Hi CNCer
Thanks for posting this - Factory QC of machine building is highly critical for us all.
If the gib is curved on one side and flat on the other that is VERY SERIOUS and a big slip up for the factory. If so Tormach should take this very seriously - I know Greg would.
I noticed my more recent 1100 was to a lower quality than the original I got in 2007 - and if what you say is correct, the factory has sunk even lower since then.
Keen
Yes, you are right, my oversight. If it is flat on one side the opposite side has to be flattened with reference to the machine to keep the proper taper angle.
With a visually noticeable concavity on the one side, a lot of mat'l may have to be removed to get that side flat, ergo the turcite since you may be undersize at that point. But remember the stress relieving. If there is residual stress in the gib, it may very well change shape as you grind or scrape making that difficult...
Please find some results of curvature measured on the discussed Y-axis gibb. It's done with a bit better methodology and with more data this time.
Results
Scraped surface: +.0065" bulge located in the center of the gibb
Non-sliding surface: right side flat; left side -.0020 depression with the center located approx. at 1/4 of the gibb length
Setup
- for both tests the gibb had the same orientation: thicker end on the right side
- the gibb was held on the 1-2-3 block with two neodymium magnets located close to center (pretty solid clamping method for this setup, verified no movement during the test)
- 1-2-3 block was slightly rotated in the vice to get the same level of both sides of gibb ends, only curvature was under scope not the taper
Videos
Setup verification, a few lines of g-code for horizontal movement with the extra dial preload on the star and end of test
https://photos.smugmug.com/photos/i-...Wv4pj-1280.mp4
Scraped surface test, manual movement (dial preloaded and zeroed)
https://photos.smugmug.com/photos/i-...JWfNd-1280.mp4
Non-sliding surface test, manual movement (dial preloaded and zeroed)
https://photos.smugmug.com/photos/i-...qS2TF-1280.mp4
I'll upload soon the video how this defect is affecting the table slop. With some details how to separate measurement of slop for the x and y axis and a few other interesting findings. I'm also looking forward to hearing your thoughts, suggestions.
LATERAL DISPLACEMENT OF WORK TABLE TEST
As per suggestions I've received from the Tormach support I was tightening the gibb to reduce the table slop. Results you can find below.
It looks really bad to me. It's a new machine, newer ever used. It's probably the gibb with geometry issue. There could be also curved support where this gibb sitting, or maybe the opposite dovetail? Any other suggestions? There is a noticeable improvement when the gibb is tightened but the sacrafice in the backlash is unaceptable. The slop from the X-axis looks pretty good when compared with the slop generated by the Y-axis.
Test Preparation
I was removing the gib for cleaning again, just before the test. It was cleaned with microfiber cleanroom cloths. All the way surfaces were also cleaned to mirror finish, no way there is any deburr or FM. The backlash of Y-axis for loose gibb is .008". The gibb limit screws are not excessively tightened. After reinstalling the gibb I was manually triggering additional way oil cycle. The oil was well distributed by mowing X, Y axes a good few times from 0 to max. Both axes were possitioned in their middle for this test.
Tightening the gibb to max. spec. backlash
Lateral displacement of work table way out of spec, even hard to measure (at the end of table) with my dial indicator.
Gibb tightened with backlash 3x max. allowable error
backlash (center of table):
0.076 mm (0.0030"), max. spec. for allowable error: 0.025 mm
Slop measured at 7.5" from the center of table
Measured for Y-axis only: 0.036 mm
Measured for Y-axis + X-axis: 0.044 mm
Calculated for X-axis: 0.008 mm
Slop measured at the end of table
0.074 mm (max spec. for lateral displacement of work table 0.04mm)
Video:
Lower Quality: https://photos.smugmug.com/photos/i-...GxNXJ9-640.mp4
HD: https://photos.smugmug.com/photos/i-...xNXJ9-1280.mp4
Tightening the gibb to 30perc over spec for backlash
backlash (center of table):
0.033 mm (0.0013"), max. spec. for allowable error 0.025mm
Slop measured at 7.5" from the center of table
Measured for Y-axis only: 0.054 mm
Measured for Y-axis + X-axis: 0.058 mm
Calculated for X-axis: 0.004 mm
Slop measured at the end of table
way above the max. 0.04mm of allowable error
Video:
Lower Quality: https://photos.smugmug.com/photos/i-...Ms3Rrj-640.mp4
HD: https://photos.smugmug.com/photos/i-...s3Rrj-1280.mp4
Hi CNCr
Sorry I am too bogged down with work to study all your findings in detail.
