Haas Mini Mill spindle/tool holder/tool balance issue?

[Stirling]

How old is the machine? Are you the original owner?

In what condition are the spindle bearings? If you put an indicator on the table touching a toolholder, and push and pull on the holder, how much motion do you see? Now if the indicator is on the spindle nose (the part that spins)?

How good is the taper? Do you see any evidence of fretting?

Does your machine have a spindle vibration G meter like mine? If so, what does it read at 2k vs 10k?

What year did they start adding the g load meter and go would one find out/access it?

 

[Milland]

I think so too Don, I've seen what these machines are capable of - I definitely do think something is up with the servos or spindle though. Of course, I could always be doing something wrong, but I have tried a lot of variables now. I'll report back soon...

I hate to be a broken record (do we have a new variant of that for our modern digital age?), but I'd love to see accelerometer traces for the X and Y motion. If your service tech doesn't have something, you can search Ebay or Amazon or wherever for something like this (no idea if the range is right for your needs):

https://www.amazon.com/Military-Grade-Inclinometer-Accelerometer-Quaternion-Compensation/dp/B07W3RQJ1M

A little searching along the lines of "recording x-y accelerometer kit" may work for you.

 

[mhajicek]

What year did they start adding the g load meter and go would one find out/access it?

My 2021 CM-1 has it, and I remember the 2015 (pre NexGen) VF-3SS had one. I believe it's under "Diagnostics".

Edit:
It's that "Vibration" window in the second pic:

Diagnostic Data - NGC

www.haascnc.com

 

[mottrhed]

As always, somebody with a Fadal has to add their 2cents worth to a Haas thread

I have a '91 4020, if I got the 'good finish' you get in the second picture, I'd take it out back and put it out of it's misery.

On the 4020, if there's a finish issue it's either a resolver gone bad, or 99.9% of the time it's the endmill. I use exclusively YG-1 Alupowers. There's a 6 month old mill in the ATC that still gives a mirror finish.

As for finish passes I leave .010" stock, then 2 finish passes net.

My thoughts exactly. When I looked at those 2 pictures, I thought, well those are both terrible. This could be a multitude of things, and it will likely take Haas drilling down to completely address it, but Ill dig in a bit.

When it comes to the spindle, accurately check the following; endplay, drawbar force, test bar runout, and vibration signatures with and w/o the tool holder at various rpms. If you dont have the equipment, bring in someone, doesnt need to be Haas.

That will do a decent job of eliminating the spindle--or bringing something to light.

After that its play in the x or y, or servo tuning / programing issues, which is where my money is.

I think you have a real problem and looking at it like its good at 2k is not right. Its an issue at all rpm, it just happens to be slightly more hidden at 2k (still not good), where in reality the faster you spin the spindle the better the finish should be, but since you are changing the feed rate as well you are inducing (amplifying) the variable that is likely the root cause. I bet if you ran the spindle at say 6-8k and kept the same feed as at the 2k, you would see a better finish than it is at 2k where you thought it was good. That is absolutely not a solution, but again would help prove its not a spindle issue.

 

[MiniMillMachinist]

My thoughts exactly. When I looked at those 2 pictures, I thought, well those are both terrible. This could be a multitude of things, and it will likely take Haas drilling down to completely address it, but Ill dig in a bit.

When it comes to the spindle, accurately check the following; endplay, drawbar force, test bar runout, and vibration signatures with and w/o the tool holder at various rpms. If you dont have the equipment, bring in someone, doesnt need to be Haas.

That will do a decent job of eliminating the spindle--or bringing something to light.

After that its play in the x or y, or servo tuning / programing issues, which is where my money is.

I think you have a real problem and looking at it like its good at 2k is not right. Its an issue at all rpm, it just happens to be slightly more hidden at 2k (still not good), where in reality the faster you spin the spindle the better the finish should be, but since you are changing the feed rate as well you are inducing (amplifying) the variable that is likely the root cause. I bet if you ran the spindle at say 6-8k and kept the same feed as at the 2k, you would see a better finish than it is at 2k where you thought it was good. That is absolutely not a solution, but again would help prove its not a spindle issue.

Thanks for your reply and for the info. Haas came in today and did some tests on the machine, and found that the spindle is just fine. The spec for the displacement tolerance is 0.100 MIL at max speed, and they found my spindle to be at 0.03 MIL. We tried a couple of different tool holders that I have too, and the worst was 0.04 MIL. The velocity tolerance was also in spec. Runout on my tool is 2-3 microns. The drawbar force was found to be 1050 lbs, the spec for this is 780lbs to 1400 ish I believe, so again in spec.

I've attached the ball bar tests from today too, the worse one is before the adjustments were made and the better one is after. Odd that the squareness is improved over the previous ball bar, and odd that the machine was a bit out versus the previous ball bar, but nothing way out found there.

Is there anything else to look at? I have tried running the same feed as the original 2K run, but the finish is still worse. I haven't yet tried the machine after the adjustments today, so I'd be interested to see what people think. Surely if there was a servo tuning issue it'd show in a ball bar?!

