3D Printed Plywood Clips

I realize that this pretty specialized, but I needed a way to connect two pieces of 5-ply 6mm birch plywood together in a perpendicular manner. Specifically, I have a Printrbot wood 3D printer and I needed a way to secure the stack of a wooden filament spool coaster/holder on top of my power tower (see pictures). Only one of the filament spools I have purchased over the last year fits into the filament coaster/holder built into the top of the power tower. So, I wanted to find a way to attach a different, adjustable wood spool coaster/holder on top of the power tower.


The concept behind my design is pretty straightforward and started with this drawing.


I just need two U channels set perpendicular to hold the plywood. The plywood measure about 5.8mm using my digital calipers and I believe from this measurement and from listening to an early Printrbot video that this is 5-ply 6mm birch plywood.

I planned to print my final parts using PLA, which does not flex as much as ABS, so I created the clips to be approximately .2mm greater than the actual thickness of the plywood. I designed the clips to use 3mm walls and a pocket 6mm deep for the plywood. I made the length of the clip 20mm as I thought it would be long enough to provide stability.

This piece requires support to print properly, so the support must be carefully removed after printing. In my case, I had to use an exacto knife and 220 grit sandpaper to remove all the support material, although it was quick and only took about 3 minutes per piece.

I started with a design in SketchUp that offset the two sides of the clip, but after printing, I decided a centered design would hold better as all the weight would sit squarely on the bottom of the bracket if it were centered. Here is the first design from SketchUp.


Then I moved on to a centered design, and prior to printing, noticed that support would be required to hold up the overhang and thought a more elegant design would include the support in the design. I was inspired by Clifford T Smyth’s book Functional Design for 3D Printing: Designing 3D printed things for everyday use – an engineering handbook to create better designs to make better 3D print models.


Here is the default support that was required for this centered design. The colors indicates the speed of the extrusion, but some of the green (green == fast) is support as indicated by the arrows. The green support on the lower left-hand side of the image below is specific support for the overhang.


Below is the improved design (STL Download of final design) that eliminates some of the support required and, IMO, looks better with the support under the overhang.



As you can see in the below image, the support for this new design is minimized, which is ideal and minimizes post print cleanup/finishing. The only support material is centered in the image below.


This is how the print looked on the print bed of the Printrbot.


In Marcus Ritland’s book 3D Printing with SketchUp, he recommends making copies of the model every time prior to make any large modifications, that way, you can always grab the model partially through its transformation to redo an operation or generally do something different. If you don’t keep a copy of the model along the way, sometimes it can be impossible (or seem that way) to adjust your model without starting over. Here is a screenshot that shows how I created copies and therefore managed my model throughout its evolution.


Below is a pictures of the stack of power tower and filament spool coaster/holder with the clips in place.


And a closer look at the black clips in place:



Occu-Pi-ed Sensor Part 1/2

I’m starting a project to sense motion in our conference rooms at my office at SPR Consulting. We recently build a mobile web app that shows availability of conference rooms at our office. Currently, the web app shows availability based on the scheduling of the room resources in Microsoft Office 365. But we have a policy that rooms can be claimed if they are empty, so knowing whether a room is physical occupied is also important.

For this project, I’m creating a motion detection device that sits in each conference room and senses motion in the room. The device checks for motion and reports whether there is motion (conference room is occupied) or no motion (conference room is not occupied) and then sleeps for a configured number of seconds. Because the sensor checks whether a room is occupied and it uses a Raspberry Pi computer, the name “Occu-Pi-ed” seemed obviously appropriate.

The device reports the occupied status to a custom web service. Our conference room web app checks this occupied status and includes the status in the web app UI.


This blog will cover the first half of the project by covering the electronics and wiring of the project. The second half (not yet written) will cover the programming on the Raspberry Pi and physical packaging into a custom enclosure.

The project wiring is very simple, you only need the following parts:

I was not able to find any data sheet on this exact board, so I’d like to share my findings to make it easier for others. The PIR motion detector board has 3 pins and they are labeled:

  • VCC – is connect to the 5v pin 4 on the Raspberry Pi
  • OUT – is the pulse out pin that is set to high if motion is detected and set to low if no motion is detected, which connects to pin 11 on the Raspberry Pi
  • GND – is connected to the ground pin 6 on the Raspberry Pi


The PIR motion sensor has two potentiometers that allows adjusting the sensitivity of the motion sensitivity and time lapse (or stickiness of sensing motion). The potentiometers can be adjusted with a small screw driver. Be gentle and be aware the potentiometers only turns about 270 degrees. While holding the board with the potentiometers at the top and the sensor dome facing away, the left potentiometer (labeled Tx) adjusts the time lapse of the sensor. Turning the potentiometer clockwise causes the sensor to continue reporting motion even after the motion stops. Turning the potentiometer all the way counterclockwise causes the sensor to report whether motion is being sensed real-time. The rightmost potentiometer (labeled Sx) adjusts the sensitivity of the motion sensor. Turning the the potentiometer clockwise causes the sensor to be more sensitive to movement (especially farther away). Turning the potentiometer counterclockwise causes the sensor to be less sensitive to motion.




