[StrawberryBoi]

Might also consider water jetting instead of laser to avoid the heat stresses in the edges.

[peteeng]

Hi Strawb - water jet cutting is a cost bracket above laser cutting in my neck of the woods and its edge quality is not as good. Yes the plates would have slots and tabs for registering parts. Having corner brackets is extra work. The more parts to align the worse the situation gets.

Craig - VSR is a valid process. I'd be careful saying its not "genuine" In your cast iron case you can't do it thermally as it would upset the metallurgy. So its horses for courses in the SR dept. Is electrical plating genuine and electroless plating non-genuine? No just different processes to achieve the same result or a different result.

The columns on Brevis2 where brazed. The RHS was cut on a tube laser with tabs and the plate parts had slots. It clipped together like a jigsaw and was true after brazing. Peter

I found a good read about flatness---

[peteeng]

Hi Craig - Just to cover the C section thing. Welding things together then drilling holes in them and heat treat and now freight between places (called unnecessary motion in lean talk) all have embodied (to use a now popular term ) cost that does not contribute to its end function or user value. For instance here's a 10mm thick part that I have made for another project. It has many holes of which several are threaded. Threaded holes cost quite a bit. It has a milled feature again milling costs and it has two bends, its stainless steel and its powder coated. It costs $150AUD. Which I think is cheap. Its bigger then Millis column part would be and much more complex. So I'm expecting 4 pieces that make up Millis column to be less then $100 each. And this is ready to assemble. I can't see this happening with buying std sections, welding, setting up for parting up then doing the finishing processes. Plus I can buy 1 or 10 no need to stock parts or over buy parts as long as you can live with the 6 week lead time.... as this company is outrageously busy at the moment. Peter

[ardenum]

Hi Craig - Just to cover the C section thing. Welding things together then drilling holes in them and heat treat and now freight between places (called unnecessary motion in lean talk) all have embodied (to use a now popular term ) cost that does not contribute to its end function or user value. For instance here's a 10mm thick part that I have made for another project. It has many holes of which several are threaded. Threaded holes cost quite a bit. It has a milled feature again milling costs and it has two bends, its stainless steel and its powder coated. It costs $150AUD. Which I think is cheap. Its bigger then Millis column part would be and much more complex. So I'm expecting 4 pieces that make up Millis column to be less then $100 each. And this is ready to assemble. I can't see this happening with buying std sections, welding, setting up for parting up then doing the finishing processes. Plus I can buy 1 or 10 no need to stock parts or over buy parts as long as you can live with the 6 week lead time.... as this company is outrageously busy at the moment. Peter

Peter can you upload that file to https://my.xometry.eu/enquiries/new ? It's the cheapest service I found and I'm kinda curious how it compares...if it's even cheap to begin with.

[peteeng]

Hi Ard - Its a clients part so I can't send it anywhere. If you have a part happy to send it to the company for a quote. PM me a step file and dwg. I'm regularly sending them parts for quote and comment so no big deal with your part.

Now I have been thinking about bonding some parts together with epoxy. This would fulfil two things: 1) a constrained viscoelastic layer which will be damp and 2) its easier to do then brazing, no heat & with the opportunity to get things straight at assembly. I have been thinking about how efficient the connection would be considering the epoxy is E=3.5GPa... So I built a 200x200 column 600mm high, adhesive bonded it with a 0.5mm thick epoxy (which is a bit thick but means the joint will flex more than a 0.3mm thick real connection) and ran the model. Turns out the connection is 98% efficient compared to a "steel" connection 0.5mm thick. I'm happy with that for the moment. Shows promise. Been epoxying things together for over 30 years so should be straightforward in steel, stainless or aluminium... May even use a methacrylate as they are less fussy.

The left assembly is the glued one. If I use 12mm plate the inside bend is 11mm which is quite tight and since I'm not having holes I can control the part size much better. If I have the separate parts zinc plated I can still bond them. If I braze hollow parts the plating won't "throw" to the inside as its a Faraday cage. Same as powder coat the charge won't be inside so won't get cover.... so maybe this is a good path.

