ok, first attempt using sketchup 8

[Jonathan]

I also can't fault your logic and argument on the cost side because that is very important for a home build.

Same here. Thought that it was obvious that you just make it as strong as you can afford. I made mine as strong as I could afford at the time, but not as well as I could and I've been waiting for the funds to improve it for some time.

One last thought:



You could weld a steel sheet (red) on to the back of the RSJ (blue). Then you've got the torsional stiffness of box section and the convenience of RSJ. It needn't be very thick, I think 3-4mm sheet would be fine. Adding several of the green triangular pieces might help, but it's a lot of effort to cut them out without a guillotine. Only problem with all this is the welding might distort the top surfaces, so it's best to do it from the start.
Other option is use box section but add plates on the top and bottom so that you can put the linear bearings off center to get the ballscrew in a good position. Not sure which would cost more...

The only thing that worries me slightly is the weight of the gantry, i know weight is a good thing, but would this be to much to get decent speeds. Now i'm not looking to break any speed records, just to get a reasonable feed etc. The motors i have at the moment are nema23-3NM and MSD542 drivers with 50v power supply.

Dean

[deannos]

Adding several of the green triangular pieces might help, but it's a lot of effort to cut them out without a guillotine.

I have a very nice angle grinder that would do the job.

Other option is use box section but add plates on the top and bottom so that you can put the linear bearings off center to get the ballscrew in a good position.

You mentioned earlier about the ballscrew being in a good position, i'm not to sure in what you mean. If you mean to bring it out closer to the Z plate, could i not use spacers on the bearings

[Jonathan]

You mentioned earlier about the ballscrew being in a good position, i'm not to sure in what you mean. If you mean to bring it out closer to the Z plate, could i not use spacers on the bearings

The best position for the Y-axis ballscrew is between the Y-axis rails so that there can be no movement of the linear bearings without the ballnut moving. More precisely the ballscrew should lie on the plane occupied by both the Y-axis rails, parallel to the rails. With the RSJ design it's close enough, with the previous design using two 80x40 box sections and 10mm plate it would be perfect.

(meant to post this in the previous post)

The only thing that worries me slightly is the weight of the gantry, i know weight is a good thing, but would this be to much to get decent speeds. Now i'm not looking to break any speed records, just to get a reasonable feed etc. The motors i have at the moment are nema23-3NM and MSD542 drivers with 50v power supply.

As long as you use RM1610 ballscrews to move the gantry it will still be fast due to the higher pitch and more importantly get better acceleration than RM1605. You'll want a 1:1 ratio with 10mm pitch, but I used pulleys anyway to reduce resonance. You can always swap the pulleys round to get a 1:2 ratio for better resolution. If you get the end machining for the ballscrew where the pulley mounts made 8mm, same as the motors, instead of 10mm then swapping the pulleys is simple. For instance use two 15T on Y and 30T on X you can swap them all round to get 15:30 if required.
Also I got the end machining for the pulley made 25mm long instead of 15mm as it makes it much easier to align and mount the pulleys.

RM1605 may still be fast enough, but when you're getting them from linearmotionbearings2008 on eBay I don't think there's much if any difference in price between the two so no point. Especially since you can use pulleys to get effectively the same pitch (if required, for the vast majority of things it wont be) as RM1605 with better v&a as the screw is only spinning half as fast.

[Swarfing]

Looks like the KISS approach is completely being over looked here? there looks like a complete ton of needless work for what will amount to very little gain in a home built machine. Welding all that extra in will no doubt make what will not be perfect in straightness even worse. Deanos if you want to make precision NASA approved telmetry parts go for. If you want to build parts that any motoring manufacture would be pleased with don't bother. The sort of numbers we have been looking at will not make a bit of difference to you i truly believe that. By not adding all this extra rubbish you will be keeping down the weight and expense, you can always make the parts with the machine once it is running.
KISS KISS KISS KISS

[Jonathan]

It's only a suggestion, take it or leave it. I post these things to benefit other people who may want more from the machine than Dean.
Adding the plate to make a box is very simple and should make a big difference but if Dean never wants to cut aluminium regularly there's no point. 0.01mm tool deflection might not sound like much, but when you start cutting metals it is - surface finish, material removal rate and tool life are affected. Running the calculation with cutting force for aluminium gives 0.058mm.

