[flick]

You would need the center of gravity vertically aligned with the single front mounting point (the mill is not symmetrical) so the center may not be the correct position.

Perhaps a good justification for attempting to acheive a plane by measuring alignment, rather than attempting to acheive alignment by creating a perfect plane.

Somehow I missed post #27. Did you check the twist with the table at both extremes of the x-axis?

I checked it in every position but hanging upside down from the cieling. The numbers that I stated were approximately centered in X & Y.

[philbur]

What was the difference in twist between x-axis centered and fully extended in one direction.

Phil

I checked it in every position but hanging upside down from the cieling. The numbers that I stated were approximately centered in X & Y.

[flick]

I did not document. I do recall that when measuring Z/X alignment with X in approximately the middle of it's travel that the misalignment was about 3 times as severe at the positive limit of Y than at the negative limit of Y. At the time I was much less interested in the effect of varying the X position. The Y axis ways are a part of the base, so I expected any twist in the base to be reflected by the Y position. My measurements seemed to support that idea.

[zephyr9900]

Perhaps a good justification for attempting to acheive a plane by measuring alignment, rather than attempting to acheive alignment by creating a perfect plane.

I've been thinking a lot about what you've posted, flick, and come to the conclusion that I've probably been over-thinking my situation.

My machine was performing well before the move. Assuming I "know" that the machine hasn't been bent or traumatized during the move (I built a sturdy pallet for shipping the machine, it was moved in the truck along with our china cabinet, bookcases etc. with glass doors which weren't damaged, I wasn't present for the loading but was present for the unloading) it is just a matter of checking whether the machine is currently twisted.

It seems that I can do that by no more than sweeping the table with my indicator mounted in the spindle, choosing a large enough radius that I can encompass the total Y travel of the table. If I null out any base twist (knowing that both the table and column are mounted to the same base casting) then by definition I should be good to go. Afterwards I can borrow or buy a good square and confirm by Z-axis sweeping.

BTW I guess I need to read more about using a cylindrical square. It seems to me that in a generic situation (squareness to X and Y are both unknown) that an error in Y-Z squareness will affect the reading for X-Z squareness, because the indicator will be tracing a helical path up the cylinder...

Randy

[flick]

I've been thinking a lot about what you've posted, flick, and come to the conclusion that I've probably been over-thinking my situation.

My machine was performing well before the move. Assuming I "know" that the machine hasn't been bent or traumatized during the move (I built a sturdy pallet for shipping the machine, it was moved in the truck along with our china cabinet, bookcases etc. with glass doors which weren't damaged, I wasn't present for the loading but was present for the unloading) it is just a matter of checking whether the machine is currently twisted.

It seems that I can do that by no more than sweeping the table with my indicator mounted in the spindle, choosing a large enough radius that I can encompass the total Y travel of the table. If I null out any base twist (knowing that both the table and column are mounted to the same base casting) then by definition I should be good to go. Afterwards I can borrow or buy a good square and confirm by Z-axis sweeping.

I think you got 'er figured There's no reason that using spindle alignment changes throughout the Y travel shouldn't be just as good for addressing twist as using Z alignment changes. Just don't get fixated on making the spindle alignment zero out until you've also checked the Z alignment.

BTW I guess I need to read more about using a cylindrical square. It seems to me that in a generic situation (squareness to X and Y are both unknown) that an error in Y-Z squareness will affect the reading for X-Z squareness, because the indicator will be tracing a helical path up the cylinder...

Randy

It's an interesting point, and worth considering. The effect is minimized by a larger diameter of cylinder. It is also minimized when the point of the dial is swept up the cylinder as closely as possible to the "crest" of the cylinder in one of the cardinal directions. You can find the East/Wast crest by sweeping the dial across the cylinder along the Y axis. By do so you may also gain a good understanding of just how minimal the effect tends to be when you are dealing with relatively small misalignments, in proportion to the radius of the cylinder. If you know how to calculate the height of a chord, you can probably already appreciate this.

