From RepRap
Jump to: navigation, search

MetalicaRap Design question feed back / discussion. Add your ideas here!

Please add

after you make a comment

General discussion

split into subpages

I think we should seriously consider doing a clean up of the wiki page for MetalicaRap, the page is very very long, and I think we lose readers underway. We should delegate some of all the designs and descriptions to subpages to cut down on the clutter, we scare people away form the page before they have even started reading it. We should also have a clear policy for how to use the wiki and discussion pages. I sugest that we use the front twiki page for a clear description of its project and its design, plus relegating the detailed descriptions of the parts to subpages. Futher we should probably use the discussion page more. Especially I think the future ideas and lose disussions like MetalicaRapFuture should be moved to the discussion pages.

I have started making some progress on the editiong of the wiki page. I have started working on the electron optics design. I will write something up on the twiki when I have figured out how to structure it.

--Kennet 20:44, 15 April 2013 (UTC)

yes I have found that people find it dificult to find pages that are linked to as their are so many other links anyway its hard to destinguish the important ones, So I have split the page in two, upper and lower with a blank gap between with message that the information below is just for builders. Upper text area should be less technical and lower can be all the technical details for the builders.

I agree -- there is so much information here that it would be better to split this up into several pages, keeping a non-technical introduction at the top of each of those pages. I recently split off "MetalicaRap Software" and "MetalicaRapBaby", but I think even further subdivision would be helpful. --DavidCary (talk) 10:20, 5 May 2014 (PDT)

Design Questions

A. Range of achievable spot sizes at target of a Pierce Electron gun running between 100W to 6KW 150KV  ? ( given cathode is 1.6m from target max deflection 7 degrees )

I'm not an expert on this subject, but maybe this paper will help in performing the simulation:

(MarkRijckenberg - Date: 20120922)

B. Possible pit falls of running an SEM at 100W in four-source photometric stereo Ruderford back scatter mode ? ( typical SEM power is 0.1W )

Q:. Target metal surface temperature measurement would be a big advantage, Do you know of a electron bombardment based remote temperature measurement approach?

A1: What about directly measuring black body radiation? IR camera system? Accuracy issues?

A2: The bombarded surface gets red-hot, right? Is it possible to put an infrared camera inside the chamber -- or an off-the-shelf webcam with its infrared filter removed -- and calibrate it using some test objects in the chamber heated to a known temperature?

Q: Some EBW machines put the workpiece outside the chamber, in open air over a distance of millimeters See Out of vacuum EB welding video here 40 Seconds in[1]. What materials are useful as a "window" -- to pass the electron beam, but hold the air out of the vacuum chamber? (plasma windows are excluded due to high power requirements, Titanium foil is only possible with rotating build platform too low resolution),

Q: how will vacuum and powder work, won't that mean some gas will be trapped in the powder and each layer you make some of it will be released and will have to be compensated for? --Robinmdh 14:03, 18 March 2011 (UTC)

This GasLoad in the powder is removed by baking in argon or vacuum prior to use in printer. Changes in size of final parts due to cooling and sight unintentional porosivity needs to be accounted for in design but it is better in EBM than other processes See 1 hour lecture which is a good introduction to the metallurgy involved in EBM[2] (Select windows streaming in main page, then Windows Streaming Video +window streaming in pop up window, some other selection options come up with the wrong video). .

Other ideas

Q: Printing in air at 1 atmosphere would mess up the metallurgy, but a high vacuum is expensive. What about a low vacuum with a high shield gas percentage?

Answer; In some existing low pressure weld EBW systems they enclose electron gun in a high vacuum box and send beam through window in to work piece in low vacuum air, keeping the work piece within 3cm and they get a 1cm diameter beam spot. Then shield the room with 1mm to 10mm of lead. So you can see that a beam will expand to 1cm diameter over 3cm of low pressure air travel. To get rid of all high vacuum the gun would not work as the electron gun is over 3cm's long, the cathode + anode+ 2 focus coils+ deflection coils is about 60cmn long may be you could get them down to less? So the electron gun needs a high vacuum at least.

A. High vacuum is a box 6/8mm steel + vane pump,(used 600E) + oil diffusion pump (used 800E) (possible make self) + Door Seals + electrical connectors + Gauge (used 80E) + circular motor feed through s +Powder deposit/electron gun/ power supply and control

B. Low vacuum with shield gas is a 2mm Box + ( cylinder of Nitrogen or Argon (XX?E)or Nitrogen plant) + Door seals + electrical connectors + Guage(used 80E) + % oxygen sensor +circular motor feed through s + Powder deposit/ power supply/ control/ electron gun vacuum box vane pump,(used 600E) + oil diffusion pump (used 800E) (possible make self)

In case A the beam can be around 100µm tolerance 20µ m, over 1m travel (max 7 degree divergence out of gun, so build area 30cm x 30cm ), the metalurgy is good enough for jet engine parts.

In case B the the beam can be around 1cm tolerance 200µ m?, travel over 3 cm (build area 3cm x 3cm ) metallurgy not so good.

I hope that's a help? ( I have not done the maths on beam spread rate in low vacuums, so the above is from what I have read. ) The match box problem will not work with multiple parts due to the time/effort to make all the seals work. What do you think of the pipe solution?

Okay so I have been doing some research for the Metalica project too. WRT changes during cooling after printing at high temperatures or due to thermal fluctuations I think this is a key issue for making reasonably precise parts and a better handle is needed on it, how to predict the changes and ultimately how predictable they even are. Great! I agree(Rapatan comment) Got some free some documents on modelling such changes etc. but have yet to read them.

