Design Help Pt2 Required for CNC design/Build

[MikeyC38]

Hi All,

Firstly many thanks to Jonathan and John S for their kind help and advice, for HiltonSteve for his inspired build log and drawings and all you others who have contributed build logs, posts and to routercnc for the fantastic spreadsheets!!. I have now bought my ballscrews, linear rails and bearings from Chai on Ebay. They turned up on time, and everything is straight and not bent. A really nice guy and fantastic value. Thanks John S for that one and the source for the stepper drivers.

Attached are various CAD diagrams for which I would appreciate your comments. You know the old adage, measure twice cut once! The ballscrew sizes are X - 1100mm, Y - 750mm and Z 300mm with 20mm supported round rails on the X and 16mm everywhere else. I plan on using Nema 23 steppers at 3.1Nm with 36V ps at 11A. The main purpose is to (mill or rout John S -:) )wood for guitar bodies , but also have the capability to do light 6060 aluminum milling. The extrusions are 40x80 (heavy 4kg/m) sections with 15mm toolplate for the gantry sizes and bed ends. The Y axis is 120x40 (heavy 8kg/m) extrusion. Z axis is 10mm toolplate. Router will be Kress 1050FME-1 with a planned upgrade to one of the 2.2Kw VFD Chineese spindles once I've got more funds. The drawings don't show the ballscrews, steppers, router mounting plate or the plate to join the gantry sides as yet - after all it is Christmas! but they will be in the ususal positions.

Comments please?

Kind Regards
Mike

[JAZZCNC]

I can make a few suggestions that you might consider having built a few machines very similiar for the same job and actually have one part built which I just don't seem to find time to finish.! I'll attach pics of it.

Looking at the drawings I presume you intended using twin screws.? If so then I would put a cross brace between the gantrys sides running under the bed, this will help make the gantry more ridged but it also then gives you the possibilty to run the ballscrews under the bed out off harmsway rather than exposed on the outside it also offers some protection from dust and rubbish.
On the machine I build the motors under there as well connected to the screws with belts. Other than the protection advantage the belts also allow geariing if needed and reduce the chance of resonance, but the main reason in my design was to keep the foot print as small as possible because the original machine was to live in a typical 8 x4 garden shed/workshop so space was precious. As you will see from the pics the Y axis motor/screw setup was inverted the same and the Z axis motor flipped to keep the height down so it could sit on a bench without hitting the roof, there are no components sticking out side the initial footprint.
I've built several versions both with single screws and twin screws, some with belts as mentioned, some with direct drive and motors on outside when space saving doesn't matter as it does ease the build cost slighlty slightly. Thats why some pics will look differant to others.

Regards the largish kink'd back gantry side, thou not a problem as such it's more material related.! . . Depending on the material width you intend to use you may want to check that you can actually get that much kink out the width's.? Unless you intend to use one large and expensive sheet then it could be tight using STD width's from the likes of Alu warehouse etc.? Most stop at 300mm width.!

Regards the Z axis I would run the rails on the Front plate and put the bearings on the backplate, this will help strengthen the front plate slightly. I would also recommend you at least beef up the Z axis with min 19mm plate. Remember it doesn't matter how well the rest of the machine is built if the Z axis flexs then accurecy goes out the window.!

Hope this helps and happy building.

[MikeyC38]

Thanks Jazzcnc for taking the time to review this. The photos are really useful. My drawings are incomplete at this stage as I will be using a single ballscrew on the long axis as per your photo with a plate underneath between the gantry sides . But the tip on adding the extra reinforcement on the long axis is taken and accepted. HiltonSteve's router/mill also had this. It also solves the problem of mounting the BF/BK blocks very elegantly and protects the ballscrew from muck! Many thanks!!

Again the comment on the material thickness on the Z axis is accepted and you given me some food for thought on the gantry side plate lay back. My original idea is to get the cg of the Z axis between the bearings on the side plate, but as you say well, cost of getting the plate for the gantry sides has to be taken into account!

I note your comment about reversing the position of the Z axis linear slides and bearings - I see in your photos you are using Hwin/THK slides which have better torque load characteristics than Supported round rail and linear slides. I'll have a think about that one.

