Latest Articles

Most Popular

BubsBuilds Projects

Assorted Frame-making Hardware

Details
Parent Category: BubsBuilds Projects
Category: Assorted (hopefully) "Useful Stuff" Projects
I'm going to make this page a catch-all for various brackets, mounts, connectors, etc. related to assembling  frames/"rigid" structures. 

90 Degree Corners

1.25" SQ Wood Frame

Designed to fasten 1.25" SQ wooden posts to make cubes with flush edges. Not intended to be particularly high load capacity, but I've now built quite a few little (500mm on a side) cubes with them and they are working just fine for my needs! Holes are designed for No. 8, flanged head wood screws.
 
Shouldn't require any supports, brims, etc. I printed the ones in the pictures from this Polymaker ASA in Army Green. But I have also printed them in PETG and they have worked out just fine.
 
I initially designed these for making these (500mm)3  cube frames for some plant growing experiments,  like the little romaine cell shown in the picture below. 
 
Curious about those funky looking red and black printed gizmos with the servos? Those are my "DrumCam Valves". Not gonna lie, I was pretty pleased with myself when  this one worked out as well as it did :)  Certainly not without its faults, but for a first generation design, I think it's definitely a pretty good alternative to the $100+ store-bought alternative!
 
 

2020 Aluminum Extrusion

I recently had a project that resulted in a small pile of different 2020 part designs, so I started a separate page dedicated to them here: 2020 Aluminum Extrusion stuff

Designed to fasten 2020 Aluminum Extrusions to make cubes with flush edges. Not intended to be particularly high load capacity, but I've built quite a few frames from this general design, and they are working just fine for my needs! Holes are designed for M5x16 BHCS fasteners, but slightly shorter (maybe down to 12mm...maybe...?) should work.

Mine have printed fine without supports, brims, etc.

I printed these in Orange Overture PETG with 0.48 layers on a 0.6mm Revo nozzle.

Find it on Printables

 

P&L Flipper Billiards

Details
Parent Category: BubsBuilds Projects
Category: Games

This game was a guaranteed part of every major family get-together with one set of my grandparents (and the exponential growth that followed from their 5 kids). Ok, well not "this" game, but the game that this was inspired by. The idea of the game is you drop balls into the center, and each player is trying to keep their pocket empty and their opponents' full.

I am just printing the last few parts for my first build, and it will be getting its first game night battle testing within the week! 

The Build

All files for the printed parts can be found at:

Printables  |  Thingiverse

BOM:

  • Printed Parts:
    • (Qty 4) Pocket_gen1.stl 
    • (Qty 4) Quad_gen1.stl
    •
    • (Qty 4) Wall_1.stl
    • (Qty 4) Wall_2.stl
    • (Qty 8) Handle.stl
    • (Qty 4) Paddle.stl
  • COTS

Assembly

MOST IMPORTANT INSTRUCTION - Be extremely organized and careful with the orientation of your magnets! You don't want to have your board flying apart! Come up with a plan for what pattern of N/S you're going to use for your parts, and just be sure that all interfaces will be attracting!!! I know it sounds trivial...but I've scrapped many-a-part over the years due to unintended maglev kinematic mounts.

The assembly is pretty straightforward, so long as you adhere to the advice above. 

  1. Glue in all your magnets. I'd recommend doing one magnet pocket at a time and let it cure fully before moving on to the next. Yes,  this one can take a while, but if you do all of the parts in parallel, it's not THAT bad. I just set up an improvised 'rack' on a tabletop (made from all nonferrous materials, of course...I love magnets, but they'ze sneaky!) that allowed me to have the Quad parts upright spaced far enough to avoid magnet interference between them.
  2. Assemble the Paddle sub assemblies
    1. Press on the first handle (if yours has a loose fit, I'd tack it in with some CA
    2. Slide on the first bearing
    3. Press on the Paddle. The paddle is split, so it SHOULD expand and slide on (I recommend PETG, or other material with some compliance to it. PLA will almost definitely break for this fella). If it doesn't, one trick you can try, prior to cursing my name...hopefully, is to freeze your shaft (this will work better if you used a polymer shaft, but aluminum might give you enough) in your freezer and heat up your paddle a bit. If your interference problem is light enough, this should do the job. Don't heat the plastic to more than 100C (or look up the glass transition temp for the plastic you choose and just give yourself some margin under that)
    4. Slide on the second bearing
    5. Press on the second handle
  3. Pop the Paddle subs into the bearing mounts of the Pockets. Not gonna lie, I was pretty pleased with myself on these bearing mounts...even though I quickly decided the bearings themselves were overkill :/ I printed mine in PETG, and the bearings click in with a nice, satisfying pop and feel rigid (thus far) in use.
  4. Let the magnets do the rest of the assembly work, and let it self-assemble!* (*may not actually self-assemble, but once you get the parts close, they do pop together in a way that I thoroughly enjoy.)

