Designing a new router called Brevis-HD

[peteeng]

Hi All - I have placed the N&B's into the assembly. Can see that they can all be done up, have clearance and don't hit anything yet...

Here's a section showing its the sdrive screw is correct length. I'll do the manufacturing details and dwg and then can move to the saddle. I have had trouble organising the N&B's in the past and have tried various ways to make this simple. Its easiest to put them into the assemblies as then its logical and a small list. But then come the general assembly and BOM you get the N&B's in several BOMs vs a single list. We'll see how this method goes. Plus having N&B's in a free floating assembly could prove fatal as they are not tightly associated with a feature as they are an "inter-design" relation, sort of like a cousin relationship vs a brother. To solve that the assembly is "anchored" but this may not be a robust method.

For instance in Inventor you can create a folder then move parts into that folder to organise them in the GA. But can't do that in Alibre.

I currently have 203mm of Z travel wondering if to take it to 260mm? The rail bolt spacing is 60mm so that's the increment.. I'll leave it at 200mm for now....Peter

[peteeng]

Hi All - Got the Z Axis detailed out ready for drawing stage. Did a quick FE to see how its travelling and its bonded stiffness works out at 91N/um which is very stiff. In reality the structure is about 50% efficient, so say 45N/um and then its still stiff. This is at full 200mm extension. If I get time I'll do a full contact analysis. The peak stress is in the rails at 41MPa and being high strength steel this is a very low stress. I've applied 100kgf at the clamp and the tool tip deflects 0.1mm in this condition. Happy with the stats so far so now to get the saddle up to scratch. Peter

[mactec54]

Hi,
yes I'm curious as to why not as well.

In a bed mill the ballscrew and servo are mounted underneath the X axis and the ball nut is mounted on the saddle. The ballscrew and servo must therefore
move with the X axis....and it works. So why not a Z axis?

Craig

You are mistaken the Ballscrew and servo would never move on a machine like that, the Ballnut and saddle move not the Ballscrew and Motor

If you use a Rotating Ballnut then the servo motor sometimes travels with the saddle or what ever it is driving but is a very different configuration than this setup

[mactec54]

Hi Mactec - Why? Works fine on my other machine and its needed for long Z movements. Peter

It makes no difference as to how long the Z axis travel is Fix Ballscrew mounting can have the same length of travel

Extra parts moving that don't need to move including motor cables, none of these parts should be moving in a normal Z axis build or any axis for that matter just a very poor design, shure it works but a poor design no matter what way you look at it

[davida1234]

The peak stress is in the rails at 41MPa and being high strength steel this is a very low stress.

Hi Peter, are you sure the rails are high strength steel? Hardened yes but high strength?

[mactec54]

Hi Peter, are you sure the rails are high strength steel? Hardened yes but high strength?

Correct they are just made with a case hardening steel which can be a little better quality steel than mild steel, the cheaper rails you see could be made from anything no one knows but they still last ok if setup correctly

Having machined mounting surfaces is key to good rail mounting for anything like this, without a machined surface's you should not waste your money, use the round rails they have more tolerance for mounting and misalignment for something like this

Quality Rails are made from Bearing grade steel and can be made from 420 SS but the cheap rails don't count on them being anymore than a case hardening steel

[peteeng]

Hi David - Linear Rail Material?? That's a good question. Bearing surfaces and near surface have to be very high strength so they do not fatigue. Particularly from Hertzian stresses. The contact pressure is compressive at the surface but at some distance inward it changes to tension and this tension fatigues the metal eventually resulting in it flaking away. Due to your question I looked some manufacturers info (Hiwin, SKF, THK, won,,,,) but they don't say what the rail material is (they say everything else , they do quote hardness in the range of HB60+. Most rolling bearing elements are made from AISI 52100 (UNS G52986) steel. This is a chrome alloy steel that can be heat treated to well over 1500MPa UTS and has excellent fatigue properties. With rolling bearings it's not just a case of hardening the surface. The hertzian stresses will fail a low strength steel very quickly as the hard balls roll over it even though the surface has been hardened.

https://en.wikipedia.org/wiki/Contact_mechanics

You do get what you pay for. Peter

https://www.azom.com/article.aspx?ArticleID=6704

I found a material note in the Hiwin ordering interpretation as "high carbon steel" This would be a 1060 type steel I expect.... still a mystery

[peteeng]

duplicate removed

[jackjr-123]

Hi All - Got the Z Axis detailed out ready for drawing stage. Did a quick FE to see how its travelling and its bonded stiffness works out at 91N/um which is very stiff. In reality the structure is about 50% efficient, so say 45N/um and then its still stiff. This is at full 200mm extension. If I get time I'll do a full contact analysis. The peak stress is in the rails at 41MPa and being high strength steel this is a very low stress. I've applied 100kgf at the clamp and the tool tip deflects 0.1mm in this condition. Happy with the stats so far so now to get the saddle up to scratch. Peter

1000N / 100µm (0.1mm) = 10N/µm. Am I missing something?

[mactec54]

Hi David - Linear Rail Material?? That's a good question.

I found a material note in the Hiwin ordering interpretation as "high carbon steel" This would be a 1060 type steel I expect.... still a mystery

You do get what you pay for.

Here is a good video to see what you are getting with the knockoff linear rails under the microscope is all you need to see as to how bad the quality is of the cheap linear rails

[peteeng]

Hi Jack - Yes my error. Thanks for the pickup. 5-10N/um is more like it. Peter

[peteeng]

It makes no difference as to how long the Z axis travel is Fix Ballscrew mounting can have the same length of travel

Extra parts moving that don't need to move including motor cables, none of these parts should be moving in a normal Z axis build or any axis for that matter just a very poor design, shure it works but a poor design no matter what way you look at it

Hi Mactec - Thanks for the feedback all opinions welcome here.