I used 8mm web thickness for the RSJ, not 6mm, in all the calculations as that's a standard size - with 6mm tool deflection is about twice as much.

[Swarfing]

Please for you

[JAZZCNC]

It's only a suggestion, take it or leave it. I post these things to benefit other people who may want more from the machine than Dean.

Great Jonathan but in a thread like this surely the target is to help the person who's building come to a design that suits his needs not others.? Thats how approach every thread.
First and foremost I try to advise with clear uncluttered information and suggestions to help them arrive at a design that will suit the intended purpose and do it the best it can. . . No more no less.!! . . . . If I can help them achieve this in a cheap and easy way then I most certainly do and get a warm fuzzy felling doing so:naughty:

Over complicating and over building past the usuall built in safety margin is just wasteful in both time and money.!

The simple FACT to DIYCNC is that it's NOT POSSIBLE to easily and cheaply build a machine that can cut every material like a purpose built machine optimised to cut the same material . . . . Belive me because I've tried very hard.!! You either build the machine to do the very best it can at that intended job at a sensible price and accept it's short falls in other area's. . . . OR . . . . Throw lots of serious money and time massively over build and have it perform in all areas average.!

Now back to the job in hand.

Dean for your MAIN need of cutting wood the RSJ will perform perfectly well without any extra bracing, welding or what ever. It will cut Ali but obviosly with restriction and greatly reduced performance compared to if it was built for that use.
Accept it's limitations and you'll love the machine which will become your faithfull servant but push it past them and you'll hate the bloody thing because it will try to kick you in the bollocks at every inopertune chance it gets.!!

Personally I'd use the RSJ but want to fill in the front and rear with thin covers, these would be serving 2 purposes with a frindge bennifit of adding strength.
Firstly functional, Secondly cosmetic, on the front they would be protecting the screw from debris with the slot idea on the previous design and on the rear it would be mainly cosmetic possibly houseing wires etc.!!. . . . but heres the fringe bennifit.?
To do this I would cut plywood braces that get bolted between the flanges and thru the web of the RSJ the covers would be just thin plywood screwed to the blocks. The bolted blocks would increase rigidity and combined with the plywood increase strength.

I've attached another pic to help explain.

Edit: Added coverd in pics for completeness.!

[JAZZCNC]

thats one nice no nonense design jazz,:tup: mind if i pinch it:naughty:

Nope go for it Chip all my ideas are some bugger else's anyway. . . Just Jazzzzz'd up. .:heehee:

[deannos]

Yes i'm going with the rsj as well. Its simple in its design and doesn't require any fancy tools. With its mass i believe this could just be bolted down on to a very sturdy table and save me messing about with a frame. I also think this would be easier to get square and level as well. I'm only interested in cutting wood not ali so the design suits me down to the ground. Jazz that design you have done in post 38 is excellent, looks like a million dollars but is cheap as chips:tup:. I'm going to base my design on that.

Cheers Dean

[JAZZCNC]

With its mass i believe this could just be bolted down on to a very sturdy table and save me messing about with a frame. I also think this would be easier to get square and level as well. I'm only interested in cutting wood not ali so the design suits me down to the ground.

Here's a bit of food for thought for ya.!!! . . . . Why not build 2 small concrete block walls with a plate securely bolted down onto it. Then with adjustment jack nut n bolts thru both to level the RSJ's. Then you could just push a strong bench under neath fasten down then surface it flat and away you go with virtually no material height restriction.
Obviosly you will have Z axis travel restriction but not material height.!!. . . The machine use's and flexibilty regards material sizing become far greater.!!. . . . Imagine some one wanting an old table surfacing or V carving. . . . how easy does that become.? . Ye I know not probable but you get the idea.
The same plate/jack system would work good on a bench thou.

Some time towards the end of the year I'll be building just such a machine that use's R&P and will be used for cutting granite slabs 3x2mtr into work surfaces. This will use the same principle but have a free standing concrete table with a hydrolic tilting frame for loading slabs from near vertical to horizontal.