BUILD LOG: Work scrap bin router build [Archive]

Vic66

10-01-2010, 10:08 PM

Hi all

I'd be interested in any comments about my ex-work scrap router design, I've not got a lot of cash to spend and am trying to use parts sourced cheaply through work, so the build is based around the parts I can get, I want to cut parts to length at work then assemble them at home in my workshop, keeping it a simple bolt together assembly. I frequently use a local laser cutting company at work and they don't mind cutting small batches of parts for me for cash so I've used quite a few 6mm ally laser cut parts in the design.

So far I've managed to get a pre machined 16mm ballscrew with a thread length of 1360mm and 12 Hepco dual V bearings, so I'm leaning towards a similar system used on Joes hybrid machine using steel angle for the V bearings to run on, all the aluminium is standard section with some commercial vehicle ally bearers (£15 for a 2.5m length) for mounting the bearing runners on. I have a 16mm ballscrew for the Z axis, I was planning on buying cheap trapezoidal for the X axis hoping to upgrade later on.

I've pretty much finished the design and want to make a start on the build but I've got a bit of a dilema, I'm thinking that only having one drive on the Y axis might make the X axis carridge 'walk' up and down, and I'm thinking about using my 1360mm ballscrew for the X axis instead of the Y (gaining 200 or 300mm of travel on that axis) and using two trapezoidal rods and two steppers for the Y axis, the rods would run in the channels (see image 3) and I would remove the cross beam from underneath.

This would be my second router, I've built a Rockliffe fixed gantry prior to this and got hooked

Cheers

Vic

Vic66

10-01-2010, 10:14 PM

I forgot to mention cutting area is currently about 720 x 1230 x 150

Ross77

10-01-2010, 11:26 PM

dickieto

11-01-2010, 12:02 AM

Vic66

11-01-2010, 08:28 AM

Hi Vic.
Nice work, Looks very well thought out. I like the idea of using ali channel instead of the profile bars. There was a thread looking at an open source mid range Ali router machine and it looks like this could be it.....

not sure on your axis terminology but i take it by y axis you mean the longest axis? if so then you could use belts or wire to control the racking and stil use one ballscrew. If you increase the distance between the bearings then that will also reduce the racking but require longer rails or reduced travel.

My main concern is why use c channel for the rail supports if the ballscrew is in the middle? it would be much stronger to use rectanglar box section, I would at least provide support at the ends and mid point.

Hi
I get confused with the axis terminology, but I did mean that the Y is longest.

I havent got the model to hand at the moment (at work) but I think the bearings are about 270mm centres I'd like to keep it around 4ft cutting area and don't really want to make them wider if I can get away with it.

I'm not certain what you mean about controlling the racking with belts or wire ? do you have any links to something that has used something like that ?

The design started off using box section, I am aware that I might get some flex with channel, I have some ideas about strengthening it if I need to, the problem with box section is bolting the steel angle to it, you cant get to the nuts to tighten them up, I thought about using large self tappers but I think they might vibrate loose so I went for channel, it's a pretty strong channnel as it's meant for supporting the floors of trucks but your right, it will twist more than box.

Vic

Vic66

11-01-2010, 08:43 AM

hi vic
looks like a very sturdy design, how long did it take you to draw up your plans?
could i ask was it solid works you used? as i am learning autocad at night school but would be struggling to draw that kind of design.:redface::confused:
what are you planning on cutting with it wood or alloy? i am no expert but i think your idea of two ball screws each side would be the best for alloy etc.
could you use one stepper and two toothed belts out to each side, just a thought!
hope you get started with the real thing soon.and keep use posted:smile:
:beer: Tom

Hi Tom

I use Autodesk Inventor 2010, very similar to Solidworks, making the model doesnt take that long it's thinking about the design thats time consuming and constantly changing stuff, it's difficult to know when to stop:rolleyes:

I still use ACAD but usually only to look at customer/supplier drawings that have been done in that format, I've been using Inventor professionally for about 4 1/2 years and really like it.

I want to cut wood with the router, model aircraft parts and general wood working bits and pieces, occasional light ally would be nice but I don't think I'd be doing much of that and don't think that this machine would be rigid enough.

