DIY Linear Rails

[High Seas]

JerryFlyGuy, Jerry you wrote:

...the rack that I'm planning on using has a length tolerance of +/- 0.003 so in the entire length of my unit the movement may be out as much as 0.018"[w/ 3 pc's of rack], this is acceptable. I would actually be happy with 1/32" accuracy, but in the thousands [+/-0.005 ish] in repeatablility. Also the rack/pinion connection is placed under load to maintain as close a mate as is possible, I've been told that a person can obtain as tight as basically a 0 backlash system.

Could you share the sources for that gear/rack combination, prices and part numbers too? Seems like a VERY good approach -- and you have some VERY long axes to deal with. Hmmm maybe the ol Shopbot wasn't such a bad idea... I've been intregued by the rack and gear approach some time and now with those numbers there may be a rush to that approach. Thanks, and
Jim

[JerryFlyGuy]

Jim, I'm pretty sure that was the tolerance number for rack/pinion's I'll have to check that number out and confirm it tonight when I get home.

Alot of spec's are available from bostongear.com they have some good online cataloges [ you have to register but thats the norm these days] as well as downloadable files for CAD.

I'm by no mean's an expert on this stuff compared to others on this forum, however if you do some search's on here you'll find a PILE of info on rack/pinion style systems. One fellow that's been especially helpful for me is 'Al' who's been there and done it himself.. Do some searchs! VERY good reading.

I've settled on a 12DP gear set and 6' long rack sections. I've also finalised my gearing to 5:1 reduction [ from the stepper down to the pinion] and hope to get a 0.003 per step resolution and then w/ micro stepping it will be a fair bit less than that. The real question is if the machine can produce those tight of a tolerance or not!

As Far as the plane is concerned.. they tell me it will get over 30mpg at over 200mph I've not flown ours yet so.. we'll see. Here's hoping! It'd be cheaper to fly out to the inlaw's rather than drive. Not to mention WAY faster

High Sea's;

If you look up Bostongear.com you'll find pretty much anything you want in the open gearing section. I have a local supplier of a different brand [ only because they came out 20% cheaper than ordering in the boston stuff] however most, if not all, of these gearing sizes are pretty standard and with some checking locally you should find a reasonable supplier. I'm planning to use 12DP gearing as it's been recommended to me by 'Al' [ he's on the forum] I had trouble getting the gearing spec'd for my application where I wasn't overloading my gearing past its 'rating'. 12DP seem's to be a good enough balance in size without being to weak or to course. You have to build some type of pre-loading to keep positive contact between the rack and pinion, otherwise you will end up w/ un-acceptable backlash. 'Al' recommends a "Rosta" tensioner however I'm planning on gas shock's due to my size/location and design. You'll have to provide some extra force to compensate for the tendancy of the pinion to want to push back out of the rack. There's some postings from a few days back that cover this.As I've mentioned to the other Jim, there's lots of good reading on the forum about this setup. Many people w/ more techical knowledge and exeriance than I, who's offered their advice and told their experiances.

Hope than helps.
Jerry

[erase42]

ive noticed that some of the buys on ebay arent as good as one would think, sometimes for a hundred or so dollars more a person could have bought new and got what they exactly wanted instead of making due with what appeared on ebay. example
thk hsr45 rail, 8 feet long, very heavy duty stuff. price new? 669 US, saw a guy pay 530 for used on ebay and one method of ups shipping was 168 dollars. not sure what the shipping would be on a new unit, but usually ebay shipping is inflated.

[JerryFlyGuy]

ive noticed that some of the buys on ebay arent as good as one would think...... <snip>
.....usually ebay shipping is inflated.

Here Here.. isn't this the truth. I've since given up on looking at things on EBAY, sellers either want to much for shipping or the end price is just to high. I would need to buy rails from several auctions to make up what I need, and waiting for them to show.. only to get sniped or out-bid due to high price has lead me to look else where. Jimbo posted a couple GREAT sites for rails and there are others. I don't think we all need the THK, IKO or whatever else HIGH level brand name rails are for sale on EBAY. There are several other manufactures who's product would work perfectly well for our application [ hobby ish stuff].

That being said.. does anyone know a good way to join linear rails so as to eliminate any kind of hang-up when the carriage/truck passes over the joint?

Weblinks?
Jerry

[jimbo]

Jerry,I believe you have to order them with that option. The ends will be custom ground and matched from the factory. I think it would be difficult to do yourself.

