Additive Manufacturing Going Round in Circles!

Following a presentation by GE at the recent Solid FreeForm Fabrication conference in Austin Texas where annular rotating AM machines were discussed I thought I'd take some time to post work completed at University of Liverpool between 2003 and 2006 by Richard Black, Carl Hauser, Mike Egan, Peter Fox, Martin Dunschen, Lawrence Bailey and I.

Way back in late 2002 I was given a sink-or-swim ultimatum concentrating the mind somewhat. Essentially I was put in a position where I had to write a couple of grants at very short order to win some funding and start an independent academic career, a real in at the deep end situation. The first one I wrote was on porous metallic structures which spawned the work on Orthopaedics and the second one was Spiral Growth Manufacturing (SGM). SGM was an attempt to parallelise the consolidation part of AM processes and to remove the layer wait time.

I came up with a machine type that allowed this by taking the multiple layers of a typical AM build and turning, quite literally, into a single layer which wound round and round in a spiral. It was a good idea and we produced two machines one using a laser for metal deposition and another for inkjet consolidation, below are some details should you be interested...sorry for the graphics it looks very dated.

The slide above shows that with a spiral machine you can deposit multiple layers at once and also consolidate those layers at the same time. It also means that the control of the system is reasonably simple but it is dependent on having a good timing signal and high data rate between the rotation and the inkjet control system. We used the then new on the market ring encoders from Renishaw to achieve this, they were very nice pieces of kit.

This image shows the machine if you look at the build drum (which is about 550mm diameter from memory) it is annular. The idea was that there is a lead screw running through the centre of this drum which can also rotate. If the drum and the lead screw rotate at the same speed then the layer thickness is zero. It the lead screw rotates slower then the drum the layer thickness is negative. If the lead screw rotates faster then the drum then the layer thickness is positive and the drum processes downwards kept in place by the linear bearings. If you think about this by setting the relative rotation speed you can achieve infinitely variable layer thicknesses, very slick but we did trash this idea mainly because we didn't have sufficient control skills with later machines simply stepping downwards for simplicity ie we made machines that built in full spirals and in circular layers with a step.

Using a drum as in this case means that you can also achieve some form of removable build unit. I did imagine a whole array of machines dangling from frames and picking up annular drums for processing.

There are a couple of problems with the system such as these but they are relativly easily solved. The inkjet heads we used had two linear rows of nozzles this is quite common. The first problem with this is that the inner nozzles see a slower bed so the image is compressed similarly at the outside the image is stretched. Scaling the image according to the radial location is therefore a necessity. Secondly as the rows of nozzles are offset from the radical line and are parallel to each other a further more complex distortion needs to be applied. The effect of the non-radial position on the image printed can be seen below. the image on the left is what you want to print and the one on the right is the one you send to the head, print the one on the right and you get the one on the left, alchemy!

So did it work, well why wouldn't it! I have a video which I cropped some frames from so I don't have to mess about with YouTube, rotation is CCW.

We also made a metal machine using galvanometer scanners and a sealed build chamber surrounding a rather smaller rotational device.

We decided to patent the invention and wrote an application which was awarded. It turned out however that several other people/companies (BEGO and EOS) had had the same idea and beat us to it. I think the idea still holds water for particular geometries and techniques. it might be applicable for large annular geometries such as gas turbine or rocket parts or perhaps small parts such as pills for personalised drug delivery.

So there you go Additive Manufacturing Going Round in Circles. If anyone wants more details just PM me and i'll see what i can dig out of my memory as to what the main problems where, how we solved them and what plans we had to do next.

From memory the university sold the patent to EOS and I never heard anything again!