What's new
What's new

My $12,000 PFH 4800 ball screw replacement.

Techguy

Hot Rolled
Joined
Oct 26, 2007
Location
Hamburg, MN
Last week we finished replacing the Z ballscrew in my Mazak PFH-4800 Horizontal. It had only been 3 years since it was last replaced albeit poorly.

The screw had developed about .001 backlash which is fatal on a horizontal. The screw started making noise six months after it's last replacement and I assumed it was just a noisy thrust bearing. I was wrong.

Here is what we found:

The end support bearings (away from the motor) had been spinning on the ballscrew because of lack of preload in the end support block. The bearing set was thinner than the housing by .001. We added a .002 shim to correct the preload problem.

The screw was not correctly captured on the shoulders of the screw. This required grinding the shim plate under the motor by .0094 to correctly capture the screw.

The new bearing set on the motor end was .0015 thinner than the set that came out resulting in no preload on the outer races. A .002 shim fixed that.

The ballscrew was .004 out of alignment vertically when assembled. Both ends allowed no vertical adjustment as the motor end was pinned in place and the end support allowed no adjustment. This was a Mazak manufacturing defect that surprised the hell out of us. This was the 3rd failed ballscrew since the machine was new and had never been caught before.

When the last ballscrew was replaced we had to change servo parameters from original to make it run smoothly. This should have been a dead giveaway but I trusted the technician.

In the end this cost me $8,500 in parts and another $3,500 in labor.

It worked out correctly this time because I personally did an analysis of the failure modes rather than trust the tech. I would have had yet another failure if I hadn't insisted on checking all variables myself.

Now it sounds the best it ever has and is set to all the factory tuning specs.

This is just a warning that you can't just drop in a new screw without checking all of the variables.

Needless to say I am a little pissed that the tech ignored all of these things 3 years ago.

Dan
 
Thanks for the write-up, what a crap situation. Are you confident the lubrication system is working correctly, and that there's no contaminants in the lines? Do you have any concerns about the .002" shims "extruding" from their placement? I try to avoid shims that thin when under high load if I can.
 
Wow. That's pretty shotty. Did you stretch the screw to the factory spec?

Way back in my early days in industrial maintenance, I was helping with the replacement of a 14" diameter spherical roller bearing on a huge drying machine. We had the bearing bore repaired by welding and line boring. When they bored it, they got the bore a little too deep and the retaining plate did not squeeze the outer race of the bearing. The new bearing lasted about 3 months and we had to do the whole process again. This time it was fixed correctly and ran for many years after that.
 
We went through the entier lubrication s3ystem and found that the first owners had put the wrong grease regulators Larger) in the X-axis to compensate for a lack of grease to X. It turned out that when I got it that they had changed grease and the grease reacted with the other and plugged some manifolds. That took out the X guides about 6 years ago.

We replaced all of the grease regulators ($600) during this rebuild to make sure we got enough grease to the Z. We weren't happy with how much we were getting to Z and X was getting way too much.

We have run Kluber MBU-15 for the last 6 years. We changed from a $3 grease to $150 per tube. That is what Makinos run.

We will soon be replacing the Y linear guides ($2800 parts and I suspect about $2,000 in labor) because of about .0002 wear due to lack of lubrication we found about 6 years ago.

Three years ago we also replaced all thrust bearings and seals about 3 years ago and rebuilt the spindle about two years ago.

Right now the machine is in the best condition it has been since I got it about 9 years ago.

In the next month it will be performing almost as new.

I also bought a bunch of new tooling and other stuff to get the best possible performance from it. Currently I am getting the best parts I have ever had out of it.

One might say I am buying a job, you would probably be correct. I can't quite figure out what else to do.

I don't have enough neighbors to properly fund a Hot Dog Cart.

Don't be like me.

Dan
 
Wow. That's pretty shotty. Did you stretch the screw to the factory spec?

Way back in my early days in industrial maintenance, I was helping with the replacement of a 14" diameter spherical roller bearing on a huge drying machine. We had the bearing bore repaired by welding and line boring. When they bored it, they got the bore a little too deep and the retaining plate did not squeeze the outer race of the bearing. The new bearing lasted about 3 months and we had to do the whole process again. This time it was fixed correctly and ran for many years after that.

This time we correctly stretched the screw to the correct spec. Last time it was stretched to the correct spec but the screw was not correctly preloaded on the support end. I suspect the stretch worked until the grease squeezed out from between the bearings. It was just done completely wrong. Stretch is about .0005 if I remember correctly.
 
Every manufacturer is different. .0005 isn't much. I have a VMC with .020 on the X screw. It takes some work to stretch a screw .020.
 
Every manufacturer is different. .0005 isn't much. I have a VMC with .020 on the X screw. It takes some work to stretch a screw .020.

Since the screw is now completely captured against the shoulders and tight in triplex bearings on both ends .0005 seems correct. This machine also has hollow chilled ball screws.
 
Seriously .001 backlash on a Z axis ballscrew and your shitting the bed?

Keep in mind that it is a horizontal so Z doesn't have gravity to work with. Z backlash affects flatness for face milling or surface finish for milling floors.

I do a fair amount of work that has to be flat to less than .0002 in 10" so an unstable Z doesn't work out well as it can drift during cutting.

Most of my parts are cosmetic and the Z was becoming a problem with floor finishes and facing finishes.

I wouldn't call it shitting the bed but just repairing a machine to how it should and needs to perform. It wasn't working out as it was and I was having problems making acceptable parts.

Imagine Mastercams dynamic milling with microlift in a pocket with an unstable Z.
 
So how did you repair the .004" vertical misalignment in the screw?

Good thing no grinding was required...??

ToolCat

The motor end of the screw had taper pins. We removed them and indicated it in to .0002. The only grinding that was necessary was the spacer plate under the motor end support to correct for shoulder to shoulder length on the screw. We had to take .0094 off of that. That took a while to fit correctly. Quite a while.
 
Does the old screw just go in the scrap or do Mazak take them back as a core and recondition them?

Mazak does not take them back for a core. This screw is a great candidate for a rebuild because it is in really clean condition but has backlash/wear. It is the nicest looking useless ballscrew I have ever seen come out of a machine in fact.

I might send it out for a good certified rebuild and sell it or sell it to someone as a core to rebuild. Otherwise it can sit along side the previously dead screw on the shelf.

Dan
 
I was asked to give an update so here goes.

It is like an entirely different machine, the floor finishes are the best I have had in the eight years I have had the machine.

I always had a problem with a circle about .0005 deep at the plunge point in a pocket and that is now finally gone.

I couldn't be happier with the end result.

I have added Schunk hydraulic holders for all of my finishing and I am producing the nicest parts ever. My old ones were pretty nice too.

Dan
 








 
Back
Top