Haas VF6 TR210 5 axis Z gouging

[Ryan Shields]

I'm having an issue with multi axis surfacing. It appears to me that the Z axis will not accelerate fast enough to keep up with A and B. At this point I'm running a test part with 2 different multi axis surfaces. One utilizes AZ&Y and the other utilizes BZ&X. Both of them are faces I used side milling to finish, and the ball endmill is simply supposed to follow the surface. In both cases, I'm getting gouging when Z should be moving up, and floating when z should be coming down.

I don't have any issues with XYZ surfacing, they all seem to sync fine.

I have a programmed feed of 100 with a max inverse feed of 3000 in the post. All the lines are posting at 3000, which tells me it isn't even running the feed of 100. If I use a feed over ride, or just slow it down, it improves. But even at a 10% feed override, it is gouging at least .003 in the worst spots. This also tells me its not the tool path itself.

Has anyone else gone thru this and found parameters that need adjusted to actually do a multiaxis surface at a reasonable feed?

It is a new VF6 (came in last October) and a P2 version trunion (with the A axis scale).

 

Thanks!

 

Ryan

[MrFish]

We never managed to get our HAAS 4th axis to cut simultaneous tool paths at a reasonable jerk free feed rate even after having the parameters tuned by a HAAS service tech !!

 

But what did help a bit was to make sure that your part dia set in setting 34 is close to what you are machining and you can also set setting 191 to 1 (rough) to allow freer movement but this can make gouging worse if there is a lot of tight direction changes

[Ryan Shields]

I'll try changing 191, the applications guy at Haas had me change that to "finish", but I can't tell that it made much or any difference. It's got the high speed option, so it does a great job of smoothly running through all the points.

It's all posted using inverse feed, so the dia setting doesn't play a role in that.

I'm very tempted to slow down the acceleration parameter on A B, see if that helps. But I'm not certain if that could have undesirable side effects lol.

[Colin Gilchrist]

I used to work for Cimtech, the WA State Mastercam Reseller. He bought a VF-4SS with the TR-160 trunnion table. We kept having problems just like you are describing. We finally had Haas send a service tech out. He spent two days tweaking the parameters for the rotary axes. He called it "tuning" the drive motors, or something like that.

 

You likely need to have Haas come to your shop and adjust the parameter settings for your trunnion if you haven't had them do that yet. It really doesn't become a problem until you are trying to do full 5 axis machining, so many shops that first get a Haas trunnion don't notice it because they are doing 3 + 2 at first.

[Bob W.]

A few things really helped the results I was getting (VM3 - TR160). First was the max inverse feed rate should be at 40,000, Haas could tell you for sure. They told me 40,000 but I did play around with it and raise it until the machine alarmed, which did in fact occur at 40k. Second was to post all rapid moves as feed moves. After that it was a matter of playing with the tolerances, step size, and max angle step in the operation to find what the machine liked. Another thing that made a difference was the G187 mode the machine was running in. There are three modes: G187 P1 is roughing, G187 P2 is medium, and G187 P3 is finishing. These make a huge difference in how the machine runs. To set it back at default you just post another G187 with no P value

[Guest]

Ryan, Did Kevin call service? He e-mailed me earlier today. It must have been a crazy day, Kyle's phone was ringing off the hook.

[GPC_CNC]

Max inverse on a Haas is F45000. I've played with the feed until I got an alarm and one of our Haas Apps guys confirmed it as well.

 

We've had a bunch of new trunnions (8) P2 versions and the scale is a gimmick. Every one of them needed to be compensated on the leadscrew page to get within the factory specs for accuracy from A-90 to A90 and if we had a laser/ball bar setup we should really have gone through and tuned for every angle in the range.

 

That being said, Maxing at 3000 could be a big problem but shouldn't necessarily affect gouging between movements. There is a setting for corner rounding accuracy that needs to be set quite high around .05" (per the haas manual) to get adequate movement, but if it is set too high could possibly cause the excessive motion you're seeing. Also check your toolpath resolution.

 

Greg

[Ryan Shields]

I'll try cranking up max inverse in the post to 40k, and throw a 187 in the mix.

