-
Welcome back Guest! Did you know you can mentor other members here at H-M? If not, please check out our Relaunch of Hobby Machinist Mentoring Program!
Is hand scraping a bench top mill worth the money?
- Thread starter ArtJ
- Start date
- Joined
- Nov 14, 2020
- Messages
- 1,761
OK. Wow I wish Richard was here.
Bolt your machine to a firm base and level it.
There can be twist in cast iron no matter whose it is.
Tram your spindle to the table and start cutting.
It easy to get caught up in the " it's a new machine and it should be .000*
You are so fortunate to be able to buy a NEW machine, as you are making things you will KNOW just what issues you have and having "commissioned" many machines you will see the iron relax to "its " position and you will level the machine again and find it only needs fine adjustment.
You are on a journey with your brand new machine I am envious.
Bolt your machine to a firm base and level it.
There can be twist in cast iron no matter whose it is.
Tram your spindle to the table and start cutting.
It easy to get caught up in the " it's a new machine and it should be .000*
You are so fortunate to be able to buy a NEW machine, as you are making things you will KNOW just what issues you have and having "commissioned" many machines you will see the iron relax to "its " position and you will level the machine again and find it only needs fine adjustment.
You are on a journey with your brand new machine I am envious.
- Joined
- Jan 25, 2015
- Messages
- 2,558
Don’t really have much to add, but 3200 for a scrape!
Holy carp!
Thats almost as I paid for my bench mill!!!!!
I know scraping is labor intensive, but man!
I’d have kittens if I had to sink that much into it after purchase! Pretty sure it would be going back to the seller if that was the case….
Holy carp!
Thats almost as I paid for my bench mill!!!!!
I know scraping is labor intensive, but man!
I’d have kittens if I had to sink that much into it after purchase! Pretty sure it would be going back to the seller if that was the case….
Matt,
I did not intend to hash this out on a public forum, that’s why I did mention a company name or brand.
yes, I was waiting for shims because you suggested that there might be space below the base that would cause warping when tightening the bolts.
Using a feeler gauge there is no gap to shim so no warping right?
When I level the table with the bolts provided, my expectation is that the table will stay level or relatively level in X and Y. It does not. At the one spot where the table is level, tram is within .0005 thou and nod is .0025 thou. If I move the table at all, this changes. Ok, some should be expected but…..Measuring just the 8” table surface, with the dial indicator contact point directly under the center of the spindle, the table moves up toward the spindle about .0027 thou and out of level when moving the table out from the column. This changes an acceptable tram and nod (at the position where the table is level) to not acceptable as the table moves out.
I may not be explaining this well enough in writing to you or your tech people, it all gets too wordy and too easy to misinterpret. That’s why I have tried to talk to you on the phone, even FaceTime so I can show you while explaining.
Even a short video is to large a file to attach here.
I did not intend to hash this out on a public forum, that’s why I did mention a company name or brand.
yes, I was waiting for shims because you suggested that there might be space below the base that would cause warping when tightening the bolts.
Using a feeler gauge there is no gap to shim so no warping right?
When I level the table with the bolts provided, my expectation is that the table will stay level or relatively level in X and Y. It does not. At the one spot where the table is level, tram is within .0005 thou and nod is .0025 thou. If I move the table at all, this changes. Ok, some should be expected but…..Measuring just the 8” table surface, with the dial indicator contact point directly under the center of the spindle, the table moves up toward the spindle about .0027 thou and out of level when moving the table out from the column. This changes an acceptable tram and nod (at the position where the table is level) to not acceptable as the table moves out.
I may not be explaining this well enough in writing to you or your tech people, it all gets too wordy and too easy to misinterpret. That’s why I have tried to talk to you on the phone, even FaceTime so I can show you while explaining.
Even a short video is to large a file to attach here.
