Latest Articles

Most Popular

Tool-related Projects

Mech Tester

Details
Parent Category: BubsBuilds Projects
Category: Tool-related Projects
This is one I've been thinking about building for quite some time, and I won't lie, I'm pretty excited to start breakin stuff!...er, I mean, quantifying the strength of my previous designs via destructive testing
 
 
 

Design 

Objectives

MT_model1_86328.jpeg
The main objective  here is a test cell for performing compressive and tensile testing. However, I don't just want to be able to measure standard test samples. I want to be able to test actual components and assemblies.  Over the years I have ended up with quite a few 3d printed designs for structural components, like the various  brackets I've made for assembling 2020 extrusion frames  or my recent caster wheel build, and I would love to get some actual data on how much load these sorts of structures can withstand.
 
So to accomplish this, I will need something that can:
  • Apply both tensile and compressive loads of up to 2.5kN (~500 lbf) - Admittedly, this was selected somewhat arbitrarily, but seemed achievable at a reasonable price and sufficient for the vast majority of my needs...at least for now
  • Withstand offset loads - Because I will be testing 'real' parts, I can't assume that loads will always be along a single line of action. Unlike standard test samples, which can be designed/held such that loads are  well-behaved and nominally symmetric.
  • Data collection
    • Load - I'd like to have a configuration that will support a somewhat modular load cell. This will allow for using load cells that are sized for the part under test, enabling a wide range of applied loads while still getting good signal-to-noise when needed  (for example, when testing a small bracket, I may want to only use a load cell with a total range of 100s of Newtons, but if I'm trying to test something bigger, I may want a full scale range of 3kN).
    • Displacement - At a minimum, I want to be able to measure the amount of strain being applied to the part under test. Ideally, I'm thinking I'd like to have at least three displacement measurements
      • One measurement of the strain in the part. For this I'm thinking  I'll consider the dsplacement between the load plates as this strain (the interfaces where the load will be passed into the part). 
      • Two measurements for the actuators (one for each). Ultimately I want to be able to run the linear actuators closed-loop from positional feedback.
    • Temp - Cause always. I've been working in precision stuff for too long, what can I say
    • Imaging -  In addition to wanting to be able to  get photos and video for my own enjoyment, I also would like to attempt to use an image-based system for some of the displacement measurements mentioned above. 

 Design Overview

The basic design concept is pretty similar to most dual-column tensile testers that are common to most material test labs. With a symmetric set of linear actuators and linear guides on either side, used to drive a cross bar. The part under test is attached to both the cross bar and the base plate, and the actuators are driven to  either pull  on or squish the part.  
 
 
The below image shows the load paths.
 
 
A heavy emphasis during the design was on making use of materials on hand wherever possible. Luckily, I had some large bits of aluminum scrap, as well as some 16mm linear rails and bearings left over from an old project.   in case you're curious where some of the design decisions, like the use of a substantial chunk of alumininum plate for the 'Base Plate', for example, came from.
 
That's enough overviewin, on to the subsystems and components...don't act like you aren't excited. 

Actuation

For the linear actuators for V1, I went with some low cost lead screw actuators with max capacities of 1kN each. I used these 12" (~300mm) travel actuators that I found for a little under $40 ea. I don't expect these to live all that long, since this sort of load profile is definitely not what they're designed for (especially the sudden load release when a part fails.), but  I think they'll be perfect for initial development, code testing, etc. at the very least. Then I can upgrade these at the same time I make any other significant modifications I make for V2. This being my first load tester build, I'm fully expecting to find some additions/modifications wanted after a bit of use.
 
 
 These actuators are intended to be pinned at each end with a 5mm pin, and are designed to only take a direct axial loading in either tension or compression. 
 

Linear Guides

The linear guides serve two purposes in my design.
 
The primary job is to take up all lateral loads in the primary load path to ensure the linear actuators are not subjected to side loading.
 
The secondary, but similar (and still important) role is to also take up any lateral loads that would otherwise result in the load cell carrying a moment. That would produce measurement errors, so by constraining/guiding the load cell relative to the cross bar, the load cell's reading is insured to be along the prescribed axis.
 
I opted for these 16mm x 700mm steel rails that I purchased extras of for a project a while back. I  mounted the rods rigidly into the cross bar (more detailed descriptions of the machined components below.) I paired these with two sets of LMK16UU linear bearings. One set mounted into the Load Plate, and the second set mounted in the Base Plate.
 
The below cross-section shows the bearing configuration a bit more clearly.
 
MT_xsec1_c63b8.jpeg
 
The fear with this sort of configuration is that the axes of these bearing holes need to be  tightly toleranced to avoid binding. The benefit though, is that if right, they can be very stiff and don't require any fine tuning or adjusting.
 