If a long gib like the X gib is just a little bent equally on both sides - but was made straight and accurate, then you can (and Tormach do) argue that it will straighten out when installed.
However if your gib is curved on one side and flat or of less of a curve on the other.
There is no adjustment that will fix this. Simply the factory have made a major slip up and you need replacement.
I am surprised that Tormach have not reacted strongly to this serious problem.
I have a background in this type of work - have measured. stripped and reconditioned several machine tools over the years.
Keen
https://www.youtube.com/watch?v=EchELgqGz-c
This is not making me want to buy a Tormach. Make a youtube video if you must to get the attention of Tormach
Sent from my iPhone using Tapatalk
Why don't you scan in and post the "inspection" report that came with your machine. I'm assuming it has some form of inspector ID on it (i.e. who signed off maybe just a number). It would be interesting to see if we can at least flag the "pencil whipper" that others should be suspect of upon delivery of their machines. Seems like yours should have failed in multiple categories upon a cursory QC final workup.
Well the quality system that was initially put into place has certainly produced a lot of functional machines with long service records. I own 2 of their mills. They have been used extensively and have had few issues over many years now. Certainly this defective machine is the result of a gross error that should be remedied by the OEM upon confirmation. The drive to zero defects is a noble goal for any manufacture, however, the glide path towards such a goal is never a smooth curve when you zoom in on it.
I don't believe that the folks at Tormach are going to have to break out too detailed of an Ishikawa diagram to put this one to bed. Taking care of the end customer first, while simultaneously making a lateral assessment of any suspect finished goods and WIP, is best practice upon confirmation of a defective unit in the field. I feel bad for anyone that has invested their hard earned dollar and fails to receive their value entitlement. Cool looking machinery that fails to perform is useless to the end user.
Everyone puts out a lemon from time to time. What gives me pause is the idea they know these machines have bowed gibs and allow them to keep going. Knowing these machines will almost certainly hold poor tolerances and wear faster.
I get the feeling Tormach has been growing fast in recent years - and with no apparent captain to hold the wheel steady, there is a real risk the QC vigilance buck gets passed around too much.
Come on Tormach, you have a great product range. Please post here and reassure us you are on to these issues.
Keen
I just viewed your videos for the first time and what I see doesn't immediately suggest an issue with the gib. When your dial gauge starts to move you see a distinct movement for each 0.0001" microstep in both videos. This suggests to me that the gib is not overly tight or causing excessive drag. If you were seeing jumps every 2-3 microsteps this would be an indication that something was resisting movement. Judging from the results you've shown I'd be more inclined to check the angular contact bearings and the ballnut for excessive play. first.
The big problem is that a remote (forum) diagnosis is often difficult - it may indeed be caused by gib - I can't rule that out while sitting on my sofa. Venting frustration on a forum doesn't help anyone, especially when there are lots of people more than willing to help you become dissatisfied with your mill. These discussions should be taking place between yourself and Tormach. They're the ones most likely to help you, and I'm sure they will.
Step
Hi TurboStep,
Please note that this thread is more about the table slop potentially caused by the gibb geometry issue. The backlash was just temporarily increased to see the relation: tightening gibb vs table slop. I do not see any issue with ballscrew nut, and angular bearing as the backlash on loose gibb is 0.0008". I'm keeping my frustration our of this board. By posting here I'm trying to share my findings, see others experiences in this matter. I'm expecting to get some corrections on my testing method, diagnosis, etc. I'm also intensively (intensively from my end so do far) discussing this issue and a few more with Tormach support. With no success so far but I still believe in them.
Hi CNCer
Can you confirm that the gib is 0.0065" convex on the sliding side and 0.002" concave on the fixed side.
If the gib is like this there is nothing to discuss - the machine is way out of spec, end of story.
Keen
Hi CNCer
The only way I can see you could indicate the gib as you did ...and get a false reading was if the gib was twisted or not level in the set up sideways.
Can you rest it concave down on a ground surface - eg your table surface, and measure the gap underneath with a feeler gauge...see if you get 0.002"
Then clamp it down flat, and use a straight edge to check the scraped side again to see if it is the 0.0045" convex.
Keen
Also, the videos showing the dial gauge running along the gib, how did you flip the gib between the 2nd and 3rd video. Did you flip front to back, like rolling a log, or from left to right?
How you're holding the gib onto the 123 block? The end of the second video appears to show 2 studs but I can't work out what they are.
Step