Is there a Benchmarks for CNC Routers?
Folks,
Allow me to preface this question with my situation. A few years ago, I started building a CNC Router table. It started out as a modest design with a 4'x4' table using a low-cost Delrin wheels straddling on a V-track linear rail. When the table was completed, I noticed that it had excessive amount of pliability. I tried to quantitate it by using a series of spring scales and a run-out gauge. I eventually upgraded the bearings on my X-axis with more expensive MGN15H bearings and rails. This stiffened my spindle considerably to the point where I needed a larger scale, as it exceeded 30 Newton limit with no noticeable reading of run-out. I am now in the process of replacing my Y-axis with the same type of bearings which will make this system increasingly stiff. When done, I may still need to replace the linear bearings in the Z axis, but I want to wait, as this may have diminished returns on invest. This project has become much worse than any drug habit.
Instead of using quantitated custom measurements I devised, I was wondering if there is a simple benchmark that exploits the quality of the design that I could carve. For 3D printers they use something called a 'Benchy' which is a cartoon-looking boat (Think of Horatio Huffenpuff's boat on Beany and Cecile) that has subtle features difficult to print to perfection. The resulting flaws exploit latent problems with adjustments and such. Is there a benchmark for CNC routers that is widely accepted?
Thanks,
Rick
1 Attachment(s)
Re: Is there a Benchmarks for CNC Routers?
Hi Rick - In short no. But look at ASME B5 attached . Peter
Re: Is there a Benchmarks for CNC Routers?
Quote:
Originally Posted by
peteeng
Hi Rick - In short no. But look at ASME B5 attached . Peter
Peter,
You are a wealth of information. I recall you sending me this document before. That is why I invested in the series of spring scales and run-out gauge. On another thread, I read of a similar question concerning build stiffness and was given a couple ball park standards of a couple of machine mills. On the lower end, it was 25N/um. The measurement from my CNC router build is ~.6N/um (30N/.002"). As mentioned, I can barely get an accurate reading with my travel gauge (.001 resolution), so it would be more economical to add a 100N spring scale than to purchase a high-resolution travel gauge. Regardless, my design has a lot of room for improvement and I expect to see this improvement when I replace my Y axis with MGN15H bearings and rails.
Thank you,
Rick
Re: Is there a Benchmarks for CNC Routers?
Hi Rick - There are many small mills and routers that test static stiffness at 1N/um or less. 1N/um is very low and people with machines of this compliance try to make them more rigid like you are doing. For hobby machines 10N/um is a good target 20 is better. It is difficult to improve machine stiffness by adding bits to it. The machines DNA is set in its basic design. You maybe better off designing a new machine, use the current machine to build the new one and then progress like that. This is typically how we have developed machines over the last 200 years. Each machine evolves into better machines. Good Luck keep making... Peter
Re: Is there a Benchmarks for CNC Routers?
Quote:
Originally Posted by
peteeng
Hi Rick - There are many small mills and routers that test static stiffness at 1N/um or less. 1N/um is very low and people with machines of this compliance try to make them more rigid like you are doing. For hobby machines 10N/um is a good target 20 is better. It is difficult to improve machine stiffness by adding bits to it. The machines DNA is set in its basic design. You maybe better off designing a new machine, use the current machine to build the new one and then progress like that. This is typically how we have developed machines over the last 200 years. Each machine evolves into better machines. Good Luck keep making... Peter
Peter,
That data is exactly what I am looking for. As mentioned, I have only replaced my X-Axis bearings and have seen a huge improvement, as I can see the flex is being caused by those cheap roller bearings on my Y-axis which will be replaced very soon. Since my design is based on 8020 extrusion that are bolted together with angle brackets and such, I don't anticipate that it will be as stiff or strong as a machine that is welded together or machined with solid billets of steel or aluminum. However, if I can achieve 5N/um I will be satisfied, especially for a 1500x1500 size machine.
