Turret alignment problems

[coquina]

Hi -
I have an older Haas HL1, bought in the mid-nineties. We have had trouble with the turret alignment. We've never had a major crash, but I had a job to turn a true radius on the end of a piece of .625 bar, and when checking with a radius gage, the radius was flat on the end.

After verifying that my program was correct, and calling Haas for other suggestions, they suggested we check turret alignment. It was off, so with Haas' help, we went through the procedure to realign it.

Several months later, we got the same job again. Had the same problem and had to realign the turret again. I didn't think the machine had been crashed in the interim, but I couldn't be sure.

So - last month I got a job that entailed machining a 20mm h7 bore. I wanted to ream the bore but it cut oversize. I checked the turret alignment, and it was off again. I know there have been no crashes, and the machine has had very light use since the previous alignment. It seems that the turret will just not hold position.

Has anyone else had this problem?

[HuFlungDung]

Would you be insulted if I asked if you were using a properly compensated toolpath for the tool tip radius? If the radius part has a flat on the end, I surely would not realign the turret for that reason.

Does the lathe cut a true cylinder over say a 4 or 5 inch length?

As for the reamer cutting oversize......well, reamers can be that way. But axial misalignment and taper could be due to turret or headstock. Have you checked the headstock? Apparently, most cnc's have a means of adjusting this (after a crash).

[coquina]

Originally Posted by HuFlungDung Would you be insulted if I asked if you were using a properly compensated toolpath for the tool tip radius? If the radius part has a flat on the end, I surely would not realign the turret for that reason. Does the lathe cut a true cylinder over say a 4 or 5 inch length? As for the reamer cutting oversize......well, reamers can be that way. But axial misalignment and taper could be due to turret or headstock. Have you checked the headstock? Apparently, most cnc's have a means of adjusting this (after a crash).

I do not insult easily, and you've gotta ask the right questions to give a correct answer, like "Is it plugged in?" But re, the tool comp, I had the programmer at Haas go over my program, he said there were no problems with it.

I don't have a tailstock, so my work is usually up close to the chuck. I have had trouble turning longer OD's and deeper bores, but thought it was due to push off. I have been able to "fool the machine" by programming a slightly conical shape to achieve a true cylindrical one.

Haas never mentioned that the headstock could be out - when you use that term, do you mean "chuck"? I'm using a Kitigawa od/id clamping chuck. I don't have collets for this machine. The jaws are in the right position and they were bored to hold the part. The part runs true in the chuck. However, if you attach the base of an indicator to the part, put the turret in its center position, put the indicator needle on the ID of the hole bored in the turret for the tool holder, and swing the chuck around, the indicator runs out by about .005". (This is how Haas told us to check it.) Moving the turret in the x dimension does not help the situation. Since the turret is limited to angular movement along its slide, I presume that the turret is out vertically. There is no easy way to move it up and down (that I know of) other than to perform the turret alignment procedure.

BTW - I'm new here, but I'm not new to machine shops. My dad started the one I now own in 1949 and I grew up in it. I learned to file a burr before I started 1st grade. I ran machines when I was younger, but don't claim to be a journeyman. I do know machining processes, and I program my machines. (I have two, the HL-1, and a VF-2 machining center.) I started programming on NC machines with a teletype machine, when I had to manually do the math to develop toolpaths and trig out tangent points for arcs.

My shop is very small, but I do have some very skilled conventional machinists working here. (One of them has been here since the '60's). On the other hand, they are clueless about the electronics of the CNC's, or programming, but they do have the mechanical skills to set up jobs properly, then we have an operator who runs the parts.

I have not had a lot of training on programming the Haas machines. I'm sure there are many things I need to learn. I use Bobcad, version 17, and just bought version 20. I am in the process of going through the training videos so I can learn about the upgrades. I don't usually make really complicated parts, so I haven't gotten into the solid surfacing, etc.

Do you have a thread where one can introduce oneself, rather than barging in as I have done?

[makrskr]

Cnc lathes are just like that. I dont care if it is a Mori, Okuma, or a Haas, they are never perfect. Indicate it in the best you can, and call it good. I never bother with reamers, boring is your friend.

[Geof]

If you have to 'fake' a cone to get a parallel bore it does suggest your headstock is out; i.e. your spindle is not parallel to the Z axis travel. One way you can check this is to mount a dial indicator on a long bar and dial in a turret hole I.D. with the Z axis near machine home then bring the dial indicator as close to the chuck as possible and again dial in without moving the X axis. If everything is parallel you should get the same tir in both positions. If there is a difference you can get a handle on how much and in which direction.

[HuFlungDung]

From 'coffee row talk', I've heard that cnc headstocks are often built on a swivel pin. So near each end, front and back, there should be something akin to push pull screws. There should be some hold down bolts to loosen off first, of course.

