Requirements to have a 1000x1000 aluminium build plate
-
Could anybody please tell me the required thickness of an aluminium plate sized at 1000x1000 which would have the smallest possible "reasonable deflection" due to it's own weight, support at three point like a triple Z kinematic bed support.
And, does anybody have any experiences of building a printer with this build area. Were any alternatives such as a glass build plate (minus any aluminium plate beneath) considered?
Thanks!
Chris -
@vistalert I think that size will be a lot of challenges. Just for the bed heater you will need more power than you have in a 1 phase electrical socket (maybe my math is wrong).
Printer: "Deltacus" Delta with 510mm bed diameter and 925mm build height. Powered by Duet 2 WiFi. Duet Smart Effector with E3D Volcano.
-
@vistalert For something that size you also have to consider the extruder or hot-end positioning mechanism, too. If you use something like coreXY or other cartesian architecture, the X axis has to span the whole plate. It is also going to sag unless it is very thick/heavy. You'll need a bed sensor on the extruder carriage and use the flatness compensation to get a decent first layer.
OTOH, with a plate that size you probably aren't going to be printing in 100um layers. A thick first print layer is more tolerant of imperfections in the bed plate and the X axis. If you're going to print in 1 mm layers, you might not need any correction for the sagging plate/X axis.
Directly heated glass is a bad idea because it won't heat evenly and there will be warm and cool spots scattered all over the place.
Back in 2013, long before flatness compensation, when I was having trouble with unflat, unsticky beds, I experimented with printing on unheated PIR foam as a sacrificial bed. I set the printer up to bury the nozzle in the foam for the first layer. I'm 99% sure that any type of plastic you print will stick to it. The real trick will be releasing the print after it's done. You'll probably want to print on a raft. Large prints will probably require cutting them free of the foam, hence the term "sacrificial". The good news is that PIR foam is cheap, especially compared to the cost of a 1m x 1m x Z mm print. In the US a sheet of 1" x 4' x 8' PIR costs $15. You can cut it to fit the print before you print to reduce waste, and for smaller prints, you might be able to reuse the foam. PIR won't burn or melt or produce toxic gasses at nozzle temperatures. Video of an ABS test print here and here . You wouldn't need and bed heater, flatness compensation, etc., and the bed and support will be pretty light weight, so the Z axis motor won't have to be anything special.
-
@vistalert I would want to support a bed that size at 9 or more points, and move the gantry in Z instead of the bed.
-
@falkia and others, thank you. I imagine there are lots of challenges. But I've seen claims from people who have built this size or bigger (e.g. 1.5m) Not sure if they used "multiple heat zones".
@dc42 - A voron style would be a good idea for sure.
I'm just keen to know the dimension of the plate. I'm aware of there being many other challenges.
-
@vistalert, If you look at the build plate in isolation I am sure that this is possible but maybe not in a conventional 'just increase thickness' approach.
One could for example drastically reduce the weight of that build plate by milling out a support structure that is part of the bed but with all unnecessary parts milled away. IE a drastic reduction in weight with very little reduction in strength.
Doing so is well beyond my capabilities but someone good at FEA could sort this out without too much effort. -
Why not use either additional supports or a frame underneath the bed? Thermal expansion would need to be considered of course.
-
@vistalert I've built several large format machines although not quite as large as 1m square.
My experience has shown that getting large, perfectly flat sheets of aluminium for this purpose is not as straightforward as it sounds. For one, most sheet metal suppliers I've found are a laser cut service, and they don't guarantee flatness, so on for example a 650mm round aluminium sheet I bought it had a bow of around 5mm between the middle and the edges. I got around this by using stove glass on top of it, carefully shimmed with rather expensive heat transfer pads. It works very well, but almost doubled the cost of the bed assembly.
On my most recent delta build, the bed circumference is 480mm, and again there is a bow - on this one I went to the effort of over sizing the plate to 12mm thick and getting a machining company to 'face' it to get it completely flat. However, when it's heated it bows and I noted a
1mm sag in the middle of the bed compared to the edges. Back to my stove glass supplier and problem solved...but the cost grated somewhat!In terms of heating, as someone else mentioned, you need a powerful heater. I like the Keenovo AC pads, and use a custom 1800w 470mm silicon pad controlled by a good quality SSR, with thermal fuses in line for safety. This pad will heat the plate (measured 11mm thick after facing) to 90c in around 2 minutes. Obviously, the glass surface plate I use takes around another 2 to 3 minutes to get to a stable temperature, matching uniformly after around 8 minutes from starting to heat.
For all of the above solutions, don't underestimate the weight - placing these assemblies on a traditional 3 or 4 mount points just won't cut it and you will get a lot of deflection as the plates heat. It's worth building a frame for the plates to sit on which support them properly.
-
Have a look at Ivan Miranda on YouTube, he is know for building large
Here is his large 3D printer: https://www.youtube.com/watch?v=KhBdMpEGH0o&t=277s -
To add to all the above, what I would do is build a frame out of aluminium extrusion with cross members at centre of each side. Use a mains powered silicone heater stuck to the underside of the bed. Drill and countersink holes in the 4 corners of the aluminium plate (I would use 12mm thick cast tooling plate), and in the centre of the plate and the centre of each cross member giving 9 fixing points. Fit two 6mm thick sheets of semi rigid insulation under the bed of the type used for underfloor heating. Make 9 off pillars of say 20mm diameter, 12mm long with a clearance hole for the mounting bolts. Make 9, 20mm diameter holes in the insulation for the support pillars. Bolt the plate to the frame with the bolts passing through the support pillars and the insulation sandwiched between the plate with heater attached, and the frame.