 

 

 

Prusa Bed Camera Mount

Details
Parent Category: BubsBuilds Projects
Category: Generative Design Projects
 
Like a lot of people, I love me a 3d printing time lapse. I've experimented with a few different options over the years for capturing them, but I've struggled to  get a camera mount for my Prusas that  I was really happy with.
 
 DSC07515-Enhanced-NR_c9c69.jpg
 
 But I think I may finally have something that'll  do the job and hold up for a while! And all it took was leveraging some generative design and metal 3d printing....I MAY  have gone a smidge overboard on the fancy tools for solving a problem that I'm sure has already been solved a good many times over, but I had me a lot of fun in the process! :)
 
 Among my previous attempts, some of my favorites have been those that are shot from a perspective that  is locked to the part, and since my main printers are Prusa bed slingers, that means they have to be camera slingers too.   Unfortunately,  the combination of the heated bed and a cantilevered plastic arm have proven to be...sub-optimal. The heat from the bed, even well below the glass transition temp for the material, is still enough to really accelerate the creep, and they tend to get a bit droopy pretty quick.
 
So, when PCBWay reached out and offered me some free parts if I'd feature those parts on YouTube, and then mentioned that included their metal 3d printing services....well, it seemed like a good time to try again!
 
To ratchet the overkill up a notch further, I figured this was definitely a task that called for using a generative design tool to get some lightweight design options
 

Design

Generative design, and it’s close cousin, topology optimization, combine some form of simulation, like finite element analysis, with an optimization algorithm to produce geometries that can maintain stresses and deflections within limits while using the bare minimum of material.

What I really love about using generative design tools, in addition to the cool shapes they make, of course, is that I feel like they force me to focus time up front on defining my requirements.

Things like:

  • What is attaching to what, and where. - “The interfaces.
  • Do these interfaces have to have any specific features on them?, “required geometries”
  • What are the loads and constraints I think these interfaces will see? - “Loads and constraints”
  • Any ‘keep out’ areas for things like tool access, cable routing, or any moving parts need clearance? - “Keep out zones”

And all that kind of fun stuff.

I feel like it pushes me to maintain more of a ‘big picture’ perspective on the design in a way that I think is great.
 
Taking a look at the screenshot showing what my end of the CAD for the camera mount looks like, we can also see that it makes for some just lovely model files :)
 
pcm_model_art1_75e4d.jpeg
 
 So  what is it we're looking at in the above? Let's start by taking a look at the 'keeper' geometries. These are the three green bodies shown below. The larger one being the camera mount interface, and the two smaller ones are the mounting interfaces to the printer bed.
 
pcm_model_keepers1_598a8.jpeg
 
The bed mount points are intended to be direct replacements for the 6mm long, steel spacers that  separate the bed heater from the Y-carriage on the Mk3.
 
 pcm_spacer1_25d4e.jpeg  model_keeps2_53ed1.jpeg
 
 Unfortunately, the center-to-center spacing between these spacers aren't symmetric, so I slotted one of the mounts in the design to allow me the flexibility on mounting.
 
The only other 'keeper' geometry is the camera mount point, which is basically just a beefy 1/4-20 washer.
 
pcm_model_keeps3_3018f.jpeg
 
It would have been cool to integrate the GoPro mount into the printed part, but because of my pretty limited experience with metal printed parts I wasn't confident enough in getting that to come out correctly without a few iterations. Plus, sticking with the 1/4-20 mount makes it compatible with a lot of other camera options, should I want to swap out the GoPro for something else down the line.
 
