Aluminium printer/milling machine build.

[peteeng]

Hi Hanermo - good to hear from you. The Space X rockets changing from carbon fibre to stainless steel is an interesting one. The main reason was production speed and availability. SS rockets can be made faster, he needed about one per week and it took a month to make a CF one. Plus there's not enough CF on the planet to build the amount of rockets he needs (roughly speaking). Peter

[G-Spot]

Reinstalled my FEA software, unfortunately the latest version is full of just as many bugs as the last version.

But after many hours fighting it I finally managed to get it to spit out the deflection of the beams I was thinking of using with 500kg at the centre point.

Even at 50 metric tons this beam(2.5m x 200mm x 200mm x 10mm) is only deflecting 3mm.

I see that HAAS have a deflection test for their spindle which is 0.0762mm with 1000lbs(~500kg)

https://www.haascnc.com/service/trou...---ad0599.html

Anyone know what the average accuracy of top end CNC milling machine is with 500Kg force on the spindle?

[joeavaerage]

Hi,
a smaller machine like a Hass mini-mill would be around 50N/um, whereas a good industrial size machine like a VF1 about 150N/um. Really top end machines
required for the most accurate parts from tough materials up to 750N/um.

A good hobby machine suitable for aluminum cutting and with care steel cutting is about 20N/um. Personal experience says that trying to design and build a 50N/um
machine is a big challenge......probably beyond what you might reasonably called hobby materials and techniques....for instance machines of 50N/um or better are all,
or at least 95% of them, are cast iron and/or stress relieved steel.

Craig

[G-Spot]

Yeah right.
.

Yeah right, not knowing that chemists in a lab decide how the biggest building companies on the planet are going to mix their next batch of cement and not some wizened old site foreman.
Builders mix concrete for profit and not maximum strength, always, if it will be in spec for the job then that's all they care about. So useless at their jobs that they have to have most of the concrete they make lab tested after it has been laid.

[peteeng]

Hi G-spot- The test is to determine spindle bearing float but if I look at the numbers the machine deflects at around 90N/um... say 0.001" at 500lbf. For the Z direction I'd expect it to be much greater than that but then at least its a number... Here's some very old data. Machines are much stiffer now. Peter

[G-Spot]

Personal experience says that trying to design and build a 50N/um
machine is a big challenge......probably beyond what you might reasonably called hobby materials and techniques....for instance machines of 50N/um or better are all,
or at least 95% of them, are cast iron and/or stress relieved steel.

Craig

I am going to weld it like a tank.
First I am going to preheat the parts before I weld them.( I think for the steel I am going to use it is 100C)
Then I am going to stick it in my furnace at 600deg for an hour then let it cool.(I checked up on this because I need to order the burners and thermocouples to suit the temp range)

The cost of the burners and thermocouples(accurate to 3 degs) are about $200 for 10 sets, so I am going to stick a thermocouple on all the pieces and make sure I am always at the right temps when I am working.

Also I have a cement truck or two coming for a job I am working on, so I intend to lay down a 500mm thick reinforced slab with holes in in to make a straightening jig.

I used to know a guy in the truck chassis straightening business, he showed me how they pull these size frames into shape with heat and chains attached to large jigs.

If things get really out of line, then I am going to put the frame down on a flat firebrick surface and heat it up to 1100 C, then it will melt flat...lol.

[G-Spot]

Hi G-spot- The test is to determine spindle bearing float but if I look at the numbers the machine deflects at around 90N/um... say 0.001" at 500lbf. For the Z direction I'd expect it to be much greater than that but then at least its a number... Here's some very old data. Machines are much stiffer now. Peter

Thanks Peter, that's exactly the info I needed. Now I can get going in CAD, already have a drawing but using those 200 x 200 x 10 thick beams which were ridiculous over-kill as the FEA has shown. So I am going to scrap that version and make version in FEA using the data you kindly shared.

[peteeng]

Hi G - You don't need to be too technical with a stress relief. If the metal is red then it has relaxed. Support it well because at 600C it will sag under gravity...
The usual guide is 1hr soak per 25mm thick section, but if it's red its done. Cherry red definitely done.

Sorry Tanks have no relevance when it comes to machine design, the more welding the more the distortion, the more hassles downstream. Minimise welding. Your journey hasn't begun yet, still optimistic thought bubbles and enthusiasm. Its called the Fuzzy Front End.... But I encourage you to keep smiling... Peter.