Measure twice cut once is the adage!!!

Kind Regards
Mike

[John S]

You can buy plate cut to size in any size, not limited to a specific width only from certain suppliers.
If the design is though about you could nest the two sides like spoons and cut wastage to a minimum.

Worth giving these people a call.

http://www.ascmetalsgroup.com/northampton.html

[JAZZCNC]

I spent some time experimenting with cutting MDF templates for the best layback and what you see in the pics is the outcome. I found it gives the best balance and can still be got from 12" x 3/4" plate from Alu warehouse with the minimum waste. Thou not exactly on centre of bearings it's pritty close.

Regards the Z axis I really would consider putting the rails on the front.! It does help regardsless of rail type used.
Remember the Z axis when extended is basicly a lever and a weak lever will bend and flex.!!
Like I allways tell folks.!!. . The machine is only has good as it's weakist part or component.!!. . . Unfortunatly I see lots make the mistake of building weak or inaccurate Z axis's. Which makes the rest of the machine no matter how well built or quality components used pointless because at the end of the day it's the Z axis which is at the sharp end taking the strain.

[JAZZCNC]

You can buy plate cut to size in any size, not limited to a specific width only from certain suppliers.
If the design is though about you could nest the two sides like spoons and cut wastage to a minimum.

Worth giving these people a call.

http://www.ascmetalsgroup.com/northampton.html

Agree John but to be honest I've always found it works out cheaper for small Qty's buying from AW. It seems unless your buying from them on a regular basis as business in decent Qty they load up the price.?

Mike. . . Something I meant to say and partly touched upon before with my material comment and I'm sure others will agree or have found out them selfs while building.?
You have to be very carefull in the design stages that you don't plan on using or design something that hisn't easily available. It's often the simple things that can be a right pain in the arse to source or may not even be available in the intended size or length. IE Timing belts lengths often catch folks out.

[MikeyC38]

Regards the Z axis I really would consider putting the rails on the front.! It does help regardsless of rail type used.
Remember the Z axis when extended is basicly a lever and a weak lever will bend and flex.!!
Like I allways tell folks.!!. . The machine is only has good as it's weakist part or component.!!. . . Unfortunatly I see lots make the mistake of building weak or inaccurate Z axis's. Which makes the rest of the machine no matter how well built or quality components used pointless because at the end of the day it's the Z axis which is at the sharp end taking the strain.

Good reasoning there JazzCNC...the placing the supported round rails on the moving part of the Z axis will add stiffness to that axis...

[Jonathan]

About the Z-axis... It is foolish to do anything other than what Jazz has suggested, namely putting the rails on the front plate. With the rails on the front plate the distance from the tool to the support (i.e. the bearings) is some constant (overhang of spindle plus length of tool) plus how far down the Z-axis is. So the minimum is the constant, and the maximum is the travel plus that constant. Whereas with rails the other way round the overhang is always the same constant plus the travel - i.e. it's always at the worst case scenario for the other orientation!

I don't really see the point of going above 20mm on the Z-axis, the gain is absolutely minuscule compared to the added cost (especially when you're using just 16mm round rails) as I'm sure you'll have found from routercnc's spreadsheet. I used 20mm for this reason.

Aluminium warehouse sell plate cut to size which is a good choice, as unlike the flat bar, it should be flat! It's a little more expensive.

http://www.aluminiumwarehouse.co.uk/cutting_calc.php

However, I urge you to consider not having gantry sides and using the style, like mine and Jazz's bigger machine where the cutter operates below the level of the X-axis rails. That eliminates the flexing of the gantry sides, as there aren't any, and puts the X-axis ballnut(s) in a much more optimal location in terms of the forces they receive.

Y-axis of 750mm sounds a bit much for just one ballscrew on X, especially considering you want to cut aluminium.

[John S]

Agree John but to be honest I've always found it works out cheaper for small Qty's buying from AW. It seems unless your buying from them on a regular basis as business in decent Qty they load up the price.?



