3 D metal printer

Fuses metal chip/particles together to create durable metal objects

Similar projects worth following
This is a new start on an older project I worked on years back that will finally resolve issues I had building a nice compact 3 D metal printer. I have discovered some new things that work, and am trying some other ideas to really improve the design.

Turn on the wayback machine to 2013. I had discovered a way to build 3 D metal objects using a method I called chipfuzing. The process was fairly simple:Use electrostatics to draw metal chips onto a metal platform and run a high current electrode set over them(like a tig torch ) and therefore fuse the chips together to form the object via CNC control.

 The "chips" were in fact bandsaw shavings from my saw, the power sources were a flyback transformer form an old TV set and the welding current came from a low voltage (well not that low actually)high current transformer(later an actual welding transformer). The "tig" (I used tig electrodes) head was mounted to a Z axis, the welding containment box was mounted on a X axis and the Y axis was a rotary table with a metal top and a drag finger to maintain good ground to the box. Here is a picture of the inside of the box:

you can see the ground strap, table and a hole in the top to allow the welding head to get inside. This was all crude as hell as you can imagine but after countless hours screwing around with it actually did produce parts.

These are some examples of rings the machine produced. Yes...rings as in jewelry type rings. The one on the far right was noticeably better as I made adjustments and tinkered. The others ere results of feed and speed errors,chip feed problems and countless other nightmares I encountered. The thing is the process did work.

Here are some later parts I made as things progressed:

Things got better sorta....This is a rod end I made for my home made actuators.

Here is another look:

a bit more refined. but still not good enough really.The parts were solid and durable enough but really had some rough edges and deformations.

Here is a video that shows the machine in operation making a ring.

One thing this thing did do well-it made parts pretty damn fast. Here is the 2nd version making a part:

Ok now that the history has been rehashed I am moving on.

I got so frustrated with trying to perfect this thing I just scrapped it all-not before making a second version of course, but generally I just gave up. What I realized long after were solutions to problems I was having with the system and after some changes(major changes) I got a lot of the former problems resolves and made the thing much more buildable and much more reliable with a nicer object output.

 I want to resurrect this oddity and perhaps make it a real thing again. Depending on finances,my current health issues etc. I am prepared to make it what it should have been years ago. 

  • Scattering, Splattering and recycled chips

    castvee804/06/2019 at 00:53 2 comments

    An interesting phenomenon takes place in the version 1 and 2 prototype printer that I called spattering and resembles the berries produced when you weld. Its a tick different however and really is better described as scattering as the metal chips are strewn about the finished object in a disturbing fashion. This photo shows a finished part right when I first open the door to remove it:

    Note the loose material. It is formed in a little row around the ring that was made. The material is loose-like it fell through the electrodes and never adhered and is still light colored like it never even got heated at all. I always swept this out and reused it but it was just weird. I always chalked it up as material misfeed but how is it so perfectly arranged around the object?

    You can note that I used an aluminum table. The finished object were always stuck a tad to the table at one or more points pretty well. I found the aluminum table allowed the object to stick well during the build by allowed fairly easy removal afterwards. My original table was steel and objects routinely either broke loose during the build or seized to it really bad. It is my thinking the aluminum spread out the heat buildup and never really fused to well to the table. I never did make any large sheet objects though, perhaps that could be an issue that may present itself later.

    The small metal particles I am using from here on out will provide a denser object which will hold more heat and not seize at all.This remains to be seen. From my small tests It all looks according to plan thus far. The smaller particles hold just enough to keep the build stable and allow an easy release. The metal particles pass through a tube just above the electrode and material control is via servo. Small vibrator motors keep the particles from clumping and clogging. I used a similar scheme in version 1.

  • Let me "pic" your brain

    castvee804/06/2019 at 00:11 2 comments

    In another purely disgraceful adventure into 3d metal printing I have to confess that I was a dedicated Pic chip fanatic. Something about that damn 40 pin version had an ugly grasp on me and I played it to the bone.

    My shortcoming(well one of MANY) was writing code for anything but assembly code.....Not a c person much to this day but I could code like it was still the old days. My pic chips served me well until something like this came along that required importing a cad file and making some go code for it. Naturally I went off on a tangent to solve this distress. The code to make parts for the metal printer at the time had no slicing, no temperature control in fact not much beyond simple move here and there statements. I basically just repeated the printing process on top of the previous layer moving only the z axis and repositioning the electrode head. There I said it now feel free to flog me.

