Print PDF
Favorite
Email
Step 12Recommendations
Now the stepper motors you find in a printer would do just fine
but after a while I would highly recommend to upgrade to some bigger stepper motors.
As for the expanded pvc I would replace that with Acrylic but do to the fact that I had a budget and the short amount of time I had no choice but to go with expanded pvc but I would really recommend you use acrylic.
but after a while I would highly recommend to upgrade to some bigger stepper motors.
As for the expanded pvc I would replace that with Acrylic but do to the fact that I had a budget and the short amount of time I had no choice but to go with expanded pvc but I would really recommend you use acrylic.
| « Previous Step | Download PDFView All Steps | Next Step » |
I really like it project, i wanted to do it 1 year ago, but... so much money ¬¬
The 1/2in x 3/4in x 1ft bar is enough to make 12x 3/4in x 1in x 1/2in thk slides (They are smaller but still adequate) plus 3x 3/4in x 1in x 1/2in thk threaded pieces. You could actually thread the UHMW instead of inserting couplers which would have a couple benefits:
1) less backlash (assuming you make a tap using the actual rod that it will slide on)
2) less friction between the threaded rod and the threaded piece, which translates to less torque needed from your steppers, and less racking of the slides.
3) UHMW is softer than the steel rod, the threaded piece will be the one to where, instead of both parts, so when you start to notice a lot of backlash you only need tor replace the one part instead of both. Though because of the super low coefficient of friction for UHMW, it will take a very long time to wear out.
The remaining parts made of acrylic and expanded PVC could be made with MDF. I recognize there will be expansion issues, but it really doesn't matter for a machine this small built with this quality of components. Please let me explain why, before I get flamed.
Please note this only applies to std. all thread, with regular couplers and nuts. General purpose ACME rods and nuts will have much higher tolerances. But unless you salvage them from an old machinem you won't come near $20 for just those components, let alone whole machine. Especially if you use precision ACME rods (Over USD 100 for just the parts for this machine_, yes there is a difference between, all-thread, general purpose ACME, and precision ACME, a difference other than the price.
A) Precision
Odds are the steppers used on this project are going to be 7.5deg per step, or 48 steps/rev, Unipolar or Bipolar motors, probably with only full step mode. Yes, going half step, eighth, 1/16, etc... will improve the precision, but we are pretending that the builder doesn't have the circuitry for such things. As even building your own board from scratch, to handle that functionality will cost a couple dollars each at a minimum, if parts need to be purchased. And as you can see below, fractional stepping won't really help for this machine other than part motion acceleration.
Now with the recommended threaded rod of 5/8"-24, that is equilalent to 24 revs/in * 48 steps/rev = 1152 steps/in which is 1/1152 in/step = 0.000868 in/step or 0.868 mil/step or 0.868 thou/step depending on how you like to think about it, (this is an estimate, I will explain why under the Accuracy header) Now consider that the typical backlash on a plain threaded rod using a regular nut or coupler (from my own experience and measurements) is on the order of atleast 1-10 mil (Making a custom nut, would greatly improve this number). With this in mind our effective precision is now 0.868 +/- 5 mil/step being very generous.
Then factor in that the nut will occasionally stick to the rod, torquing the whole slide mechanism. This will impact precision in that the theoretical distance traveled in this case 0.868 mil/step will in fact be 0 assuming it did not move at all. It is unlikely that it will stick for more than a few steps, but that is still a couple mil precision lost. A low friction nut would help in this case as well. Now we are 0.868 +/- 7 mil/step. Not to mention the pitch along a rod may vary to some degree.
Thrust deflection in the screw/stepper, can also be a major headache. If you have ever pushed or pulled on the spindle of a motor, you will have noticed there is quite a bit of play. This will translate into additional backlash (not technically the correct term, but the motion is similar). This can be eliminated through the use of thrust bearings. Unfortunately this project doesn't include any. The play on average in many of the steppers I have pulled out of printers and scanners is on the order of a 1/16th or 32nd of an inch or 62.5 and 31.25 mil respectively. That is huge. So we are left with 0.686 +/- 35 mil/step.
Does MDF expand yes, I am not going to contest this fact. However, it is _never_ going to expand to such a degree over the course of a single session (barring spilling something on it on it, or generally introducing it to a 100% relative humidity environment) that you will be able to measure an appreciable deformation from start to finish. Next because of how it is formed and used, it is probable that any expansion that does occur will happen proportionately along the x and y axis of the machine. The z dimension (typically the thickness of the base) will expand more but still at an extremely slow rate. MDF if exposed to a 30% to 90% relative humidity environment will expand ~0.3% in length and width, and in thickness up to ~5%. However, this often takes weeks or months to occur. for a 24"x24" piece that ends up being little more that a 1/16th of an inch change over the whole machine.
B) Accuracy
Accuracy on a CNC machine comes down to a few things. The theoretical distance traveled, 0.868 mil/step in this case, vs the actual distance traveled. The squareness or orthogonallity of the axis'. There are other accuracy points but for our purposes they are inconsequential, if I ever get around to writing this whole thing up as an instructable, I'll include them there.
As far as 24 rev/in on the 5/8"-24 rod, you should be skeptical as to that value. Precision ACME rod has a tolerance of +/- 9 mil travel per foot, not to mention the tolerance between each thread. This stuff is extremely expensive in relative terms and is far more accurate. I have seen regular all thread that was nominally 5/16"-18 with 18.1 or 17.9 threads per inch. That works out to be a difference of ~ +/- 66 mil travel per foot. This can be accounted for over the lenght of the whole machine by telling the machine to move a set number of steps then measuring how far it actually went to calculate a steps/in or in/step, Depending on what the cnc program you use. However, there is no easy way to account for the fact that while it may be 18.0 tpi in one area it could be 18.1 in an other area. I have see this happen on a 2 foot section of rod.
Right angles are very important when doing cnc, I'm sure someone could write a program that allowed you to enter the relative angles of your axis' and it would calculate the g-code to account for the skew. But as far as I know functionality for this is almost none existant. Now it seems that the main concern for using MDF is its propensity to expand. Warping on this scale, with a reasonably thick piece of material is for practical purposes non-existant. Now one way to handle this issue is to simply only attach each axis slide assembly on one end and allow the other to be free. You could also have a slot and bolt on the other end that would allow for it to be secured while cutting and when need be it could be adjusted. The Z access is suspended, and can easily be calibrated for each different job (usually it would be anyway), so as long as the 4 legs supporting the top are identical they should in theory have expanded/contracted at the same rate, thuse only lifting and lowering the z axis slides. Any deviation from vertical that occurs is going to be less than the runout of the tool and spindle.
This really belongs under precision but it has components here as well. The finally runout on the tool and spindle assembly, which in this case is the dremel and z axis slide. Because of how the z axis slide is supported. A cantilever beam for any engineers out there. There is a propensity for the whole assembly to deflect/bend. So when the cutter is moving through a piece force is exerted against the arm and it moves.
_Conclusion_
At the end of the day MDF for this type of project should be perfectly acceptable. The only structural parts that wood or mdf couldn't replace is bearings, bushings, and rods. As there is simply too much friction. Yes it will expand, and maybe warp a tad, but it won't happen while you are using it. And if you build your machine with that in mind you should be able to adjust the frames and and slides to acount for it, when the humidity changes. There is always sealing the MDF also which essentially lock your machine to one size within tolerances for motion I listed above.
I hope this is helpful, good luck with your project.
1. With the measurements you have used what is the useable cutting area?
And
2. Is there any reason you couldn't screw and glue the joins for extra stability and strength?
Any help would be muchly appreciated