VF-3SS Repeatability Issue

[Kazsaw]

We have a VF-3SS that we use to cut steel. We have had some accuracy and repeatability issues that may be either application, operator, or machine related...or some combination of the three.

>We cut steel (S7, A2, D2, ...)
>We see repeatability of 0.0009" on holes about 1" apart, measuring center-to-center. The holes are less than 0.250" in diameter, are bored through a 0.300" thick plate using a four step process (center drill, drill, rough ream, finish ream).
>A Haas rep says that the published accuracy and repeatability specs of 0.0002" and 0.0001" respectively, are for machine movement when not cutting metal and when cutting metal a true position of 0.0005" should be considered really good for the machine.
>So, at 0.0009", we still appear to have a repeatability problem (calling repeatability the range between minimum and maximum measurements of the same features, part-to-part).
>We have had the backlash adjusted two times within a couple of weeks by two different techs.
>We changed our tools from HSS to carbide and are going to all screw length bits per Haas recommendation to try to acheive better repeatability results.
>The backlash adjustments and change in tooling occurred at the same time and led to our repeatability results improving to the 0.0009" mentioned earlier. (Whether it was the backlash adjustment, carbide tools, or combination of the two, unkown.)
>We had a compensation test run using a step gage that showed the x-axis was off 0.0025", the Y was off about 0.0002". The X was adusted to be off by 0.001" (considered very good and not worth trying to improve further by the Haas rep) and the Y was left alone.

So far, the results are the same.

So my questions are:

1. Are we expecting too much since our application is running steel? (A lot of people say the machine is great for aluminum, but none say it is great for steel AND we ran our process with the HSS tools on a Haas VM-3 and the repeatability was within a tenth or two over a number of parts.)

2. Is it normal to need the backlash adjusted? Does a need to adjust backlash signify something important?

3. Could it just be that we need to continue working on refining our process?

All input will be appreciated. Thanks!

[big_mak]

I'd think with the tolerances you are trying to achieve, you may want to try and bore your holes for position and size. Reaming twice will not necesarily get you in true position as well as single point boreing your holes.

[diggityds]

I'm Wondering About That Vm3 Your Talking About Might Have Scales On It Standard. The Vf Series Scales Are Only An Option. If Your Running Scales You'll Have True Repeatability With Themal Growth In The Ballscrew And Heat In The Shop Wont Take Any Effect To The Positioning Of The Cnc. I Would Start From The Beginning And Get The Machine Releveled And Maybe Get Lazered, The Pitch Comps Could Then Be Adjusted For Any Wear In The Ballscrew That Might Have Occured. There Might Be Some Mechanical Issues In The Ballscrew And Bearing Setup. Your .0009 Deviation Sounds Kinda Large To Me.

[HuFlungDung]

I'd agree with big mak, you really must bore a series of holes to determine if there is a repeatability issue.

Lathe center drills should not be used unless you drill no deeper than the end cone of the pilot. A real spot drill works much more accurately because it steadies the drill point immediately upon entry, that is, there is no chatter and wobble interval which can send the drill off the spindle centerline.

In this case, boring might consist of drilling to 7/32, then redrilling (boring) with a 6mm endmill (.236") then, ream.

[Kazsaw]

Thanks for the feedback.

Relative to diggityds statement:

Originally Posted by diggityds Your .0009 Deviation Sounds Kinda Large To Me.

The two holes that we monitor should be 1.1058" apart center to center. Using carbide tools and the lower end of the speed and feed rates recommended in the machinery's handbook, we see actual distances of 1.1040 to 1.1069" (after accounting for measurement system error with the CMM).

Haas has checked the machine out from top to bottom, and says that everything is within specifications. Haas recommended the tools that we are using.

Should we consider this reasonable results for a VF-3SS using our process?

[JPMach]

I think the SS mills have coarser pitch screws in order to achieve the faster rapids. If this is the case I would think that .0009" is within reason but I agree it should be within .0005" or so. I agree with big mak also that in order to get the accuracies you want on any machine you would want to single point bore them. A reamer will only follow the hole that was drilled not correct it, a boring head will correct its posistion and make it round.

Still though .0009" ain't bad for a machine that doesn't say Mazak or Mori on the front.

JP

[gar]

060815-0949 EST USA

Kazsaw:

My following comments are from memory, and other a priori information.