This will do several thing. The pillars and insulation will prevent most of the heat from reaching the frame (especially if they have low thermal conductivity such as stainless steel). The bolts will alow the sideways thermal expansion or the plate but the frame itself will be isolated from that expansion, contraction. The insulation will hold the silicone heater in place if and when the adhesive gives out. Another frame under the bed frame will alow 3 point levelling or you could use lead screws connected to the bed frame and use these for levelling
-
@vistalert said in Requirements to have a 1000x1000 aluminium build plate:
Could anybody please tell me the required thickness of an aluminium plate sized at 1000x1000 which would have the smallest possible "reasonable deflection" due to it's own weight, support at three point like a triple Z kinematic bed support.
And, does anybody have any experiences of building a printer with this build area. Were any alternatives such as a glass build plate (minus any aluminium plate beneath) considered?
Thanks!
Chris
@slimshader said in Requirements to have a 1000x1000 aluminium build plate:
@vistalert I've built several large format machines although not quite as large as 1m square.My experience has shown that getting large, perfectly flat sheets of aluminium for this purpose is not as straightforward as it sounds. For one, most sheet metal suppliers...
Sheet =/= plate.
@vistalert
You need to contact a local aluminium supplier and ask about sourcing some MIC6 plate. That is cast aluminium that will be reasonably flat.
Typically 1/4" is the minimum thickness that will stay 'flat' at around 12" square. If you want to go to 1000mm, you'll either need to be able to adjust the bed with many many supports to take out any curve (there are a couple of commercial printers with beds like that) just from sag, or you'll probably need something like 3/4 to 1 inch thickness. Maybe both.
Taking to a local supplier is a very effective way to find out who can supply something suitable. The first people you call might not be able to help, but they will usually be able to suggest the right people to call.There are plenty of companies out there that supply large pieces of MIC6 for things like injection molding tools, its just expensive.
Glass is not a reasonable alternative, a piece that large is going to be either very thick or very fragile. Either way people expect glass to be much much flatter that it actually is (float glass is smooth, not especially flat or of consistent thickness, at least compared to other things, like MIC6).
If you want properly flat aluminium, you'll want at least 1 inch thick and then get it surface ground. Although that will be expensive given the size.
The old joke in engineering: How do you make 8mm expensive? 8.0, 8.00, 8.000...
Getting MIC6 in stock form will probably be just fine, getting it ground to be really flat could be quite expensive (and the company doing the grinding might want a really really thick piece to start from).@jens55 said in Requirements to have a 1000x1000 aluminium build plate:
@vistalert, If you look at the build plate in isolation I am sure that this is possible but maybe not in a conventional 'just increase thickness' approach.
One could for example drastically reduce the weight of that build plate by milling out a support structure that is part of the bed but with all unnecessary parts milled away. IE a drastic reduction in weight with very little reduction in strength.
Doing so is well beyond my capabilities but someone good at FEA could sort this out without too much effort.That would greatly increase the cost and adds the risk of deforming the plate. If you really really really wanted to do something like this (which also complicates heating the bed) you would use a custom casting and then just grind the top surface flat.
It would be insanely expensive though.Thermal expansion also needs to taken into account. A piece that large will expand by ~2mm when heated to 100C. If the bed is rigidly mounted you'll get quite a bow in the surface at that point.
You could also just make the bed from a granite surface plate. Granite works really well as a build surface. Takes FOREVER to heat up, but I never had PET chip it the way it does to glass. And its thermal expansion is so low that 100C is nothing.
Kinda heavy.
I seem to remember that you can get granite surface plates fairly cheaply these days though.Isolate, substitute, verify.
-
@theruttmeister said in Requirements to have a 1000x1000 aluminium build plate:
I seem to remember that you can get granite surface plates fairly cheaply these days though.
Get's expensive once you add the cost of the hydraulic lift to raise and lower the plate
On the bright side, it would be a real solid printer approaching a ton ........ and you are right, I had forgotten that we want to stick a heater on the bottom so carving things out to reduce weight is not an option.
-
MIC 6
aluminium cast tool & jig plate [theruttmeister]. I am building 650 x 1000mm x 13mm. However I am removing some of the mass with a grid of 12mm holes underneath which will hold neodymium magnets. The bed surface will be supported at multiple points using Peek insulators. -
Cut me, I bleed Duet3D. 😳 And I love the people here. 😍😍😍
www.MatterHackers.com - https://discord.gg/Ked7GREqux -
@jens55 said in Requirements to have a 1000x1000 aluminium build plate:
@theruttmeister said in Requirements to have a 1000x1000 aluminium build plate:
I seem to remember that you can get granite surface plates fairly cheaply these days though.
Get's expensive once you add the cost of the hydraulic lift to raise and lower the plate
On the bright side, it would be a real solid printer approaching a ton ........ and you are right, I had forgotten that we want to stick a heater on the bottom so carving things out to reduce weight is not an option.
Just out of interest, I checked the cost of granite surface plates here in the UK. 1,000mm x 1,000mm comes at a thickness of 150mm, cost is about £1,850 and it weighs in at 450 Kgs.
-
Pocket change .....
-
@tinken said in Requirements to have a 1000x1000 aluminium build plate:
Get's expensive once you add the cost of the hydraulic lift to raise and lower the plate
That's cheating if you leave the bed plate stationary .....
-
@deckingman said in Requirements to have a 1000x1000 aluminium build plate:
Just out of interest, I checked the cost of granite surface plates here in the UK. 1,000mm x 1,000mm comes at a thickness of 150mm, cost is about £1,850 and it weighs in at 450 Kgs.
Ot of curiosity, how long does something like that take to heat up with ummmm a 3KW heater ?
-
@jens55 About 1 and a half hours. Source: https://www.warmlyyours.com/en-CA/countertop-heater
You didn't specify the temperature.
-
I get impatient after a couple of minutes