Preheat is useful if the parent is some funny alloy that has cold or short cracking,. If its mild steel then preheat is useful if you can bring the entire part up to a temp say 200C or hotter. This evens out the thermal stress created by the hot puddle cooling. I would recommend small mirrored welds done in sections and let cool. Take a long time to do the welding vs getting the whole thing done in short time. This allows the structure to settle in between seasons. Straightening cars to 1mm is a bit different to machines at 0.01mm. I have hot corrected large lathe drive screws with oxy in my youth, steel is like butter at the correct temp... Peter

If you have an oven then tack the assembly together, preheat in the oven to 200C then pull out and start welding. Use heavy protective leathers., I have several scars from bumping hot metal... Personally would not go down this path but you have time to figure that out... get busy in CAD and FE... Peter

[G-Spot]

Here's the sensor info if anyone else wants to make their own De-stressing oven. $ 4.25 each including thermocouples, you can get the boards by themselves and then use a higher temp probe if you want to forge or melt steel. But they are more than good enough for 600C.

Correct me if I am wrong but structural steel requires 600C for an hour for each inch.

These sensors can be bought as cheap as buttons everywhere, they hook up to an Arduino and you can read and log the temp within 3 degs across the range.

Cheaper solution than getting your stuff milled first then hoping you can bolt it together straight.

I costed the fire bricks and the insulation for my machine which is 2.5m x 2.5m and it is still a lot cheaper and easier route than bolting.

Cast a flat surface with self leveling cement, jig your machine to it, then weld up. Best of all you leave the sensors in your machine when it is built. Then you can preheat your machine to the same exact temp each time you do machining, or you can log all your machining with temps then make corrections to your accuracy as you machine based on the existing temps.

I also want to be able to de-stress anything I print so having this ability will really save me a lot of headaches in the future.

You cant use a handheld thingy for something big, because unless you wrap something big up well in insulating blankets you will battle to get it up to 600C, super easy to burn a hole in whatever you are heating if you can only see it to measure the temp after taking 5 minutes to remove blankets. Arduino connected to a phone, you can watch the temps without losing all your heat by opening it, you will see rapidly rising temps and lower the gas at the burner in that area. Yes I am aware that propane is not good for forging or smelting because of the carbon in the burnt gas but no such problems with 600C. If I had to do this commercially I would rig up an induction heater and then use the Arduino to do the whole thing, bring it up slowly to 600 then hold for an hour then slowly bring it back down to atmos temp.

[peteeng]

Hi G - You have to be very careful with the concrete. It explodes due to the trapped water. Its a very powerful explosion. So you have to insulate the floor very well so it CANNOT get above 100C. This is not trivial when your ovens going to be 700degC. Heat Treat ovens are usually 500mm or more above the floor... steel framed then firebricks etc etc. Peter

Heres an axle I designed coming out of heat treat at 650C. You can see the oven floor is off the ground... two reasons - protect the floor from the heat and allow the creation of a trolley. I don't think you realise the energy required to heat the oven is huge. Firebrick insulation etc is very leaky - good luck. I have designed 3 ovens for heat treating aluminium. The energy required is more about heating the oven then the part... Peter

I was involved in fire tests for building products when I was a student. The lab had a specially made house which used a particular fire hardened concrete. We did a test one day involving plastic chairs and the floor started exploding. It was like a machine gun, chunks of concrete flying at whizzing velocity everywhere. Very scary. We were close as we were reading the instruments but not for long, everyone ran! Peter

[joeavaerage]

Hi,

I costed the fire bricks and the insulation for my machine which is 2.5m x 2.5m and it is still a lot cheaper and easier route than bolting.

Fire bricks are not the right thing for a heat treating oven. Fire bricks are thermally stable, can usually resist flame products, but are moderately dense and moderately thermally conductive.

For a heat treatment oven you need insulating bricks. Here in New Zealand they are $12.50NZD (plus tax) each. Not too bad for small ovens but at the large sizes it becomes extreme.
Insulating bricks are very low density and highly thermally insulating, quite a different beast to fire bricks.

https://thermalnz.co.nz/products/ref...on-bricks.html

Is there not a company or companies that can do it for you? I suspect that the cost to have it done will be well less than building your own oven.