If your equipment is new, the positioning components are preloaded, and friction is small, then there should be virtually zero backlash.

On one of our HAAS machines I checked backlash, which was zero, but HAAS had overcompensated by about 0.000,1". Probably a stock default value because they assume in time some backlash will develop.

To run the backlash check we used an LVDT to measure machine table position. The LVDT resolved better than 0.000,01". The output of the LVDT was plotted vs time, and the machine on one axis was incremented by 0.000,1" every 1 second over a range of about 0.001,5" back and forth. The HAAS encoder quantizes in the range of 0.000,03".

You could run a similar test with a 0.000,1" indicator. Mount the indicator on the spindle housing to minimize shock to the indicator. Probably use a longer time between steps. Run this test with backlash compensation set to zero. When your machine makes its last increment in one direction, then the next increment (which is in the opposite direction) should take you back to where the previous increment started. If not you have backlash. You can determine how much backlash by the amount of peak-to-peak travel measured vs the amount expected. This you would see over several +/- travels.

With zero backlash circular milling is quite accurate.

Before you look for other problems independently measure your backlash on both axes.

If you have zero backlash, then the suggestions from others above are very good.

The actual variation from your last post is -1.8 to +1.1 thousanths. This is very bad. Are there any clamping forces that may be causing a problem?

Note in a preloaded system (implies no backlash) your side displacement under load will be a function of the spring rate of the preloaded system and the applied load.

.

.

[Kazsaw]

Thanks for your comments. Since I will be at the IMTS show in Chicago, any recommendations of whihc machines I should be looking at if our main work is high precision steel parts...we are starting to get into 5-axis work as well.

Originally Posted by gar 060815-0949 EST USA The actual variation from your last post is -1.8 to +1.1 thousanths. This is very bad. Are there any clamping forces that may be causing a problem? .

Parts are pinned and bolted to a subplate that is bolted to the table. Indicators show zero movement during operation. Should I be looking someplace else?

Originally Posted by gar 060815-0949 EST USA Note in a preloaded system (implies no backlash) your side displacement under load will be a function of the spring rate of the preloaded system and the applied load. .

This sounds interesting and of importance since we are cutting steel. Please expand on this. Thanks.

[Geof]

Originally Posted by gar 060815-0949 EST USA..........The actual variation from your last post is -1.8 to +1.1 thousanths. This is very bad.......

This is bad from a machine precision point of view but in my experience is well within the range of drill wander particularly when working with small drills and tough steels which is the case here.

[Kazsaw]

Originally Posted by Geof This is bad from a machine precision point of view but in my experience is well within the range of drill wander particularly when working with small drills and tough steels which is the case here.

All opinions welcome...

We are putting in two thru-holes on each part, a .200" and a .125". We need the holes to be with 0.0002" center to center to mount on a grinding fixture. The material is S7. The plate thickness is 0.300". Are we dreaming to think that we can do that on the VF-3SS? Or, are we not approaching it the right way?

We are open to trying different things and posting the results here.

Thanks!

[Geof]

I think you are pushing it because you are right on the machine specs. I suggest considering a change to the fixture design. You only need one pin for positioning the part on the fixture the second pin provides orientation. Slightly modify your fixture pins; have one round pin and one flatted on the sides. The round pin gives position while the flatted pin gives orientation from the rounded sides while the flats accommodate slight variations in the hole spacing.

[gar]

060815-1213 EST USA

Kazsaw:

Do you mean center-to-center distance of X+/-0.000,2" or +/-0.000,1" when you say "within 0.0002 in."?

You can only position HAAS in 0.000,1" increments, but as I previously mentioned HAAS, on standard machines, quantizies at about 0.000,03". This you can calculate from the lead screw pitch and the encoder counts per revolution.

External factors, such as drill wander, are likely bigger factors than the machine positioning if the machine is fully preloaded (meaning no mechanical backlash). These external factors would be true on any brand of machine.

You need to bore, and maybe grind your holes if you really need +/-0.000,1" accuracy.

In your tooling this part fits on --- how much clearance between the pins and the holes? How does this relate to your specified center-to-center spacing?

Goef's suggestion of the diamond pin is standard practice.

I doubt that you have a clearance of just 0.000,1" between pin and hole.

If you have dirt or lack of flatness of the plate when you clamp the plate to the table, then you can distort the shape of the part.

.