As misalignment decreases in the 3rd axis, the effect it has on the reading of the other 2 decreases even faster. If you have twist resulting from imperfect base mounting both axes will improve simultaneously as your shimming gets closer. The effect will likely diminish and disappear without you having to worry about it at all... especially if you start by aligning your spindle, and it happens to be very close to your Z alignment, as it probably is in most cases.

[zephyr9900]

It's an interesting point, and worth considering. The effect is minimized by a larger diameter of cylinder.

You are not kidding, flick. I did a little 2D CAD and found that, on a 3" diameter cylinder, it takes .0548" Y error to make a .001" X error, when the indicator starts on "top dead center" of the cylinder. It takes .0173" Y error to give .0001" X error under the same conditions.

Randy

[WallaceLau]

I hate to dig up an old thread, but I've had a lot of problem tramming my 1100 (bought used, just a month ago) and this is the closest thread I can find. For now I'd like to see if there is a consensus among the experts? Right now I am following the factory manual, shimming the front legs. I can easily get the X axis to be perfect, but the Y axis is 0.004" off at only ~3 inches apart. There is noticeable stepping when using the 1.5" face mill. Realistically, how bad is 0.004"?

By the way, before I shim the legs, Y error was "only" 0.003". Bringing X from 0.002 to 0.000 negatively impacted Y axis tram. Is it better to have some error on both, or have at least one of them being perfect?

Thanks,



W.

[philbur]

The certificate of inspection says 0.001" in 8" so you are way off. Your error sounds too extreme to be fixed by a bit of shimming of the base. As it is not a new machine I think you need to go right though the machine (in the right sequence) rather than jumping on one particular possibility.

The certficate of inspection procedure and numbers can be found here:

http://www.tormach.com/uploads/7/TD1...0112A-pdf.html

Phil

I hate to dig up an old thread, but I've had a lot of problem tramming my 1100 (bought used, just a month ago) and this is the closest thread I can find. For now I'd like to see if there is a consensus among the experts? Right now I am following the factory manual, shimming the front legs. I can easily get the X axis to be perfect, but the Y axis is 0.004" off at only ~3 inches apart. There is noticeable stepping when using the 1.5" face mill. Realistically, how bad is 0.004"?

By the way, before I shim the legs, Y error was "only" 0.003". Bringing X from 0.002 to 0.000 negatively impacted Y axis tram. Is it better to have some error on both, or have at least one of them being perfect?

Thanks,



W.

[BobWarfield]

One thing I found that made a huge difference in squaring my mill was to level the table first. Whether this is done by adjusting or shimming the feet on the base, or by leveling the stand is immaterial. But there is a big advantage in starting from a level table.

I tried squaring by shimming both before leveling the table and after, and it was dramatically easier, required less shimming, and was more stable ( I checked the squareness weekly until it settled down) after leveling.

In retrospect, I concluded that a slight lean meant a lot of cantilevered weight was levering things out of alignment as the axes moved. This bending motion was making it hard to get square axis motions.

The tools used for the process where a 4" x 12" cylindrical square, tenths DTI, and a Starrett precision machinist's level.

I'll be going through this all again shortly when I move my mill off its current stand and into a flood enclosure, so I'll try to get more pix of the process.

Cheers,

BW

[btu44]

If when facing does it cut deep on the Y- side?
Maybe the Z ways are very loose and are allowing the head to tip forward?

The Series 3 manual has a good procedure for adjusting the ways.

[philbur]

It's also possible to pull the dowel pins that pin the head to the Z axis saddle. This allows you to swivel the head a little bit in order to correct x axis tram.

But again, with a machine of unknown history and serious misalignment issues, you should check everything, otherwise you may correct one issue only to introduce another.

For example having the spindle axis at 90 degrees to the table surface doesn't necessarily mean the spindle axis is parallel to the z travel. There is a right and a wrong sequence for checking and adjusting a mill.

Phil

[WallaceLau]

Thanks for all the feedback - looks like I am gonna be busy for a while! I wonder why Rory from Tormach never send me the link to the Certificate of Inspection...