I also think that the twin ballscrew or some sort of drive on both sides would be better but built this around my single ballscrew, hoping I can pick up some ideas about how to improve things before I start :smile:

Ross77

11-01-2010, 10:55 AM

Hi Vic
Cant find any pics at the mo. there here some where......The belts are basically mounted each side and conected will a bar, the same setup as if you where going to use belts for the main drive. The belts work indipendently of the ballscrew and just control the racking. If the ball screw is in the middle and your aiming for 800mm O/A width then it might be worth trying with out it (only 400mm each side)

If youve got ACAD Inventor why dont you you run a FEA test for stesses, Im trying to get my hands on that program so I can do the same. Ive got Vannilla cad at the momment.

irving2008

11-01-2010, 11:00 AM

... the problem with box section is bolting the steel angle to it, you cant get to the nuts to tighten them up...Think back to those old cars with box section chassis, how did they get round the problem? SImply drill/bore out a hole in the box opposite the nut so a long socket or socket+extension can be passed through the hole withthe nut held in the socket with a dab of grease.. If the rear of the box isnt accessible then cut a slot in the wall of the box sufficient to get a spanner and nut through. Third option is to create captive nuts by drilling/tapping out a piece of steel that is a sliding fit in the box section (only really works for narrow section). A fourth option, if the wall of the box section is thick enough, is to tap it directly. And a fifth option is to cut a rectanglular slot where the hole should be, slide in a tapped piece slightly narrower but longer than the hole and weld in place.

ptjw7uk

11-01-2010, 11:09 AM

Using box section why not bolt right through your only problem then will be collapsing the box with to much tightening!

Peter

irving2008

11-01-2010, 11:27 AM

Using box section why not bolt right through your only problem then will be collapsing the box with to much tightening!

Peterif you do that you never get the bolts to stay tight. There is a solution to this.. drill through the bolt size then get some tubing whose ID is the bolt size and open up the reverse side hole to the tubing OD. the length of the tube needs to be the inside dimension of the box plus one wall thickness. Then you can bolt through with a washer under the bolt head, the tubing is strong in compression compared to the box section and prevents it crumpling.

irving2008

11-01-2010, 11:31 AM

...I'm not certain what you mean about controlling the racking with belts or wire ? do you have any links to something that has used something like that ?

He means like this (http://www.cnczone.com/forums/showthread.php?t=51485)

Vic66

11-01-2010, 02:47 PM

Thanks for the suggestions, I had considered using counter holes in box section but theres an awfull lot of holes, and I want to keep it simple and as quick as possible to build.

I did run some stress tests when I went for the channel, cant remember the results but it didnt deflect to much, I've no idea what sort of forces are involved, I just put a force of 500N in various directions at the tip of the router, I will re run them and post the results.

If the channel twists to much I have thought about closing the open end with some folded ally channel as a work around.

Thanks for the link and suggestion to the cable anti-racking system, I like the look of that, it appears straightforward to do and could be done afterwards if I get problems so I think I'll stick with the single center ballscrew and use something like this if it becomes a problem.

Vic

irving2008

11-01-2010, 05:21 PM

Generally cuttting forces can be estimated as 5N (light wood), 20N hardwards/light alloys, 75+N cast iron, 100N+ steels, 150N+ stainless steel/hardened steels. A first approximation can be taken by using one of the many web-calculators for cutter spped/feed rates/cutting power based on a given material. Knowing the required power and revs and the cutter radius the (max) cutting force that can be achieved without stalling the cutter can be estimated from Force (N) = Power(W)*60/(cutter-radius(m)* 2pi * revs(rpm))

Vic66

12-01-2010, 09:10 AM

Irving, thanks for that, I am unfortunately just a basic steel and ally basher and maths is not a strong point, I would appreciate it if you could help me out with the calculation, I currently use a rotozip for a spindle, I want to change this as soon as I can afford to, I think this is about 800w 20k rpm, I imagine the largest diameter router bit I would use would be 10mm and the worst material would be MDF which I understand needs a fairly quick feedrate.

I didnt have much spare time last night but quickly ran a test applying a force of 500N in the Y axis direction at the tip of the router with it at its lowest point produced a deflection of about 2.5mm !! much more than I remember although I have changed things a bit since the last run, I think that the channels on the side of the machine will be ok, but I need to change the X axis channel to something more substantial, I think that the 500N force is possibly over the top it is a stab in the dark !!

irving2008

12-01-2010, 10:49 AM

800W rotozip, 60% efficient (guess but typical) = 480W cutting. Ity probably doesnt maintain 20krpm cutting, say 15krpm

then Power = torque * revs, torque = Power/revs where power in Watts, revs in radians/sec (= rpm*2pi/60)

so torque = 480/ (15000 * 6.28/60) = 0.3Nm... for a 10mm cutter (radius = .005m) max force therefore = 0.3/.005 = 60N although you'd never actually get to that in practice...

500N as a test force (how did you measure it BTW) is probably overkill!

Vic66

12-01-2010, 11:13 PM

Thanks again !

I have applied a force of 100N in the Y axis direction and X axis.

Visual representation is exagerated for clarity (I hope ! :rolleyes: ) Red area denotes maximum deflection, arrow direction of force.