Also, I did note that there appeared to be 1 micron of backlash correction in the settings - is it worth asking Haas to remove this? I just wonder if this could improve diagonals. Doubt it though. The D-Gain value appeared to be 500 - does anyone know if this is correct? I found a thread where a guy said his was set to 500 too, but he thought this might be a factory value to get the machine up and running as all the D-gain values on his other machines were 2000-4000

 

[Houdini16]

After re-reading this thread instead of bits and pieces, It seems on the surface to be servo tuning, we have seen it happen a lot, machines are sent incorrectly tuned, machine was sent out poorly or non-tuned.
But..., ball bar looks ok, does that negate servo tuning issue? maybe, but if it does what is left? the spindle.
It cuts good at sub 2k, regardless of tooling or cut params.
Then question, have you tried the same cuts, with same params, but using a conventional cut instead of climb cut? if this was a belt issue this might change it.
Did you remove the cover from the head and observe the belt with a flashlight to see if you see the belt wave as you increase rpm above 2k, usually if it gets a wave or harmonic in it you will see it widen in its path.
Have you watched the belt during a cut.
I would say have Haas tech check belt tension but he will probably not have a tension meter, and will have no idea what tension it needs to be anyway.
Couple things that did make a difference for me, recheck the machine level, but also make sure none of the feet are floating, objects like to balance on 3 points you may have a 4th floating.
On finish passes on non-linear geometry use G187 G187 Video Link
and last on a brand new machine, this is the only things I could see causing this, and it did come incorrectly programmed.

 

[MiniMillMachinist]

After re-reading this thread instead of bits and pieces, It seems on the surface to be servo tuning, we have seen it happen a lot, machines are sent incorrectly tuned, machine was sent out poorly or non-tuned.
But..., ball bar looks ok, does that negate servo tuning issue? maybe, but if it does what is left? the spindle.
It cuts good at sub 2k, regardless of tooling or cut params.
Then question, have you tried the same cuts, with same params, but using a conventional cut instead of climb cut? if this was a belt issue this might change it.
Did you remove the cover from the head and observe the belt with a flashlight to see if you see the belt wave as you increase rpm above 2k, usually if it gets a wave or harmonic in it you will see it widen in its path.
Have you watched the belt during a cut.
I would say have Haas tech check belt tension but he will probably not have a tension meter, and will have no idea what tension it needs to be anyway.
Couple things that did make a difference for me, recheck the machine level, but also make sure none of the feet are floating, objects like to balance on 3 points you may have a 4th floating.
On finish passes on non-linear geometry use G187 G187 Video Link
and last on a brand new machine, this is the only things I could see causing this, and it did come incorrectly programmed.

Thanks for the reply. My tech had ordered a different type of belt for my machine but Haas sent the wrong one (the original one), so he didn't fit it. He thought the spindle belt sounded fine and vibration was good too. My tech said the belt makes a really obnoxious noise when it's not properly aligned so mine was likely fine, however the new rubber one is softer and makes the belt quieter. I do wonder if it's worth fitting for that reason, a real shame it didn't turn up. I haven't looked at the belt in a cut but I can do.

I can't spin any of the levelling feet by hand so they are all in contact, I did note that my tech originally put slightly less weight on one foot to get some of the twist out of the axis (looks like it worked according to latest ball bar), so recently I tried adding weight to one of the two feet with less weight on them but this made no difference to the cut whatsoever, so I put it back to exactly where it was and checked level again. It is within 1 division on a 0.02MM/M level.

When you say incorrectly programmed, do you mean the servo motors? Yes, my tech had to use loads of linear compensation to get the result in the ball bar above. I believe he originally had to take away 250-300 PPM, and then the latest ball bar said he had gone 50 PPM too far, so he added 50PPM.

 

[MiniMillMachinist]

More shots... what on earth? One is 10K, the other is 5K, this was manually jogging this time so I'm not sure what the chip load was, but the feed was roughly the same for both runs.

 

[MiniMillMachinist]

Something I just thought - if I turn my spindle slowly it will fall back on itself slightly at certain points, but these points aren't at the same point in rotation on the spindle, which suggests to me that it may be the belt, or the pulley? It's only VERY slight, but there is a bit more resistance at certain points. Also, in the Haas instructional videos I've seen where the presenter spins the spindle by hand at the tool end, they tend to continue to spin for a while, whereas mine is quite quick to stop. There's not a ton of resistance, but there is definitely some. I wonder if this could be my issue?

 

[Wsurfer]

More shots... what on earth? One is 10K, the other is 5K, this was manually jogging this time so I'm not sure what the chip load was, but the feed was roughly the same for both runs.

Man, thats horrible. This is on a new machine and HAAS is leaving it up the the customer to diagnose ?. Yikes !, hope you get it sorted soon.