I'll check and see how it's doing with angle accuracy, I had to fix that on a different trunion a while back.

When I talked to the ap guy at Haas, he had me set the corner rounding setting at .06, but it didn't seem to hurt or help the issue.

I did talk to Selway yesterday and they suggested getting a tech out. Not sure if it got scheduled, but I'll follow up on that here in a bit.

Maybe I can hassle Kevin to figure out the ball bar on a trunion, we just recently invested in one, but haven't had time to apply it to the trunnions.

Thanks for all the helpful responses, I've got faith in Haas, I think it's just a matter of getting ahold of the right person to get it dialed.

Thanks again!

[Guest]

Ryan, do you have a Renishaw spindle probe on that machine?

[GPC_CNC]

It's definitely about getting the right person, but angle increment/vector length can play a big part in how a part gouges, especially in transition moves. I usually have to bring the angle increment down to .2 or so to keep from gouging. I don't tend to do the vector length as much, I think it tended to dump a lot more code (more than the machine/inverse time could keep up with).

 

As far as Haas goes it's definitely about getting the right person. You'll need a good Apps guy, not a service guy. But as far as the error I was getting in the trunnion, it was about .001-.0015" taper along 6" (milled flat) between A0 to A90. If your part is smaller, it may not matter as much, but the parts on these particular machines were long (9") with tolerances of +/-.0015" between planes, so it really mattered. I doubt that would have anything to do with gouging. maybe mismatch to blended surfaces, but not gouging.

 

Greg

 

 

I'll try cranking up max inverse in the post to 40k, and throw a 187 in the mix.

I'll check and see how it's doing with angle accuracy, I had to fix that on a different trunion a while back.

When I talked to the ap guy at Haas, he had me set the corner rounding setting at .06, but it didn't seem to hurt or help the issue.

I did talk to Selway yesterday and they suggested getting a tech out. Not sure if it got scheduled, but I'll follow up on that here in a bit.

Maybe I can hassle Kevin to figure out the ball bar on a trunion, we just recently invested in one, but haven't had time to apply it to the trunnions.

Thanks for all the helpful responses, I've got faith in Haas, I think it's just a matter of getting ahold of the right person to get it dialed.

Thanks again!

[Ryan Shields]

No Renishaw It would make my life so much easier at this point..... But I am equipped with a Interapid .0001 indicator.

So I started with basics this morning. Pulled the fixturing and started with backlash. There was no value for A, and it certainly needed one. Now machine (G53) A0 is true to the Y axis regardless of coming from A+ or A-. I don't believe it was enough to cause my issue, but progress is progress....

I'm now looking at how true it is at 45 and 90. It seems to be progressively off in A+ (A.010 at 45 and A.018 at 90) which I know the steps per unit parameter would fix, but when I go to A- its different (A-.005 at -45 and spot on at A-90) so I am wary of correcting + and throwing off -. I'm gonna double check those figures, I'll try using feed moves as apposed to handle jog, but any thoughts on that dilemma?

I'll try tightening up the angle increment, but I have all the tolerances in the tool path real tight, takes forever to generate the tool path. I did repost with higher max feed, one toolpath runs a smidge over the 3k it was at, and the other kicked up to a little over 7k. I'll try running it in a bit.

[kccadcam]

Ryan, Did Kevin call service? He e-mailed me earlier today. It must have been a crazy day, Kyle's phone was ringing off the hook.

Crazy days is right,, we have just moved into a new facility and still trying to get everything dialed in.

Ryan and I are going to put the screws to the boss to order up a Matsuura 63v Cublex and get serious.

[crazy^millman]

Awesome machine. Did a project for a smaller one and with Camplete standard on those machines one sweet package. Get James' company to show you how it it the solution you need and it will be an easy sale.

[Ryan Shields]

Cublex 63 for the win!!!

 

I got it working! There is a Notch Center Frequency and Notch Depth Factor parameter for each axis, and they all have to match for simultaneous movement. I ran into a problem with a zero return margin on A after setting them, but an automated grid shift took care of that! Works pretty good, and I think with a little fine tuning it will work great!

 

I really want to thank you guys for the help!

 

Ryan