Craig

[joeavaerage]

Hi,
issue to tease your brain: how are you going to hold the fire bricks/insulating brick up in the ceiling of your oven? Having some sort of frame or columns to hold them up
is only going to interfere with the working volume. Boring a hole and riveting/bolting to some external frame is going to allow heat to conduct via the rivets/bolts cooling the
working volume and heating your frame.

You might try shaping the brick so that they form a dome.

Maybe some bolt let into the surface of a brick to about half the depth of the brick and then cemented in place with a refactory cement.

As you can guess I have given very considerable thought to this matter, with a view to making an oven myself. Truth is that to make an oven of useful size becomes expensive, and then
you might use it once or twice and seldom ever again. The prospect of getting commercial custom for such an oven is pretty limited. Most heat treatments involve carburising or nitriding or
quench and temper cycles. This is not an area where you can wing it, industrial customers are demanding and exacting.

Craig

[peteeng]

Hi G - I understand the ingrained attitude of doing everything yourself, I tend the same way. But in this instance building a heat treatment oven is way off the critical path for this project. It's a many years project within itself. If you have not built a large oven yourself there's so many pitfalls and rabbit holes that when you turn on the switch and it does not work then your whole project is 2 years behind and many, many $$$ behind. There are several ways to do this without stress relief. You need to investigate various designs and methods first. You do not even have a general arrangement. A good GA is the usual official starting point for a project. It can be looked at reviewed and critiqued. Round 1 designs usually end up in the bin... I think around version 10 you start seeing the real path forward and by V20 your starting to polish the design. Keep at it - Peter

[G-Spot]

Design update

Fixed my design size based on the following;

1) Max width allowed on the public roads here is 2.6m
2) Maximum height of room where machine will be operated is 2.4m
3) Maximum room space is 3m x 3m
4) Rail and screw sizes are 2000mm, 2500mm, 3000mm (the ones I have any chance of being able to afford)
5) Sheets that I intend to weld are plentiful in 2m widths

Fixed my design type as moving gantry based on the following;

1) Welding sheets of metal 2m wide at ground or near level gives me the ability to weld up sheets to makes up a single piece of 2m x 15m or any length if I stick the machine on a truck.
2) Any other type of design means I have to have a table or supports to lift the sheets onto and I would only be able to weld 2 sheets.
3) No clamping required for many welding printing jobs.
4) Safety, your gantry weight does not change so you can dial in your servos to not exceed limits, fixed gantry means moving bed and work-piece, rapid accelerations or deceleration could cause work-piece to be catapulted.


Design philosophy I am going to use based on;

1) Work experience at the top of masts has taught me that trying to lean out means everything goes to ****.
2) If you are moving an x and y axis up high and using M20 rails and 20mm connecting plates then you are always going to be off centre by a minimum of 140mm plus half your beam thickness. Which means on a machine with a beam 200mmm wide you are off centre by a minimum of 240mm. Then half the diameter of your spindle on top of that.
3) The problem with movement(vibrations) is because you are working so far off centre 300mm not that you are working so high.
4) Tilting your beam doesn't solve much.

I am going to use FEA to optimise a design with perfect symmetry, where the centre of the cutting tool is aligned with the centre of the beam.

I started out with a work area of 2m(x) x 2m(z) x 1.2m(y). I am using 2500mm rails and screws on this design and sacrificing 500mm to symmetry

Constraints I am using is that the maximum size beam/pillars can be is 300mm x 300mm (cost, availability and weight)

I used a 100mm x 100mm beam with the spindle at the end that travels up and down my gantry beam for the y axis

First image is stress analysis the gantry 300mm x 300mm beams and pillars with 2m between them and 1.2m below them.

Second image is the optimisation based on 5000N of force pushing up from the spindle column

[G-Spot]

Oh WOW! this is the type of AI that impresses me.

This is the bare minimum of steel you need from those original 300mm x 300mm solid blocks to get the job done. Works out at 15% of the weight of the original solid beams.

I now know what my minimum cost making it from steel will be.

Also what is super interesting is how it has put the legs on opposite diagonals of the bases.

Dont forget this design is operating as a press. The final design will factor in the torque and forces from the tool in the horizontal plane, the acceleration of the gantry forces etc, so it will need more steel than this and will definitely not look anything like this when those are factored in.

Oh yeah the spindle column is 200mm x 200mm not the 100 I put in my previous post


And the outer dimensions of this machine are 2.6m x 2.6m x 1.5m the work area is 2m x 2m x 1.2m when I talk about dimensions in mm I am talking about cross sections and not length.

