[cnc2]

Hi guys !

Can anyone help with some FEA of HEB beams for a mechmate like router ? (sideways & torsional (moment) loads)

I think about using two HEB 100 beams for the X axis & one for Y axis, the HEB beams are laid out like H rails, in the gap of the H will be a room for a roller chain drive system, & the H will get the loads form the sides(like if it was an IPE, i hope you understand what i mean)

Well, HEB 100 is made of two 10mm thick vertical plates (the sides of the H) & one 6mm thick horizontal plate.

Why HEB ? because HEB beams are cheap, available, relatively accurate & can act like rails & beams at the same time.

Why FEA ? to know if it is possible to use the HEB as unsupported rails & beams at the same time & having a machine rigid enough & capable of 0.1mm accuracy minimum.

The table will be 270cm X 150cm X 30cm, yes 270cm unsupported beams if possible, say the Z+Y gantry are around 120 Kg & feed force is around 170 Kg...How much will it flex ?

Let me know if you need more info.

Thanks !
cnc2.

[wutzu]

These calculations are relatively easy for any 3rd-year mechanical or civil engineering student, by hand. That being said, you have more problems than (presumably) not knowing how to do these calculations.

What support will the two X-axis beams have? Are they lying on a floor, supported only at the ends, supported with some kind of welded frame? If you expect calculations or finite element analysis to be predict better than 0.1mm, the information you feed in gets very important. Material properties, exact section properties, loading conditions, support conditions, etcetera, etcetera. The precision to which you know these properties will dictate the precision to which FEA can predict deflection.

[harryn]

Hi, I looked into the material a bit more and changed my answer.

HEB 100 is just 100mm depth. This might be sufficient for a 150cm distance, but not even close for the 270cm distance. For the 270cm distance, you will need more like 200 - 300mm depth. HEB 240 - 300 sort of range.

The good news is that you have a starting point for the forces, so this can be used to in some simple testing. A laser pointer can help with the deflection measurements.

[cnc2]

These calculations are relatively easy for any 3rd-year mechanical or civil engineering student, by hand. That being said, you have more problems than (presumably) not knowing how to do these calculations.

What support will the two X-axis beams have? Are they lying on a floor, supported only at the ends, supported with some kind of welded frame? If you expect calculations or finite element analysis to be predict better than 0.1mm, the information you feed in gets very important. Material properties, exact section properties, loading conditions, support conditions, etcetera, etcetera. The precision to which you know these properties will dictate the precision to which FEA can predict deflection.

Thanks for the reply wutzu !

Well, I'm a computer guy, so this kind of calculations are not in my knowledge base if i can say so, and I'm in the design stage of a cheap/easy to build router.

The X axis beams are preferably supported(bolted) only at the ends (the simplest way to build it, if viable) The H beams are not used in their original intended way, they are laid on the side like an H not like an i. They are bolted at the ends on a preferably all bolted frame made out of some kind of IPE80 or tube, i'm not sure yet about the whole design, that's why I need to know if the starting point is ok.

I've found a CM66 excel sheet for calculating steel beams but it calculates them like they are used in buildings which is not what I want. In buildings the H is used horizontally & I want to use it vertically.

I hope you understand what I mean

Thanks !
cnc2.

[cnc2]

Hi, I looked into the material a bit more and changed my answer.

HEB 100 is just 100mm depth. This might be sufficient for a 150cm distance, but not even close for the 270cm distance. For the 270cm distance, you will need more like 200 - 300mm depth. HEB 240 - 300 sort of range.

The good news is that you have a starting point for the forces, so this can be used to in some simple testing. A laser pointer can help with the deflection measurements.

Thanks for the reply harryn !

HEB300 ??? the half of the X axis will weight more than a whole mechmate

Look at the little drawing to understand how I intend to use the HEB beams, it only shows one side of the X axis, intended to be used as an unsupported rail.

What do you think ?

Thanks !
cnc2.

[Khalid]

just post the 3D model of Beam... Provide us the distributed/concentrated Load ...and I will do the FEM for you.

[cnc2]

just post the 3D model of Beam... Provide us the distributed/concentrated Load ...and I will do the FEM for you.

It's very kind to you Khalid, but I didn't make a 3d model yet & I have no idea on how to make one.

I'm not sure on the design of the base of the machine, but I know that the table's legs/feet will support the X axis above the table surface, from there i'll have to find a way to secure the legs together to get rigidity on the frame.