One more newbie question (and hopefully not one that's been asked a million times): I need to get a few parts out that doesn't require very tight tolerance, is there any long-term ill-effect if I run the mill as-is (for a month or two)? The table / stand is relatively level, I am still shopping for a machinist's level but my 4-ft carpenter's level is about as dead-center as I can get it to. And the ways seems to be getting enough oil. I think someone once told me it is not healthy to run a large mill until everything is properly adjusted, not sure if the 1100 is considered "large"...

Thanks!!



W.

[TurboStep]

W. Level is relative! I think if you have your machine leveled with a carpenters level then that should be more than sufficient.
As far as a machinists level is concerned I bought the metric version of this one:
12" MASTER PRECISION LEVEL FOR MACHINIST TOOL NEW .0005 | eBay
I would say that you don't need one of these to level the mill but I found it useful to ensure that the Y ways aren't twisted (I think this is the answer to philbur's question in post 27). Before I purchased it I wasn't sure whether it would be sensitive enough or maybe even so sensitive that I couldn't use it. I'm pleased to say that it works out really well.
It is often been recommended to shim the base to correct for tramming but this is quite likely to ADD twist to the base and may only be accurate for one Y position. I first shim the base to remove twist in the Y ways (using the level as reference), then shim the column to be perpendicular to the X/Y axes and finally shim/adjust the head to remove tramming errors.
Step

[rowbare]

You have to be careful tightening the bolts that hold the base to the stand. You can easily induce a warp in the base doing that. I would suggest loosening the bolts then leveling the stand before you do much else. Then do your testing and shimming. Once that is done, you can tighten the bolts up making sure that you aren't throwing anything out of whack.

bob

[TurboStep]

Bob, wouldn't it be better to shim the base (to the stand) first to ensure that Y ways have no twist with the bolts already tightened and then adjust the column and head?
Step

[rowbare]

Bob, wouldn't it be better to shim the base (to the stand) first to ensure that Y ways have no twist with the bolts already tightened and then adjust the column and head?
Step

According to a post by Greg Jackson on the Yahoo groups some time ago, the mill is in a 3 point stance when it is assembled i.e. the two back pads are on a surface plate and the front sits on a round bar. See the section "Shim the Column or the Base?" here: Tramming a Mill

Before doing something as radical as shimming the column, the first thing one has to do is evaluate the real error. Thus set the machine up on a level surface in the three point stanch and determine what the errors are. If they are within spec, then the problem is in how the machine is mounted to the stand. If they are out of spec at that point (after double and triple checking) then tramming the column may be called for.

That is why I suggested levelling the stand and leaving the bolts loose while figuring out what to shim. That most closely approximates the three point stance. Torquing the mill down down to an out of level stand sets up stresses in both the base and stand that will be very difficult to resolve by levelling alone.

I see it like taking two sheets of paper, stapling them together in four corners then trying to line the edges up. Anything you do in an attempt to align them will stress one sheet or the other. But if you align them before stapling it is much easier.

bob

[TurboStep]

Bob,
My very first measurement of the tram on my PCNC1100 appeared to be quite impressive. There was very little error in the middle of the Y travel. However, the error at the ends of the Y travel told a different story. This is quite easy to measure using the machinist’s level that I referred to in my previous post. It didn’t change very much between a 3 point stance and sitting with all 4 feet on the stand because the base is actually quite stiff as Greg implied. I say that the Y ways were twisted because it could have been a twisted base or an alignment error between the base and the ways. If you only shim one corner of the base in order to tram in the X direction (as many suggest) then you may be adding or subtracting twist but it’s very unlikely that you will be removing it completely. You will only be able to cancel out all errors to improve the tramming at only ONE Y position. It’s like a clock that isn’t running: it will be very accurate twice a day. Your link to the cnccookbook page quotes from philbur “Twisting the bed will most probably mask one error by introducing a second error”. I fully agree. IMHO this is not the way to do it.
I’m not talking about twisting the bed, I’m untwisting it!
Philbur also stated “I think that shimming the bed must be the last resort”. I’m not measuring tram at this point, only the change of slope of the table as I travel between the extremes of the Y travel. In this case it isn’t the last resort because it’s the just the first step in setting up the mill. Once the Y rotation has been corrected you will have a reference for the remaining operations. It’s then very likely that the column will need to be shimmed. This isn’t “radical”. If the column isn’t perpendicular (enough) to the reference then it must be adjusted. The only difficult part in shimming the column is getting access to the screws! The last step is to adjust the head.
Your last step appears to be “tighten the bolts up making sure that you aren't throwing anything out of whack”. What if tightening up the bolts does throw the tram out of whack? What do you do then? With my approach the bolts are already tight from the beginning. The errors are removed from the bottom-up and when the last adjustment is made to the head then you’re done.
Step