So with that question answered its time to tackle the base. Peter

[peteeng]

Morning all - I looked at the base and since its not geometry constrained in its depth it is a candidate for casting in CSA. The current base is 100mm steel so I calculate that it needs to be 188mm deep in CSA. Its also half the weight. So I modelled that and it looks OK at 16N/um. But I look at the Milli-Steel frame config and it has 4 rails on the bottom, lots of seemingly wasted space behind the table and the column is thin and tall. Do I visit a gantry again? I can put half of the "deep" geometry on the other side so the footprint doesn't change yet gain the gantry/bridge double column stiffness. So many decisions to be made. There is a small concrete high rail design here somewhere. I can't find its images. Really like that.... found it.. see attached. Going back thru my sketches and images to see what's next. Peter

[peteeng]

Evening All - I have started putting the high rail design together. We've been there before but with a different head space. This is just the rough model to see if it has merit. Prior gantries that were triangular did well. These were on the lifting gantry design. But it allows the forward stack of the Z axis to be minimised. So I'm using it here. Been clearing out space in the workshop today to move YaG in. Sorting a bench design/build or buy at the moment.... Will get the Z axis on asap. Peter

[ardenum]

Evening All - I have started putting the high rail design together. We've been there before but with a different head space. This is just the rough model to see if it has merit. Prior gantries that were triangular did well. These were on the lifting gantry design. But it allows the forward stack of the Z axis to be minimised. So I'm using it here. Been clearing out space in the workshop today to move YaG in. Sorting a bench design/build or buy at the moment.... Will get the Z axis on asap. Peter

Really curious how you're gonna handle the z axis. It being the weak point of the design and all.

[Zorbit]

High rails look great, until you imagine yourself trying to fix or remove work from the bed. My favourite features on the Bridgeport VMC1000/22 are those big double doors and a table that comes right forward for loading, so you can unscrew thirty parts from a fixture without cramping up your back.

[peteeng]

Morning Zorbit - Good to see your out there. Milli is a Hobbyist machine (Maker, DIY, creative level) Will be small single parts I imagine. I'm shrinking the table to 500x500 on this one. I thought a robot loader would be in order? Ask Elon he's developing one. Peter

[peteeng]

Morning All - Here's No1 attempt. X 13 Y 6 and Z a huge 80N/um great for plunging. I can see the issues, so will get onto No2. The envelope is 400x500x350mm. This model is at trade study level to determine if things are on weight and stiffness. The walls will be separate parts Weights are OK. Z axis is steel saddle is aluminium, gantry can be steel or aluminium and base is cast CSA. Base is 280kg other parts are quite light. The z axis in steel is 30kg and will get a bit heavier as this one is 12mm thick and 16mm will be better. The cars have a slight geometry issue which is giving me grief. They have a 0.16deg error in the track part which means things don't line up properly and won't constrain without using tricks., So I'll fix that in No2. Not sure why that has not shown up before... I think I'll go back to a squarish gantry ??? that may fix Y... thinking about that...Have a great day out there. Peter

[peteeng]

Hi All - I have been thinking about the relationship between the saddle and gantry. This has been discussed in other threads but here it is again. For a load to be efficiently transferred to a member it should be applied through its geometric centroid. For a moment to be efficiently transferred to a member it should be through its shear center. These are generally internal to the section so thats hard to do. But using two bearings provides a "crank" and the center of this crank (center of action) can be organised to pass through these centers. This removes secondary moments in the structure and eccentric forces. So arrangement 1 is a good solution. Arrangement 3 is the usual solution and it can be seen that its centre of action is not through the centroid. But A3 has the bearings nearby each other which is a plus. Since the main loads are in the XY plane the reactions are in the XY plane so the bearings should be arranged co-XY planer to minimise eccentric loads. My latest router uses A2 and I think this may change in future. A5 is good but it has a high fwd stack distance like A3 which is good to minimise and this is possible with A1. A1 is harder to do but since this will be machined lets go with that. I also think A1 is a good solid saddle arrangement. If you have used a tapping wrench to tap a thread you will know its double handle is well balanced. If you use a std spanner to tap its hard to balance even though you use the same torque but the force is eccentric. This is the action centre I'm talking about... Peter

[StrawberryBoi]

Unless I misunderstand, A1 is rails on opposite sides, correct? You are concerned about overconstraining but machining it removes the concern?

You could use side bolted cars as well instead of face bolted.