So, interfaces/keepers out of the way, what was that big ass red thing? The geometries that the software shows in red are the 'obstacle' geometries, which I tend to refer to more as "keep out zones". These are where I'm telling the software that it's not allowed to place material.
 
pcm_model_obs1_49776.jpeg
 
The big, blocky bit  provides the keep out areas that cover the heated  bed and Y-carriage parts above and below the spacers. The only other additional bodies are the  posts that sit inside of the clearance holes for the fasteners that will need to go through my mount points.
 
And that's pretty much it for the 'CAD' portion of the design
 
 Now that the base geometries have been defined, it's time to apply some loads and constraints to them.
 
To get some estimates for the expected loads, I started by getting the mass for my GoPro and it's mount(s)....yeah, I realize the heft of these kind of wipes out the benefits of my optimized mount, but cest la vie.
 
pcm_weigh_gp1_ac18f.jpeg
 
 I then estimated some accelerations I expect  that mass to be subjected to and used some of that Newton goodness (F = ma) to get the forces.  The accel limit in the Prusa Slicer  software (which I'm guessing is the default, because I don't think that's one I've ever tweaked) is 1000mm/s^2, or 1 m/s^2. That's about 0.1g. So I rounded that up to about 1g for the Y, and I went with 2gs in Z to cover gravity plus some margin for the dynamic load. I also added about 0.5g to X, mainly just to ensure that it has SOME stiffness in this direction to handle some light jostling and such.
 pcm_model_loads1_7449d.jpeg
 
 There are quite a few issues with how I've decided to do this, including the fact that this  approach doesn't account for the additional offsets to the CG of the mass due to the GoPro's mounting.  I decided to just add some additional margin to the load cases, in addition to the built-in Safety Factor in the solver, instead of creating a more 'accurate' loading scenario.   While I still think this is a reasonable trade-off for this project, I'm hoping as I get more familiar with the software I'll learn better ways to structure these within a larger model, allowing some of this to be fed in parametrically...but we shall see
 
Now that my loads are applied, it's time to set some constraints. I set both of the bed mounts to be constrained in both Z and Y, but only  the non-slotted one was constrained in X. This wasn't strictly necessary, since technically the slotted one is also constrained in X due to the friction of being sandwiched between the heated bed and carriage. But I wanted to also ensure there would be some stiffness in the part along this axis, and if it throught these mounts were both  rigidly constrained, it could conceivably have the whole mount extend from only one bed mount and still satisfy the criteria.
 
pcm_model_loads2_bfcd4.jpeg
 
Now that the loads and constraints are set, all that's left is to configure the settings for the solver. Starting by setting the 'Objectives', shown below. I left the Safety Factor as 2, but given the low loads we're talking about here, I'm guessing it's likely to hit the minimum thickness limits before getting anywhere close to this (this minimum thickness parameter can also be changed, but in a different setting window). I opted to also require displacement limits in addition to the stress limit. The thought here was that I don't want my camera bouncing around on me as the bed whips back and forth. So  I targeted a somewhat ambitious 10 microns of max movement on all axes...this is only going to make that Safety Factor even less important :) 
  pcm_gd_obj1_629c7.jpeg
 
 Next is setting up any specific considerations for the manufacturing process you intend to use....this is the area of these software tools that I most hope to see advance significantly in the coming years.... Although it does 'support' multiple fabrication types, in my  (admittedly limited) experience, it is not particularly good at actually  incorporating Design For Manufacture into the solutions.  But, this is where I can set things like overhang angles and minimum thicknesses.
pcm_gd_man1_939ce.jpeg
 
 And, last but not least I just have to set a material, if not already configured in the solid model. I selected the only aluminum flavor Fusion lists under its Additive Materials list (AlSi10Mg).  I'm not sure which specific alloy PCBWay uses, but I'm assuming it will be similar. I know that for the metal AM that I have worked with previously, they tend to use alloys that I've never heard of but that are well-suited to the process.  
 
--Update from the future, I went and looked on PCBWay's site and they are indeed using this AlSi10Mg alloy. So tip-o-the-hat AutoDesk :)
 
Now that everything's set up, all that's left is to hit 'submit' and wait....and then wait a bit longer...it's not the fastest thing...but then again, neither am I.
 
Although, it is kind of enjoyable watching as the solver iterates through, getting ever-cooler looking designs. And when it's done (assuming it converges) it reveals all of the different variants that resulted. My assumption is that each of these "Outcomes" represents some difference in starting conditions in the solver, but it's not obvious exactly what these changes are.
 
pcm_gd_res1_78501.jpeg
 
 But, regardless of how it got there, we now have our parts to do with as we please! From here, I could export them to a  Design environment, to modify or integrate into a larger design. But, since one of my goals with this project is to look at the generative design to metal printing pipeline, I just exported them directly to STEP files  to send em off for printing!

 Ordering

A huge thank you to PCBWay for sending me these parts for free! But even without getting them for free, I am really excited by the thought of having access to metal printed parts like this! The only metal AM parts I've had access to in the past were through my day job, and even internally they cost at least $1k just to get them to put the powder in the machine. I was blown away when I went through the quoting on these and saw that  would have only cost me less than $40 each. And they aren't paying me to hype this, I promise, just the comp'd parts for credit on YouTube. I just genuinely think it's wild that garage tinkerers like myself can now get these kinds of parts for less than the cost of going out to dinner....the future is awesome :)
 
But anyway, I digress. I took the four 'Recommended' designs from Fusion360 and uploaded them to PCBWay's quoting system. I clicked on Titanium and tried desparately to justify to myself why I needed it. After failing at that, I clicked the aluminum option and skimmed the other options, but the defaults all seemed ok by me. So I sent em off for fab and got back to work on my Mechanical Tester to try to have it ready for some fun once the parts arrived. More on that below.
 
....two....weeks....lata....ish
 

Measuring

Mass

47.9g
 pcm_model_weigh1_7e275.jpeg
  pcm_weight1_20f1d.jpeg
43.1g 
pcm_model_weight2_02fa0.jpeg
  pcm_weight3_cdbc3.jpeg
41.2g 
pcm_model_weight3_87fe2.jpeg
  pcm_weight2_10802.jpeg
 43.0g
pcm_model_weight4_6f08a.jpeg
  pcm_weight4_2daaa.jpeg
 
Wow...I'm actually pretty surprised by how closely the actual masses came to the predicted. My take aways from this are:
  1. There is not substantial porosity in the part. This does not necessarily mean that all of the powder was fully melted, but  since even the high packing factor powder mixes used in these processes has  a few percent void space, it does bode pretty well. Bases on the bit of deformation visible on a couple of the camera mounts, my hunch is that they have the laser powered cranked up a bit. That will give some margin on getting density, but it makes me wonder what sort of issues they may have with risidual stresses in some parts and how they're handling that.
  2. There is not appreciable over or under printing increasing the total volume of material.
Of course, it's possible that some combination of the above things could just be canceling out, but the odds of all four parts being so close to nominal makes me think that's less likely.
 

 Dimensional

I   by no means did any sort of 'real' inspection, but I did put the calipers to a couple of key dimensions for each part to compare to the nominal "keepers".
 
The one that's most likely to impact performance is the thickness of the bed mounts. The existing spacers  are nominally 6mm....well crap....as part of putting this write up together, I discovered I somehow messed up here! Turns out, these are actually nominally 5.5mm (mine all measured between 5.52 and 5.54 millimeters).  Well damn, I guess this speaks to how good bed leveling is! I've now made several parts with these mounts installed, and apparently they aren't off by the 0.1 - 0.2mm that I thought they were (based on the measurements shown in the pictures below)...more like between 0.6 and 0.7mm.   For the models that I upload to sharing sites, I will update the model to reflect this 5.5mm length.
 
But anyway, my poor quality control issues aside, what I wanted to show with the below was supposed to be that it does indeed appear there is a little bit of overprinting, about 100-150 microns for the thicknesses of the bed mounts. 
 
 pcm_meas_bm1_81621.jpeg  pcm_meas_bm2_041e0.jpeg
 pcm_meas_bm3_e39b2.jpeg  pcm_meas_bm4_74ea3.jpeg
 
Taking a look at the camera mount, we see a comparable amount of overprinting. The extra thick boyy down in the bottom right isn't an overprinting issue. I decided not to add obstacle geometries on that face, and the solver added an interesting little ring of material here.  Since the only thing touching this face will be a fastener head, it isn't an issue. But next time I'd probably go ahead and add the blocking geoemtry just so it can be sized for a specific fastener length.
 
 pcm_meas_cm1_21f0b.jpeg  pcm_meas_cm2_baf44.jpeg
 pcm_meas_cm3_1f2c2.jpeg  pcm_meas_cm4_f5466.jpeg
 
One additional dimension I'd be curious to take a closer look at is the diameter of the camera mount, but I've noticed that they all have at least some level of circularity...challenges. Unfortunately, I'm not really equipped to get a good measure on that, so I'm just going to save that for another day.  But as you can see from the  one measurement on "Outcome 4" below, they did come out  generally close to the nominal. 
 
pcm_meas_cmd4_49258.jpeg
 

 Mechanical Testing

This is, quite honestly, the part of this project that I've been most excited about....it's also the part that definitely still has the most to go before I'll be happy with it. 
 
While I've been playing with generative design tools like the Fusion360 one that I showed above for a couple of years, I haven't had the opportunity to get real data to compare the software's prediction on the strength of the part to that of its real world counterpart.
 
DSC07529-Enhanced-NR_92ac5.jpg
 
The metric that I'm  equating to "strength" here is actually stiffness. That's something I can measure fairly one-to-one between the simulation and the physical tester. But to make that possible, I first needed a way to measure the deflection of the part as I apply load (the Mech Tester thus far has only been instrumented for load). So I rigged up a  displacement sensor that mounts underneath the mount in the tester (that's that hideous yellow fella with the super-safe razor blade stuck to it.)   
 
The big delrin ball is just there to try to keep the force applied as close to a pure Z-axis load as possible, since the ball can't transfer a moment (other than  from whatever friction there is between it and the aluminum parts on both sides).   
 
pcm_stiffness_sims1_02b0b.jpeg
 
 
pcm_stiffness_outcome3_1_aab9b.jpeg
 
 
 
 pcm_plot_out3_8b81a.jpeg
 

Design Notes

2024/02/14 - Initial stab at it

To start, I'm going to see what I'd get if I design for a GoPro, but I don't really intend to go with anything larger than a Raspberry Pi HQ camera. So we'll see how it looks with a GoPro mass as a conservative first pass.
 
According to the GoPro site, the GoPro 12, with battery, weighs in at 153g.
 
For the model, I started with just defining the contacts  for under the bed. The right one is slotted to allow the bracket to be used on either the right or left side. And, frustratingly, these are not symmetric for some reason. So it is slotted to allow for either the 105mm (left) or 110mm (right) position.
 
pcm_model1_242ea.jpeg
 
I added in the disk for the camera mount. It's basically just a bit 1/4-20 washer. I offset it 75mm from the centerline of the stage fasteners and with its bottom surface 15mm above the bottom of the bed spacers. This should put the plane of the camera mount 10mm above the print surface. This seems like a reasonable spot for a nice wide angle timelapse.
 
pcm_model2_2a069.jpeg
 
I then added in a first swing at a keep out zone. I think I covered everywhere needed, but I'm sure I'll be wrong :)
 
pcm_model3_c3668.jpeg
 
I set one face of each space as fixed.
 
pcm_model4_d4812.jpeg
 
I approximated the load of the camera mass at 2N. 
 
To get a rough number to use for the side loading due to acceleration, I got an accel rate of 1000mm/s^2 from a quick perusing of forums. So I put that with the 153g GoPro mass and got 0.153N....to be fair, I put it in MathCAD before I knew the number would be as easy as 1000...
 
pcm_math1_50ecc.jpeg
 
Again, to be conservative, I'm going to round this up to 0.5N. And I will apply two load cases, one in each direction of Y movement.
 
I'm ultimately hoping to send this to PCBWay to take advantage of their metal AM, and have this printed in Al. So I set the only Al option in Fusion's additive materials library and started the solver runnin!...now I wait...
 
 First batch of solutions are done! Overall, they look pretty good, other than I forgot to include keep outs inside of the bores. I'm also a little worried about that spindly little link to the left mount.
pcm_result1_2db87.jpeg
 
 I added some material to the keep out for the holes. I also  changed the constraint for the left side mount to be only constrained in Z. My hope is that this will force a little bit of cross-bracing.
 
I also updated the 'design objectives' to include a maximum deflection criteria. I set it at 0.01mm globally, which may be a bit too optimistic. So once I see what comes out I can adjust if needed.
 
 Well I didn't get the cross brace that I was hoping for, but this definitely looks  abit more substantial:
pcm_result2_4cc37.jpeg
 
I'm going to  call it there for today, but I exported two of the results from that last series to see how a quote looks.
  1. Generative Design Caliper Stand
  2. Generative Design SD Card (and Guitar pick, it turns out) Holder
  3. HydroHub - 4 in / 4 out distribution hub
  4. High Flow Peristaltic Pump
  5. Precision Dosing Pump
  6. Stackable/Modular Peristaltic Pump Concept
  7. Initial/Manual Pump
  8. Cam Valve - Gen3
  9. Peristaltic Valve - gen1
  10. Four Channel DrumCam Valve

Subcategories

Assorted (hopefully) "Useful Stuff" Projects Article Count:  12

Like the bucket of assorted fasteners on that bottom shelf, this category is for stuff that I didn't know how to group...oh, and speaking of those fasteners, check out the little sortin fella!

2020 Aluminum Extrusion Hardware

Quick Bolt Sorter

General Purpose Turntable - Gen1

Games Article Count:  1

Generative Design Projects Article Count:  3

During the good financial decision-making times of Covid lockdowns, etc. I decided it was a good idea to buy a license for the Fusion360 Generative Design extension...Good news, I did finally pay that off :) I had worked around, and been somewhat involved in a handful of Topology Optimization/Generative Design projects through my work, and I've found the tech super interesting for some time. So after the free trial, and feeling like I was just starting to gain some level of competence in Fusion360's tool....I done did it, and bought the year. Ok, now that I'm done justifying that to myself...I mean you...

<engineering/design> What I really like about Generative Design is that it forces the designer to think about the thing they are trying to design from it's core requirements: Forces, Interfaces, and Keep Outs. I think it's far from perfect, especially given the still very primitive Design For Manufacture capabilities these tools have (among other shortcomings, but this one is certainly a big one to me.)

<precision engineering> One last also (for now), but ALSO, what I find exciting about these tools from a precision engineering perspective, is that the above-mentioned focus on forces and interfaces, these tools are extremely well-suited to kinematic/exact constraint designs! I think every one of the Generative Design projects below features at least some aspects of kinematic constraint (I say, "I think" because I may or may not be writing this before I go through my files and remind myself what all I actually made vs what I just thought about making ¯\_(ツ)_/¯ )

Generative Design Projects

 

 

 

JFS Projects Article Count:  14

AgTech (aka finding a way to complicate and digitize gardening) Projects Article Count:  12

Hydration and Hydroponics Projects Article Count:  8

Peristaltic Pumpin Article Count:  4

A lot of projects I work on/have worked on seem to involve the controlled movement of fluids. Below is a bit of a history of my builds involving attempts at obtaining this controlled movement for incompressible fluids. I haven’t done much myself with making custom solutions for the compressible stuff, but if you’re interested in such things, I thoroughly enjoy Major Hardware’s “Fan Showdown” series :)

This article/section is by no means intended as a thorough overview on the design and operation of pumps. While I will try to give some overview on operating principles and design considerations as I go, this is mainly just going to be a wander through my personal builds and experiences.

Peristaltic Pumps

What is a peristaltic pump?

I’m sure there are innumerable sources online for (much better) detailed discussions of the workings of peristaltic pumps. So I’m just going to hit the highlights, and I’ll try to remember to find some promising links and add them below, should a deep dive seem intriguing to ya.

Basic Operation:

The fluid being pumped is carried into the pump in a compliant tubing. This tube is routed around some portion of a circular/cylindrical path around the axis of the pump and then exits the pump. This is one interesting/attractive aspect of peristaltic pumps, the fluid never has to leave the tube that it is in, making these pumps well-suited to situations where contamination and/or leaks are highly undesirable. The housing that features the cylindrical wall that the tubing is being routed along can be considered the Stator, and that is generally the nomenclature that I tend to use.

So if there’s a Stator, there must be a Rotor…? Yup, the rotor includes some set of features that extend out to some defined gap between this feature and the Stator wall. These features, which in many peristaltic pumps are rolling element bearings, pinch the tubing to the point of sealing (ideally) the tube. As the rotor turns, this contact point proceeds around the circumference. Because the pinched point of the tube is sealed, the volume of fluid in the tube ‘ahead’ of the pinch point are, as a result, pushed forward. So, keep rotating, keep pushing….pretty much as simple as that!