[rowbare]

Step,

I was thinking radical in terms of moving 600 lbs of column and head around to do the shimming. To some I am sure that is trivial, but I would try and avoid that if I could.

My concern about bolting the base tight to the stand before shimming/adjusting is that it might lock in some stresses and you will be constantly fighting those. Think of surface grinding a piece of metal that isn't flat. If you put it on magnetic chuck as is, the chuck will pull it flat. When you release the chuck after grinding, it will spring back and you now have a nice shiny warped surface. If you then flip it over to grind the other side, the same thing will happen and you will spend a bunch of time chasing the error. Whereas if you had shimmed the part appropriately the first go round, you would have generated a surface that would have remained flat when the chuck was released and you could have ground the parallel surface from that reference.

As for the tightening, if the machine is square before it is tightened down and goes out of square afterwards, then the interface between the base and the stand needs to be corrected since you have already proved the machine to be square.

If the machine isn't square before being tightened down then that is a whole different situation. Here hopefully you can manipulate the stand/base enough to square things up. However I am not sure that I would simply tighten the machine to the base and then try to square from there. I would be more inclined to do it iteratively and try to minimise the number of variables. Probably something like this, tighten the base to the stand using a torque wrench, measure, determine what correction to try, loosen the base nuts (might not need to loosen them all), apply the correction, re-torque and measure again. Loosening the nuts between adjustments avoids locking in the stresses, and the constant torque value gives you a repeatable reference.

I could be completely wrong though...

I hadn't clued in that your table was changing slope over the Y travel though... I assume it isn't a gib issue?

bob

[zephyr9900]

Hi all, back to our regularly scheduled program of "I really have no idea what is going on here"

I have been living with my out-of-tramness, which is the spindle axis out of square with the table surface laterally by a little less than a thou per inch. The parts I've been cutting lately have not been very thick (1" at the most) and when I flycut a surface, I'm taking passes in X that are about 1/2" apart with a 2" swing cutter, to minimize the cusping.

But I was doing some cleanup the other day and noticed that the taper pins keying the head to the Z-axis slide are not inserted the same deptth. The LH pin is noticably proud of the surface. If this side of the head was dropped very slightly, that would be in the correct direction to account for (at least part of) the observed spindle lean.

Do this left-hand pin protrusion look anamolus? I've never dealt with reamed-in-place tapered pins and am not sure how to even check its snugness. I "know" I can't just bash on it for fear of just indenting the cast iron pieces rather than realigning them. I'm thinking to have the machine fired up, lower the head until the spindle is supported on my "keep the head from dropping after powerdown" support, loosen the head bolts just a smidge and then try tapping the dowels and see if either seats more, then retighten the head bolts. Those of you who have had your mill apart for moving, how did you reseat the head?

Then I'm back to properly mounting the mill. I have one of the early, rough, non-coplanar-mounting-pads stands with the cheap rubber pads the mill actually sits on so any jacking/shimming/adjusting is moot. I'm trying to figure out an elegant way to support the mill on the stand adjacent to the actual mounting pads, leveling the stand relative to the floor, leveling the mill, and then some way to polymer-bed the mill onto the pads. (Remember these pads are not necessarily either level or coplanar so I'm not taking anything for granted, and any adjustments I make, I need to make with the stand bearing the mill's weight.)

Thanks,

Randy

[twocik]

If you're in a pinch to get a few parts out before you open her up, just tram and level a vise or fixture to the spindle. I've shimmed my vise for a few parts before, but it is recommend to properly adjust your machine before you use it.