Thank you for all of the great advice everyone! To answer a few questions i was moving the Z up and back down each time because i wasn't sure about dragging the tip of the depth gauge, i will repeat the tests with an Interapid dial i have. with the Z backlash at <.0005 i don't think it would be too far off to move it up and down (at least compared to the 0.006 i'm seeing).
I think for me a surface plate would probably work best, i am thinking about shimming it up with as thin of shims as possible to keep good contact with the table and preserve Z height, and then plug all of the holes with set screws. I like the idea of making my own but that limits me to just a little over the 9.5" x 18" work area of the tormach if i want to keep it true (i fear if i move it and resurface it will not be consistent due to the warp in the table. Tosa sells a aluminum tool plate for like $500-$600 but i wonder how much i would ding it up setting vices, turn tables, and other heavy fixtures on it.
Looks like i should start saving up for the Tosa Tool Plate lol in the mean time i am going to tram it again and double check the shims under the feet, spindle squareness etc and see if i can find any other culprit.
Thanks again everyone, this is one of the most helpful forums i have been on!
Did you use set screws with slots or hex sockets? I'm wondering if hex sockets would tend to get packed with chips that are difficult to remove even with an air gun. That's a minor problem on the depth gauge for my 4x6 bandsaw, but only one set screw gets filled on that.
Mike
Tolerances under discussion are here:
http://www.tormach.com/uploads/7/TD1...0112A-pdf.html
Test no. G2 and G10.
Are all the gibs adjusted correctly? Have you repeated the right hand front to back measurement (y axis) with the table in different left to right positions (x axis) ?
Phil
PS: Weight transfer with loose gibs can produce some interesting results.
That's exactly the problem I've had. I find it easier to leave the holes open, and just push a punch through the clear them. Hence my suggestion to user risers and leave bottoms open. I've had some Allen screws that were a real bear to get out, because chips got so tightly packed into the hex holes it was near impossible to get them out so I could get an Allen wrench in.
Regards,
Ray L.
Don't torus pros need to use risers regardless because their spindle can't go as low as a tormach, 5 inches from the table correct?
You can buy GOOD PARTS or you can buy CHEAP PARTS, but you can't buy GOOD CHEAP PARTS.
You guys knoow that you can cover the holes with cling wrap quick easy and cheaply. When you change the setup, change the cling wrap. That will keep probably 99% of the chips out of the holes.
Lee
Thank you for pointing that out, i have read through that document but i seem to have missed the relevant tests,. when you told me to look at G2 and G12 i can see it now, and i guess my machine is borderline in spec (probably could be measured to pass i would bet) but those are kind of loose specs in my mind. I would have thought the center T-slot to be much better than that (and henceforth will NOT refer to it as a Precision T-slot lol).
I think the machine is capable of much more and i have my sights set on some kind of Tooling Table (maybe i will make one out of aluminum first until i can afford a tool steel one), i may run an endmill down the top edge of the T slot also to make it at least a little better. And i think i will repeat the left Y axis measurement again at different positions, maybe with a vice or two on the table as well just to see if it changes.
thanks again for all of the great input
You're introducing all kinds of possible issues when you move Z up and down. It's even possible you have a Z axis issue and not a table one. There is no problem with dragging the tip as long as it only drops down slightly or you carefully move it up with your finger when crossing a T-slot. The few grams of force the plunger is exerting or the tiny force on the plunger housing isn't an issue.
That is good to know, i think i will try with the indicator and just lift it up over the T-slots, I hadn't really thought it could be a Z height issue because i have made a lot of small parts and hit the Z depth to +/-0.001 pretty routinely, but it does make sense now that a disconnect in the Z DRO and stepper could show the same error. I'm going to check it out again not moving the Z.
Z axis isn't good enough repeatability to be moving up and down with DRO to be indicating the table
even a tool change introduces some difference
Hi All.
Grinding the table will only fix a problem if the top surface is not parallel to the X slide way under the table. You can check that with a micrometer - (about 4" from memory) - easily enough without dismantling. If its parallel there is no need to grind.
keen
Why do you consider the center T-slot not having a precision enough alignment specification? The spec reads that the T-slot is parallel to within better than
0.6 mils (0.591) over a travel distance of nearly 8" (7.874") anywhere within the X axis travel limits. Also note that only the edge closest to the mill column has
been precision ground for alignment.