Okay next it is also advertised that electron beam printers can make stress free parts. From the documents that are available that appears to be marketing speak. Great question ; Nasa says they get residual stress in their weld process only in the first layer? will our process behave the same ? An important question See 33min in to 1 hour lecture metallurgy involved in EBM see here [3](Select windows streaming in main page, then Windows Streaming Video +window streaming in pop up window, some other selection options come up with the wrong video). How is this achieved is it the vacuum insulation effect as opposed to argon gas atmosphere used in most laser DMLS processes? or wire versus powder?Rapatan comment

There are minor stresses and sometimes even microcracks. However that is better than all the other freeform methods apparently. It is not clear why this is the case, as the process is essentially much like DMLS except with higher powers (4 point somethin KW ebeam) and secondly with very high temperatures during printing, in the range of 800 to 900 or even more deg C (yes C), which is probably well in the range where e.g. tool steel will creep. Some authors say it is due to the 1m/s or so scan speeds, which is higher than in DMLS. I'm skeptical though as it is probably possible to achieve that scan rate in DMLS too.

WRT using mats instead of powder I think this coudl be quite helpful but the problem is levelling the bed, as the density of the mat is lower than the steel so over the print process otherwise the top level of the mat stack would overtake the level of the object being printed. It could simply be machined off as the excess mat would be easy to remove that way. conventional machining need takes us backwards to heavy cast iron non self-manufacturable machine frames Rapatan comment

Thirdly from the research I have been doing it looks like it might be altogether possible in fact to use less than a high vacuum in the print chamber, which might come in useful as it may turn out that sensors etc. are desirable in the chamber and any purchased sensors will not likely be made of materials that outgas at rates low enough for a high vacuum. Even the stepper motors for the gantry etc will likely be a problem. great idea Rapatan

To do this the conventional way is to just use a really tiny hole (probably laser drilled) in a plate, a micron wide or so. The electron beam comes out the hole, and gas of course gets in there but it is so little that you can still maintain the high vacuum on the other side of the beam. Obviously the deflector coils are on the other side. Well you might want some on the other side too so it goes through the hole maybe?

The gas can only get through the hole so fast actually because of so called necking, the gas cannot exceed the speed of sound. This way the main chamber could perhaps be only a medium vacuum. With regards to beam divergence, that needs to be investigated but I seriously doubt it is a problem as focal area is on the order of 0.9mm anyway. However suppose it is only 100 pascals on one side of the hole 1 square micron in cross section, and the gas gets through at 330 m/s the speed of sound that's 330*10^6 cubic microns, if it has to be 10-4 torr on the other side that is about 1.33*10^-2 pascals, so you'd get roughly 330*10^10 cubic microns that the high vacuum pump needs to remove at 10-4 torr, it's 10^15 cubic microns per liter, so you only need to pump 10^-5 liters. A diffusion pump might get: gets 500 liters at 550 watts or so (independent of pressure presumably as that is how diffusion pumps go basically at low pressures) at 550 watts so that shoudl be fine. Indeed the hole can be a lot bigger than a micron... So that's handy.

Another method is to use a plasma window but that might require more development work so might be better left for future versions and it appears to be unneeded. The plasma window requires 8kw/cm2 of power to function so it would limit a build platform area badly in the domestic power context. Unfortuanetly the beam spot size ballons if the beam goes through gas for more than a few millimeters. Rapatan comment

I could not find a quick answer to how much the beam diverges or looses energy in a partial vacuum but maybe you could ask on a physics forum or something. Or just calculate it from the high energy electron collision cross section and the density of gas molecules in the chamber. If we could quantify beam spread vs vacuum pressure that would be great Rapatan commment( also they traditionally put an intermediate chamber with a hole in and out and a huge vacuum pump attached ie a differentially pumped orifice see paragraph 2 of paper [4] Rapatan comment

Lastly, I think it might be a very good idea to just build a small one first. It will cost a fraction as much, will still be quite useful and will allow you to get the process parameters etc down. IMO it makes no sense to try to do this big one when there is unlikely to be enough money to deal with all the unforseen issues... Greenatolsecondtry 03:38, 23 August 2011 (UTC)

The plan is after power supply is finished we work on gun and finally then powder mechanics and build platform. Current mechanics are for argon gas we need to also develop a High Vacuum version, we have yet to quantify at what vacuum pressure the powder will start clumping, once clumping we need to include depositing rotating textured cylinder under hopper, and other powder transport methods, ideally we really want to apply a known clumped powder working solution.

Documentation approach

I suggest creating a user manual and the bill of materials (1 for the US and 1 for Europe) the way that the guys at did.

I think their documentation approach is brilliant!

However, there is still room for improvement in their documentation on the PCB's. I think they should not use a SVG file, but share the schematics using a tool like to fully document the PCB schematics and electronic components.

(MarkRijckenberg - Date: 20120922)

Current Status (2019)

What's the current status on this project? There's a whole boatload of things written on the main article, but my impression is that no real tests have been done, and the last edit was made 6 months ago. Is anybody still around? I feel like a self-replicating means of printing in metal is important for the future, and this project seems like one of the most thought-out, at least. I'd be interested in helping. My Master's is in Computational Mathematics, so I'm not currently going to be much help on the advanced physics side, though I might be able to learn if that's what's needed. I might be able to help buy components, if needed, or set up prototypes in my friend's garage, haha. Send me an email or something if anybody's still working on this.

--Erhannis (talk) 01:24, 19 March 2019 (EDT)