Dave's Router

I have started building my router. The plan is that it will be a Kress (at least to start with) based 3-axis job with a cutting area of about 900mm x 600mm x ~220mm. I am attempting to over engineer so that it should be able to cope with aluminium, and be as flexible as possible.

The design involves two steppers and ballscrews on the X-axis so that I can unbolt some of the baseplate and put large objects (and flat objects on their side) into the router through a hole in the table.

The X and Y are on 20mm supported rail, and my plan is to use 20mm unsupported rail for the Z axis.

So far I have bought a scary amount of aluminium and other stuff and built some of it into a Y axis gantry - the photos are looking at the back of the Y axis
Attachment 1786 Attachment 1787

Looks nice dave, keep the pics coming mate :)

Hi Dave. It looks like you are building the exact machine I would like to. I'm a fair way behind you, as I still trying to work out the terminology. Are you following a plan or is it your own scratch plan?

It is very much out of my head, the plan tends to change quite radically up until I actually start cutting metal - it is more fun this way. And as soon as I actually see a part I immediately think of a better way of doing the next step.

I am not entirely sure of the stiffness of the 50mm square Al beam with the supported rail on top, but it probably isn't bad. I think a bigger problem is fighting it turning into a parallelogram which is why I made the plate which is sitting on the top in the photo. If I started again I would have made one of the Square Al beams 5 inches longer so I could put a good triangular cross bracing plate overhanging the end.

Progress is going to be slow for a couple of weeks, as it is the Science Festival here in Cambridge and I have many many other things I have to build for it.

After a long interval - due to having many other things to build (most of which could have done with a CNC router - but the timings didn't work), I am starting to think about getting back to it.

I was going to use 20mm unsupported round steel rails for the Z axis. I was going to attach the rails to a plank of 12mm aluminium at each end. But just having received the rod and looked at it, it seems pretty rigid, the aluminium is going to involve more machining and use up space (I can probably get the router 20mm closer to the Y axis without it.

So I was wondering if I could get away with just clamping the router support at one end of the rails, and a stepper motor support at the other end, and just use the rails and a ballscrew for the axis, with no scaffolding. It is going to be a bit less stiff but not hideously so as the rails were unsupported anyway, possibly a bit more prone to twisting, but this isn't going to move the router very much, and the rails have got 4 well spread out bearings.

I haven't seen a design like this so was wondering if anyone else has done this and have I failed to think of a critical bending mode?

How you getting on with this one dave?

It has been very quiet for a while - I do a lot of science outreach, and science week is a killer for me, and the subsequent month has been catching up on the stuff I should have been doing in the runup to it.

I now have the Z axis bits, and I have just spent the weekend playing a sliding puzzle game with racks of dexion shelving to get an extra rack in so I can try and absorb some of the pile of junk that was accumulating on the floor of the workshop. But at some point this week I should be able to get back to it. I think I might have to bite the bullet and start on the x axis... Starting maching big expensive pieces of ali is a bit scary.

Have finally managed to do something to the router, and set up the base. Though quite how drilling 16 holes in the profile took most of a weekend, I am not sure...

At some time soon I am going to have to build some form of flying motor support cantilevered out over the rails... It seemed like a good idea at the time

I was doing really well, I have managed to mount the Y axis on slides and it all seems to work well.... Then I decided to try turning a ballnut around on its ballscrew...

I got it off ok onto the nice piece of tube Zapp sends with it. I think I then made the mistake of trying to put it back on the motor end not the floating end...

result balls all over the floor which I am never going to be able to find. I really hope I can get some more... Any suggestions?

Hi,

Have you managed to find them or replace them yet, if not let me know.

I emailed gary from zapp and he said to ring him when he was back from his holiday. So that sounds positive. I have plenty to keep me occupied in the meantime.

Hmmm refilling ballnuts. Fun lol

I got some from ebay for mine. have you got enough to check the size?
I would suggest a tray and lots of caffine :smile:

Yeah I have about 3/4 of them and have managed to put 2/3 back in without too much difficulty. I will then have to try threading it back on the screw which might mean starting again... I guess there is nothing special about the balls, just their size is important.