To all who replied, I appreciate your perspective on hand scraping. It’s a new skill set I’d like to learn. The $ would be better spent taking a class. Thanks
Art
- Joined
- Mar 22, 2013
- Messages
- 498
Not what I would call hashing out, public forums are great, the more people involved the more may have good suggestions (Or confusion, after all its the internet and ideas and opinions are all over the place)
I sent a few things to check by email, when you have some time check those out. I'm off to the airport for vacation, I'll get back in to it when I return. For now check that nod again and make some cuts. Ill send you a drop box link in a sec before I go to upload large files video, no limit there. I can assure you that if there is a problem, it will be taken care of.
I sent a few things to check by email, when you have some time check those out. I'm off to the airport for vacation, I'll get back in to it when I return. For now check that nod again and make some cuts. Ill send you a drop box link in a sec before I go to upload large files video, no limit there. I can assure you that if there is a problem, it will be taken care of.
- Joined
- Feb 1, 2015
- Messages
- 9,640
Table lift can be expected due to the center of gravity shifting. Details are here in post #2. https://www.hobby-machinist.com/threads/g0704-accuracy-question.60622/#post-499447 The fix for the lifting problem is to tighten both v and y gibs as much as practical consistent with ability to move the table and wear considerations. Of necessity, there needs to be some clearance for the gibs. A thousandth clearance provides almost a a thousandth vertical movement, As the table is moved, the center of gravity moves with it and eventually it is past the ways causing the table to pivot about the bear end and lift the far end. This occurs for both x and y axes.
Tormach's recommended method of adjusting table gibs is to tighten until an increase in backlash is observed, the backlash being caused by twist in the lead screw due to increased frictional forces. Adjusting by this method still leaves an excessive clearance, as indicated table movement when pushing or pulling on the table.
The Tormach mill has no locks for the x and y ways but using the locks on a manual mill should show an improvement in tram consistency for tram in different locations as it pulls the table and saddle into close contact with the dovetails on the far sides.
All table movement is relative to the ways. An assumption is made that the table surface is always parallel to the ways so it becomes a de facto reference for tramming the head. That isn't a good assumption. When I tram my Tormach mill, I place a sacrificial plate with the center as close to the center of gravity. I then mill a flat surface using a 1/4" end mill. The reason for the small end mill is to minimize any effects of the head being out of tram. This presents a surface as close as possible to the parallel to the ways. I then tram the mill to that surface.
Hi @ArtJ
So do two things first:
1) repeat the level measurement on the reference surface. This is the surface at the top of the dovetail that stretches from one dove tail to the other . This is a ground surface and is what is most likely used to fit the tooling to when the dove tail was cut. It should be really flat all the way across and from the front to back (y-axis).
2) The more important property is the backlash and this is highly dependent upon how tight the gib is. The following will tell you if your dovetails for the way are parallel. First tighten the gib up a little too much so that you see backlash. (As @RJSakowski just said tighten the gip up and measure the backlash as a function of the y position.) Now just crank the handle and see if you can feel the required torque is different as you move in the y-direction. (Here torque is just how hard you have to turn the handle. Not an actual torque measurement. Although that would be great if you had that ability.) This drag will tell you if the ways are parallel or not. Not uniform then they are not completely parallel. Per your measurement they may not be parallel because some portion of the dovetail is not flat or if the dovetails diverge... or other non-ideal things. When the gib is slightly over tighten the lead screw and couplings will tend to act like springs and you will be able to feel the torque differences along the y position. Having done this, find the point where the torque is the min. and re-adjust the gib to give you just a bit more backlash than you would have if the gib were very loose. If it is too loose the backlash will be due to other things than the way problems. Then repeat the torque versus y position . Then move on to actual backlash measurements as a function of y axis motion. Write them down as a function of y position.
I did all of this and found that my ways were way out of parallel. Then, because my reference surface was really flat I used it and a digital dial gauge to measure the variation of height along the area of the "flat" part of the dove tail. Mine was pretty bad, several thousands off. Furthermore the backlash varied by many thousands. In fact, at one end I was concerned that it might be stressing the saddle too much. I also found that the "flat" rolled off as you went from the dovetail corner outward by a lot... in some places and this was not consistent as a function of y distance.