I have just checked the motion of my axes. For what it is worth here are the average results from multiple measurements:
G2 - Parallelism of axis motion to the table surface:
bottom left 0.00 mm
top left +0.04 mm
top right +0.01 mm
bottom right -0.05 mm
G7 - Lost motion of Linear Axes:
X-axis 0.015 mm
Y axis 0.025 mm
Z axis 0.025 mm
G8 - Axis repeatability
X-axis 0.005 mm
Y axis 0.005 mm
Z axis 0.01 mm
This is after 2 years of light use since last adjustment. (the results of which I should have kept!).
Phil
We made some more measurements tonight and i am very puzzled. we tried putting weight on the table in different spots and that just made the needle bounce but it always read accurately. When measuring the center T-slot we pushed the end of the table away from and towards the column (in case the Y gibs were loose) and that had the same result, the needle would vibrate a little bit but it wouldn't change the result.
We ran the Interapid Dial indicator across the table WITHOUT changing the Z height and the results were startling.
Setting the end of EACH row to zero and tramming right i got the following over 18" (x direction)
Top rail of table (closest to column) = -.0035
Second rail of table = -.0005
Third rail of table = +.0045
Fourth rail of table = +.0075
And setting the top to zero EACH time and tramming down in the y direction i got the following over 9.5" (Y direction)
Left edge = -.005
Middle = +.0055
Right edge = +.007
We took videos of the measurements and i will post that soon so you can review and let me know if something looks wrong.
These are WAY outside of the initial machine specs, and i even have the inspection packet saying that the machine passed QC. SO i hope i can adjust it to be better than this. My question is:
Do you think this is something that could be corrected by re-shimming the base of the mill?
To best determine what is happening you need to alter your measurement method since your current measurement effectively loses some
important information.
Move to the center of the mill table in both X and Y and adjust your Z dial indicator so that you can view both positive and negative Z distances.
Now take Z deviation measurements at 12 different locations[*] without making any Z axis changes. DO NOT REZERO the indicator after the
initial zeroing in the center of the table - that would only lose information. After recording the data you may want to repeat the measurements
again and record them separately to assess the repeatability. Now you can plot the deviations from zero in the Z direction by using Excel and
that will show what is happening. If you don't have Excel or some similar package just post you numbers here and I will plot them for you. This
method will clearly show any table twist or other height non-uniformity with the table XY position.
*NOTE: 12 locations is just following your above example. However, I would select 20 locations as an alternative; 5 in X and 4 in Y, and all are
equally spaced within their respective axis.
In a word, no. Twist due to the base casting not being flat would explain tramming errors, which would be pretty much the same regardless of where the table was positioned. But what you're seeing is clearly that the table is not flat, and not a consistent height above the X/Y ways.
All of your readings (both your original measurements, and the ones you took today) are actually quite consistent, both with your original measurements, and with the ones you took today. You didn't record enough data to say it's absolutely consistent, but it is certainly in the ballpark. It appears to me your left-rear corner is about 0.0035" high, and your right-front corner is about 0.0035" low, relative to the center of the table. It's possible it was mis-machined, or it's possible the whole casting twisted after being machined. If the former, the test I suggest below should show it. If the latter, testing for clearance between the horizontal ways on the table and the saddle using feeler gauges should show large clearances in some locations, and tight clearances in other locations. The clearance should be the same everywhere, within a small fraction of a thou.
Keep in mind the one factory test that might have caught this problem (Test G9, laser interfrometry, though it's not clear if this tests X, Y and Z, or only X,Y) is performed on only ONE machine per production lot. It is NOT performed on all machines. So, it's easy to see how this could have slipped out unnoticed.
One thing I would do is measure the table height using a micrometer or caliper, from the bottom horizontal way surface (the oiled surface where it rides on the saddle) to the top of the table surface. Do this at several points along the front and back edges of the table, ideally near where you took readings today. I predict you will get readings that mirror your measurements today.
You REALLY need to talk to Tormach about this. I can't imagine they would find that much deviation from flat acceptable, regardless of root cause. No point having 0.001" position accuracy in X/Y if the table surface rises and falls +/-0.005" based on where it's positioned. It might be possible to replace the table. It would certainly be possible to have this one re-ground, assuming it is not twisted.
Regards,
Ray L.
Seems like you're sorta torn between whether the table is actually 'crooket' or if your measuring technique is flawed in some way...
Do you have any long straightedges? Even a carpenter's level would probably be OK. Slap it on the table and see what it looks like.
Perhaps a laser (or even just a small light) shone from one end to the other would let you quickly visualize whether the table is flat or not.