[G-Spot]

Hi,
Fire bricks are not the right thing for a heat treating oven. Fire bricks are thermally stable, can usually resist flame products, but are moderately dense and moderately thermally conductive.
Craig

I already have a lot of ceramic fiber cloth, enough to cover it half a metre thick, the frame is only 200mm high. The gantry will lie flat at ~200mm also. Will use at least 10 propane burners and compressed air + a leaf blower or 2. If I was trying to melt it I would be worried about heat escaping quicker than I could put it in.

I have an electric heater already, I had to build a test load when I was designing my welder, so I had to be able to sink 300 Amps. bought the roll of heater wire direct from the manufacturer and made a large heater that I stuck outside, was more like a light.....lol fastest heat source ever. if you stood in front of it and powered on it would hit you like a wave.

Lots of experience with fire, used to be bush fire fighter, lived off grid before solar power was a thing where everything was done with fire, water for house was heated with a fire boiler. I have been trained in forging and loaded and unloaded kilns when I was studying.

[G-Spot]

Hi G - You have to be very careful with the concrete...

Hey concrete can withstand up to 500C without a sweat, they build all the public fire places from the stuff here.
Just gets the surface worn down a bit. Anyone charging for chemicals to put in it to make it fire proof must be in on a con, I not surprised there were unpredictable results.

Admit it Peter you are so anti concrete

Awesome oven design where you combined the door opening mechanism with the trolley to remove the work piece.

[peteeng]

Hi G - I have put a lot of time and effort into concrete for machine parts. It has a place in the CNC universe. Your projects are very aspirational and I'll follow them with interest. Your AI work I expect is based on stress optimisation not stiffness optimisation. Keep at it. Peter

[G-Spot]

Hi,
issue to tease your brain: how are you going to hold the fire bricks/insulating brick up in the ceiling of your oven?
Craig

I have so much building materials here, If I could melt all the scrap steel I have I could probably cast a machine. Would make a net from high tensile steel fence wire, tie the back of the blankets to that with glass fibre and then lower and higher it like the circus does with the big top. If it's a once off then I am only going to be concerned that the blanket cannot get close to any of the burners.

I have already sourced fire proof mats that are used in industry to hold back fire for a short while and protect concrete and other structures. They work for up to an hour and are like $100. Would have laid that down on top of the concrete, then firebricks on top, then a layer of fireproof mat on top first, then multiple layers of ceramic fiber insulating mat put a sheet of stainless steel in between a few layers.

Get some of my hippy friends round and we can throw turf and sods on the top like it's some sort of medieval earth cleansing ritual.

Concrete is an OK insulator, especially at thicknesses more than 100mm, that fireproof mat will be more than enough to keep the concrete below 500C especially when there are firebricks too, I would be using a minimum of 10 thermocouples, not all of them could ever be faulty, the only source of fuel is the propane, there is next to no chance that the temps would exceed 600C.
I sure as hell would be making sure there were 2 main valves, one far away next to the gas cylinders, and one in the work area.
All work operating burners, carried out to the side and not above. Been there, found myself in the middle of fireball before lol....

[G-Spot]

Keep at it. Peter

Wow I reckon I will be cutting sheets before the end of the weeks, this FEA has just answered questions that I think would take me 10 years to work out manually.

OK I reckon this design is as close to the machine I will make. Basic shape, curved stiffeners on the sides. This is with only 15% of the base material, I will up that to get within top machine specs.

The height of y axis of 1.2m was ridiculous because to achieve symmetry with a 500mm bearing spacing on the base of my gantry I needed to be within an equilateral triangle whose height is 420mm, so I made the max machinable height 400mm so the bottom of the rails on the beam will be exactly at the top of an equilateral triangle with the 2 base bearings the other points. This means that the tool head will always be operating inside the triangle.

The red ball in one the pics is the exact centre of mass of the machine. Means when you operating at a height of 200mm, the mid point the tip of the tool will be at the centre of mass.

I don't care about the only 400mm y axis. This machine will be what I lift up using screws on posts on each corner if all goes well, so it will have an extra y axis at a later date.

So I am almost fixed on a height at 400mm unless I splurge extra and get 3000mm long rails and screws which would allow me a height of somewhere around 900mm.

These optimisations are amazing, can make the thing the stiffest just using thick plates welded in strategic areas.

I am going to try and get to a smoothed and double checked optimization then make it out of balsa wood and test it.