So I decided to put some real faith in my little Precision Matthews mill, and go for it. However, to give myself the best shot possible, I wanted to make sure the spacing of the bearing holes was within the travel of the mill so that each set could be cut in a single setup.  
 

Load Cell

For now I'm using one of these 300kg load cells (it upsets me deeply that they class these things by mass...it needed to be said).  
 
mt_loadcellstockphoto_91504.jpeg
 
I had a lot of trouble trying to find a datasheet for the "PSD-S1" model number marked on the load cell itself. But I found this TAS501 datasheet that appears to match up in all of the metrics I have available (and looks identical). So my assumption, until proven otherwise, is that this is just a relabeling of the same generic load cell.
 
mt_loadcelldatasheet_97c15.jpeg
 
 

Axial Flow Compressor Concept Tester

Details
Parent Category: BubsBuilds Projects
Category: Tool-related Projects
I frequently run into situations when working on things in the lab (aka my basement) where I want/need compressed air. But I've long hated  the heart stress testing that is  the standard compressor starting up. I've considered things like putting the compressor in a shed and plumbing air to the shop and lab (garage and basement, respectively :) .) Or building an enclosure that could isolate the noise and vibration, while still managing the heat and air intake.
 
The main use cases that I currently have in mind are 1) cleaning/drying resin parts after wash and 2) air bearings. The first does not need a ton of pressure, but does love some flow rate. Currently I'm using an airbrush compressor that tops out at around 30psi, and no clue on the flow rate. The latter would ideally have very low flow, and high pressure, but because the air bearings in question are largely experimental and 3d printed....kinda would love both :)   But I think if I could supply 50+ psi at a couple cfm it would be enough for testing purposes.
 
But I've also been kinda kickin around an idea for a continuous-operation compressor that may not be quiet, per se, but should be much quieter than my current options. The article title kind of gives it away, but the idea is based on the axial flow compressor stages in a jet engine. So basically just try to duplicate(ish) the left half of the below (although I just noticed that their sketch is missing the stator blades...so pretend those are there too.)
Components of jet engines - Wikipedia
 
 

Objectives

Ultimately, what I'd  like from this compressor is the ability to provide 50+ psi at a flow rate of at least 3 scfm. But in the shorter term, I'm just curious if a 3d printed (or mostly) axial flow compressor can achieve any meaningful pressure increase at all.
 
So objectives for this concept tester are:
  • Modular, axial flow compressor testbed that allows for relatively easy exchange of rotor and stator components to test geometries
  • Data outputs
    • Pressure measurement(s) - Would be great to  find a way to have pressure taps at each stage, but this may not be feasible
    • RPMs - Positional measurement, nor direction of rotation are needed, so this can be a pretty simple encoder
    • Power Consumption -  To ensure this isn't just absurdly wasteful. My expectation is that it will be significantly less energy efficient than my current piston-driven counterpart, but no clue what to expect really. Doesn't seem reasonable to baseline against standard axial flow compressors, since their tolerances, aero-optimisation, and operating regime are just SOOO far from what I will be working with.
    • Flow rate - Measure of flow rate out of main tap.   Ideally, a very low impedance  measurement at the exit valve of the compressor to try to maximize range of pressure to flow rate trade-off.
 
 
 
 
 Design Notes
2024/02/10 - First pass
Starting out with  trying to make a basic (although I'm sure extremely inefficient) airfoil.
afc_airfoil1_sketch1_5da7f.jpeg
 
My plan for the first iteration is to use my trusty print-in bearings, with BBs (4.5mm steel balls) as the rolling elements. The rotors will be fixed to the bearing at the OD, instead of the traditional ID hub. So the bearings will be around the perimeter, with one race of each bearing being in a rotor, and the other in a stator.  The below shows the sketch cross-section I've got as a first pass, although these may change as I try to actually make this thing 3d :) I think a big perk of going with this style of bearing arrangement for this project is that the gap between the rotors and stators will be set at each bearing, and so won't be subject to stackup through the system.
 
afc_housing_sketch1_df08f.jpeg
 
 I realized I probably should have gone ahead and included the stator in the initial sketch. I am definitely swaggin it here, so I just copied the rotor airfoil and positioned it in a relative location that...looks reasonable 
afc_airfoil_sketch2_6dd0e.jpeg
 
 Revolving those, hopefully it makes it a little easier to see what I'm goin for here. The left side will be the rotor, and the right the stator. The little tab at the top of the rotor ring is intended to offer the option to put an encoder on it, but that doesn't make sense on all rotor stages, so I may revise that. My plan is to have the stators rigidly fixed to each other with bands around the exterior to :
  1. The main purpose is mechanical. It will be taking the axial loads trying to separate the stages, the torsional loads from the drag on the stators, as well as any radial loads due to balancing and such.
  2. Sealing. These outer bands will keep the pressure gradiant across the bearings to only be the gradiant between stages, instead of allowing it to vent directly to atmosphere. This should help reduce losses (although surely it won't stop them entirely....assuming I'm lucky enough to get any pressure to build up in this little thing anyway)
 
afc_t1_model1_ee95f.jpeg
 
 
Random thought, but I wonder if it might be worthwhile to try putting a standard PC box fan on the intake of this thing. They are quite good at providing air flow and maybe having this as a feeder to the first stage would be worthwhile...somethin to consider if it seems like it's being starved at the inlet (which is something I'm a smidge concerned about.)
 
Ok, so back on track. I struggled with the Fusion360 Project on surface feature for quite a while and decided to give up and go cludge. The only real issue is that now I have little bits of blade sitting in the V groove. But I'll just do a subtract revolve at the end to clean it up. Not ideal, but it'll do. In hopes of helping to improve packing, I made a smaller, higher angle of attack sketch  of a similar airfoil shape, and then just lofted the larger one  to the smaller for both Rotor and Stator.
 
afc_t1_model2_7890a.jpeg
 
Ok....I'm officially excited....
 
afc_t1_model3_4dc8f.jpeg
 
No clue if this thing will work, but I already love how it looks...my bar, she's low.
 
Ok, got the Vs cleaned up. I'm going to consider these kind of the base 'unit cells' for all of the stages. The tangled mess in the middle will be replaced by either the hub on the rotor or an aperture on the stator. The stages will need to have a progression in the diameters of both of these along the direction of flow. My hunch is that this change in diameter is going to be the main thing I'm going to need to tweak down the line. 
 
afc_t1_model4_2711d.jpeg
 
I am far from a fluids person (as you have probably already seen clearly from my approach to airfoil 'design' :) ), but my fear is that since I don't really have any idea what kind of pressure ratio I might be able to get across a stage, I suspect I'm going to get the compression in the volume wrong and stall the flow...I think that's a thing...right? While I'm speculating, the other parameter I'm already thinking about wanting some flexibility in is the attack angle of the stators.  I'm wondering if I could figure out a not-too-complex option for getting some ability to adjust this...but, that's for a future rev (if ever).
 
Time to start crackin on some first attempt stages!
 
afc_t1_fs_model1_96e13.jpeg
 
Since I'm building off of the Final stage, and since I should be able to stack any number of stages, I'm going to use "fs-#" as the naming scheme for the stages. So the final stage, with the motor attached is fs-0, the one right before that is fs-1, etc.
 
Starting at fs-3, I'm switching to a slope of 2mm per stage. 
afc_t1_fs-3_sketch1_bd227.jpeg
 
Alright, I tossed fs-0 through fs-4 into an assembly even though they're still incomplete....I'm impatient and thought it was gonna look cool... I wasn't dissappointed
 
afc_t1_model5_cdcd0.jpeg
 
 The stairstepping along the inner diameters inside the hubs throws off the look a little, but that isn't in the flow and I still need to add the couplings/hubs to the rotors...in hindsight, that would have been smart to do before making the other parts....oh well, I'm gonna make the rest and close it   ¯\_(ツ)_/¯
 
 I'm also going to increase the pitch again to 3mm per stage starting at fs-5.
 
And with that, this config  seems to run out of room at fs-6.
afc_t1_model6_f50a1.jpeg
 
The total length is a little over 350mm
 
It's gettin late, and I'm runnin out of words...you're welcome, but I got a first pass at the hubs and simple linkages.
 
afc_t1_model7_274ab.jpeg
 
The hubs just have a "+" shaped cutout, and the linkages are matching extrusions with relatively loose running fits. The thought is, all I need is to transfer the torque, I don't really care about backlash, etc. (I don't think). And the fastest motor I have in this form factor is 550 RPM, so I don't expect balance issues to be too significant.
 
The next thing I'll need to do is add the structural elements to hold it all together, but that's gonna have to wait for tomorrow.
 

Rotary Table

Details
Parent Category: BubsBuilds Projects
Category: Tool-related Projects

WIP - PAGE/PROJECT IN THE WORKS  - WIP

 I've come across more than a couple of occassions recently in which I found myself wanting to controllably spin stuff while tinkering on projects. The best tool at my disposal for this task was my "little" 6 inch rotary table for my mill.

 While  it's certainly up to the task, it's a heavy little bastard! So to keep me from lugging it up and down the stairs from the shop to the 'lab', I decided I should take a stab at a DIY, lighter-duty variant. 

Design

OnShape CAD files available to do with as you please here.

RotTable ann1Asset 2 100 314fd

The above image shows the major externally-visible components. The Base is a printed part that  provides not  only the structural support for loads from all of the other components, but also  is responsible for controlling the alignments of the motion components. 

On top of the Table is the kinematic mount, formed by bar magnets set at 45 degrees from the table surface (so a 90 deg V.) If you haven't come across this style of interface before, you can find general info if you give a Google to "Maxwell clamp" or "kinematic coupler."   I included a very similar coupler on my passive rotary / overkill papertowel holder build  , as well as a great many of projects in my day job over the years...they're just tops :) hopefully seeing the repeatability numbers down below will start bringing you around to sharing my fondness for 'em.

 One thing I decided to add that my machining rotary is missing is a power feed!  In addition to the hand wheel for manually driving the table, a DC motor can be coupled to the other side of the drive mechanism.  There also should be plenty of space to swap out this DC motor for a stepper, should I  decide  to down the line. 

 The Table is driven by a worm gear drive housed inside the base, as shown in the below image. The worm gear is integrated into Hub component, which attaches to the Table via three M5 fasteners.  Sandwiched between the Hub and Base is a needle thrust bearing.  This thrust bearing's counterpart is the main bearing  that is integrated into the surfaces of the Base and Table. V-groove channels in the top of the Base and the bottom of the Table form the bearing races, and the rolling elements are 9.5mm steel balls (aka 3/8 inch slingshot ammo....hey, it's quite round and damn cheap). 

RotTable xsec1Asset 3 100 92f41

rt_model_underside1_84303.jpeg

 

 On the hand crank side of the table, a printed coupler sits inside the "hand wheel horn", held in place with a roller skate bearing (I used these. They're overkill for this, but they run really smooth.) The hand wheel presses onto the coupler with a light interference fit. It's really just intended as a close running fit, but with 3d printed parts it ends up being more like  a bit of an interference fit.

 

RotTable xsec1Asset 3 100 1877e

 

On the power feed side is a similar printed coupler, but with a d-hub  hole for engaging the motor shaft and no support bearing (the motor has a bushing that's serving the same purpose.) For a motor, I'm using one of these 24VDC, 150RPM motors.  Something a little faster might be nice, but I might worry about the wear on the worm. Plus I don't really need this thing to be  speed demon, so  I don't have any plans to swap it.

RotTable exp1Asset 5 100 48786

 

 Build

BOM

 

Measuring runout:

Just for the sake of making sure there's no uncertainty with what I mean by runout, allow me to elaborate...probably to much. Feel free to skim. When I say, "runout", what I I'm referring to is any movement of the Table, relative to the Base, other than the desired spinny bit. There are three types of runout that I want to measure on my rotary table:

 Radial Runout

Radial runout is the side to side movement. It is the error motion radially out from the rotational axis. The below gif is an attempt to show what I mean, but just imagine the base isn't floating all over the place. Apologies for the seemingly lazy animation...I say seemingly, cause it sadly took me a fair bit. So that's as good as you're gettin! 
 

Axial Runout

Axial runout  is error motion that is along the axis of rotation, or 'up and down' as I'd probably refer to it if just pointing at it and talking   ¯\_(ツ)_/¯.  But hey, in good news, my gif-making skills definitely improved a bit for this one. So by Axial I bet it's gonna be fuckin epic. 

Angular Runout

Ha, just kiddin, it's even shittier...Oh, right, angular runout. Yeah, that's the tip/tilt one. Basically either of the two rotations that you didn't request of the table.

 

Measurement Setups

In order to obtain the measurements I'm after, it will require nine sets of measurements. Each set of measurements will include measurements recorded at consistent intervals throughout the 360 degree travel of the table. So why nine? What seem like an extra six will (hopefully) make a bit more sense shortly, but they're going to help us make sure we're measuring our actual motion and not the surface/shape of the surface we're measuring.

The nine measurements will require the indicator/sensor being mounted in three positions. Or, if you're flush with cash, just get yourself three sensors and a nice frame for em...while you're at  it, send me a set too :) 

Position 1 - Low on the post, near to the mount face

Position 2 - 50mm up the shaft from Position 1

 

Position 3 - On top and centered on end of rod.

 

The remaining six measurements, are repeats of these positions at  each position of the kinematic mount. I'm using the kinematic mount here to perform a variation on the "reversal technique" to cancel out part geometry. If you aren't familiar, but are interested, or just want to see if I'm totally bullshitting you, you can find plenty of info out there by giving something like, "reversal technique metrology".  The kinematic mount allows us to repeatably position the test sample at 120 degree offsets.  I'll show below in the data analysis how I then used these offset datasets.

 

 

Mount Repeatability

 

 

WIP - PAGE/PROJECT IN THE WORKS  - WIP

 

 

 

  1. Heat set Helper
  2. Toolbox Prints

Page 1 of 2