If the ballscrew is fixed to the saddle then this means the toolplate is as long as the screw and the saddle is very long. In fact it's longer as it needs some stickout to get the spindle down past the gantry/saddle bottom. So now the machine will have two long heavy parts vs one long part as per Brevis-HD. The only "moving" part that's different is that the motor goes up and down and the Z screw will easily accommodate that. It still has to go sideways so the gantry/saddle/z axis motion requirements are the same. Running a wire to the motor is a trivial event as there has to be wires there anyway. So I contend that its one large part less and the compromise of the moving motor is as such a good compromise. All machines have many compromises and many choices have to be made. "Normals" change often and "should" is a stance vs a requirement driven outcome. As you also say "shure it works" and that's all it has to do. Cheers Peter

[machinedude]

if it works and does what you need to then i don't see why it would matter other than extra parts moving and possibly extra motor requirements to handle that. one plus side to the design is i think it keeps your spindle center line tighter to the gantry cross beam which i think will help keep things more rigid. so i would agree it's ok to do as long as everything works as it should. i would think about sealing the back up somehow if it was my circus but that's me

i think you probably have more than most on the design testing of the basic functions of the frame. so nobody can honestly say your not trying to sort things out. every machine has it's up side and down side to it i think. i would agree that if you want precision you need machined surfaces to get there but that could be in the mix and we have just not seen it yet? i think someone on here once told me the proof is in the pudding guess we have to wait and see how it turns out

[peteeng]

Hi MD - This machine competes with extrusion machines so no budget for machining lands for rails or anything really. It will be as good or better then extrusion machine accuracy. Dust ingress is always a biggy. My first machine Scoot had a similar drive assembly but uses a belt and its capped bottom fills up with sawdust and stuff very easily. That's why the bottom is not capped. Sorting out the saddle now. The bending allowances are too big so a bit of a rethink in the saddle area. If I use the correct allowances the parts get too big. More thought bubbles... Stay tuned. Peter

[machinedude]

seems like you could use a U channel rather than custom bent pieces and save some time but you would have to true some surfaces up before mounting anything to it. what kind of process were you thinking? laser cut pieces then bent from there with all your mounting holes pre cut in the laser process?

[peteeng]

Hi MD - Yes laser cut and bent parts. Std sections limit the design to std dims and you can't have flanges on 4 sides etc. Laser is very accurate and the CNC presses are very good as well. Not saving time using std section on a semi automated laser line with robotic handling of plate and robotic picking etc. Made 3 machines like this already plus 2 commercial machines that where not CNC but they bolted together perfectly for the job. Cheers Peter

[machinedude]

very accurate is a stretch. sort of accurate is probably a bit more realistic. but i would say you can hold +/- .01 imperial with the laser process on bigger pieces. in assembly you can rope that in a little with some indicating work setting your linear motion parts.

the laser process throws heat into your sheet so you pick up some stress from the process as a result. not something i would call precision being in the trade for 30 years.

laser cut parts bent are going to require some messing with to get to very accurate not impossible done right just some more processes to get there

[peteeng]

Hi MD - At the end of the day this machine is for timber work. So +/-0.1mm is good enough. The laser cut approach has built machines that can hold a bit better than that so I'm happy. I should say I use "very" in terms of users of these machines not metal shop CNC accurate. 0.01" is 0.25mm and the laser company works to +/-0.15mm on their jobs. My machines are aimed at Makers, DIY and hobby guys and gals. They won't pay the $15k AUD plus for a commercial machine of this size. So its horses for courses.

If someone asked me to design a machine at commercial/industrial level the whole process would be different. I work with general engineers, heat treaters and toolmakers in my daily work so all of those would come into play in that arena. For one client I work at >0.001mm for special coatings and parts for ultrasonic devices at the other end I design trailers that carry 600T trucks and are tennis courts on huge wheels. And I'd build a carbon fibre router not a metal one.... But in this router design space its trying to get the best bang out of every buck, every part and even down to cost of freight and paint is considered. Its a mission and I'm up against Chinese machines as well. Cheers Peter

[machinedude]

close enough for wood working no doubt the laser process is much more accurate on thin sheet. the thicker you get the the more you drift away from very accurate. it's just the nature of the process is all

[mactec54]

Hi Mactec - Thanks for the feedback all opinions welcome here.

If the ballscrew is fixed to the saddle then this means the toolplate is as long as the screw and the saddle is very long. In fact it's longer as it needs some stickout to get the spindle down past the gantry/saddle bottom. So now the machine will have two long heavy parts vs one long part as per Brevis-HD. The only "moving" part that's different is that the motor goes up and down and the Z screw will easily accommodate that. It still has to go sideways so the gantry/saddle/z axis motion requirements are the same. Running a wire to the motor is a trivial event as there has to be wires there anyway. So I contend that its one large part less and the compromise of the moving motor is as such a good compromise. All machines have many compromises and many choices have to be made. "Normals" change often and "should" is a stance vs a requirement driven outcome. As you also say "shure it works" and that's all it has to do. Cheers Peter

I agree that there are more than one way to do things, but what I said did not cover everything, how do you plan to machine the rail and ballscrew bearing mounting points, your material 6mm would be left too thin after machining, this needs to have high accuracy machining, which is required for mounting linear rails and the bearing mounting blocks, your extra cable is a big deal as it will have to be quality cable with a flex rating, regular cable will not last very long with the up / down of the Z axis, you don't just have the motor that is moving, there is the motor mount, motor, Ballscrew, Ballscrew bearing mounts, coupling, Etc, this is a lot of extra moving parts, the ballscrew is also subject to bend loading which they do not like

Stress reliving before machining would be required on a steel plate like this, or any steel plates that require machining