I would do a simple turning test first, though. Turn a hefty piece about 5 inches long with a sharp tool (Duh! ). If it is only out a couple of thou, you might be able to fix it by relevelling the machine. If the part tapers small at the outer end, then elevate the levelling screw underneath the back side at the tailstock end.

[Geof]

"relevelling the machine" Hu triggered a thought!

Have you checked that the machine is level and that it has not gone out of level at the same time the turret goes out of alignment? In particular do forklifts drive past the machine and do you have frequent vibration from passing traffic and things like that? The HL1 puts a good load on a small area of floor with those little cast iron pucks; if the floor is marginal in thickness or cracked movement does occur.

[little bubba]

We had a crash a while ago on a Mazak, with help from another board, got the turret all happy and straight. Figured everything was OK, but still programming tapers, problems with reamers, espeically the big ones. Turns out it really isn't that hard to knock a headstock out of whack. If your programming tapers, you're headstock is knocked out.

Take a piece of decent sized stock, at least an inch and a half or 2" and start taking skims on it, somebody already said this, but if your as stubborn as me, you'll need to be told several times. You should come up within a thou without programming a taper. For me, I was advised to leave a small taper in for deflection, and since the headstock settled on .0007 over 5.5" (probably the factory setting), that is where I left it, and running that lathe has been nothing but sweet ever since, reams straight, drills straight, turns straight, got the presetter to within +.0002 when cold, -.0001 when going full tilt, absolutely beautilful. I'm not running a Haas, but a headstock is a headstock.

[coquina]

Thanks guys, I had not thought about the floor, the lathe is in an older part of the shop and I'm not sure how thick it is there. Perhaps it has settled.

I recently sold an 8' Gray Planer for scrap and cleared out a big area in my new building where I know the floor is 8" thick - intention has been to relocate lathe and mill eventually - maybe sooner than later would be a good idea.

I'll get with my mechanic on Monday and check things out and let you know what we found.

[gar]

051210-0910 EST USA

coquina:

We had an HL-1 serial 0009. This was traded in on an SL-20.

The original HL-1 manual had some information on turret and headstock alignment. A newer manual for the HL-1 did not have this information, nor do I see any such information in the SL-20 manual.

The turret rides on ways that are at a 30 deg angle to the floor, and is the X-axis. The only turret alignment that HAAS discusses is the angular position of the turret relative to the X-axis. The technique is based on a straight bar placed across a tool holder in pocket #2 and #5 with the turret in position #1. Then the angular position of the turret is adjusted so that the bar is parallel to the X-axis ways. Why did they pick on 2 and 5 instead of 1 and 6? I do not know.

Note: the turret alignment is basically a one axis adjustment and relatively simple. That is unless you get into the relationship to X and Z.

What is assumed here is that if a crash occurs the most likely movement will be in the rotational direction of the turret, and not in the Z-axis mechanism or the mounting of the turret to the Z-axis.

Once the turret is aligned, then the headstock has to be adjusted. The headstock rests on a surface 0 deg to the floor. Thus, the surface you are moving the headstock on has to be parallel to the Z-axis ways. If not you can never achieve alignment. We have to assume this is OK, or the entire Z-axis mechanism would have to be adjusted, if it is possible. I do not have drawings that tell anything about this.

If your turret alignment is changing, then you need to find out why. The bar across pockets 2 and 5 should be a relatively independent means to test for rotational error. (edit 051210-1155 This alignment test does not require the headstock to be aligned. If the spindle axis was parallel to Z and centering was slightly off, maybe a few 0.000,1", then you might correct centering by rotating the turret. But this is not the correct way. )

The headstock is a two axis adjustment. Make it parallel to the Z-axis, and then adjust in and out for centering. (edit 051210-1155 This must be done after the turret is set correctly. )

The rotational axis of the spindle must be made parallel to the Z-axis motion, and the headstock must be moved front and back to center with the tool #1 pocket.

We had some alignment problems on the HL-1. We have never realigned the SL-20.

We ordered the SL-20 with a tool setter, seemed like a good idea. Not so. It is not real accurate, meaning to a .000,1", because of the poor pivot point. That is not the worst part. It is too easy to destroy the mechanism. A totally different concept needs to be used that is accurate and is unlikely to be damaged.

We also got a tailstock. Very useful.

What is lacking in all low cost lathes is a milling machine style tool changer that can change the tool in one or more pockets of the turret. This could solve the the long tool problem and the shortage of positions on the turret.

We also have a hydraulic closer on the HL-20 spindle, much better than the air operated spring unit.

We have a Hardinge 22J collet head and seldom put a chuck on the machine. This provides up to 2.25" dia.