Thanks again,
Rick
1 Attachment(s)
Re: Is there a Benchmarks for CNC Routers?
Hi Rick - Post images of what you have and what your doing. Always good to know what machines at this level are doing. Its been on my list to test my latest build (Scoot-S3) to see where it's at and compare actual to FEA results. Better bring it up the todo list. Stay away from welding more trouble than it's worth..... Peter
A mill - is usually intended for machining metal.
A router - is usually intended to cut plastic, timber, foam and other soft materials. Often it can cut aluminium, rarely intended to cut steel then it's in cross over territory.
The attached image is a survey of machines some 20+ years ago. I don't know where I snipped this from. A machine thesis in 2000 was chasing 70N/um and this was described as a very very stiff machine back then. Some recent VMCs have been measured at 150N/um and I would think that this is a typical VMC static stiffness now. I also came across a special titanium mill at 650N/um which is the stiffest machine I have found.
Also - The machine will have some static hysteresis or slack. So if you do your test add a small preload and then measure say 3 points of load vs deflection. Plot the results in a spreadsheet and do a linear regression through the data. The line will not go thru (0,0) this will show up how much slop is in the machine. Its important to remove this initial slop as it could be a big problem as well as machine compliance. You have done this by going from wheels to linear bearings on one axis. Linear cars come in several preload specs. In future get medium or high preloaded cars these have no hysteresis. Cars bought from ebay have clearance and hysteresis... keep at it Peter
Re: Is there a Benchmarks for CNC Routers?
Quote:
Originally Posted by
peteeng
Hi Rick - Post images of what you have and what your doing. Always good to know what machines at this level are doing. Its been on my list to test my latest build (Scoot-S3) to see where it's at and compare actual to FEA results. Better bring it up the todo list. Stay away from welding more trouble than it's worth..... Peter
A mill - is usually intended for machining metal.
A router - is usually intended to cut plastic, timber, foam and other soft materials. Often it can cut aluminium, rarely intended to cut steel then it's in cross over territory.
The attached image is a survey of machines some 20+ years ago. I don't know where I snipped this from. A machine thesis in 2000 was chasing 70N/um and this was described as a very very stiff machine back then. Some recent VMCs have been measured at 150N/um and I would think that this is a typical VMC static stiffness now. I also came across a special titanium mill at 650N/um which is the stiffest machine I have found.
Also - The machine will have some static hysteresis or slack. So if you do your test add a small preload and then measure say 3 points of load vs deflection. Plot the results in a spreadsheet and do a linear regression through the data. The line will not go thru (0,0) this will show up how much slop is in the machine. Its important to remove this initial slop as it could be a big problem as well as machine compliance. You have done this by going from wheels to linear bearings on one axis. Linear cars come in several preload specs. In future get medium or high preloaded cars these have no hysteresis. Cars bought from ebay have clearance and hysteresis... keep at it Peter
Peter,
Thanks. I plan to complete my retrofit this weekend. After making the necessary adjustments I'll take these measurements and publish them here. Perhaps these measurements can serve as a benchmark or baseline for future builders.
Regards,
Rick
1 Attachment(s)
Re: Is there a Benchmarks for CNC Routers?
Folks,
Sorry for the delay. Long story short, after converting my bearings to a MGN15 Linear Rail I had discovered that I had to fortify my gantry by adding web gussets and replacing my belt anchors. I soon found that my HTD-3M Timing Belts were too short and replaced them too. Unfortunately, the belts I replaced them with were lower quality and hence they stretched easier than the original belts.
measuring backlash was very difficult, as the static hysteresis was less than +/-.004" for both X and Y axis when moving each axis using the stepper servos under no load. Under a +/- 50N load the X axis .007" and .008" for the Y axis.
As for Newtons/um, this was very difficult to measure using the equipment available, as my analog dial indicator only shows .001" which is 25.4um per tick. Secondly, holding a spring scale steady and reading the dial did not lend to repeatability. Regardless I was able to perform a few measurement points:
