Printable View
Hi All - Moved the bearing and ran the model, then ran it without the support bearing... Answer 1.08mm with bearing, 1.05mm without. So at the bottom position seems the support bearing is a not needed. Hmmmm will have to set up a couple of models at 25% 50% 75% extension to see whats happening. But its all happening at the 0.01mm area so maybe chasing tissue paper. Move onto finalising drawings for quotes better time spent. Peter
This was my Idea to make the Z axis compact and also ridged ,by having the spindle mounted where it is there is no flex at that point when driven down and the lineal rails add also.
https://www.cnczone.com/forums/attac...d=423642&stc=1
Hi Inmesh - everything flexes. What is the distance from your tool centre to your gantry face? Peter
Hey pete. I probably read it but cant remember but are your plates getting cut and bent by the same manufacturer?
Its good you got their die specs.
I got my motor mounts laser cut from 3mm steel and when i designed them i didnt take into consideration the radius needed to bend the plate 90 degrees. ( had plates bent at another place ) so a few holes were elongated, luckily they weren't critical but lesson learned for next time.
Yes true everything flexes ,my tool centre is 130mm from the gantry .When I was building the carriage I thought the 80mm dia metal case on the spindle would give a certain amount of support to keep it all solid along with the rails .There is a 3mm piece of steel in there to make up the distance for the screw bearing support clearance ,the machined ally was just not thick enough .Remembering these were just my Ideas as it was all built from scratch (mostly my head).:D
Hi V2 - I only use shops that do the lot. I also only use shops that have unfolding and folding software so can check the models. Every shop has its own set of "K" factors which is the fudge factor for springback and internal radius that also depends on the tool shape. Most shops doing this will "air bend" ie the back of the plate never touches the die. The other method is bottoming which means the plate goes to the bottom of the tool and removes wrinkles and edge effects. But this means the tool is also generally for 90deg bends. So I have a table from two companies that specifies the companies achieved internal radii and the minimum flange length for SS, Al and mild steel for their tool sets. Even with all of this info there is debate over some parts so I just specify the critical things on the dwg and its up to them to achieve it...
Cheers Peter
attached is an image of one of the tube lasers used.
Hi All and Sundry - Big ticket day!! I have redesigned and re-established the drawing stack based upon the tooling allowances. Looking at the dwg register, for this project I have made and drawn 31 assemblies of which only 11 have made it to the end. In the parts I have made 124 parts and only 39 have made it to the end. The stack is about 11mm thick including the dividers. So now its hole check time and then organise the quote packages. So hopefully end of this week it's out to quote. If anyone is interested in a maximus kit or some of its parts let me know (just express interest) as this will change the order qty. The more parts ordered the cheaper it is. I typically get 2 or 3 sets for first run but if 10 units it would halve the price nearly. Cheers Peter S
Started this on the 15th march 2019 now its July so its 4 months of a slog....
Plus I need to generalise the gantry design so people can put a ballscrew or R&P on it. More to do...
Hi All & Sundry - I have been detailing the drawings and when working on the gantry had the thought is the current design with the brace overcooked? The front and rear plates are 5mm and the brace is 2mm. What if I deleted the brace and only used the outside parts but make them 6mm thick? By weight it would be 2kg lighter as well. So I set up a model without the brace and it deflected a little more than with the brace. But the model is slightly different in how the bearings are supported. 1.2mm deflection without brace vs 1.03mm with brace. Its difficult to determine which bits are contributing the deflection components without doing a fair bit of detail work but I think I'll have to do this. If I look at the gantry itself it only is 8% diff. If I look at the tool holder its 12% diff so there is a delta in the assembly performance vs a global delta. So I may have to do an assm by assm calculation to figure out what is deflecting where.
Its worthwhile if I can get rid of a part. Plus then the Max-HD will be 6mm basic thickness and I'd make the Max-M 3mm thickness basic. Peter
Edit - I whittled it down to the gantry itself. With brace it deflects 0.33mm (5mm plates and 2mm brace) and without 0.35mm (6mm plates). This is a 6% delta which I'm happy to do at the moment. It will make construction easier and be less cost. So the HD will be 6mm basic plate and the M will be 3 or 4mm, lets see how that goes.