Jazz,
To be honest I don't buy large quantities but buy just what I need unless it's small section and then it doesn't pay to get a bar cut.
I have to rely on their delivery service as my nearest branch is Lincoln, delivery is a reasonable £7.50 no matter what but their rep tells me that if I'm ever in the area to call in and take advantage of offcuts they have lying around.

Not going to happen with me but others who are not so tied to a workshop all working hours could score on this.

[JAZZCNC]

However, I urge you to consider not having gantry sides and using the style, like mine and Jazz's bigger machine where the cutter operates below the level of the X-axis rails. That eliminates the flexing of the gantry sides, as there aren't any, and puts the X-axis ballnut(s) in a much more optimal location in terms of the forces they receive.

Y-axis of 750mm sounds a bit much for just one ballscrew on X, especially considering you want to cut aluminium.

Seen as Jonathan agree'd with me I'll agree with him. . Lol . . . . Thou Ive actually built the same type machine as shown, it was with the specific intension of cutting just Wood which it does very very well and not to disagree with jonathan it also cutts Ali ok.
That said If building an allround machine then I'd go with what Jonathan says and actually I will be building just such a machine over the next few months ounce I've got other commitments out the way.
I've attached the first ruff models thou it may change in dimensions before it's built with ballscrews covered and rear of gantry enclosed along with Cable chain etc. The bed will be adjustable because a forth axis will be needed at some point also allows for milling vises etc.

Edit: Oh by the way It's going to sit on a bench not floor mounted.!

[MikeyC38]

Hi John S & JazzCNC

You guys have got me going now and its late (got work tomorrow :cry: ) but I have revised the drawings with your suggestions in mind... its begiining to take shape and make sense.

Thanks guys

Mike

[wiatroda]

Worth giving these people a call.
http://www.ascmetalsgroup.com/northampton.html

But please check price first.
I bought 500mm long 2" aluminium round bar from ASC wolverhampton branch and paid £34 for it. AW price £62 for 2000mm long. So much overpriced. As an addition it was the worse piece of metal I've ever had .Totally un-machinable. Still have this piece of c.... in a garage.
It was first and last time I bought from them. :cry: Very disappointed:thumbdown:

[JAZZCNC]

As an addition it was the worse piece of metal I've ever had .Totally un-machinable. Still have this piece of c.... in a garage.
It was first and last time I bought from them. :cry: Very disappointed:thumbdown:

It wouldn't be sat in my Garage.!. . . It would have been bounced back at them. . . . Like I tell my customers If you don't tell me about problems I can't fix them.!
Thou I prefer customers don't have cause to complain in the first place I much prefer customers to complain in a polite manner rather than not so I can give a replacement has well as highlighting any potential problems which then get corrected.

[wiatroda]

[QUOTE=JAZZCNC;26511 Like I tell my customers If you don't tell me about problems I can't fix them.!
Thou I prefer customers don't have cause to complain in the first place I much prefer customers to complain in a polite manner rather than not so I can give a replacement has well as highlighting any potential problems which then get corrected.[/QUOTE]

You're absolutely right. Like your approach.
The fact is I, as a customer agreed to pay this price (mostly bcos of unawareness of market prices) so there is no reason for price promise return. I used a bit of it, turning it first. If I wanted to knit aluminium jumper swarf would be perfect for this purpose.
Then I broke few 2mm drill bits trying to drill it -different speeds, feeds various lubricants then gave up. When making my timing pulley video (drilling way) I used free-cuting ally- no probs at all.
Anyway I'll keep it in any case- may be useful some day.
Next time I'll check prices first before making any hasty shopping. Just didn't expected price to be so sky-high-different than from other suppliers.

[JAZZCNC]

Next time I'll check prices first before making any hasty shopping. Just didn't expected price to be so sky-high-different than from other suppliers.

Ye I know it really feels like a kickin the nuts don't it when you find it much cheaper else where.! . . . On this point when it's so blatantly over priced I agree with not going back. . . we all need profit to survive but then there's profitering pirecy.!! . . . .That's unacceptable to me too. . . .I was mainly referring to the bad quality of the material.