Almost certainly you have the original style HAAS tool holders. I suggest eliminating the snap ring that holds the nut in the turret. Then whenever you change a toolholder completely remove the nut and washer and clean them and the turret pocket, and the toolholder thread. A nut should always run perfectly freely on the holder thread. Test every one before putting back in the turret. (edit 051210-1155. Also make a spanner wrench that can be driven from a socket wrench ratchet. This saves a lot of time. )

There were software bugs in the early machines, probably still are in newer machines, but many fewer.

Our business is RS232 communication. Our I232 Isolator system provides 115.2 kbaud up to 4000 ft and approximately 2000 v peak electrical isolation. This means that in most applications you could actually operate at 115.2 kbaud at 4000 ft and have virtually no bit errors. HAAS machines have either a 38.4 kbaud or 115.2 kbaud maximum rate depending upon age. Between 96 and 98 the increase occurred. 115.2 kbaud is 12 times faster than 9600 baud. This is a more dramatic change than going from typical dial-up to DSL. Isolation means you get nearly zero error bits even at high baud rate and long cable length. Visit our web site www.beta-a2.com. My miscellaneous photo page shows some parts made on HAAS machines.

We have a mold maker customer that installed our I232 System and he estimates about a 30 to 40% increase in efficiency, and better surface finish. Mold makers generally work with multi-megabyte files that must be drip feed.

His estimate is qualitative but results from faster operation of the machine on short strokes, and no bit errors that cause one to have to restart the CNC part program. This becomes especially important when the machine is started on Friday, everybody goes home, and you expect the part to be done on Monday. Better surface finish was achieved by going from 9600 baud to 38.4 kbaud which basically eliminated machine stuttering.

.

[Technical Ted]

coquina - It sounds like realigning the turret cured your problem so it's a fact that the turret is moving. So it's either moving too easily or is being crashed more often than you know. Did Haas give you a torque setting to tighten everything to after the alignment?

My other question is how much has the turret been off each time? Dial indicators and magnetic bases are not a very accurate way to align the turret. Gravity can affect the reading a lot. Put your base on a sturdy piece of steel & zero the indicator on the steel. Now turn everything upside down & see how much the reading changes.

[gar]

051210-1254 EST USA

Technical Ted:

It appears that coquina was told by HAAS to put an indicator on the spindle and use this to check the center position of a turret pocket to do turret alignment. This is not a good way because you do not have an independent measurement of the alignment of the turret and therefore can not independently track changes in its alignment. The bar across two pockets provides this capability.

coquina:

(edit do not read this paragraph.)
You do not use the indicator mounted on the rotating spindle to align the turret, but rather to zero position the headstock once the turret has been aligned, and the headstock axis of rotation is parallel to the Z-axis. This may have to be an iterative process.

(edit 051211-0921. Rewrite of previous paragraph. You do not use the indicator mounted on the rotating spindle to align the turret, but rather after the turret pockets #2 and #5 are aligned parallel to the X-axis, and the spindle rotational axis is made parallel to the Z-axis, then the indicator is used to adjust the spindle in-out position (front to rear and maintain parallel to Z-axis) and the turret X position to center the two together. )

Also there may be differences (there will be) between different pockets. Thus, one must always use specified tool positions to do alignment.

As Technical Ted said it is very important to know the magnitude and direction of the shifts that you are encountering.

I do not think it is clear cut that the only problem is turret alignment. In all probability you have both a headstock and turret alignment problem, as pointed out by Geof, if not other problems as well. But nut torque may be an important point.

I am not presently where the manual is, but HAAS said something about needing a 250 #-ft torque wrench. That is a hefty wrench.

(edit 051211-0852 The reference to 250 #-ft is under section 2.3 SPINDLE ALIGNMENT. Quoting from a part of 2.3 "-- torque the eight spindle mounting bolts to 250 ft-lbs carefully --". Note HAAS's incorrect units for torque. The unit of measure foot-pound is work, not torque. In section 1.4 on TURRET ALIGNMENT there is no reference to torque on the 10 bolts. In section 1.5 reference is 25 #-ft. Very likely HAAS wants the turret to slip in a crash and not the spindle assembly. That makes good sense. Note: torque wrench manufacturers tend to incorrectly label their wrenches in English units and correctly in metric. )

(edit 051211-2034 A way to visualize the spindle alignment procedure is to make a drawing in a plane perpendicular to the Z-axis. Draw a line parallel to the mounting surface for the spindle assemby thru the spindle rotational center line. This is parallel to the floor. Next, parallel to the X-axis, this is 30 deg to the floor, draw a line thru the center of the currently selected turret pocket. The intersection of these two lines is where the spindle needs to be positioned. Here the spindle axis and turret pocket axis are coincident. )

.