Project Logs

Collapse

• MakerFaire San Diego: In Action!

  doctek • 11/15/2015 at 05:17 • 0 comments

  Too many projects, too little time! But I finally got back to this one and have made some good head way. Using a milligram scale, I was able to match the feed from the syringe and the output of the Moineau pump. The Moineau pump now has the longer chamber mentioned in the last posting. I built a little frame to hold the Syringe Pump and the electronics that controls them. Here it all is at MakerFaire San Diego in early October.

• Pumping Peanut Butter

  doctek • 12/07/2014 at 04:40 • 0 comments

  Pumping peanut butter through the Moineau Extruder is a much better test of the system than yogurt. Quick summary: Worked great! I was concerned that the gearing wouldn't handle thicker material, but it was no problem. The leakage was also a lot less. You can see the video here: http://youtu.be/4vtTN_aZ8iU

  As mentioned in the video, clean up is a mess. But it's not a disaster and will get easier with practice. Probably easier with clay since it would be more water soluble. Delivering the material through a tube worked ok, but the large syringe (from my Oogoo project) needs to be improved.

  There is one problem that needs to be dealt with before trying to actually extrude a shape. This occurred to me as I was doing the peanut butter test. The Moineau chamber is less than 360 degrees. That means that the progressing cavities never completely block the pressure from the supply system and thus do not independently control the pressure and flow of the material being pumped. The expected advantage of the Moineau design is not fully realized!

  The answer is to extend the Moineau chamber so it has more than 360 degrees of twist. That's next. In the mean time, enjoy the extuder in action.

• Power for the Pump

  doctek • 11/30/2014 at 00:45 • 0 comments

  To complete the extruder, I needed a stepper motor to drive the Moineau impeller through the Schmidt coupling. Wishing to keep the package as light as possible, I wanted to use a NEMA 14 stepper. Since the pump turns fairly slowly but requires a moderate amount of force, some gearing was required. I turned to an earlier project I did and posted to Thingiverse: Hypocycloid Gear Reduction.

  The hypocycloidal gear reduction can be used with little modification with the Moineau pump. The upper case needs to have a flange added to mount the Schmidt coupling cover and the Schmidt coupling disk nearest the gear box needs to have a nut pocket put into it for an M5 nut. The Hypocycloidal gear reduction can be easily configured to give several ratios from 30:1 down to 4:1 in the same size package. This is explained in the second part posted on Thingiverse. By changing any or all of the impeller, the lower case, and the output, the final drive ratio can be adjusted. I feel this is a big plus given that I'm not sure what the final torque requirements will be.

  Now we're at a critical stage: Time to pump something stiff through the extruder and see what happens! Stand by for Peanut Butter!

• What's a Schmidt Coupling?

  doctek • 11/26/2014 at 05:08 • 0 comments

  "So what's hard about a Moineau pump?" you ask. "OK, you need some math, but big deal. That should do it." Seems like all that's left is to create the impeller and chamber using the profiles, then connect a motor and away you go.

  There's just one small problem. If you look at the animations on the Gravesen site, you'll notice that the impeller follows an eccentric path as it rotates: there is no fixed center of rotation. So you can't fix a gear or shaft to the impeller and simply rotate it. Ugly! Any center you define and try to align with a motor shaft (for example) will move out of alignment before one pumping cycle is complete. I spent some time trying to come up with a way to use elliptical or eccentric gearing. While I was able to create a way to follow the correct path, the impeller needed to turn more than one revolution while the gearing would only turn one revolution. A frustrating failure!

  What is needed is a mechanism that can accommodate shafts that are not aligned. Fortunately, such mechanisms exist. The one I liked best is called a Schmidt Coupling. I found a nice design for one on Thingiverse. Printing one of these helped me (a lot!) to understand the operation. The design was modified slightly to fit my Moineau pump. You can see it in action in the video.

  Now how do we drive the Schmidt coupling ?

• Design details

  doctek • 11/23/2014 at 23:45 • 0 comments

  The source code for the designs, the python scripts, and the dxf profiles may now be found on github: https://github.com/doctek/Moineau-Paste-Extruder

  If you want the stl files for printing, they are posted here: http://www.thingiverse.com/thing:548390

  There's a bit of video of the pump in action as well. Not pumping any paste yet, but the pump has been tested with yogurt and pumped that (easy target!).