 

[crossthread82]

Something I just thought - if I turn my spindle slowly it will fall back on itself slightly at certain points, but these points aren't at the same point in rotation on the spindle, which suggests to me that it may be the belt, or the pulley? It's only VERY slight, but there is a bit more resistance at certain points. Also, in the Haas instructional videos I've seen where the presenter spins the spindle by hand at the tool end, they tend to continue to spin for a while, whereas mine is quite quick to stop. There's not a ton of resistance, but there is definitely some. I wonder if this could be my issue?

That's pretty common for a belt driven spindle especially when the belt is new. The belt has some memory from when it was folded up in the package, so those "tight" spots are basically the "humps" in the belt. It usually goes away after you run it for a bit.

 

[MiniMillMachinist]

That's pretty common for a belt driven spindle especially when the belt is new. The belt has some memory from when it was folded up in the package, so those "tight" spots are basically the "humps" in the belt. It usually goes away after you run it for a bit.

That makes sense! It's only got 100 hours or so of spindle time on it, but I'd have thought this would be enough to reset this? How much resistance should there actually be?

 

[MiniMillMachinist]

Man, thats horrible. This is on a new machine and HAAS is leaving it up the the customer to diagnose ?. Yikes !, hope you get it sorted soon.

Haas have been pretty good with me, I just like getting into the nitty-gritty of things. However, now that they've tested drawbar force, spindle vibration and the machine accuracy, they assume it is a programming error. Because I've tried so many recipes, even hand jogging the thing, I really doubt that, but if they want me to get programming tips in order to get to the bottom of this, fine.

 

[crossthread82]

That makes sense! It's only got 100 hours or so of spindle time on it, but I'd have thought this would be enough to reset this? How much resistance should there actually be?

You should be able to easily turn it by hand, but if you let go it will sort of rotate forward or back until it settles on one of the spots in the belt. It really depends on how you run the machine to determine how long it takes the belt to fully break in. High rpm and lots of spindle load will do it faster than if you are only running a couple thousand rpm taking light cuts.

 

[Houdini16]

More shots... what on earth? One is 10K, the other is 5K, this was manually jogging this time so I'm not sure what the chip load was, but the feed was roughly the same for both runs.

The strange thing to me is, usually the machine caused poor finish issues leave "facets", Like in the image you said was a "better finish".
The one with the variable elongated fish scales you said was "poor", looks to me, with 14 years experience, as recut, or too high rpm with too small amount of material in cut.
Either material isn't getting out of the path and being cut a second time, or the tool is vibrating because of to high of rpm and too little of material and causing a vibration.
The one you are saying has a "poor finish" has a organic look to the finish, this isn't usually the type of poor finish when your machine is causing issue, this is usually a operator error.
2 cents.

 

[MiniMillMachinist]

The strange thing to me is, usually the machine caused poor finish issues leave "facets", Like in the image you said was a "better finish".
The one with the variable elongated fish scales you said was "poor", looks to me, with 14 years experience, as recut, or too high rpm with too small amount of material in cut.
Either material isn't getting out of the path and being cut a second time, or the tool is vibrating because of to high of rpm and too little of material and causing a vibration.
The one you are saying has a "poor finish" has a organic look to the finish, this isn't usually the type of poor finish when your machine is causing issue, this is usually a operator error.
2 cents.

Hi Houdini, yes this was just manually jogging, so it could well be a recut, however it looks an awful lot like the issue I was experiencing to begin with. The WOC was 0.2MM though, certainly enough to take a cut from the metal. I'll have some photos up soon of progress I've made with diving into this...

 

[MiniMillMachinist]

Ok guys, I took the cover off and had a look at the belt. It looks fine in motion and the tension seems good too. I did notice it seems to be a more rubbery type of belt than the one Haas sent to replace it (they actually said they wanted to fit the rubber belt rather than the more plasticky type). Does anyone know if this is correct? I'm running out of ideas here, the spindle looks good, the parameters for the motors apparently should be 500 for D-Gain according to someone in the factory, so it looks like my motor parameters are probably OK too. Is it worth trying to really take out any error from the spindle balance? The error did seem very minor - 0.01-0.04 MIL where 0.1 MIL is the max allowed. The only thing I haven't looked at is pre-load as some people have suggested, but pushing the table in either direction by hand, it seems to return to center within 1 or 2 microns. Is this enough to cause this kind of issue?

 

[plastikdreams]

Have you tried a ramp tool path to see if you get the same results?

 

[Houdini16]

Hi Houdini, yes this was just manually jogging, so it could well be a recut, however it looks an awful lot like the issue I was experiencing to begin with. The WOC was 0.2MM though, certainly enough to take a cut from the metal. I'll have some photos up soon of progress I've made with diving into this...

The .2mm isnt the issue explained before, it is that the cut chips are not being removed out of the way either by coolant or air, and is being cut a second time, a .2mm WOC is fine, but at what rpm and what feed rate, and with what flushing, coolant, air?

 

[Houdini16]

Sorry bud, this does look more like operator error than something wrong with the machine. Not to be rude, but Are you a machinist? Like have you worked at a machine shop and have been trained by a machinist?