Saturday, May 14, 2005


The Icing on the Cake

Icing sugar turns out to be a wonderful material for doing extrusion experiments, because:
  1. You can work at room temperature,
  2. You can get any viscosity you like by changing the sugar/water ratio (mix in a little glycerin too),
  3. It is completely benign,
  4. It holds its shape perfectly when you lay it down,
  5. Fine strands set very quickly because they have a high area:volume ratio for water evaporation (especially if you hit them with a few seconds from a hair dryer...), and
  6. It sets like a rock.
In addition it is a near-perfect support material for FDM, because it is water-soluble.

So I dug an old Cartesian robot out of a cupboard at the University (it was made in the 1980s...), upgraded its electronics so they were the size of a dictionary, not a suitcase, and put one of our syringe pump heads on it (the pump contains the first part out of AoI Vik mentions below). The pump has one of our heating jackets round it, but I didn't use that for these experiments - hence the unconnected angled hole at its front. Here it is (with the head enlarged top right):

I filled the pump with icing sugar and set it scribbling. It quickly became obvious that the secret is flow control (note panic button added as an afterthought to the syringe pump...). By adjusting the pump voltage, however, it was quite easy to get a good consistent track:

The good track is actually two, one laid on top of the other after the first dried. They stack well, and are about 1mm wide. This is too fat because I used too fat a hypodermic needle, but the process actually seems to work better the finer you go. However, like any good cook, you have to avoid lumps when you mix the icing sugar - sieve it first... To get the good track I had to anticipate the starts and ends with the panic button, making the motor run before the robot started its movement, and reversing it for a few seconds before the end to avoid blobbing. The problem is the elasticity of the syringe-pump components, which store pressure energy and cause the flow to continue after the pump motor stops.

So finally I made a crude valve on the end of the syringe by interpolating a 2cm length of silicone tube between the syringe and the needle. The result would no longer fit in the syringe pump nor on the robot, so I just pressurized the syringe permanently using a spring, and clamped the silicone tube in a pair of long-nosed pliers. This gave perfect control, and you can sign your name with the thing.

So. We may be able to replace the syringe pump with something much simpler:

On the left is a micropipette tip. This has a hole in the end with a diameter of 0.4mm, which is about the same extrusion diameter as commercial FDM machines. It is a throw-away item costing pence, and for RepRap may represent a better bet than syringe needles as its tapering form offers less flow resistance. In the middle is the cut-off end of the tip in a silicone tube, with a diagram showing a pressurized reservoir of deposition material and a simple clamp (A). The clamp is all that is needed to make a very effective valve. This could be opened and closed with a solenoid, or maybe a radio-control servo (right).

The rule in the picture is not really bent - that's the wide angle macro setting on my camera...

A simple string-weight-and-pully experiment (below) shows that the force needed to shut off flow in the tube is 13N, though this may be a bit higher if the upstream pressure is greater than the 300mm water head that you can see in the picture, and a bit lower if the tube were to be squashed using a sharp edge instead of the flat face of a pair of pliers.

For other materials (polymorph, Wood's metal) the pipette tip can withstand temperatures of about 100 degrees C (maybe higher, that's just as far as I got), and the silicone tube is happy up there too.

This comment has been removed by a blog administrator.
I recently tried pushing a plastic similar to polymorph through a syringe with a 1.5mm opening and found it very difficult. Have you had any luck pushing polymorph through the pipette tip?
Not at 90 deg C, which was as hot as my experiment could go. Next I'm going to try a glue gun heater (as opposed to hot air) which will take me up to about 140 deg C. I think that a big no-no is long parallel-sided tubes (like syringe needles), as then the viscosity really becomes a killer. Better to have a fat pipe reducing to small diameter at the last minute. I've been able to find drills as small as 0.2mm in my local hobby shop.

- Adrian
nice update. the viscosity issue makes sense. now wondering if our earlier offline(?) discussion on injection molding style internals is still applicable. i'd venture it isn't.
I was going to try embedding a piece of stainless steel wire in solder applied over a hole in some brass. In theory, if I pull the wire out straight, I get a neat hole equal in diameter to the wire.
The solder trick works. I was trying to attach NiTi ('shape metal') wire & did this by accident. I'd try titanium wire.
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to
Posts [Atom]