Like I said to John-S I've found with larger metal suppliers like ASC that if your not a company or don't buy on regular basis in decent Qty's they load the price.!
Fortunatly I am Vat registered and thou my main business hisn't engineering related I do get away with out price loading for awhile but because I don't buy often it soon starts to creep up.!! . . . I don't mind paying slightly higher prices for smaller qty's or tariff Qty based charging but it should stay constant at least.!. . . This piss's me off.

[MikeyC38]

Thanks Jonathan for the additional comments and feedback. More than 1 person saying the same thing is always encouraging!
I am pretty much committed down the existing gantry route because I already have the ballscrews in hand and can't afford to buy anymore at the moment. My earliests thoughts on this was a metal version of the twin ballscrew- belt drive Solsylva design. I accept that the machine may have some limitations with aluminium milling but as this is a tertiary requirement, then I can live with that compromise for now. Wood, GRP, Plastic and CFRP are the priority at this time, hence the size of the machine in the Y axis. I am looking to source aluminium toolplate (6082 T6) and have found a reasonably priced source in Milton Keynes (MetalFast) but that was awhile ago so I will need to recheck.

Kind Regards
Mike

[Jonathan]

I am looking to source aluminium toolplate (6082 T6) and have found a reasonably priced source in Milton Keynes (MetalFast)

Aluminium warehouse has tooling plate which may be cheaper? It's on the website in the same place I linked to earlier. They call it EcoCast.

[JAZZCNC]

I am looking to source aluminium toolplate (6082 T6) and have found a reasonably priced source in Milton Keynes (MetalFast) but that was awhile ago so I will need to recheck.

AW is good like Jonathan says but you really only need tool plate quality in just a few key places, like the Z Axis front/rear plates or for the Bed etc.

Things like gantry sides and general end plates or bearing brackets, motor mounts etc can be got from cheaper flat bar stock, still 6082 T6 just not machined perfectly flat.
The wider width's of flat plate stock can some times come slightly cupped but for none critical stuff or if machining several parts or brackets out of the width then it's not really a problem. . . Often just a quick surface will take away any cup. . . .Thou I must say the last few 300mm wide lengths I've had to deal with have been quite bad. The one I'm using at the moment is particulerly bad and if it wasn't for the fact I was cutting it up into smaller components then it would have been going back.!!

[JAZZCNC]

Mikey now Knowing a bit more about your plans to use a single screw and looking at your design again then I think you may want to reconsider the width of the gantry and make the bearing spacing wider.! . . .That looks awfull narrow for a single ballscrew design making racking a big possibilty.? . . .How wide is it.?

[MikeyC38]

Mikey now Knowing a bit more about your plans to use a single screw and looking at your design again then I think you may want to reconsider the width of the gantry and make the bearing spacing wider.! . . .That looks awfull narrow for a single ballscrew design making racking a big possibilty.? . . .How wide is it.?

It is currently 125mm. I have updated the Z Axis design as you and Jonathan suggested and I was coming to the same conclusion as well. So the gantry sides are up for redesign. Perhaps 175 - 200mm width and up the material thickness to 20mm should do the trick. Eliminating the 'S' shape to make manufacture easier. My thoughts are that the router tool bit should ideally lie between the gantry bearings to reduce racking. Is this true? In the current design it does not (see side shot).

Thanks for the tip on using the toolplate in the accuracy-critical areas - should save costs

RegardsMike

[wiatroda]

I think you may want to reconsider the width of the gantry and make the bearing spacing wider.! . . .That looks awfull narrow for a single ballscrew design making racking a big possibilty.?

Everything what Jazz said plus I would insist two ballscrews. It is worth to invest a bit more ££ and have machine free from any distortion/ racking. You may regret one ballscrew some day.
One stepper per ballscrew, perfectly one stepper with its own driver but I run my machines with one shared driver per two parallel steppers on X axis and never had any problems with this at all.

[Jonathan]

You may regret one ballscrew some day.

No, you will regret one ballscrew some day. So many times people start off saying they'll use one ballscrew and a couple of weeks later we manage to persuade them to get another!
Just imagine what happens when you apply a cutting force with the spindle near the gantry side (i.e. at limits of Y-axis travel). It's going to be at least 300mm from the support (the ballnut) so you've got a large turning moment which is only resisted by the stiffness of the rails. Combined with having the X-axis bearing blocks close together the deflection will be quite a lot, especially since the fixing of the supported rails to the rails is relatively weak. A 0.2mm deflection (for instance) may not sound like much, but that will affect the finish you obtain with wood and stop you machining metals.