I'm afraid to overbuild it & I'm more afraid of having too much steel without achieving rigidity.

I saw your machine, the one that has the tubes for Y axis & the 4th axis it was green I think, it looked like a light machine, is it rigid enough, what can it achieve depth@IPM.

Thanks !
cnc2.

[Khalid]

OKay.. I will calculate for you HEB300 and 270cm length... How much the Gantry load?.. I suppose 50Kg... I defined the material A36.. I fixed both ends and aided 50 kgforce at the one side of surface throughout the 2700mm length...

Will show you the results soon...

[Khalid]

The model of H Beam was created as per standard dimensions given in the following Pdf:
http://www.millstockstainless.com/documents/doc_34.pdf

Both ends were fixed constraint.
On one edge where the roller will move i applied 50kgf throughout the length (270 cm).. Following are the Von mises stress and displacement results..

Maximum deflection 5.6mm...


PS:
You can add plate Braces to strengthen the beam edge.

[Khalid]

I added 10mm plate triangular ribs at both sides equally spaced at 27cm.. but i didn't find improvement in the Displacement readings..

If someone else calculate it and show us the results for comparison. I took the Material A36.

[cnc2]

Wow, thanks Khalid !

HEB300 is not what I want, it's too big too heavy, I want to use HEB100 a bit differently from how you analyzed it, sorry I thought it was obvious for the side loads.

Look at the following drawing in order to understand what I mean, do you understand now, why I want the rail beams to be supported only at the ends ?

It'd be cool if you could do the same for loads from the sides & for the load of the gantry which will weight up to 100Kg max. If you can test with side loads needed to flex the beam by 0.1mm it'd be great.

Thanks !
cnc2.

[Khalid]

Okay.. but the bearing you will use at top will be Qty:04Ea and the same amount at the bottom... This will distribute the 100Kg load /8= 12.5 Kg at all the supported locations!!!!

The same will true for the side bearings, i guess two each at both sides..Total 04 ..so side load will be170/4 = 42.5Kg...

Now as the Y-Gantry is supported at two sides so total load at one side will be:

12.5 Kg/ 2= 6.25Kgf Vertical Load

42.5Kg/2= 21.25 Kgf Side load

Tomorrow i will make some calculation on above basis and will show you the results..

All above posted pictures are considered wrong...

[cnc2]

Khalid, you've got it but I don't think the beams will be facing 100Kgf from the bottom.

That's how I want it to be & it seems to be viable, but I'm not sure it can achieve 0.1mm max deflection, I'm not sure also about the Y beam because it'll face moment loads when the cutting bit is on the table 300mm below.

I'm not sure about the length of the bearing block (distance between the two sets of two bearings) I only know it'll have a minimum of 100mm to be wide enough for bolting the Y beam, I'm not sure if it can be bolted from the inside to reduce the length of the bearing block.

Look at the drawing, I placed the Y beam on the same level as the X beam to reduce moment loads on the X beams.

The bearing block should have an opening from the top to make room for the chain drive, I'm not sure if the bearing block will be like you drawing(will the plates flex?) or if I'll have to make a cast Al bearing block of the shape of an H, what I know is that I want the chain to sit on the groove of the H beam this way it'll be possible to fill that groove with oil & have the chain permanently lubricated.

Well I'm not sure how to accomplish this but I want a chain drive that looks like the one in the video, but how to get the chain to sit flat on the groove of the H beam so it'll be protected & lubed ?



PS: No, I'm not an agent of the "office for complication of otherwise simple affairs"

Thanks !
cnc2.

[wutzu]

Khalid, I don't think you can assume that the ends are fixed. It's a best-case for deflection, but there's no reason to think that cnc2 will be able to mount it to a stiff enough frame. What about if the beams were lying on a fixed foundation like a concrete slab?

[cnc2]

Khalid, I don't think you can assume that the ends are fixed. It's a best-case for deflection, but there's no reason to think that cnc2 will be able to mount it to a stiff enough frame. What about if the beams were lying on a fixed foundation like a concrete slab?

Does it make sense if the legs supporting the beams were bolted to the floor ?

cnc2.

[wutzu]

Does it make sense if the legs supporting the beams were bolted to the floor ?

cnc2.

Look at it like this: The beam is an index card.