Part of the reason the linear rails go so high on ebay sometimes, is that vendors make it difficult to find out much they cost in the first place and many don't want the business of small time hobby people.

[erase42]

well im having to peice together more than one rail for my machine which is a commercial size foam cutter. Im hoping that due to the extreme precision at which the rails are ground, that joining end to end will have little effect. Also hoping that the length of the bearing slide will help, having only 4 balls on the break at any moment while perhaps 200 balls are on a good smooth section. Im ALSO hoping that if there is a noticable difference between rail ends(which i highly doubt) that the cunning sacrifice of a few feeler gauges as shims will make the rest go away. Any opinions?

[JerryFlyGuy]

Jimbo.. I have a confession.. [whats the hardest three words for a man to say ] I WAS WRONG.. When I told you that the tolerance for a rack was +/-0.003 over a 6' length.. If I'm reading my little chart in my trusty spec book [ anyone else have a huge 3 ring binder full of print-out's off websites w/ all the technical spec's you need for easy reference.. or am I just a nut and the only one who does this?] I find that for my 12DP rack which has a 1" face and THE FACE has a width tolerance of +0.000 & -0.003. The length of the rack has a tolerance of.. get this.. +1.000 & -0.000 [ends not machined] there is no place that I can find that tells you what the tolerance of error along the rack is.
What does that mean? Well.. here I am planning on building a big fancy system made of all the finest components [for the most part] but with a rack of questionable dimensions and accuracy. Oh well, what is it they say about measure with a caliper, mark with a chalk and cut it w/ a axe? I'm going to have to delve deeper into this rack/pinion accuracy thing. I'm still gonna use it, I just need to figure out how to adjust my system [map it] for any possible error's I can find.. [OR something like that.] I've got a feeling that on a system this long.. the termal expansion would be in excess of the previously [ believed] rack tolerance.

Erase43, how big is your system going to be? How many rail joints do you plan to make? How are you going to handle the Colinear-ness [ not a word I suspect but you get the idea] of all the rail segments? What kind of structure are you mounting the rails to? any Pic's?

Jerry

[jimbo]

Jerry, I did a little research. Found this site, http://www.qtcgears.com/ and this guide in their library documents, http://www.qtcgears.com/Q410/QTC/Q410P392.htm. I don't understand the terminology. The various errors can be calculated based on what grade it is and other parameters.

[JerryFlyGuy]

Jerry, I did a little research. Found this site, <snip>

The various errors can be calculated based on what grade it is and other parameters.

Wow.. need my friggin microscope to read it! It sure is pretty technical. I read farther onto the next page and it starts talking about tolerance in um which is way outside my tolerance limits. I think what I'm going to do is just map my system as I need to, to get it to a decent tolerance. I think I mentioned it before but, I think that the thermal expansion in this system is going to be a larger issue over some of my equipment tolerance's.

Making HASTE .. slowly..

Jerry

[JerryFlyGuy]

Hey guys just wanted to mention These Guy's have some pretty good pricing on rails. Anyone looking for rails that look pretty decent for our hobbyish level of use might be very happy with these.
Cheaper [or as cheap] as the stuff on ebay in some cases, the up side? You can order exactly what you want.

I haven't purchased any yet but others may be interested.

Jerry

[JerryFlyGuy]

Ok this thread is getting away from the DIY side of rails but I do have 1 question. What is a realistic linear bearing life to spec when deciding what size of rail to shoot for. Now I do understand that alot of used stuff might have some of its useful [or alot]life used, but say you were spec'ing a new rail system. What would be a good distance to shoot for? 100,000 km? half a mill km? less? Just looking at the numbers at this point.

Jerry

[Hobbiest]

not sure what the shipping would be on a new unit, but usually ebay shipping is inflated.

Many people do charge way too much on shipping, but much of the time, the honest people just break even after their time is computed, if they even break even (I have taken a bath, when I thought my flat rate would be enough to handle packaging, etc). If you don't like it though, definately don't buy it!