Why the adjustments?

  • Time lapse – Most PIR sensors will be connected to a microcontroller (or circuit) that will be “listening” to the PIR board for the pulse pin to be raised high, to indicate movement. Depending on the implementation, the connected microcontroller may poll the pulse pin periodically. Time lapse allows the identification of motion to be propagated through the pulse pin after the motion is sensed, for a configurable amount of time. This is very useful if your microcontroller, is for example, checking for motion every 30 seconds and then sleeping. You could miss the motion if you are just checking at a given microsecond every 30 seconds, so the board can propagate the sensing of movement for a longer period of time to increase the chance that your board catches the movement. As you turn up the time lapse, the board increases the amount of time the pulse pin is high after sensing movement. From some playing around, it seems the potentiometer for time lapse is exponential and not linear, meaning that the time lapse gets dramatically longer as you turn the screw little by little. In my experimentation, if I turned the time lapse sensitivity to high, I waited over 15 minutes and the sensor was still sticky. Because I am impatient, I turned the sensitivity back down and unplugged the 5V jumper and plugged it back in to get back to a reasonable time lapse potentiometer setting.
  • Motion sensitivity – From my experimentation, I found the sensitivity effected the distance more than any other aspect of its sensitivity. With the sensitivity set at its highest, it seems there may be some false positives, meaning the sensor every long once and a while pulses the output pin even though I was not aware of any movement in the room.

For my final configuration, I have the sensitivity turned all the way up (turned to High location on image above) and time lapse set to 1/3 of the way between Low and High. I have found this creates the pulse out that is high for about 1.5 minutes whenever motion is sensed.

Stay tuned, the next blog in this series will include the python programming on the Raspberry Pi as well as the physical packaging into a custom enclosure.

Simplify3D Makes Cooling Tower Unnecessary 2/2

On my Printrbot Plus (using Repetier v0.56 for Mac and Slicer for Mac v0.9.9) 3D printer, the height and width of my prints are very accurate to their intended dimensions, but I struggle with small parts or small details that extend out from the faces of my models, especially toward the top of my prints. See this image of the 5mm calibration cube as an example.


As I mentioned in a previous blog, these printed models become deformed because the hot end that is extruding is running in such a tight/small area, that the filament never gets a chance to cool properly. Using Repetier/Slic3r, I don’t have any extrusion settings that allow me to slow down printing in small areas, such as the top of the cube and therefore the filament does not have a chance to cool properly. Sure, I could slow down the entire print to be a speed that is okay for the top to print properly, but that would make a 20 minute print take hours, which is not acceptable.

Recently I evaluated Simplify3D , integrated software that supports importing, repairing, slicing, previewing and printing. I have read online, in more than one place, that Simplify3D can result in higher

I am happy to report that Simplify3D provides settings that allow me to print the 5mm calibration cube as well as I was able to print it using a cooling tower approach. See the image below that shows side-by-side prints using a cooling tower and a print using Simplify3D.  The left-most two pieces were sliced and printed using Repetier and Slic3r. The far left piece was printed without a cooling tower and the middle piece was printed using the tower in the photo. The piece on the far right was sliced and printed using Simplify3D. As you can see in the picture (click it to see a larger version), the quality is almost identical for the middle piece using a cooling tower and the piece on the right, that leveraged Simplify3D settings to print properly. The simplify3D piece printed much more quickly and did not waste filament to print the cooling tower. As a result, I am declaring that Simplify3D’s abilities are superior to Repetier and Slic3r.


As of this writing, the only downside to Simplify3D is that is costs $149, whereas Repetier and Slic3r are free.

The images below capture the key Simplify3D settings that allowed the piece to be printed properly without a cooling tower. Inside the Simplify3D process (detailed settings that govern the slicing and printing), the cooling settings needs to include a speed reduction override for layers that take less than 15 seconds to slow down the extruder and the fan to be enabled to ensure the filament has time to cool properly.


Additionally, on the Other tab in Simplify3D, the default speed settings were way too fast for me to get a good print (my config was defaulted to 3000 mm/minute). I experienced good results with my speed set to 900 mm/minute.