If you set the index card on a flat table and press down in the center, it doesn't deflect downward at all. This would be equivalent to a fully supported beam. I say go with laying the two beams on a floor (bolted down so they're not going to slide), but this will keep deflection to the absolute minimum you can get.

If you set pencils under either end of the index card and press down, the card deflects downward. While the card deflects downward, the ends of the card move into the center, since the card stays the same length. The ends of the card also bend upward, because they aren't supporting a moment. This would be equivalent to a beam with simply supported ends. This is similar to what you propose with legs that are bolted down. The legs may support a moment at the end of the beam, depending on how they're attached. However, the legs can still bow in, even if they're bolted to the floor.

If you take that same card with the pencils, and then clamp down the ends of the card to the pencils, you get what khalid analyzed. The ends of the beam are now supporting a moment, which means that the center will deflect less because it has less of a moment on it. The ends of the beam are also unable to move inward, which means that in order for the center to deflect downward, the card has to get longer (stretch). This is an extremely unrealistic end condition.

The tolerances for the HEB that khalid used (which I'll assume are standard) list +/- 0.5mm for the width of the dimension that will be controlling your Z. Do you plan on uniforming this somehow?

[cnc2]

Look at the drawing, the X beams are in RED, I'm not sure about the best design for a rigid table that'll act like a big torsion box.

What do you mean when you say "The tolerances for the HEB list +/- 0.5mm for the width of the dimension" do you really mean the HEB100 beams are 100mm +/- 0.5mm wide & high ? so it won't be able to draw a "perfect" strait line ?

If it's really the case & if the width or height of the beam vary along the length, I have no idea on how to work it out a 3m 60Kg beam won't be easy to machine or grind to the required tolerances.

Other questions arise, is 0.5mm acceptable for wood/Al/marble working/carving what's the accuracy of the chain drive, 0.5mm seems to be big especially for small/fine carvings.
What kind of accuracy/tolerances are you guys happy with, what can a mechmate achieve ?

Thanks !
cnc2.

[wutzu]

Look at the drawing, the X beams are in RED, I'm not sure about the best design for a rigid table that'll act like a big torsion box.

What do you mean when you say "The tolerances for the HEB list +/- 0.5mm for the width of the dimension" do you really mean the HEB100 beams are 100mm +/- 0.5mm wide & high ? so it won't be able to draw a "perfect" strait line ?

If it's really the case & if the width or height of the beam vary along the length, I have no idea on how to work it out a 3m 60Kg beam won't be easy to machine or grind to the required tolerances.

Other questions arise, is 0.5mm acceptable for wood/Al/marble working/carving what's the accuracy of the chain drive, 0.5mm seems to be big especially for small/fine carvings.
What kind of accuracy/tolerances are you guys happy with, what can a mechmate achieve ?

Thanks !
cnc2.

Exactly. The top surface (side surface the way you're mounting it) of the beam is spec'ed to 100mm +/- 0.5mm total width. Which means (probably) that the surface finish of the beam is +/- 0.25mm, which is about 0.01". This means that the line it draws will be straight, but it will be wiggly along the way. That ought to be accurate enough for woodworking, if not high precision machining.

I can't speak to the chain drives, or mechmate stuff

[cnc2]

Exactly. The top surface (side surface the way you're mounting it) of the beam is spec'ed to 100mm +/- 0.5mm total width. Which means (probably) that the surface finish of the beam is +/- 0.25mm, which is about 0.01". This means that the line it draws will be straight, but it will be wiggly along the way. That ought to be accurate enough for woodworking, if not high precision machining.

I can't speak to the chain drives, or mechmate stuff

Ah, but what about the side of the beam that I'm using on the top ? do you have any idea about its accuracy ?

Is there any precise rulers I could use as guides for grinding the rolling surfaces ? (you know, an angle grinder mounted on a bearing skate sliding on an accurate ruler, just in the same way they grind mechmate rails) Maybe aluminium extrusions are accurate enough to use them as rulers?


Thanks !
cnc2.

[RICHARD ZASTROW]

Can you add triangulated bracing to your structure? That would go a long way to stabilize it. Adding bracing from the legs near the floor to the X & Y horizontal structure and securing the legs to the floor would be fairly rigid.

If you didn't secure the legs to the floor you could add bracing around the perimeter near the floor to prevent the legs from spreading out (splaying).

Just a thought.

Dick Z