Hi All - The last question in my mind now is the extra support bearing I put in, so I ran P5 with and without the support bearing and the difference was 0.01mm so I think the bearing can go as well. So lets get this puppy detailed and to the sales managers for quotes.. Peter
Allicdata can be found similar: https://www.allicdata.com/
Hi All- Placed a bulkhead in the top of the tool plate tube and this dropped the deflection from 1.23mm to 1.16mm so all's good. I had noticed the top was splaying a little. Tube ends need support as they aren't very stiff. I keep hacking at the dwgs... I've worked thru the assemblies and corrected the hole centres etc now the brace isn't in there. Tomorrow may get the RFQ emails out.... Peter
I have also run some modal analyses and the first vibration is at 200htz which is the tool tube rocking backwards and fwds. This is the usual first mode as its a cantilever. Then its 300htz which is the Z motor waggling around, then its 309htz the top of the Z axis waggling, then at 336htz the tool tube twists on the spot. The target was to have first vibs above 300htz so we'll look at those in the polishing phase. The support bearing moves these up a little but probably will add some mass to the motor and this will move all of these past 300htz. I'm going to leave the holes in the Z plate for the support bearing as an alternative as well. Peter
Hi All & Sundry - I've done quite a bit of FEA for this project and there is still some to come. For those that are doing FEA out there or for those that want to know more about it here are some pointers:
1) Linear analysis is done by most simple FEA programs. This means that things like buckling, snap throughs (like an oil can ), following loads, which means the load changes direction as the structure deflects and several other effects are not considered. The software assumes that the shape of the loaded structure is identical to the shape of the unloaded structure, but it deflects a little
2) Non linear analysis applies a small load, figures out the deflection and load direction, resets itself then applies a small load again etc. It does this until the full load is achieved. In this way if the structure changes shape like a buckle it gets captured by the solver as the structure or load changes shape or direction
In regard to CNC machines we design them to be very stiff so all of the linear rules apply. It would be unwise to design a machine that had non linear behavior as we would get different positions depending on loading, direction and perhaps speed of the machine.
So in the linear world we have:
1) Stress analysis which predicts if the material will deform permanently, this we do not want
2) Deflection calculation to decide if the machine or its components are stiff enough
3) Modal analysis which calculates the way the machine will vibrate. It will figure out a series of frequencies the structure will vibrate at and where. This allows the designer to change the shape, add stiffness, add mass or friction or mechanical damping to remove these vibrations if they are deemed to be an issue
4) Linear buckling prediction, which calculates the load and place the structure will buckle at. We design or redesign the components so their buckling modes do not occur in the operational load range
Simple FEA has trouble with assemblies as these are held together with bolts or fasteners and these use friction. Friction is tough to model except for very sophisticated software. Welds are considered to be continuous so their deflection performance is predicted quite well with simple systems. Welding is a very difficult area for stress for various reasons and there are several strategies to predict welding stress in more expensive software.
Overall FEA is a useful tool and is becoming more and more available in basic CAD systems, But be aware of the old maxim, garbage in is garbage out. It's quite easy to set up models that break elastic rules leading to errors that cannot be immediately identified. Sanity checks using manual calcs is a good way to filter out some of these issues. Cheers Peter back to Maximus' drawing stack....
My brain hurts.....
Hi Sterob - But does it feel better when you stop thinking? Peter
Yeah.....eventually....lol
Sent from my SM-G920I using Tapatalk
Hi All and Sterob - Been trolling through the dwgs and model looking for clashes and adjusting bolt lengths. My heads spinning (right hand threaded) as well so it's time to call it a night. So far so good... On Scoot Alpha I missed a spot where a bolt head had to pass through and had to grind the head a 1mm so it would pass. Fixed that on Scoot Beta when I upgraded some of the bearings. So I look at parts and gaps very closely....Peter
So you have factored in the fatigue through rigid vibration that's a good thing ,what spindle are you using .
Hi Inmesh - no specific spindle yet. The model has an Dia80mm clamp although other style mounts can be used. Vibration is considered from the point of view of poor performance not fatigue. It's unlikely anything will fail from a high cyclic fatigue condition as the parts are extremely low stressed due to the high stiffness of the machine (P5 is the current version). The design load is 1000kgf and the machine would not permanently deform under this load. In reality tool loads will be less then 50kgf maybe 30kgf so the service stress is very low on the parts. Peter
I looked at the gantry design and compared it to a std SHS. Its stiffer then a 200x6SHS and similar weight. The next size up is 250x6mm which is 28% heavier and 50% stiffer. Cheers Peter S
Hi Sundry Lurkers - I was about to say all the N&B's were in then found a rail that needs some, damn. But I must be close!! Been realigning the belts due to the new bending radii. Drawings are done and organised so last troll through and it's done!! Peter
Edit - got the Z bolts in!! I think that's it!! Peter
Edit - In another thread there was alot of discussion and work on vibration dampers. Discs with holes in then filled with a washer or heavy round object. What was the final conclusion on that? They worked? or didn,t?
Hi All - Checking the cutting envelope I found I couldn't quite get to 1220mm wide. So I made the gantry rail one hole wider 60mm. The gantry did this easily but the rails had some odd relationship and I had to sort that to get it to play properly. Now the gantry has 1263mm centre to centre so will take a 4ft wide sheet easy. Getting close to quotes now. Peter
Edit - Maximus has 273mm of Z travel. Have to think about where this sits in relation to the bed. Maybe set the bottom at 60mm below the bed and allow for an average tool length. Will have a think about this. The gantry height is set by the bench walls height. cheers