Should they be lubricated? And how much space should there be?

I've only rebuilt one so I'm defiantly no expert. I had an old worn ballscrew so thought I'd give it a go. I don't think the balls are anything special except that they should be softer then the ball screw and nut so its the balls themselves that wear out. the ones I got where advertised as suitable for ballscrews so I just had to trust them..........

Reloading the balls isn't to bad . I assembled mine on the screw itself tho. i took off the cover and dropped each one down the same feed tube and then every 2 or 3 pushed them down and around the screw with a small bar. Eventually they appear at the other side of the return tube, that way you know there are no gaps. use plenty of oil. Setting the preload is the difficult bit and will require a box of different sizes to get it right. you might be lucky as yours is new and just be able to fill it with them all the same size.

NSK do a good guide for ball screws, (cant find it at the mo tho.) You will want to check the sizes as I believe some are set up so it is only ever other ball that is the correct size (drive ball) and the intermediate ones used to set the preload. This lowers the drag but also reduces the load capacity.

Like i said mine was basically scrap before I started so didn't have anything to lose. If in doubt wait for Gary....

Good luck

I seem to actually be getting somewhere after last weekend.

I have mounted the Y axis on the X-axis rails, and built a Z-axis minus a motor mount which will have to be bodged somehow.
Attachment 2127Attachment 2129Attachment 2128
The z axis ballscrews are binding slightly, I think because the plastic inserts (where the balls are returned) are getting squashed by the linear bearings when I do up the bolts on the ballscrew mount. It is fine when they are loose, so for the moment I might get away with some shim under the mounting block and moving the whole ballscrew out a touch.

I am getting to the point where I want to get the thing working, at least so I can use it to make more accurate replacement parts for itself...

Gary is apparently sending me some replacement balls, free of charge... thankyou!

If you look closely at all the unimportant bits which are just left with rough sawn edges it would possibly look less tidy. It is big and shiny though.

Right I think I have fixed the Z-axis binding problem - it was just not aligned right - rather than adding shim I shaved a couple of mm off the ballscrew support block and it is now good.

However it is running so smoothly compared to the Y axis, it rather indicates that there is something wrong with the Y- axis... I need weekends with less work in them.

It moves - have finally got it wired up and moving! I have even routed something - the emc default gcode, so not very imaginative.

It seems quite noisy though, should you be able to hear the steppers over the top of a Kress router going?

Next job is some limit switches and probably some kind of optical home switches, then I might be a bit less stressed when driving it around... Has anyone any experience in accurate home switches?

Hi Dave,

I can hear my MDF router over the kress, especially on rapids. When the vac's on too though I can't even hear myself think!

Looking good... What's the first project going to be?

Ok I will worry a bit less then, I am slightly worried that some of the ballscrews are not quite aligned, which probably isn't very good for them...

Most of the first few jobs are going to be building bits of itself - I need to make a baseplate for it, and now I have a way of drilling accurately positioned holes it seems stupid doing them by hand. I also want to drill lots of holes to tap in the plate. Which may as well be automated. I will also experiment with milling some pcbs.

After than, some science demos, probably some bits of the van de graaf I am building, after that there will be many others...

Have spent the weekend drilling and tapping a scary amount of M10 holes in half inch aluminium plates, for the base. In the process discovering various issues. The oldham couplings were slipping on the Z axis, so I added some grub screws which has largely fixed that problem. Plus the nuts holding the ballscrews into the bearing got loose... which was not ideal! So had to take it all apart and tighten them up. If it wasn't for this though my forearms would be significantly more painful from tapping...

I also discovered that one end of my X-axis ballscrew supports was 0.7mm further out than the other and the ballnut, which might have been making the couplings work harder than they should. There is a definite wobble to the screw now...

I have been hiding for a while, due to getting distracted by work and building other things. I have had a request for photos of how far it has now got so here are some...

I really ought to get back to it, but it will probably have to wait until after science week. I need to track down why it looses its X position sometimes, and I want to mount a webcam onto it to make zeroing easier..

I also need to work out how to make gcode more efficiently.