1200lbs for the gantry? I know that you have said less now. Still...800lbs for the gantry? Is it made out of a volkswagon? Your molds are going to be made out of foam right? Isn't that what I am understanding? Then they will still have to be shot with gelcoat at least. Then the gelcoat would hae to be finished to give smooth enough operating surfaces. I spent some time doing fiberglass negative work, and it is really awful, hard work, but sometimes the fastest way is to do a lot of really awful, hard work

[JerryFlyGuy]

1200lbs for the gantry? I know that you have said less now. Still...800lbs for the gantry? Is it made out of a volkswagon? Your molds are going to be made out of foam right? Isn't that what I am understanding? Then they will still have to be shot with gelcoat at least. Then the gelcoat would hae to be finished to give smooth enough operating surfaces. I spent some time doing fiberglass negative work, and it is really awful, hard work, but sometimes the fastest way is to do a lot of really awful, hard work

Hobbiest, your right.. 800lbs is.. well crazy. But, as I've said before, there are two reason's it got this heavy, I wanted a gantry that was as stiff as I possibly could make it and yet as light as possible. However Stiffness took precident over light. There isn't really that much material in the gantry but it all adds up. Its made of 40' of 3 x 3 .250Hss and a bit more than that in 1 1/2" sqr tubing. It's trussed so its very stiff for its weight, yet it meet's my rather tight size restrictiouns [ it has to fit in my garage for now]. Also, I work at a shop designing stuff in my day job and have access to this mat'l at a very good price so the steel in my complete system is actually the cheapest part. It's going to limit some of the performance, but at the same time performance is a compromise also. If I had built the gantry and it only weighed 150lb's when I cranked up my acceleration rates I could have accelerated from 0 to 400 ipm in 2 tenths of a second and to a 300ipm feed speed in 16% of a second [those are calculated numbers. not guess's]. But.. the gantry would have been wobbling all over the place. To combat this, I would have had to turn down the rates of accl which would have put me back at the accl rates I currently have.. or even lower. I figure for a couple hundred bucks more mat'l and a few days more time it's worth having the gantry I do have at the stiffness I do have. I can do some HSM and at my max accl rates o to 400 ipm rapid will deflect my entire gantry only a couple thou[and thats still accelerating to rapid speed in about 1 sec-or less] I was planning on these steppers regardless of other issues due to the cutting force they generate, this way I can do wood or other things [ possibly even alum under the right conditions].
Also it should be worth noting that the 800 lb's includes the weight of the z axis.. and the z axis... is a whole other bag of worm's. I won't get into it today.

As far as molds, this is where I placed my requirement for highest accuracy. Some of the people who I expect will want to use my services have their own [smaller size] machines to make their molds, there're of the store-bought variety. I don't want any issue's with any of my delivered product's when it comes time to match a molded part up from my system and their system. And you've seen molds so you know there is very little chance of accuratly checking your work when it comes to some of these 3d parts. You do it, it looks 'about right' and your done.. no pulling out the caliper's to check it. You can use 2D templates at stations.. but thats only as accurate as your eye and template [not to mention how accuratly you place it] I do plan to also do the finishing of the top surface's after the foam is finished w/ the mill. It will still require doing some hand work but that should be cut back to hour[s] vs 100's of hours to do it all by hand.
I'm curious to know how heavy some of the big 'store bought' iron is doing the same job, I imagine I'm probably at 1/10 of their weight? For a hobbiest [or is that a pro/am] level this is sufficent however..


wow.. I just read back over that.... sorry about the 'book'

Any idea's on what an acceptable rail life should be??
Jerry

[Hobbiest]

what is the size of the gantry? You could easily shave 1/3 to 1/2 the weight just by using .125 tubing. More strength? Go with 1" .125. The amount of triangulations you have illustrated is sufficient for a very strong cardboard bridge. You say though that the other tubing is cheap from work, but if it ultimatley cost you more time in the machining process than it costs you more money to build it with the less expensive materials. Are you going to be running steppers or servos? What size? Open or closed loop? Controllers? Have you settled on a guide system and positioning system (gear rack?)? As far as bearing life, it would depend on what bearings, what surface, smoothness of surface, hardness of surface and bearing, weight being supported, distance between bearings, height of gantry, cutting forces put on the bearings, machining and traveling speeds, debris, operating (ambiant) temp., duration of run...I think that may be all the factors I can think of (yes, I am a bit sarcastic). They should last plenty long enough, however, you should be able to expect enough life out of whichever guides you choose, that you can make the money to buy new ones

[JerryFlyGuy]

Woa! talk about bombarding
Ok I'll see if I can answer all of this...



>>>>>what is the size of the gantry?

The basic's of the Gantry are 10' long [ wide] it's 16" high and 16" deep [ along the x direction]


>>>>>You could easily shave 1/3 to 1/2 the weight just by using .125 tubing. More strength? Go with 1" .125.

Your absolutely correct, I could. However I'm concerned about the amount of weld-flexture that is going to take place. I'll most likely go w/ a 1 1/2 sqr .125 wll for the truss members.1" would work fine, however it would make all of less than 40lb's difference in the final weight. However the 3 x 3 will probably stay at 1/4 simply so that if I have to I can do some heat work and or grinding like this fellow did here
to get flat enough [ and parrellel] faces for mounting the rails. While I do plan to use many laser cut jig's and fixtures I still expect some deviations during the fab. and to add fuel to the fire... I don't claim to be an expert welder.. so that will just make it worse


>>>>>The amount of triangulations you have illustrated is sufficient for a very strong cardboard bridge.

Again your entirely correct if strength was my sole concern cardboard would probably he strong enough, however strength isn't my major concern. Stiffness is and to promote stiffness it takes mat'l. I figured out that this gantry would take something like 15000lb's in the center of it before it would fail, but 500 lb's vert loading and 1200lbs twist loading will deflect it less than 0.006".

>>>>>You say though that the other tubing is cheap from work, but if it ultimatley cost you more time in the machining process than it costs you more money to build it with the less expensive materials.

I'll be purchasing whatever size I need so, the 1x1 and lighter tubing sizes would be cheaper.. I'm just not convinced that they would be entirely better.

>>>>>Are you going to be running steppers or servos? What size? Open or closed loop? Controllers?

I plan for now to use steppers, they are 640 oz-in w/ a 5:1 reduction then put through a 12DP pinion of a 1"PD. these should put out about 800 lb's force at stall [ two combined on the x axis]. They will be open loop [ another reason for the over-kill on power size and gearing.. there is no way I need 800 lb stall-force for most of my machine app's] but I do plan to sometime in the future convert to servo, once I get familiar w/ the whole system and see if there is any lack for performance. I'd love to set up a closed loop stepper system if I have the nesc power and speed and am happy with that side of things. Possibly move to Mach4 and a Ge100 if/when that happens. I plan on Ge202's and Mach 3 for now.


>>>>>As far as bearing life, it would depend on what bearings, what surface, smoothness of surface, hardness of surface and bearing, weight being supported, distance between bearings, height of gantry, cutting forces put on the bearings, machining and traveling speeds, debris, operating (ambiant) temp., duration of run...I think that may be all the factors I can think of (yes, I am a bit sarcastic).

Yes, there is alot of parameter's to concider when spec'ing linear rails. I've done several of the calculations and depending on which size I was to go w/ I could get anywhere from 34,000km's right up to 500,000km's this is a fairly large range.. and I was hoping someone could say.. 'well for most of the typical use's [hobby and pro/am stuff] I'd try to get at least 100,000 km's from any rail system I spec'd" or something like that.. I've never gone through the exercise of spec'ing rail's before and don't know what a realistic number would be. I did some ruff numbers and found that If my system was to work at 300ipm for 17 hr's it would travel 5 km's a day [ thats from memory.. don't have the exact number infront of me] so for a 100,000 km system assume 20% error [in the life calculation] or a 80,000 system. 80,000/5=16000 working days.. hrm..... thats PROBABLY longer than the fatigue life of my structure..
From my calc's that would mean a 20mm rail system would work just FINE. When I went to do the actual modeling of the 20mm rails.. they are TINY! it just seem's odd that they would work.. of course appearanc's aren't always what they seem.

>>>>>They should last plenty long enough, however, you should be able to expect enough life out of whichever guides you choose, that you can make the money to buy new ones

Thats the ultimate reality.. however for say $250 to jump from a 20mm to a 25mm rail.. if it has 10 times the life for your system.. is it worth it?

Ahhhh... the life of a designer... lots of thought provoking question's there Hobbiest... anything else that comes to mind?

Jerry

[Hobbiest]

Stiffness doesn't always require mass. Some of the best race car chassis in the world use a birdcage style space frame. Some of them are even built out of 3/4" 1/16" wall square tubing, but because of the amount of triangulations, they end up stronger than any other frame. This type of gantry would require a jig whil welding, to minimize movement. The smaller the tubing, and thinner the wall, the more distortion is telegraphed to the material, but the thicker stuff is both harder to keep straight (in a jig), and harder to straighten (with proper, reverse application of heat). I hate to say it too, because you have mentioned that you would like to maybe cut other materials with it, but if it were me (I know it isn't, thank God), I would build this machine as a foam only, and use it for such. That seems like it would be the biggest money maker so far, and allow you to build a smaller machine for aluminum and wood. In fact, it might not be a bad idea to build the smaller machine first. In all likelyhood, you will probably find that the smaller machine for those uses, will be better suited to your requirements. Besides, you gotta cut some ribs for the foam molds anyway, right?

[JerryFlyGuy]

No your right stiffness doesn't require mass, the only good thing about mass is simply the reduction in vibration while machining... I don't think that is going to be a huge issue with this system. As far as building a smaller system for other things, this is a definate possibility. My plan in this regards had been to use the steppers and what not out of this system [ when I moved up to servo's] and build a smaller system for doing alum and heavy wood stuff. However my problem is this, room. I seem to be lacking in it in a huge way.. Well not really if I wasn't making a monster mill/router. I do have plans to move to a larger building in the future if this all works the way I need/want it to. However at this point I'm in enough crap already [the wife's V-dub in on the street] and building two system's would be... you get the pic.

I don't know that I'm planning to do alum on this system. It's just a possibility if it comes out significantly stiff enough and can take it. Thats still up in the air. However as far as wood and sheet foam, there won't be much point in building another system that would have a fairly large foot-print [ 4 x 8 at least] when I already have a huge mother in the works. The 4 x 8 system would be a good dust-collector and thats about it.

On other thing I forgot to mention before.. was simply that I need the stiffness of the system because I need the highest accel rates and feed speeds I can get and sustain.

example;
If I was milling a block of foam.. 15' x 6' x 4' high and I was going to essentially mill 1/2 the foam away. Using a 3/4" cutter with a 75% step over and 1 1/2" DOC. That cutter at 300 ipm could mill out 253 cubic" of foam per minuted. Now if we are going to mill out 1/2 the foam then 15' [ 180"] x 6'[72"] x 2' [24"] = 311040 cubic" of foam to remove. 311040/ 253=1229 minutes or 20 hr's[this doesn't include the 1 or 2 seconds to decel and accel again from an exact stop at each direction reversal].
Now this is only the ruff cut but if we could only sustain half that speed do to accel and flexture issue's we could take 40 hr's to do the same job. and thats just the ruffing. Imagine the final pass's where you have 1/16" or less of Step over and way slower finish speeds [50 ipm or something] and you get the pic. [btw that would be 58 hrs at 1/16" and 50ipm to do the finishing -just for ref]

So.. I decided to take the Army's line and 'be all I could be' And of course build the biggest baddest thing I could [ something that would make Mr. Tim the Toolman happy ]

Ahh well.. who know's you may convince me to go w/ the lighter gantry yet Hey.. btw do ya have any answer's to my earlier question about acceptable rail life expectations... any opinon's??

Thanks again...

Jerry

[Geof]

Jerry;

You asked a question a few posts back: "I'm curious to know how heavy some of the big 'store bought' iron is doing the same job, I imagine I'm probably at 1/10 of their weight? For a hobbiest [or is that a pro/am] level this is sufficent however."

A 10' by 5' by 2' envelope off-the-shelf machine with 800ipm X rapids, 2100ipm Y rapids and 1200 Z rapids with accuracy and precision around 1 to 2 thou, maybe a bit better weighs in at around 16,000lbs for the whole machine with a gantry/Y-axis/Z-axis weight of around 6000lbs (my guess). Your machine is four times the size so yes you are still well down in weight.

[JerryFlyGuy]

16,000lbs for the whole machine with a gantry/Y-axis/Z-axis weight of around 6000lbs (my guess). Your machine is four times the size so yes you are still well down in weight.

Thanks Geof, thats re-assuring.. at least I'm not over building it compared to the OEM store bought stuff.

Jerry

[Geof]

I think it is difficult to overbuild no matter what anyone says. I am a big fan of BSH engineering; "built like a brick ****house."

I have been following your thread for a while and have noticed you have discovered something that I did a while ago when I started to build a 8' by 12' machine. This was back in the dark ages of the early 1980s and I wanted a gantry drill press for drilling thousands of holes in 3/4" and 1" steel plates; these are strainers plates used in hydraulic debarkers. To take the reactive force from a big drill I needed a stiff gantry and to make sure the gantry didn't "ring" when I advanced it four inches in something like 100 milliseconds I needed an even stiffer gantry. I found I went around in circles getting heavier and heavier and eventually I had the main cross structure built; no legs, no Z axis just the bare beam with machined and ground ways, around 700 lbs. Then the bottom dropped out of the economy; I went broke and had to go work for someone else for a couiple of years and the machine never got finished. I still have the cross structure; want a good boat anchor?