling but it is when the gib is released the gib is likely to bend. Anyone machining long precision parts will know about this difficult to control issue.

However as long as the gib is to spec ie correct angles/straight surfaces when it is held in the factory fixture, this is not so serious providing: In assembly the bent gib needs to be adjusted so that it is almost straight in place. If you try to flex a long gib you will be surprised how easily it flexs - and so I have always assumed the friction caused by adjusting it into an almost straight position should not be too great. That said a bent gib is more difficult to adjust to the optimum and I am not excusing lazy gib manufacturing habits.

After reading this interesting thread I am now not sure If my previously held conclusions (above) need to be reviewed. Either the friction caused by adjusting a bent gib correctly is more than I had accessed, or the gibs in question are not just 'bent' but are also out of spec in terms of angles etc, or the torque measuring procedure above has been carried out with the gib setup/installed incorrectly.

[flick]

Thanks for your comments keen. Your opinions w/ respect to the significance (or rather insignificance) of gib flatness seem to be widely held - including by the support person that I dealt with at Tormach. I give those opinions a great deal of weight, and I agree that my gibs had more than one problem. The tender manipulations of an overzealous, scraper wielding masochist left the surface of my gibs in a state that reminded me of a 3D relief map of the Rocky Mountains.

I would like to add that after working the mill for a bit, I realized that my issue with missing steps had not entirely gone away! It had shrunk to a degree that made it much less noticeable, but indeed, after running a program that had a nearly 2 hour finishing cycle, I found that the zero of the finishing tool was "drifting" slowly in all 3 axes. It turned out that the control computer shipped with my machine had multiprocessing turned on in the BIOS. After correcting this apparent oversight and doing a lot of jogging around with a dial in the spindle I have not been able to duplicate the problem. I think it's fixed for good!

There is now no way to be sure whether the larger quantities of missing steps that I experienced while tightening my gibs was due to the gibs themselves, or just a coincidence, a spurious event due entirely to the misconfigured BIOS. I can, however, be 100% sure that the BIOS had no effect on the degree of table deflection that I measured and I'm pretty confident that the factor of 5(+) MRR improvement that I achieved was directly due to the improved stiffness.

[keen]

Hi Flick - You may be aware that position loss on Tormach's may be not step loss but coupling slip. I had this problem on the Y axis. See pic for my positive drive solution.

[stevespo]

I suppose this is another topic for discussion, but given how important the gibs are to the performance of the machine - shouldn't Tormach QA be paying a lot more attention to how things are done at the factory?

I would assume the worst ordering an inexpensive manual mill. The Tormach is not an expensive machine by any means, but I guess I would expect more care for $7500...

Steve

[flick]

Hi Flick - You may be aware that position loss on Tormach's may be not step loss but coupling slip. I had this problem on the Y axis. See pic for my positive drive solution.

Nice solution! I checked mine by spraying on some layout bluing into the split in the coupling. Doesn't seem to be any rotation happening. If it ever does become a problem I'll remember your approach.

stevespo: I'm not a Tormach fanboy by any means, and I'm accustomed to working on real industrial machines, but having owned the Tormach for over a year now I've become realistic about the level of quality and performance that can be expected at the price point. When I first bought it I expected it to be the next best thing to a VF1. The reality is that it's much closer in performance and quality to the other machines that would be considered hobby class. It's a big benchtop mill, not a small machining center, though it shares some characteristics of both. Even if it is the cadillac of the chinese hobby machines, don't forget that's only 1/3 "cadillac", and 2/3 "chinese hobby".

Having said all that, I would still consider my gibs flawed. From the performance seen in the videos on Youtube I think that the average Tormach owner CAN expect better and I hope that Tormach continues to take note of these forums and work to improve things, so that others don't have to share my experiences.

[David Bord]

Guys, hoping to get some help and felt this the best place to ask my questions.

I have run out of adjustment on my X gib. Like others, I found that I had to keep adjusting it in to keep the machine tight.

When taking the gib out to shim it with shim stock and get some adjustment back, I found that the side that slides was the rough side of the gib. It had what I would call a rough grinding done on it in something like a cross hatch pattern. The side that does not slide is considerably smoother.

I assume the smooth side should be facing the surface that slides?? Is it possible the rough side has been grinding away at the table requiring the frequent adjustments?

If th smooth side should be facing the sliding side, can I flip the gib keeping the narrow end on the same side?

I also found the gib not to be flat. Based on the previous posts Im going to leave it as it and see what happens when I put it back together.

Any suggestions? I dont want to get into a lengthly scraping process if I can avoid it. I want to get my machine up and running.

David

[keen]

Hi David. I am interested as to why your X slides have worn so much - how many hours would your machine have done?

Re your question on rotaing the x gib - Is the X gib symetrical and able to be rotated?

Have you tried stepping the X axis one way (with a dial indicator on the movement) then changing direction and counting the lost steps before it moves the other way - I would be interested to know how worn your High milage X screw is also.

Cheers.

keen

[flick]

The rough side has been subjected to power scraping. It is intended to serve as the sliding surface. Nicely scraped surfaces have a lot of positive qualities in flat bearing applications - oil retension, reduced "stick-slip", improved planarity. As has been mentioned in other threads, it seems that chinese factories employ the most ham-handed butcher available to scrape the gibs and fit them to the dovetail. The resulting surface has lots of big pockets to retain oil, but it is innacurate and has very few contact points, resulting in unpredictable friction and rapid wear. Sad, but I think the only good way to fix the scraping job is with a better scraping job. Please don't lap your ways. No matter what anyone tells you there is no way to improve the situation by tossing abrasives in there.

[David Bord]

Thanks guys... appreciate the feedback. Sounds as though its the way its supposed to be. The sliding side is crude at best... there are very large areas that have voids or no possible contact. I measured the thickness side to side and there is no way to flip it. its several thousands different side to side.

What I did, was take wet dry sand paper on a small granite surface plate and go from 240 to 400 grit in 3 steps to at least smooth the high spots so that the gib would not saw away at the table. I dont think I took any appreciable material off, the idea was just to smooth the high spots evenly a bit so it would hopefully wear as much. Was this a reasonable thing to do?

I added a .005 thin shim to the length of the gib(non sliding surface) and it brought the adjusters back near perfect with lots of adjustment.

My Z and Y are near the end of their adjustement as well so maybe similiar quality gibs are the reason.

The machine is about 4 years old I think, but has pretty low use. I would say on average I play with it every other weekend for about an hour.. I'll go a month or two without touching it as well. So probably less than 100 hours use.(maybe much less...) I cut 99% alum with smallish end mills.

I measured my X backlash maybe a year back and it was .0005, today it was .001. When measuring y about a year ago, it was .001(same as the Z) Today it was around .003.... so something has changed... I wonder if my Y axis is out of adjustment or something and its rocking a bit when jogging?

Speaking of backlash, are you guys compensating in Mach for it? If so, are you editing a config file to do this? If so, what file is it and how do I find it?

David