Two ballscrews also lets you use the design I suggested earlier, without gantry sides, which saves a bit on materials I reckon and is definitely much stronger.


I run my machines with one shared driver per two parallel steppers on X axis and never had any problems with this at all.

Ooh that's interesting / unexpected. Are the steppers mechanically linked (timing belt or whatever)? What steppers and drivers are you using?

[wiatroda]

Ooh that's interesting / unexpected. Are the steppers mechanically linked (timing belt or whatever)? What steppers and drivers are you using?

Now it does sounds like one stepper is shared between two machines. Does it?? or it just me. One stepper per ballscrew per machine on X axis. I have 2 CNC machines and each one have it's own set of motors and drivers. One is for "wet" another is for "dry" job. I use PM752 from zapp. Motors on X axis are not linked together by any mechanical way apart of Yaxis on my wet machine- 2 steppers, one driver common timing belt. Please see my post no.23 (http://www.mycncuk.com/forums/showthread.php/3192-Few-pics-of-my-homemade-CNC-mill/page3)

[JAZZCNC]

Mike I agree with the others completely that 2 ballscrews is much much better and prefered. . .BUT . . . I do know from experience with pritty much the same size machine it will handle wood, even hard woods no problem. Yes you won't be able to be has aggressive as if you had twin screws regards DOC etc but it will do the job ok. It will cut Ali but very light duty with shallow cuts.
You will need to widen the distance between bearings, more the better, and make the gantry as stiff as possible. The gantry width on the pics I've shown is 220mm with 240mm of actual bearing spread because it use's profiled rails not round and the bearings are slightly longer than the mounting pads so effectively 240mm wide gantry.

If you can stretch to twin screws then it's a no brainer just do it.!. . . but if not then I recommend you design the end plates for twin screw upgrade in future. The machine I showed in the pics is done this way.
The pics don't show it.!. . . but because I use Timing belts to keep the motors on the inside, the end plates are drilled at either side on the inside to accept twin screw setup and can be upgraded at any time.
It's a simple case of removing the existing single screw remounting the BK/BF blocks on the inside of the outer profiles and mounting the motor bracket setup in it's new locations on the end plate. Obviuosly then you add the extra screw along with a extra motor brackets etc. It also means a new ballnut mount for both sides so the existing one is not re-used but that's no big deal or great expense. All in all it takes 2 to 3 hours to change from single to twin screws.

In your case with direct mount motors just have the end plate machined at either side to accept the motors then cover with blanking plates untill ready, then just swap when ready and cover single screw mount with one of the blanking plates.!! . . .Simplizzzzz.

[MikeyC38]

Mike I agree with the others completely that 2 ballscrews is much much better and prefered. . .BUT . . . I do know from experience with pritty much the same size machine it will handle wood, even hard woods no problem. Yes you won't be able to be has aggressive as if you had twin screws regards DOC etc but it will do the job ok. It will cut Ali but very light duty with shallow cuts.
You will need to widen the distance between bearings, more the better, and make the gantry as stiff as possible. The gantry width on the pics I've shown is 220mm with 240mm of actual bearing spread because it use's profiled rails not round and the bearings are slightly longer than the mounting pads so effectively 240mm wide gantry.

If you can stretch to twin screws then it's a no brainer just do it.!. . . but if not then I recommend you design the end plates for twin screw upgrade in future. The machine I showed in the pics is done this way. . . .Simplizzzzz.

JazzCNC - this is a brillant suggestion :tup: !!! Improve what I have but design in an upgrade path so that the existing investment is not totally lost!!!. A question springs to mind for everyone - in a twin ballscrew setup with 2 motors on the X-Axis (and I presume 2 stepper drivers) how do you ensure equal power output and speed to avoid crabbing the gantry? The Solsylva solution uses a single motor and belts so you don't get the problem.