    I quickly realized this was not going to work beyond very simple printing and needed a way to get commands from an object file of some sort....Then I built this:

    Again-I am prepared for a good flogging at your leisure. This is a mechanical 3d scanner that outputs code to the metal printer. I must have missed years of meds or something to go this far to get what I wanted, but as clunky as the damned thing was it did the job. I in fact later built another version I use to this day that outputs a real file format, so your flogging should be less this time around.

    This is a simple calibration example of a small cube and its printed copy in metal:

    The crusty metal copy is obviously on the left. Note the rounded corners..typical metal printer problem caused by chip flake size.

    Here is a video of it operation:

    Its goofy and crude bit it worked....mostly,somwhat...As you can see I was really big on rotary tables during that time or simply just batshit crazy.

    As it turned out this did become important later in my other machine building.

    Thankfully I did shed myself from the pic and jumped aboard the Arduino train. I never used a pic again, but as I write this I have a sudden urge. I intend to sit very still for the next few moments until that thought passes. It's not that there is anything wrong with the Pic I just have serious shortcomings implementing them nowadays.

    The new version printer will have another controller...I will get anything but a pic and that's a promise.

    Cad to file is so much more simple nowadays. This will improve all the aspects of finer control for the new version.

  • Arcing,fusing and an old screwdriver

    castvee804/05/2019 at 23:10 0 comments

    So if you have been reading all this I know you must be wondering...How did the idea to melt chips all come together? Wait for it......It was all just a crazy accident.

    I was and always will be a serious high voltage junkie. I liked to build very high voltage power supplies, add a bunch of capacitors and put a big spark gap across the thing. I just the loved to see and hear the snapping, watch the arc and stick shit in between to watch it fry. Better than television!

    On this particular occasion I was opening the arc tips so the gap was too wide to jump across with me putting something between it to start it. It was a pretty high voltage so I chose and old beat up screwdriver laying on the bandsaw that was used to make adjustments. What it didn't realize right away was the tip was magnetized and had some metal sawdust flakes sticking to it. Of course when the inserted it between the electrodes it snapped and made a nice tiny shower of sparks flying about....It was awesome.....I then did it a dozen more times...Better than television!

    I soon readjusted the spark gap so the arc would run continuous and would load the screwdriver tip up with chips and stick it between the electrodes. When I would clean the tip off later I noticed some has fused onto it and had built up areas on the edges. It was stuck really well. I wasn't sure at the time, but I knew that meant something or was useful. A week later I was trying to build a machine that did that chip welding in a more controlled fashion. A year later I was cursing the day I ever started it....

    A week ago I set up the spark gap and the power supplies again. Same screwdriver only now with some abrasive metal powder. I am greatly encouraged.

    One of the things several people had pointed out on the early version days.The chipwelding was doing so without flux and that was the problem producing the "crusty" results. Back then I did in fact try flooding the welding cabinet with argon. It made no difference whatsoever. I think perhaps this may be true because the welding process I was using isn't welding at all-the metal never reaches a state of fluidity. Just the corners of the metal flakes were actually bonding to each other. If you examine the objects I made under a microscope they are clearly not melted fully to each other. It looks more like a series of hundreds of tiny spotwelds. The parts it made were pretty durable but you could smash the a few times with a hammer and the results appeared crumbly. Not an expert...just my observations. In any case the parts would never be used in structural or critical applications. This is more a machine for making hobby parts.

    The new tests on the metal abrasive looks very similar but the surfaces are so much smoother looking(unless you super magnify them)

    I will be renaming the next 3d metal printer particle fuzer.

  • Time, Technology and better quality junk

    castvee804/05/2019 at 21:53 0 comments

    Both versions of the chipfuzer got robbed of many of their vital components pretty quickly but I kept what was left on a shelf in my shop(I made the damn watch me use their parts in other things...) although Those parts are now sitting on the shelf next to their respective carcasses.I never throw away junk even if its not good junk.