I have been working on these surfaces and have found that by sanding off some of the higher parts of the dovetail flat I can adjust the distance between the dovetails and significantly reduced the backlash!!
Anyway, flat is a relative term and is far more important than level! Also, you cannot fix what you cannot measure and I tend to think one of the better metrics is the backlash measurement. So I have been making my self some measurement tools which hold the dial gauges so that I can determine what needs to be "adjusted". I started out thinking I would learn to scrape, but found that sanding works pretty well and, actually, maybe better. Yes, even on a hardened way.
Dave L.
There have been a lot of useful comments here. But let me just say, I have been dealing with similar issues for some time and have learned a few things. WRT to your level pictures. I would suggest that you repeat these measurments, but by putting the level on the reference surface. You measured the "flat" part of the dove tail of the way. Who even says that the matching surface on the saddle even touches at the point where the level is placed. If not the level measurement may not be relevant. i.e. who says the flat part of the dove tail is actually flat!
So do two things first:
1) repeat the level measurement on the reference surface. This is the surface at the top of the dovetail that stretches from one dove tail to the other . This is a ground surface and is what is most likely used to fit the tooling to when the dove tail was cut. It should be really flat all the way across and from the front to back (y-axis).
2) The more important property is the backlash and this is highly dependent upon how tight the gib is. The following will tell you if your dovetails for the way are parallel. First tighten the gib up a little too much so that you see backlash. (As @RJSakowski just said tighten the gip up and measure the backlash as a function of the y position.) Now just crank the handle and see if you can feel the required torque is different as you move in the y-direction. (Here torque is just how hard you have to turn the handle. Not an actual torque measurement. Although that would be great if you had that ability.) This drag will tell you if the ways are parallel or not. Not uniform then they are not completely parallel. Per your measurement they may not be parallel because some portion of the dovetail is not flat or if the dovetails diverge... or other non-ideal things. When the gib is slightly over tighten the lead screw and couplings will tend to act like springs and you will be able to feel the torque differences along the y position. Having done this, find the point where the torque is the min. and re-adjust the gib to give you just a bit more backlash than you would have if the gib were very loose. If it is too loose the backlash will be due to other things than the way problems. Then repeat the torque versus y position . Then move on to actual backlash measurements as a function of y axis motion. Write them down as a function of y position.
I did all of this and found that my ways were way out of parallel. Then, because my reference surface was really flat I used it and a digital dial gauge to measure the variation of height along the area of the "flat" part of the dove tail. Mine was pretty bad, several thousands off. Furthermore the backlash varied by many thousands. In fact, at one end I was concerned that it might be stressing the saddle too much. I also found that the "flat" rolled off as you went from the dovetail corner outward by a lot... in some places and this was not consistent as a function of y distance.
I have been working on these surfaces and have found that by sanding off some of the higher parts of the dovetail flat I can adjust the distance between the dovetails and significantly reduced the backlash!!
Anyway, flat is a relative term and is far more important than level! Also, you cannot fix what you cannot measure and I tend to think one of the better metrics is the backlash measurement. So I have been making my self some measurement tools which hold the dial gauges so that I can determine what needs to be "adjusted". I started out thinking I would learn to scrape, but found that sanding works pretty well and, actually, maybe better. Yes, even on a hardened way.
Dave L.
- Joined
- Feb 1, 2015
- Messages
- 9,640
I developed a simple means of actually measuring the torque, which was subsequently used by John Saunders in one of his videos.
Measuring Ball And Lead Screw Torque
I have recently begun the task of tuning up my Tormach PCNC 770. Part of the process is adjusting the ball screw thrust bearing preloads. Two methods for adjusting are recommended: measuring backlash or free play as the bearing play is removed, and estimating the amount of torque needed to...
www.hobby-machinist.com
