Bat House Project

[reality_boy]

While waiting on the new temperature sensors I thought I would do a little bit of a design sanity check. I modeled the house in one piece as 3-4 cylinders at the final height with no internal bracing, no textures on the walls, and no seams at the interface. The idea being to measure how much cost is involved in my design work, vs just the cost of printing anything of this relative size.

To analyze the results I created a fake printer with a 800 mm cubed build volume, big enough to print all the parts out. The numbers below are what cura reported for this virtual printer to print the design. The times are all quite long because I used a much slowed down profile that I'm currently using with my ender3, a better tuned printer could take less time. However what matters is the relative change not the total time taken.

Here are my results for both the 2 chamber and 3 chamber houses:

3 chamber
- Simple 4 walls, slow speed
315 hour print time
3633g / 1218m material
$70

- Full model 4 walls, slow speed
412 hour print time
4044g / 1356m material
$77

- Increase by
31% more time
11% more material

2 chamber
- Simple 3 walls, slow speed
136 hour print time
1569g / 526m material
$30

- Full model 3 walls, slow speed
183 hour print time
1801g / 604m material
$34

- Increase by
35% more time
15% more material

Basically the cost of finishing the design is not a big deal, it is the overall size and shape that dictates the cost and print speeds. This was all done with 4.5 mm thick walls. I could go down to 2.4 mm thick walls to see if that speeds things up any. With a 3 shell thickness that essentially results in solid walls with no infill. Infill is set to 20%, so the savings can't be too great, but it may speed things up by a larger amount.

---

Edit, I ran a quick test and reducing the walls to 2.4 mm reduces both the material cost and print time by about 20%, basically the amount of infill in the walls. I will do some strength tests to see if reducing infill has a significant impact, a 20% savings is not amazing but definitely worth something.

[reality_boy]

My thermisters came in, I picked up 8 thermisters with 3 meter cables that are housed in metal pipes and epoxied into place. That should be safe enough to leave in the house even with bats (if I ever get so lucky to have bats). The up side is these are dirt cheap, it cost $9 for 8. The down side is you need a MUX or a device that can log analog voltages from 8 or more inputs, and they are uncalibrated so I have to go through that step.



I'm still waiting on my i2c temperature probes. I went ahead and picked up 8 of those as well, at around $30. The nice thing there is they can share a single wire for the data bus, and are easily to gaing together. They are also calibrated and have there own 12 bit A/D converter built in. The down side is the cost and the slightly larger size. Also they are exposed, so I would need to print a housing if I was to leave them permanently in the house. For now they are good enough.

I also picked up several light sensors to see if I could work out the thermal load on the house. My light sensor from above is not good for much more than a binary on/off value indicating the sun is up.

[reality_boy]

Our agave plants are starting to bloom. They rely heavily on bats to help them go to seed.



The rest of the hardware came in, and I managed to wire things up. I'm just finalizing the logging code and working on a more permanent mount.

[reality_boy]

I finished up my arduino code and calibrated my analog temperature sensors. I still need to finish wiring up the other digital temp sensors and see how well they compare to each other.

At the same time I designed a small prototype box that is hopefully large enough to measure some benefits, yet small enough that it is inexpensive to replicate. I plan on printing it out with multiple wall thickness and line counts. The hope is to work out what the minimum wall thickness is for both temperature control and structural integrity. Finally I hope to paint these to see how that effects things.

I'm currently struggling with my 3D printer, I consistently get under extrusion at times and nothing I do seems to help it along. I have tried new nozzles, new hotends, new extruders, new bowden tubes, hot/cold pulls, rolled back the software, and so on. Nothing seems to cure the problem. Im going to experiment with drying out the filament, maybe that is the issue? Something is causing too much back pressure at the extruder, at times I can hardly push the filament through manually. Anyway I need to work through this issue before I can do more serious testing.

Finally my test bat house is cooking along (literally) outside and showing no signs of warping so far. We have only hit 107 F, but there is plenty of summer left!

[reality_boy]

I went ahead and published my code, it is still a bit rough and lacks proper documentation, but it should work. This is built around an arduino uno and a Adafruit data logger shield. Add on top of that some analog thermistors and up to 8 MCP9808 i2c temperature sensors and either a VEML6075 UVA UVB and UV Index Sensor or VEML7700 Lux Sensor to track the thermal radiance.

https://github.com/reality-boy/tempLogger

I have been down on my health for a while, but I think I'm turning a corner. Hopefully things here will pick up now that I'm on the mend.

[curious aardvark]

just out of curiosity, now you've got all this data - have you considered going back and actually making something that fits your original brief ?

Ie: something cheap and easy for kids in a primary school to make ?

You're clearly having lots of fun :-)
But it's so far from your original starting point, it'll probably never see light of day in an actual school.

[reality_boy]

I redesigned my sample prints to be both skinnier and taller, I think that more closely resembles the real thing. I then made them with a 1.6mm, 2.4mm, 4mm and 8mm thick outer wall, and 2-3 0.4mm walls when printing. The first two effectively have solid walls when printing with either 2 or 3 walls (that is getting confusing to say!). The other two have some percentage of infill between the walls.One of my questions is how much plastic is needed to make a solid outer wall. Ideally we use the bare minimum that gets the job done, it makes printing faster and reduces the cost. Anyway the first two samples while stiff can still be flexed with moderate pressure while the 8mm sample is rock solid. The 4 and 8 mm prints also block significantly more light than the thinner samples. I am curious to see how much paint will improve that. I still need to finish printing out the full samples and stick them outside to see how well they stay cool.

[reality_boy]

I wanted to test out how well different wall geometries block out light, so I printed a 100x100 mm square that was built up out of 4 different thicknesses. Then I painted the flat side with two different types of white paint and finally used a light box to view how well the bare plastic and painted plastic block the light. The results are interesting and highlight that I need to switch to a different plastic, this one is much to translucent. Even the thickest walls and the best paint combined failed to block out all visible light.


I tested out some generic white spray paint and some rubberized spray on roof sealant. The roof spray is super thick and very flexible, however it does a poor job of blocking the light and does not stick well to the plastic. I can scrape it back up easily with my nail. It may stick better to a primer, maybe a coat of black would help with blocking the light as well.Anyway I'm picking up some black PLA plastic, and will repeat the experiment with that. Hopefully that blocks the light completely. If not I will try some black primer as well. I also want to try scuffing the paint to see how well it sticks, and placing it over a heat lamp and measuring how well each wall and paint combo blocks heat.

[reality_boy]

I also finished printing my test containers. The above test brings there value into question, but I will test them out as is and after painting them just to see if we can detect a benefit. The test will need to be re-run using the final plastic and paint as well.

[reality_boy]

I got some black PLA, it blocks the light 100% even with only 2 perimeters. I also tried painting my green PLA black and even with one coat of paint I was able to block out 100% of the light. So basically use black as a base paint before painting the exterior of the bat box, or go with black PLA.


Interestingly enough both black pieces warped in the sun on a 100 degree day here in Tucson. The test print I have outside in green PLA has not warped at all, I need to do some more investigation to see what temperature black PLA can handle. Of course in 100 degree weather we want the house painted white, and I suspect that will survive even a 120 F degree day.


I'm planning on printing my thinnest test tube twice in black PLA, and paint one white. I will put those two up with the green PLA to see if any will warp in the sun in a stronger circular configuration rather than as a flat sheet.


I also need to finish putting my temperature probes together. I'm 90% done, but I have a bit of work left to do. Just need to get off the couch and do it. Then I can quantify how much the color of the PLA affects the interior temp relative to the air temp.