Pros:

  • Positive Displacement Pump
    • Because the pinch point is (ideally) fully sealing the tubing, the amount of fluid moved is directly proportional to the movement of the pump. This makes them very good choices for things like dosing pumps or other applications where the desired volume of fluid to be moved needs to be deterministic.
    • This is a large driver for my initial interest in using peristaltic pumps. Their deterministic flow is/was very attractive for my plant growth experiments. They can give very repeatable watering volumes and nutrient concentrations.
  • Fluid Isolation
    • Because the fluid never leaves the tubing, these pumps can be suitable for moving hazardous materials. For example, I have been using a peristaltic pump for transferring 99% IPA
  • Relatively Simple Construction
    • Because the fluid does not have to be sealed within the pump, these pumps lend themselves well to DIY builds. No shaft seals, gaskets, etc. or complex (at least to do well) impeller design needed.
  • Self-priming and Head height
    • If well-sealed, these pumps are capable of self-priming (and even pumping air) and of achieving pretty impressive head heights (the measure of how high above the pump it can pump a column of water)

Cons:

  • High drive torque
    • Because of the preloading needed against the tubing, and the rolling friction, even with good rolling elements (more below on this), it can be quite easy to end up with designs that require quite a lot of drive toque.
  • Tubing wear
    • With the relatively large deformation and high number of cycles, the tubing will eventually fail, either due to material wear, fatigue cracking, or who knows what else. Because this failure mode can cause fluids leaking into your pump not designed to experience this fluid, this failure can potentially be quite problematic. So the use of high-quality tubing material and a plan for periodic maintenance, are worthwhile.

Test Build 1

A couple of years back, I had a concept for an in-line-mixing hydroponics system. The idea being that the supplies to the system would be just pure water and nutrient concentrates, and a series of pumps and valves would allow precise dosing mixes to each target plant in a system (I refer to this concept as Rail Yard Hydro, since it moves the fluids around the tubing network quite like rail cars are moved around a rail system. I’m planning to add a separate page diving into that one a bit deeper since it is the design scheme I am using in my current projects.)

Well, to facilitate this plan, I wanted to find an option for a dosing pump that I could integrate in to my control system (aka Arduinos and Raspberry Pi’s :)). Unfortunately, I quickly found that a servo-driven peristaltic pump could easily set me back north of $100….so I set out to spend many multiples of that making my own!

Actually, when I saw the pricing, I decided I should see if I could make myself a cheapo, manual version that I could use to just test out some basic questions on the Rail Hydro idea (mainly verifying that I could induce good material mixing in-line and that there was no cross-contamination between fluid reservoirs.) And so, ‘twas this endeavor that resulted in the pump I’m apparently referring to as “Test Build 1”

Design Objectives:

  • Be a peristaltic pump
  • Provide a full seal (at 100mm head)
  • Be hand-cranked
  • Not require any parts that would have to be ordered (I’m impatient)

The Build

She ain't pretty (especially after a good while of getting knocked around), but the pic above shows the dual pump setup I rigged up for my testing needs. I was VERY pleasantly surprised that, other than a tweak to the hand wheel, these things worked pretty damn well!

I decided upfront that I was going to go with a resin printed build, because I thought the high stiffness and good surface finish throughout the 'pinch region' would give me a better chance. Since I was already going to have the good surface roughness, I might as well also integrate the main bearing into the printed parts.

In the image of the model, below, the Stator is the part shown in green, and the Rotor is shown in blue(ish.) Riding on the rotor are roller skate bearings to provide the contact with the tube. Race 1 has v-grooves on both sides of the race, providing the main constraint for locating the rotor, and Race 2 has a v-groove on the Stator, but only a single plane of contact on the Rotor side. This keeps from over-constraining the bearing.

The absurdly overkill bolt running through the center is a real showcase of "using what I had on hand" :) in that these were the only bolt/nut sets I had on hand with the length I was looking for.

Valving Article Count:  3

System-level Projects Article Count:  4

Project Fireplace Article Count:  4

Just Playin and Concept Demo Projects Article Count:  8

Print-in-Place Designs Article Count:  1

Flexure Fun Article Count:  5

Pet Stuff Article Count:  1

Printer Stuff Article Count:  3

Tool-related Projects Article Count:  6

WIP Article Count:  3

A temporary home for projects I'm currently working on.

Page 5 of 16