Now that you guys have planted the seed of doubt in my mind :confused:, I will work up a twin screw solution as you've all suggested :redface: and do the numbers to see if my budget and timelines can take it !!!

Thanks All for your patience and interest

Mike

[MikeyC38]

No, you will regret one ballscrew some day. So many times people start off saying they'll use one ballscrew and a couple of weeks later we manage to persuade them to get another!
Just imagine what happens when you apply a cutting force with the spindle near the gantry side (i.e. at limits of Y-axis travel). It's going to be at least 300mm from the support (the ballnut) so you've got a large turning moment which is only resisted by the stiffness of the rails. Combined with having the X-axis bearing blocks close together the deflection will be quite a lot, especially since the fixing of the supported rails to the rails is relatively weak. A 0.2mm deflection (for instance) may not sound like much, but that will affect the finish you obtain with wood and stop you machining metals.

Two ballscrews also lets you use the design I suggested earlier, without gantry sides, which saves a bit on materials I reckon and is definitely much stronger.



Ooh that's interesting / unexpected. Are the steppers mechanically linked (timing belt or whatever)? What steppers and drivers are you using?

Well Jonathan, I think you managed to plant the seed of doubt in my mind :confused:...:wink:

[m.marino]

Mike,

Jazz and I designed a single ball screw machine and when it all came down to it the cost for the pulleys and belts where actually more then going the twin screw route. Now I have to get a second parallel port card to allow me to bring the 4th axis on line and set up a second BOB and driver set up. But outside of materials this is not really a problem when using Mach3. Good luck and hope all things go well.

Michael

[routercnc]

Hi Mike
Good to hear someone getting use out of my spreadsheet. Nice to put something back into this forum.
Just a comment on the x axis bearings. When mounted on their side it does make for a neat installation but be aware that because the blocks are open (sort of C shaped) that they will open up slightly due to the gantry mass and to a less extent the cutting forces. I've not done the calcs but suspect the effect is small so don't panic. Interested to hear if anyone has worked out how much they deflect for a given side load (or vertical in this case) or has experience of this orientation.

As for twin screw - I have a gantry about 600mm wide with a single ball screw and I can rock the ends of the gantry back and forth by hand up to about 1mm if I push hard. Having said that I can happily machine balsa, liteply and ply and they come out at the right sizes. But I would still say twin screw is better and would like to upgrade when funds allow because I could then up the feedrate.

[MikeyC38]

Hi Mike
Good to hear someone getting use out of my spreadsheet. Nice to put something back into this forum.
Just a comment on the x axis bearings. When mounted on their side it does make for a neat installation but be aware that because the blocks are open (sort of C shaped) that they will open up slightly due to the gantry mass and to a less extent the cutting forces. I've not done the calcs but suspect the effect is small so don't panic. Interested to hear if anyone has worked out how much they deflect for a given side load (or vertical in this case) or has experience of this orientation.

As for twin screw - I have a gantry about 600mm wide with a single ball screw and I can rock the ends of the gantry back and forth by hand up to about 1mm if I push hard. Having said that I can happily machine balsa, liteply and ply and they come out at the right sizes. But I would still say twin screw is better and would like to upgrade when funds allow because I could then up the feedrate.

Thanks routercnc for the spreadsheet and the comment re the X axis bearing placement. I've seen them placed like this on another design, but somehow felt that vertical placement would be better for carring the vertical weight of the gantry and resisting racking. For the gantry I will be using 20mm plate sides with 120x40mm heavy extrusion (8.8 kg/m) for the y-axis on the basis that the box section resists twisting from the cutting loads better than plate and is lighter, reducing the forces required to accelerate and stop the gantry.

BTW I too am a model flyer (Gliders and power) and have in mind to build a 1:3.5 scale (4.25m) "Reheinland" glider with all the bits CNC cut - once I have done my brother's guitars!

Kind Regards
Mike

[Jonathan]

I've not done the calcs but suspect the effect is small so don't panic. Interested to hear if anyone has worked out how much they deflect for a given side load (or vertical in this case) or has experience of this orientation

It's going to be tricky to calculate as it relies on the strength of the bolted joint between the round rail and the support. The aluminium surface at the joint is a V-shape, so the rail only makes two line contacts. There's not many bolts (and if you don't threadlock them they can work loose) which overall means the round bar can, and do, bend in that direction upon the rail. Check it with a dial indicator and you should be able to see the effect.


For the gantry I will be using 20mm plate sides with 120x40mm heavy extrusion (8.8 kg/m) for the y-axis on the basis that the box section resists twisting from the cutting loads better than plate and is lighter, reducing the forces required to accelerate and stop the gantry.

See you mention making the frame very strong, which is great, but with only one ballscrew you're kind of defeating the object of doing that. Not saying don't make the frame strong, as you can't change that later (without wasting material/money) unlike adding an additional ballscrew which is obviously easy enough.


BTW I too am a model flyer (Gliders and power) and have in mind to build a 1:3.5 scale (4.25m) "Reheinland" glider with all the bits CNC cut

I bought a 500 size helicopter recently, so I'm also almost a model flyer. Not dared actually fly it yet! Couple of bits to cut from carbon on my router first. So far I have just stuck to model cars.

[JAZZCNC]

A question springs to mind for everyone - in a twin ballscrew setup with 2 motors on the X-Axis (and I presume 2 stepper drivers) how do you ensure equal power output and speed to avoid crabbing the gantry? The Solsylva solution uses a single motor and belts so you don't get the problem.


The control software IE: Mach3 handles this side and so long as you don't over tune the motors then it's not usually a problem.

That said PERSONALY I don't like slaving motors for several reasons.!! . . . Main one being that when it does go wrong, which more often than not is at high feeds, the damage thats done can be massive. Imagine the one of the motors stalling coming to a dead stop while the other keeps on going.!! . . .It's like a train wreck and scarey shit, esp on larger machines.!!
To be sure to avoid this you have to effectively de-tune and run slightly lower acceleration/velocity than you can using belts.

There are other reasons like keeping the gantry absolutely square is fiddly to setup and requires good quality accurate home switch's to be absolutely sure it's square every time you home.
Basicly What your doing with the switchs is setting the start point or zero position for each screw and by positioning the switchs you set the gantry square. . . .This is less than ideal to me.!! . . . Mainly because every time you home the machine the gantry gets twisted and co-jolled into square by touching one switch then the other potentially putting strain on the machine.
If you don't use home switch's to square then you can't be absolutly sure the gantry is square and if the motors start dropping the odd step here and there, which can and does happen if you push the machine too hard or over tune motors or it's not perfectly aliagned or setup causing binding.
Basicly it slowly loses position putting the screws out of sync and the gantry out of square.! . . . or if you stop the machine abruptly from high feed one of the motors can slightly run on more than the other due to inertia and position is instantly lost with no easy way to get back in position. . . . With belts this never happens.

Like I say it's a PERSONAL thing and there are lots that use slaved motors and are very happy with them. I've built machines that use slaved motors without any problems.!
It's just MY PERSONAL PREFERENCE to use belts on my own machine because I feel it's safer and more accurate and ounce setup it's pritty much done forgot and never changes.
Yes belts can snap but I've only snapped 1 in 4yrs and that was my fault not belt failure.!

(I'LL SAY IT AGAIN TO STOP ALL YOU SLAVED MOTOR USERS JUMPING ON ME.. . . . . IT'S MY PERSONAL PREFERENCE )

[routercnc]

Hi Mike,
Good to see a fellow flyer on the forum. I'm sure you know this already but Chris Williams writes a good 'big glider' column in QEFI magazine. Once you cnc cut parts from CAD designs there's no going back to carbon paper and all those hand methods!

Hi Jonathan,
Good point you make about the rail itself deflecting - I was think only of the block itself which is of course open to allow the supported rail underneath. That makes things more complicated and either needs FE, or someone to do a quick test!
If you've never flown a model heli before then join a club and get trained because otherwise you will crash it in seconds! Lots of PC simulators out there which are worth a go because you can often connect your own Transmitter (using the buddy lead).

[MikeyC38]

Thanks everyone for the very kind and useful advice. Here are the revised plans with most of the stuff hopefully taken into account:

1. Revised Z axis with bearing blocks fixed to the Z plate against the gantry
2. Gantry sides beefed up to 200mm width to reduce racking . 20mm thickness
3. More detail added in drawings

The endplates will be extended to allow for a future upgrade to twin ballscrews as everyone is suggesting! and belt driven to avoid synchronisation issues as discussed earlier.

Thanks for the guidance on sources of aluminium plate.

For the eagle-eyed viewers you may notice that the y-axis bearings are floating... this is not deliberate but something I realised that a 750mm ballscrew is the total length, not the threadlength and as I want to bolt the BF/BK blocks to the gantry sides then I will have to reduce the length of the y-axis round rail and support. Question here - is it best to dismantle the supported round rail then cut the rail and support seperately and re-assemble? or try to cut it all in one go?

I'm really getting into the 3D solid CAD in a big way - it really does help you work out how things need to fit together and how on a 300mm ball screw, the maximum travel you can get on the Z axis using supported round rails would be no more than 150mm at best!

Thanks again everyone!

[JAZZCNC]

Question here - is it best to dismantle the supported round rail then cut the rail and support seperately and re-assemble? or try to cut it all in one go?


Cut it in one go. . . . I do it by first cutting the round rail using 1mm thin disc in the angle grinder, cuts like hot knife thru butter then finish Alu base using mechanical hacksaw but hand held will work just same.

[Jonathan]

The distance measured parallel to the Y-axis between the spindle and the Y-axis ballnut/screw is currently quite large. This will cause the cutter to deflect parallel to Y as the rails offer little support in that direction. Between the Y-axis rails is a good place, hence why a lot of people mount them on separate pieces of extrusion.

[routercnc]

The vertical plate on the back of the Z axis (to which the Y axis ballscrew is attached) is trapped between the small top and bottom plates. This means it needs to be machined accurately to ensure the Y axis bearings sit correctly. It would be better to make this plate full height and have narrower top and bottom plates so that everything can be set before tightening the bolts.

[MikeyC38]

Thanks Jazzcnc, Jonathan and routercnc for the comments on the Y and Z axis. Re-design in process to include using smaller seperate extrusions to reduce
alignment issues and enable the Y axis ballscrew and nut to be located under the Y axis rails and reduce the distance to the Z axis assembly.

Talking about the alignment of the axes for a CNC, I had the idea to buy a large thick piece of float glass (10mm x 1500 x 1000) to provide my "flat" surface from
which I could then make accurate measurements using a dial micrometer. By measuring the surface first, one could reasonably then account for the variations in the surface when measuring...

Regards
Mike

[MikeyC38]

Hi Guys

Been busy at work but have re-worked the Y-axis gantry design, taking into account all your comments. Have shifted the Y axis ballscrew closer to the Z axis gantry and used seperate extrusions to allow this and maintain torsional stiffness I think. I guess the main issue is the use of supported round rails - profiled rails would reduce the height and thickness. However given my other constraints of time, budget etc..lets see. The other approach is to use the extrusions in a L-shape as per Jazzcnc's photos. This has the advantage of mounting the BF and BK blocks on the Y-axis and not having to cut the rails. Will probably work up that design as well and make a decision. Now I just need some more money for the ali, steppers and electronics!

Regards
Mike

[MikeyC38]

Hi Guys

Been busy at work but have re-worked the Y-axis gantry design, taking into account all your comments. Have shifted the Y axis ballscrew closer to the Z axis gantry and used seperate extrusions to allow this and maintain torsional stiffness I think. I guess the main issue is the use of supported round rails - profiled rails would reduce the height and thickness. However given my other constraints of time, budget etc..lets see. The other approach is to use the extrusions in a L-shape as per Jazzcnc's photos. This has the advantage of mounting the BF and BK blocks on the Y-axis and not having to cut the rails. Will probably work up that design as well and make a decision. Now I just need some more money for the ali, steppers and electronics!

Regards
Mike

Well finally after all this time, I plan to start my build in August. Attached is the final design. I got all the bits now (Plate, extrusion, Zapp Stepper kit 1, ballscrews) so watch the build log for pics and progress. Again many thanks to all especially Jazz for the redesign. Here goes!

12862