I need to either replace the channel for the X axis or strengthen it somehow, I've no idea if .5mm deflection is acceptable on a router or am I being to fussy ?

irving2008

12-01-2010, 11:43 PM

0.5mm is OK if cutting wood as you'll never get close to 100N, 10 more like it... .05 would be a problem for metalwork. I'd still use box section, but i like to overengineer anyway...

ptjw7uk

13-01-2010, 11:12 AM

Does anyone know how much forces those v bearings are able to withstand as they do not look taht capable.

Good for taking thrust in one direction but not at 90degrees.
Just an observation on the mechanical loadings put on the bearings as all side load will be dependant on the fit of the ball races in the v bearings, to my mind the design lacks something.

Just my 2p worth.
Peter

irving2008

13-01-2010, 11:56 AM

Peter,
I agree, If this was box section I would put a couple of skate bearings at 90 degrees onto the side of the box to contain lateral forces. Or better still use pairs of v-bearings at 90 degrees onto the supported rail as per diagram... (obviously not to scale)

1461

Vic66

13-01-2010, 12:46 PM

According to Hepco, The dual V bearings I have (W3) are 5900N raidal and 1701N axial, literature here... http://www.hepcomotion.com/en/view-pg-21-view-6

I aquired 12 of them from work for nuppence they were left over from a job, and based their use on Joe's Hybrid design here... http://www.joescnc.com/themachines-hybrid.php

I was a bit skeptical about running them on steel angle rather than the ground V they are intended for but Joe seems to have had success with his design.

I originaly went for box, attached pics of the unfinshed box section model below, I could return to box section, I've been thinking about what Irving said in a previous post about cutting access holes for spanners etc, this would mean re thinking the position of the lead screw on the X axis though, which could make the X axis carridge more complicated.

I would rather not buy any supported rail, financially I have to run with what I have but I take on board the comments made about the side load on the bearings.

With the forces Irving calculated would this be an issue though ?

Vic

irving2008

13-01-2010, 01:43 PM

I see Vic, hadn't looked that closely. How thick is your steel angle and are you planning to prepare the edge in any way? The side load will translate into a vertical and horizontal force due to the shape of the vbearing. If they are rated for 1700N axially then I would have thought that was fine - I suspect the bearings would force the rails towards each other (or the studs apart) and pop-off the rails long before the bearing gave up!

Vic66

13-01-2010, 03:19 PM

Apart from cleaning them up I wasn't planning on preparing the edges, they havent in Joe's design, the V is running on the natural small radius of the edge of the steel angle, the V bearings can be run inside an angle which seems a better way but I could'nt think of a simple way to mount an angle in a suitable attitude.

Angle I've shown is 25mm x 3mm I would rather use 4mm or 5mm 20mm angle but that isnt a size that we stock at work.

Ross77

13-01-2010, 11:46 PM

Hi Vic
FEA looks good, I need that program.....looks so much quicker, dose it also generate a list for validating? How does it calculate the fixing type at the bearings or is it asuming a rigid connection?

From the analysis it looks like the z axis plate is the weakest as its only a 2d plate. Id mount some angle on it parallel to the spindle for strength.

For the C channels on the long axis you could flip them 180 deg so the web is on the out side and directly below the angle also it might be worth getting some unequal angle for the vee bearings, something with only 10 or 15mm upstand would be much better. Alternatley you could stick with the original config but add a plate to close the channel, this could extend up and bolt in to the angle as well to help brace them, two birds one stone etc.

This would also be a good idea for the x axis (?) as the open C channel is weak in torsion, need to move the ballscrew tho. As a final observation the vee bearings are only bolted to the end plates, it might be worth incorporating their mounts with the C channel uprights.

Anyway good luck.

Vic66

14-01-2010, 08:28 AM

To be honest I'm not that experienced with the stress analysis side of the software, I design specialist vehicles, mainly steel framework, sheet metal with some mechanical elements to it www.jsfraser.com (http://www.jsfraser.com) all the CAD stuff on the site is mine. We've only had the version of the software with FEA for a year now and so far I've just used it for testing framework bending over distance and seeing what sort of lightening holes I can get away with, so I'm not really qualified to answer your question, I have the disks though which will install as a 30 day trial version, you want a fairly decent PC though.

I keep having different ideas for my router, I settle on one idea get it modeled then think of something else, I've got to draw a line soon and go with what I've got or it's never going to get built ;-)

I know what you mean about the z axis plate, one of the other designs had angle along the side, this one has two 16mm supported rails bolted to the back of it, I bought them from a member on here, the bend on it is extremely slight, it's not to clear from the still pictures but when you can spin the model round you get a much better idea of what is bending and what isnt, its mostly the channel deflecting, I'm thinking about going back to box for this.