  OK. Let's talk about the design of the Moineau profiles. The equations I use are those from Wikipedia for the hypo- and epi- cycloids. http://en.wikipedia.org/wiki/Hypocycloid

  http://en.wikipedia.org/wiki/Epicycloid

  The values for R and r in the parametric equations come from http://www2.mat.dtu.dk/people/J.Gravesen/MoineauPump/

  For example, let's consider this design: http://www2.mat.dtu.dk/people/J.Gravesen/MoineauPump/HypoEpi3_2.html

  The impeller will use the two-lobe version, so that r = 1 and R = 4.

  The equations for the two sections are:

  The only trick is to switch between the two curves at the appropriate time. The python code to implement the equations looks like this:

  def xhypo(t):
      return 3*math.cos(t)+math.cos(3*t)
  
  def xepi(t):
      return 5*math.cos(t)-math.cos(5*t)
  
  def x(t):
  	if t >= 0 and t < math.pi/2:
  		return xhypo(t)
  	elif t >= math.pi/2 and t < math.pi:
  		return xepi(t)
  	elif t >= math.pi and t < 3 * math.pi / 2:
  		return xhypo(t)
  	else:
  		return xepi(t)
  

  As mentioned, the complete source code is at GitHub. The 3-lobe version makes the chamber for the pump. The appropriate values are r = 1 and R = 5. The source code for the 3 lobe profile is also in the repository.

  The parameter t varies from 0 to 2 pi. I divide the interval into 100 steps. The sdfx.py library is used to create the dfx files that will be used in OpenSCAD to create the 3D pump chamber and impeller.

  In OpenSCAD, the profile is imported and extruded with rotation.

  module impeller()
  {
  				linear_extrude(height = H, center = false, convexity = 20, twist=phi)
  				import("moinDual.dxf", layer = "STATOR");
  }
  
  

  For the impeller (two lobe), phi is 360 degrees. The height I use is H = 30mm. Both of these values could be varied for experimenting. The rotation of the three lobe profile for the pump chamber is only 2/3 of 360, or 240 degrees. Seems like something is wrong, but that's what it takes to make the impeller fit in the chamber. The chamber profile is scaled up 20% - another value that could be varied experimentally.

  Next time, I'll discuss the Schmidt coupling and why it's necessary.

• What's all this "Moineau" stuff anyway?

  doctek • 11/18/2014 at 01:58 • 0 comments

  The Moineau, or progressing cavity, pump was invented in 1931 by a Frenchman named Rene Moineau (imagine that!). The principle of this pump is best explained and illustrated by Prof. Jens Gravesen here: http://www2.mat.dtu.dk/people/J.Gravesen/MoineauPump/

  Select one of the designs, say "Hypo- & epi-cycloids 3:2", to see the details. Excellent illustrations show both the principle of the pump and the geometry creation method. Consult the Wikipedia for the equations for the hypocycloid and epicycloid.

  The advantage of this pump design for 3D printing of pastes is that it is rigid and positive displacement. Why is that an advantage? Because the pumping chamber and impeller are both rigid, unlike a syringe or peristaltic pump, the only elasticity is in the material being pumped; usually easier to compensate for. Effects like pulsing are also absent. Positive displacement means that the pumping action starts and stops when the impeller starts and stops. The idea is that these factors result in more precise control of the paste printing process.

  "OK, fine," you say. "So how do I create a Moineau pump"

  Glad you asked, 'cause that's what this is all about. A Python script will do the math to create the profile for the Moineau chamber and impeller. I will then use the linear extrude function in OpenSCAD to create the Moineau pump impeller and chamber in three dimensions from the profiles. By providing both the Python and the OpenSCAD source, other sizes and geometries of the Moineau may be created. So if my design doesn't suit your fancy, here are the tools to design your own version. Since this is my first time using the HAD tools, I'm not sure yet where to put my code. When I figure that out, I'll post it and add my code snippets. (Maybe github?)

View all 6 project logs