Thanks Mat-C, I was curious as I never heard of multiple starts before.
Jason
I'm Imperial
My plan is to use 1" ID open linear bearings on a 1" supported hardened shaft for Y; Then .5" ID closed linear bearings on .5" oil hardened drill rod for X. My Z will be .75" ID closed linear bearings on .75" hardened shaft. (weird size choices I know, but those are what were available on E-Bay) I'll build the machine out of MDF and am guessing in the neighborhood of 20 pounds of weight to move around. I'll probably get it lower than that, but am trying to be conservative in my estimate. I plan on using thrust bearings on either end of a 1/2" thick delrin block for my all thread support at my driven end and a simple bearing at the far end. For my nut I plan to use the all thread to cut threads in another block of delrin. I just found the tutorial with formulas to estimate component sizes in the FAQ and am going to put my numbers in tonight and see what comes out. I plan on cutting mostly foam, MDF, Delrin, a little light wood now and then and one small part from 6061 Aluminum. I plan on using a small end mill and going extremly slow. I will probably end up making it smaller than my dream size, but want to make it big enough to upgrade later without buying new linear bearings and shafts. To think this all started because I wanted to turn an old printer into a stencil cutter. It just got bigger and bigger
Thanks again for the quick reply. This place is great! If I could just figure out how to stop reading thread after thread until early morning.
Randall
Thanks Mat-C, I was curious as I never heard of multiple starts before.
Jason
Or it will increase the efficiency. (Actually it's the pitch that matters, the multiple starts just increase the load carrying capacity.) With a high pitch, the screw will rotate slower for a given speed. So less energy is wasted in accelerating and decelerating the ball screw. A high speed machine with a low pitch screw can waste more in accelerating the screw than the slide it is moving. Then you need a motor with higher torque, but slower speed.Originally Posted by Mat-C
To get a "seat of the pants" feeling for the efficiency, try to backdrive it. That is, try to get the screw rotating by pushing the nut along it. The low efficiency of an Acme screw then becomes quite easy to understand. So does the better efficiency of a high pitch ball screw.
A word of warning, before everyone starts looking on eBay or firing off enquiries for multiple start screws; looking to increase load capacity and speed.
Preloaded Assemblies
1 Multi-start Rolled Ballscrews; I can guarantee that most multi-start rolled ballscrew assemblies are only loaded with balls in one track, or if more than one track is loaded most of the balls in the other tracks are doing nothing, just recirculating!!!
Just think about it; Rolled screw, manufactured by the "plastic deformation" process, it can be hard enough to preload one track!
To increase load capacity you need bigger balls, no pun intended!
2. Multi-start Ground ballscrews; Again we have the same issues, even with precision grinding in a controlled environment I guarantee that the second, third, fourth and fifth track are not taking there designated load (the first track is doing the majority of work) Take one apart, the balls in one track will be different sizes to overcome this issue!
One advantage of this is when the assembly requires a rebuild, just rotate the ballnut one track, fit new balls and your away!
Regards / Kevin
www.marchantdice.com/linear
I'm wondering if we can use software to bring cheaper rolled ball screws to within ground ball screw tolerances. Theoretically, if you had high repeatability, but poor accuracy, all you need is just the right amount of compensation (facilitated by stepper motors with enough "steps"). If you could temporarily mount a high precision measuring device (at C0 or better tolerance) to the carriage that your ball screw is moving and then take measurements throughout the length of the ball screw for each step that your stepper motor takes. You should be able to get a map of where the stepper motor thinks it is and how far the ball screw actually moved.
Step
1_____2_____3_____4_____5________6___7______8_____ 9____10
|=====|=====|=====|=====|========|===|======|===== |====|
0_____6____12____18____24_______33__37_____44____5 0___55
Distance in microns
If you reverse this mapping, you'll know how far to turn the stepper motors to get the exact desired position. For instance (using the example map above) if my stepper is at step 6 @ 33 microns and I want to move the cutting head 10 microns to the left to a distance of 23 microns. Since that is closest to step 5, I would only move the stepper motor one step left to step 5 @ 24 microns. In a non-compensated system, I would have assumed that each step was = 6 microns so i would have moved 2 steps over to step 4 @ 18 microns.
I'm thinking the problem here might be having to re-zero the stepper motor with your calibration scale before zeroing your piece with your cutting head. For this you would have to define your "calibration zero point" as a spot that you could always move your cutting head to with PERFECT accuracy. That is the point that aligns your map with your ball screw so if you get it wrong, your accuracy might be horrible. Then after this step, your system will have to be able to count the number of "steps" it takes to actually zero the object.
This lets the software know where along the map the you actually are. Because the calibration is relative to your "calibration zero point", the software correction will need to take place where you can get access to the number of steps between the calibration zero and the zero of the object. This puts it in the realm of the stepper motor driver software and not at the g-code level.
The beauty of this solution is that once you have stored your "map" you can remove your high precision measuring equipment. My guess is that these laser measuring devices are quite expensive so if they are only needed for calibrating then a single device could be used to calibrate several cnc machines and the cost could be distributed amoung them.
What do you guys think? I'm only a newbie at this so forgive me if I got this all wrong.
Mach3 can do screw compensation. The hard part is the measuring.
Gerry
UCCNC 2017 Screenset
[URL]http://www.thecncwoodworker.com/2017.html[/URL]
Mach3 2010 Screenset
[URL]http://www.thecncwoodworker.com/2010.html[/URL]
JointCAM - CNC Dovetails & Box Joints
[URL]http://www.g-forcecnc.com/jointcam.html[/URL]
(Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)
I just bought one of these for $29 + shipping off ebay
If aligned to your axis of motion 1 inch incremental steps could be measured with a dial indicator. Also they are highly accurate for lesser distances. More than adequate for mapping the first order errors in the screw.
I just love a bargain!
I was thinking of something more high-tech like the AR600 shown here: http://www.acuityresearch.com/products/index.shtml
I would imagine that a computer interface is a requirement since you don't want to manually make the thousands of measurements it would take to measure the entire length of the screw and then input them into a computer. (Ex 200 step motor x (12" screw/.200" travel per turn) = 12,000 steps/1' of screw, w/ 1/8 driver thats 96,000 steps!).
With that device you could zero your blank and cutting head, measure the distance of the cutting head to some reference pt, then step your motor through the various positions it will make during its cuts and make measurements to the ref at each step and input them into the computer. Still a very tedious task.
I think laser displacement measuring is the dogs danglies as the brits would say, but most of the DIY machines here hardly support that kind of accurate geometry that would indeed make use of the linear displacement accuracy. Certainly not for $29
Here's another idea I came up with. I'm not sure if my math here is correct so let me know if its not.
If you have a 1600 dpi optical mouse (like this http://www.logitech.com/index.cfm/pr...ONTENTID=10121) then I'm assuming that you can accurately measure movement as little as 1/1600in or .000625".
C5 is .0005" error but its over 12" so the mouse may not be very good at correcting screw errors.
Especially since mice are probably not so good with absolute positioning to begin with, but I bet they do real well at relative positioning.
That is if you move the mouse a small distances it will be fairly accurate, but over long distances it will probably loose track and start accumulating error.
With this in mind, you could use the optical mouse to create your own backlash compensator! It already senses in 2 axes and you already got your computer interface.
As long as your stepper motor's steps translates to movement >= .000625" you should expect to sense at least one tick from the mouse driver for each motor step.
If you don't sense the tick, then the stepper motor is taking up play in your system and you should take another step until you sense the appropriate number of ticks.
Is it backlash or screw compensation your after?
What IS that device?Originally Posted by DieGuy
Hi Swede,
thanks for such a good write up on ballscrews. I have a question. I have allways been taught that when tensioning bearings (like the front wheel bearings on a car) you tighten the nut until it bottoms out and then back it off slightly to allow clearance in the bearing for the grease to move between the race and the rollers, otherwise in a very short time the bearing will be damaged. Doing this allways left a very small amount of movement.
When tensinoning the bearings on the fixed end of a ballscrew should this rule still be applied or are shims used to control the clearence between the two inner races?
Cheers
Splint
I'm posting a reply to get this to show up again as a current thread.
It bears bookmarking.
Regarding post number 32, is that a master height guage? Also called a Hommel if I remember corectly.
If so, how in the hell was that highly accurate instrument only $29? ( or did they charge $1000 shipping )
I used to use them in the toolroom on a surface table for component measuring. Using a dial indicator on a normal height guage you would zero the indicator on the part being measured then move the indicator over to the Hommel and place it at the nearest ledge ( proper name???) and use the dial on top of the Hommel to move the column of ledges up or down untill the dial indicator was back to the zero point.
Each ledge was 25 mm apart and the micrometer dial at the top only had a travel of about 25 mm. Obviously if the dial read 12.7mm and you were on the 100 ledge you would be at 112.7mm from the feet/table
Please correct me if I am wrong but it has been a while since I last used one.
(And if it's not a Hommel the jury will disregard what i just said)
Regards all M
Hi all. I have come looking for info on ballscrews for my next machine, and this thread jumped out and bit me!
I know the thread is fairly old, but I have one question that maybe Marchant Dice ( Who I intend to buy from ) or another knowlegeable person can answer:
On the Marchant Dice website they have Ground, Rolled, and Trapezoidal ballscrews. Im thinking that Trapezoidal may just be science talk for 'Multi-Start'.....is this correct?
Thanks.
Trapezoidal is the metric equivalent of ACME.