I've also thought about using steel box for the long axis, with 25mm angle tacked top and bottom with the ends of the angle laid on the box so it looks like a pyramid when viewed from the end, this would give the V bearings much more contact area and the angle is less likely to deflect, I've been trying to avoid welding because of distortion though.

Wobblybootie

14-01-2010, 09:35 AM

I know diddly squat about stresses torsion ... etc etc etc so if I am out of line here I'll shut up. The Mechmate Machine is a real beast and that runs a heavy steel Y axis Gantry on Hepco style V wheels (size 3) and they grind a V onto the top of the Angle iron leaving a 1mm flat surface on the top. There are a whole load of those beasts cutting 8' X 4' sheets and bigger in commercial operation

Vic66

14-01-2010, 09:42 AM

Hi Tim

The Joe's Hybrid machine seems to have quite a following with the V bearings (smaller than the ones I have) setup the same as my model, have you got a link to how it's done on the Mechmate ?

Vic

Wobblybootie

14-01-2010, 10:47 AM

Vic,

http://www.mechmate.com If you sign up to the forum (no charge) you have full access to all the drawings for a machine ... the Forums are a wealth of information but only about these machines ... There is a massive section on the pros and cons of V wheels and masses of info on grinding the angle iron ...

If I had the space for one of these, I would build it without hesitation. For a machine that can run all day machining 1" MDF with a huge router some of the stuff produced is stunning some very good fine detail 2.5D work etc.

Vic66

14-01-2010, 11:16 AM

Thanks Tim, I have seen that before but did'nt notice they were using V bearings, have registered will browse around later.

Like the Eddy avatar by the way !

Vic

Ross77

14-01-2010, 08:11 PM

I keep having different ideas for my router, I settle on one idea get it modeled then think of something else, I've got to draw a line soon and go with what I've got or it's never going to get built ;-)

Unfortunatly the best way to find out if it works is to just build it. if its only your time then I guess it dosnt matter and is part of the fun eh. :smile:

I've had a quick look at the hepco site and those bearings should be fine as when used in opposing pairs the double vee forms a 4 way equal loading. not sure about relying on the load capabilities if using the rounded edge of the angle as a rail tho. as the contact area is reduced to only 2 or 3 points.

I've also thought about using steel box for the long axis, with 25mm angle tacked top and bottom with the ends of the angle laid on the box so it looks like a pyramid when viewed from the end, this would give the V bearings much more contact area and the angle is less likely to deflect, I've been trying to avoid welding because of distortion though.

Using the corner of the angle would be better from a load point of veiw but I feel you may have problems aligning them, the reduced contact of the rounded angle maybe more forgiving. (test if you can).

An easier option IMHO is to use a rectangular box and rotate it, as the pic, No bolting requred and both corners will be parrallel, there is also space for the ballscrew. The box section is better at resisting the torsion from the z and y axis's.

I still think that just adding some closing plates to your original idea is the best. its still modular as you wanted and the smaller contact from the edge of the angle should make alignment and adjustment easier

I have the disks though which will install as a 30 day trial version, you want a fairly decent PC though.

Are you offering to loan them to me? I could probaly get most of what I want to do done in 30 days, Ive got a dell precision 470, 6.4 Ghz 3mb memory and nvidia quaddro 540 (I think) will that do?

Vic66

15-01-2010, 08:24 AM

Wobblybootie

15-01-2010, 08:31 AM

It may not be much use but Pacific Bearings have a 'square' bearing using hi tech plastic as the bearing surfaces, each side can be adjusted for clearance .. but I think these may have to run on steel box section ...

Ross77

15-01-2010, 07:23 PM

Thanks for taking the time to do the drawing, the closing plate idea was going to be my strengthening measure if the channel flexed to much. Using box section as a rail but rotated had occured to me, but I was put off by the radious on the corner of steel box section, Ally has a nice square corner but wouldnt last long with the bearing running on it, I'll try running a bearing along the edge of some steel box and see how the V matches the radius.

No probs with the drawing, probably took me less time than trying to explain in words.

Well at least you should have have enough options now.......looking forward to seeing some pic's

It may not be much use but Pacific Bearings have a 'square' bearing using hi tech plastic as the bearing surfaces, each side can be adjusted for clearance .. but I think these may have to run on steel box section ...

Not sure I would bother personally, looks like theres alot of surface contact so high friction and being fully enclosed means the rail cant be supported. I would also guess that the rail would have to be solid as box sections often have slightly concave faces.

Still i'm sure someone will prove me wrong :heehee: