PDA

View Full Version : Looking for some help designing my gantry please



psiron
18-06-2010, 05:15 PM
Hi,

Im following a design from automatedwoodworks.com and have reservations about the brace used in the gantry. A few people have said the gantry may twist using this design and I would like to do something to eliminate this possibility as much as I can.

The original design uses 1no. cross piece of 100x50x6 aluminium channel but I have updated the images here to show the use of 2n0. 100x50 aluminium hollow section.

Original machine can be seen here http://www.youtube.com/watch?v=7XwM_azm1so

Thanks guys.

2304 2303 2302
http://www.mycncuk.com/forums/images/misc/pencil.png

Ross77
18-06-2010, 06:15 PM
Hi psiron
Not sure I understand the how the y axis will work the same.

The obvious problem I can see is that it looks like there are only bearings on the top beam and plate arrangement and the belt drive is then down a lot lower. if this is the case then you will have problems with the bearings binding due to the lever arm (cantilever) effect. I 'd move the driving part of the belt up to the centre line of the bearings. having the side plates drop down so far will also cause a lever arm in the y axis direction and cause racking that way to.

Another weak point is the belt drive shafts at each end need more support, either a central strap in the middle or the best solution is to support the free ends. it may not be to apparent but once the belts are tensioned and its in motion even the smallest flex will cause problems.

Have you got any more drawings or plans, is that CAD or sketch up? its all good in principle just dosnt look very well executed. :smile: to me the whole thing could do with more diagonal bracing, if its done in opposing pairs then they can be considered to act in tension so can be quite small sections with a thin wall.

If it was mine I would turn the x axis bearings through 90 deg, centre it on the plate and then run 2 rails/flat bars for the y axis (1 above and one below the bearing) again with the belt driving half way between the 2, that way it's all braced up with no cantilevers or skewing forces.

Anyway thats my 2 peneth

routercnc
18-06-2010, 07:49 PM
Hi Psiron,

I've been an Engineer for about 16 years, and into DIY cnc for about 2 years, so you can decide how to take any of my thoughts on this one.

I'm struggling to imagine the Y gantry since alot of it is missing, but it sounds like an open section 'aluminium channel' which is bad for torsion. Closed sections are much stiffer. Any chance of a picture to help explain your idea?

Linear rail for the Z axis will be fine. Tried and trusted solution. Consider mounting this in reverse with the blocks stationary and the rails moving. This allows the lowest block to give maximum support all the time. The conventional way round with the rails fixed means that the lowest block is always further up rail to allow for the plunge. Sorry, this is difficult to describe.

It looks like the only thing stopping the gantry falling off the rails is the belt tension, and it's own weight. Not sure about that one. A high speed router coming off the rails could be nasty. Watch out when plunging! Maybe there is a lower bearing restraint missing from the CAD.

As for the binding, there might be more to that than Russ suggests. In theory there will be a cantilever effect (binding risk), if the product of all the forces leads to a net rotation. During high speed traversing and changes of direction when not cutting, the force applied by the belt is likely to be well below the C of G of the heavy bits on the gantry, therefore there is a net rotation, and a binding risk. This could be overcome with gentle acceleration values.
But during cutting, things are a bit different. The router inertia would have less effect (not moving so fast) and there is now a reaction force at the cutter. This might be quite close in Z to the current belt height. If they are the same height there is no net rotation and no risk of binding. If you move the belt up to the bearings, then you would be some distance from the cutting, and a rotation would occur. I've given this lots of thought as you can tell! This could do with a good picture, but the football has started so I'm off . . .

Agree with Ross about the support on the shafts, they are loaded by the belts and may start to whip.

Also agree that a bit more CAD would help . . .

Thanks
Barry

psiron
18-06-2010, 08:27 PM
Hi, thanks for replying.
Firstly the carriages have bearings below. The carriages are mounted to flat bar, there are 6 bearings in total in each carriage. Full details are available at www.cncrouterparts.com.

With regards to the pillow block bearings I agree, the final design will include a central block bearing on each shaft to stop flex when applying tension to the belts.

I will re-model the frame to include the Y but I will include rhs instead of channel. The designer has a video if the machine here http://www.youtube.com/watch?v=7XwM_azm1so&feature=youtube_gdata

I'll get some other drawings put together, thanks guys, much appreciated.

Ross77
18-06-2010, 10:21 PM
careful bud .....that utube vid looks like welded steel and more importantly has 20mm precision rail and carriages (thk, etc.) add the power cable cable management and that is one expensive machine.

You Will not be able to achieve the same results with the carriages you propose to use. dont get me wrong I'm not trying to put you off. i want to help :smile:

maybe take a step back and ask your self exaclty what you whant to cut and performance you need.

how much do the carridges work out at? they dont seem that cheap. are you going to grind the contact surface of the running bar?

anyway look forward to the following drawings....

routercnc
18-06-2010, 10:29 PM
Hi Psiron,

The double 100x50 box section gantry looks much better. If it is well tied in to the upright side pieces, then this may be OK. The next thing to flex would probably be the raised side rails upon which the gantry runs. Can these be joined across to each other at the far ends? This would block loading of the workpiece a bit, but is worth considering at least at one end. Or a big shear panel at one end to tie down to the under structure.

The video shows linear rails, which are a great help if you can afford the long lengths. The video example does not bind with the belts low down on the sides, and looks like it would cut OK too.

What do you mean by rhs instead of channel? If this is RSJ then this is also poor in torsion and would not make the best gantry. Run with the box section.

Good luck with it, and send out more when you have it.

Barry

psiron
18-06-2010, 10:46 PM
Hi,

Thanks for helping here, much appreciated.

RHS is the term for rectangular hollow section, that's what is in the drawing. With regards to welded steel, my frame is the same as the original one in the video, welded 50x50x4 box section. The sides of the original gantry are drilled and tapped, the same as mine will be. I have the actual plans from the original machine, I'm just trying to improve the gantry really, I would also like to make any changes which would eliminate any problems down the road.

The linear carriages run on drawn 75mm bright bar, drawn steel is extremely accurate and the carriages are designed to run on this material. I have seen a machine run on these rails and they are impressive to say the least. They are used in Joes CNC hybrid too. The flat bar is drilled and secured to the top of the table rails, by clamping both sides of the carriages together I can obtain a precise fit on the flatbar, once this is done on both sides the gantry braces can be cut to the precise size.

I think I'm close to getting the cutting list now, I just want to avoid any serious problems by getting as much advice as possible, once I have the frame sorted I just need the electrics and I'll be away.

Thanks again guys, much appreciated.

Ross77
18-06-2010, 10:55 PM
The video shows linear rails, which are a great help if you can afford the long lengths. The video example does not bind with the belts low down on the sides, and looks like it would cut OK too.


Linear rails are a completely different animal and can resist huge loads, so any additional loading due to poor design can be accommodated. with the simple bearings this is not the case.


What do you mean by rhs instead of channel? If this is RSJ then this is also poor in torsion and would not make the best gantry. Run with the box section.



I've been an Engineer for about 16 years, and into DIY cnc for about 2 years, so you can decide how to take any of my thoughts on this one.


bit worried now:surprised:... rhs is rectangular hollow section and although an rsj (not been called that for a while, UB or UC now) is poor in torsion, even the smallest size would be perfectly adequate for diy cnc applications..........:smile:

how to make Friends and influence people eh:heehee:

routercnc
19-06-2010, 10:12 AM
Hi Ross,

Yes that was badly worded. I'd assumed RHS (but it was written as rhs), and since you don't know what other people know the only open section (since channel had been mentioned) I could think of with similar letters was RSJ. Wanted to make sure, and suggest the closed sections as the ideal.

Will do better next time :smile:

Psiron,
There is a good following for Joe's 4x4 machine, and videos show it cutting well. My next machine may go that way for some aspects.

Barry

Ross77
20-06-2010, 12:37 AM
Hi Barry

I don't mean to cause offense and all opinions are welcome :smile:

If you knew me you would know I'm a joker :joker: but structurally I know what I'm talking about.....so if I said you dont know your '"RHS from your elbow" then tell me too F off......:whistling:

routercnc
21-06-2010, 12:55 PM
Hi Ross,

No harm done :smile:. Forums can be a bit impersonal. I have a Mechanical Engineering Degree so think I have something to offer people who are not sure, but they are only my views.

I'm new here, so am feeling my way around, and upon re-reading my initial post it came across as a bit critical, so apologies for that. Over the past couple of weeks I've been thinking alot about these forces and moments, and wondering what the ideal layout was. Psiron's post came at a time when I thought I was onto something and if anything every comes of it in the way of a useful spreadsheet, I'll post it up. Don't hold your breath though . . .

Anyway, time to hand this post back to Psiron . . . .

psiron
05-07-2010, 06:18 PM
Hey guys, just getting ready to post the build log. I have added a piece of 100x50 aluminium box section behind the 100x50 channel, it's solid not and hopefully eliminates any possibility of twisting when increasing the cutting speed.

routercnc
05-07-2010, 07:57 PM
Hi Psiron,

We've both now posted spreadsheets to help. No open channel formula (yet!), but the RHS formula would give you some idea of the twist on that section. See this link :

Bending (simple or fixed ends) - post#1
Bending (simple) and torsion - post#3

http://www.mycncuk.com/forums/showthread.php?2017-Unsupported-Precision-Round-Rail-Calculator

Thanks

psiron
05-07-2010, 09:52 PM
Your spreadsheet is awesome! It shows marginal deflection without the aid of the channel which I trust will almost reuduce this to next to nothing.
That spreadhseet really is something else, thanks for the link.