[peteeng]

Hi Strawb - I am concerned that the machining operation has to be done in two setups in A1 because they are on opposite sides of the part. Or if the part can be machined in one set up using a side cutting tool then thats OK. Its not a case of over constraint its a concern of complex ($$$) machining. If the rails are on the same side they can be machined in one op by nearly anyone. Although there are tales in the threads of people having lands machined for their machines that are not correct even though they are one one side...Peter

[StrawberryBoi]

Hi Strawb - I am concerned that the machining operation has to be done in two setups in A1 because they are on opposite sides of the part. Or if the part can be machined in one set up using a side cutting tool then thats OK. Its not a case of over constraint its a concern of complex ($$$) machining. If the rails are on the same side they can be machined in one op by nearly anyone. Although there are tales in the threads of people having lands machined for their machines that are not correct even though they are one one side...Peter

Ah yes, I see, you're talking about the lands to locate the rails then? One rail being the master and the other the follower in terms of alignment, right?

Could you only machine a recessed land on one side and have the other rail land be non-recessed?

EDIT: Or orient them parallel instead of normal-opposed and have them mount on the same plane on opposite sides, so it's a single facing+lands operation across the two extensions.

[peteeng]

Hi Strawb - Its not about the specification of the detail its about the operator competency (and attention vs thinking about the football) and quality of the machinery. I'm sometimes amazed how poor work can come from excellent machines due to operator slackness. I'm confident the machinist I use can achieve the detail. The question becomes how much does it cost to do this vs a simpler detail cost and is the extra cost worth the benefit. Have to model it both ways and cost it both ways to find out. But for now I go with the technically better solution and back track if I have to... if you mount as "edited" this means the car holes are horizontal and the saddle will have to have difficult mounting details. One solution creates other dilemmas... Peter

if you look at #1204 all the commercial machines have the rails so they can be milled in one OP ie the land is facing the same dirn, Granted these machines are working to 0.001mm or better. They would even do the Ops in matched sequence so tool wear and tool changes are equal in the finishing cuts...But nothing ventured nothing gained... and anything worthwhile is never easy... someone once said

[peteeng]

Hi All - I updated the model No2 and its still down on stiffness in the gantry even when its solid steel 150x150mm so I've made a hollow 200x200 one but it went awol on me so will need to rebuild. .... Peter

[ardenum]

Hi Strawb - Its not about the specification of the detail its about the operator competency (and attention vs thinking about the football) and quality of the machinery. I'm sometimes amazed how poor work can come from excellent machines due to operator slackness. I'm confident the machinist I use can achieve the detail. The question becomes how much does it cost to do this vs a simpler detail cost and is the extra cost worth the benefit. Have to model it both ways and cost it both ways to find out. But for now I go with the technically better solution and back track if I have to... if you mount as "edited" this means the car holes are horizontal and the saddle will have to have difficult mounting details. One solution creates other dilemmas... Peter

if you look at #1204 all the commercial machines have the rails so they can be milled in one OP ie the land is facing the same dirn, Granted these machines are working to 0.001mm or better. They would even do the Ops in matched sequence so tool wear and tool changes are equal in the finishing cuts...But nothing ventured nothing gained... and anything worthwhile is never easy... someone once said

You don't need the rail reference edge touching the rail. You can use any other edge on the part. That way all you'd care about is the opposite surfaces being parallel. It's briefly mentioned on one of those nsk/misumi diagrams about aligning rails and carriages if I remember correctly.

EDIT it was thk

https://tech.thk.com/en/products/kai...Q==&tmp=433211

[peteeng]

Hi Ard - yes agreed but parallel in two set ups can be tricky (eg how parallel is parallel?). Some cut a tight gutter for the rail to sit in. But you have to have the rail in hand to fit it. The rail width is not a controlled dimension. There are lots of controlled dims from the rail to the car. But I'm sure the machinist will get them parallel and a register correct as well. Not a big deal... Peter

[peteeng]

Evening All - Milli no3 is good. X 31N/um Y 25N/um and Z 111N/um. The Z axis is steel and weighs 40kg (16mm thick) The saddle weighs 12kg and is aluminium, the gantry weighs 67kg and is 16mm thick steel 200x200mm and the base is CSA and weighs 270kg. All up about 406kg definitely benchtop material! Its footprint is 700mm deep and 1000mm wide. So shrink it a little more and then detail the car bolts. They are usually hard to get access to... Peter