    A year or so after I had given up hope on the contrary bastards I was sifting the halls of a website and found powdered metal abrasive. This was rather exciting as my thoughts turned to the failed experiments using chips...and started dreaming of dust. I made this dust in my own shop on a small scale-my grinders and chopsaws made this tuff all the time, I never gave it much thought. But this place offered it in various sizes. That was rather exciting and I ordered a bag, got it and sat it on the shelf with the other stuff...and promptly just went on to other things.

    Another bone to pick was the rf garbage the version 1 and 2 produced. It was so bad the pic microcontroller I was using at the time freaked out every time the thing started. I built the welding cabinet out of sheet acrylic  and had to use that nasty rf emi spray to coat the cabinet and ground it to reduce the noise coming out of the thing. That did do the trick but it was still a basic arc transmitter between the flyback high voltage and the electrode arcing. If I could feed the metal powder without using electrostatics it would be so much better and consistent. The feeder would be a real bugger to build though and needed to be non conductive. That's about the time I got my first 3d printer...that made abs parts.....I designed a version of the feeder in cad....But never printed it. I was way too busy building 3d printed lathes mills etc.

    The next piece of tech fell into place a bit later. A new microcontroller....the Arduino. I set up programs to track axis, control servos and run my other machines, but never even thought much about the metal printer getting an upgrade. And I was still pretty pissed about having to rebuild one of them to test it.

    Even later(4 or 5 months) I discovered transformers being made in China that would easily replace my giant heavy welding ones.....And the shipping cost more than the transformer...but they were still stupid cheap. I got a couple and promptly used one for another project, placing the spare on the shelf.

    I started designing lots of 3d printed axis for my other machines. I did print a set for the metal printer but never assembled them. Stepper motors started to be plentiful and cheap on ebay and I bought a few handfuls of them for my projects. I stashed six on the shelf...the shelf was getting full. I feared it might even take on a life of its own......but I digress.

    I didn't realize until about 6 months ago that the time was coming for a resurrection. I needed to run some tests first to find out if what I had been thinking about in terms of design changes would be realistic. So I slowly began to tinker a bit.

  • The chipfuzer 2.0 prototype

    castvee804/05/2019 at 21:41 0 comments

    More history on the original project-Sorry but it is important...…

    The version 1 had a number of problems mainly because I am very poor and it was built from junk. Version 2 would also be made from junk but I was much more selective about the junk I used, parts were made more precise, and everything I learned was cleaned up before building number 2.

    I wanted the printer to be small and compact. Something you could carry around and take with you when you went to other hackerspaces etc. Version 1 was big, heavy and just ugly as hell. I know that's a bit vain but this was going to be big and I wanted it presentable. It took three grueling weeks to make version 2 and another few months to fix its own unique quirks. Here it what it looked like:

    It was more compact, better shielded(nasty RF leakage) and a slightly better controller. It made some nice parts!.....sometimes. I had power supply issues mainly from the parts(transformers) I was using and the chip feeding was still having some issues. But it did produce some parts that actually were usable although still pretty ugly. At this point I resigned myself to the fact that I was not going to make this a real working product. So much so I even tried to give it away...…..But I digress...

    This version probably would have been so much better had I been able to just buy what I needed and build everything right. Money was so tight I could barely even buy hardware to make the thing. I has so many weird nuts and bolts holding it all together it was just embarrassing.

View all 5 project logs

Enjoy this project?



Daren Schwenke wrote 04/07/2019 at 04:45 point

I don't see a vacuum chamber or hear a lot of shielding gas.  How are you handling oxidation of the metal between printable layers?

  Are you sure? yes | no

castvee8 wrote 04/07/2019 at 10:46 point

Tried shielding gas as described in the logs. Not reaching temperatures that it makes a difference. The metal is not reaching anywhere near melting point. While some oxidation is taking place it's not horrific enough to address just yet.  Once I get other factors smoothed out a bit I will revisit it though.

  Are you sure? yes | no

aquaticedge wrote 04/07/2019 at 03:37 point

How big is the work surface?

  Are you sure? yes | no

castvee8 wrote 04/07/2019 at 03:44 point

Version 1 and 2 were 5 inches.The newer version I am shooting for 6 inches.

  Are you sure? yes | no

aquaticedge wrote 04/07/2019 at 03:46 point

not bad! bigger than I was thinking

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates