1. For downloading SimTools plugins you need a Download Package. Get it with virtual coins that you receive for forum activity or Buy Download Package - We have a zero Spam tolerance so read our forum rules first.

    Buy Now a Download Plan!
  2. Do not try to cheat our system and do not post an unnecessary amount of useless posts only to earn credits here. We have a zero spam tolerance policy and this will cause a ban of your user account. Otherwise we wish you a pleasant stay here! Read the forum rules
  3. We have a few rules which you need to read and accept before posting anything here! Following these rules will keep the forum clean and your stay pleasant here. Do not following these rules will lead to permanent exclusion from this website: Read the forum rules.

Stress Testing 3D Printed Parts

Discussion in '3D Printing' started by Hoddem, Feb 20, 2018.

  1. Hoddem

    Hoddem Well-Known Member

    Joined:
    Sep 27, 2016
    Messages:
    358
    Occupation:
    Engineer
    Location:
    Minnesota, USA
    Balance:
    3,047Coins
    Ratings:
    +599 / 2 / -0
    I didn't know exactly where to put this, but I wanted to document the process so I am going to put it here. When I started building my linear actuators I had a lot of people questioning the strength and safety of the 3d printed parts. I didn't have a scientific way to strength test the parts, but I felt they were plenty strong enough. I have since decided that I would like to know how strong the parts are and determine how to make them stronger.

    My plan is to convert a cheap hydraulic press into a makeshift compression tester.

    Goals
    1. Safety
    2. Keep project cost around $200.
    2. Pressure and positional data.
    3. Data logging for review later.
    4. Retain full use of the press for actual shop work

    Parts
    1. 12 ton hydraulic press ( $100 with coupon at harbor freight)
    2. Arduino Uno R3 (Free or 9.99 amazon)
    3. protoshield (Free $4 amazon)
    4. Elegoo 2.8" TFT touch Screen ( $14.99 amazon)
    5. ADS1115 16bit ADC converter ($6.99 amazon)
    6. TE M3031-000005-05KPG transducer ($68.32 mouser.com)
    7. 3590s-2-502L 5k 10 turn pot (11.99 amazon
    8. Retractable reel with metal clip (Free or 5.99 amazon)
    9. 8 x AA battery box with on/off switch (Free or $4 amazon)

    I picked up the press a few days ago and have that already assembled, I did my first test just to see and I was quite surprised. This was a PLA part printed nearly solid and when it failed it did so very violently. Luckily I was protected and anticipating the failure, but a safety guard will be a must for future testing.




    ADS1115.jpg Proto.jpg Batt.jpg Sensor.jpg Reels.jpg Pot.jpg press.jpg TFT.jpg uno.jpg
    • Informative Informative x 4
    • Like Like x 2
    • Winner Winner x 1
    • Creative Creative x 1
    Last edited: Feb 20, 2018
  2. Ads Master

    Ads Master

    Balance:
    Coins
    Ratings:
    +0 / 0 / -0
  3. RandomCoder

    RandomCoder Active Member Gold Contributor

    Joined:
    Feb 19, 2017
    Messages:
    185
    Occupation:
    Control Engineer
    Location:
    UK
    Balance:
    1,626Coins
    Ratings:
    +138 / 0 / -0
    @Hoddem can I also suggest videoing the failure of the part so that it can be viewed in slow motion. It might be interesting to see where the part fails.
    • Agree Agree x 1
  4. Hoddem

    Hoddem Well-Known Member

    Joined:
    Sep 27, 2016
    Messages:
    358
    Occupation:
    Engineer
    Location:
    Minnesota, USA
    Balance:
    3,047Coins
    Ratings:
    +599 / 2 / -0
    absolutely, I am hoping I can setup a dual camera system where I can video the part up close and then superimpose the digital readout so all the information is easy to see.
    • Like Like x 2
  5. Hoddem

    Hoddem Well-Known Member

    Joined:
    Sep 27, 2016
    Messages:
    358
    Occupation:
    Engineer
    Location:
    Minnesota, USA
    Balance:
    3,047Coins
    Ratings:
    +599 / 2 / -0
    I got the first batch of parts today and started writing the arduino code. I just did a quick mock-up with the ASD1115 and a couple of 1/4 turn 5k pots. I'm not a programmer so I worked through the sample sketches for each function I will be using (LCD,Touch, SD Card, ASD1115). Each worked well enough on there own, however when I tried to merge the LCD sketch and the ASD1115 sketch I had some problems. turns out that the ASD1115 uses analog pins 4 and 5 and the lcd used analog pin 4 as the reset pin. so every time I try to read the values from the ASD1115 the lcd resets. I ended up clipping the lead on the LCD and rewiring it to the arduino reset pin. I also had to change the reset pin in the lcd sketch. Other then that everything is going pretty well, just working through some formatting issues before I merge in the touch and SD logging.

    Here is what the screen will look like. Black background and red text, its hard to tell in the picture.
    1.jpg


    I am still waiting for a few parts (mainly the pressure sensor), in the mean time I will clean up all of my wiring on the arduino and keep plugging away at the code.
    • Like Like x 6
  6. Zed

    Zed VR Simming w/Index Gold Contributor

    Joined:
    Apr 4, 2017
    Messages:
    867
    Location:
    USA
    Balance:
    4,726Coins
    Ratings:
    +869 / 3 / -0
    My Motion Simulator:
    2DOF, DC motor, JRK
    From the looks of it though, that part took a lot of pressure before it failed. Will be interested to see the numbers. Very cool!
    • Agree Agree x 1
  7. Hoddem

    Hoddem Well-Known Member

    Joined:
    Sep 27, 2016
    Messages:
    358
    Occupation:
    Engineer
    Location:
    Minnesota, USA
    Balance:
    3,047Coins
    Ratings:
    +599 / 2 / -0
    I have all of my parts now so I soldered up my uno shield to accept the ADS1115 board. I also added some screw terminals for the battery box and the two sensors.

    12.jpg

    13.jpg
    I also made the final modification to the LCD shield to eliminate the reset pin issue.

    11.jpg

    I think I have the software side finished up, not going to lie it was a struggle over several nights. Here is the final display configuration.

    14.jpg

    When the log button is pushed a text file is created on the SD card and Time, Pressure, Max Pressure and Inches are all logged every .1 seconds. A second push stops logging and a third will start again with a new file and so on. The files are easy enough to bring into excel so that I can plot the data for future use.

    This was the part of the build I had planned on struggling with, the rest is mostly mechanical stuff so if I can find the time it should go pretty smooth.

    Things to do
    Print an enclosure
    Build a 3d printed string pot (I will be designing my own so that I can adjust the diameter of the wheel)
    Install the pressure sensor in bottle jack
    Destroy stuff!
    • Like Like x 2
  8. Hoddem

    Hoddem Well-Known Member

    Joined:
    Sep 27, 2016
    Messages:
    358
    Occupation:
    Engineer
    Location:
    Minnesota, USA
    Balance:
    3,047Coins
    Ratings:
    +599 / 2 / -0
    I completed the dirtiest part for project, installing the pressure sensor.

    First I disassembled the jack by removing the large nut on the top. It was on pretty good, but I have 24" pipe wrench that made short work if it. This is with the piston, cylinder and housing removed. I drained as much of the oil out as possible before disassemble.
    1.jpg

    Then I drilled a port with a 1/8" drill bit, I had to drill down about 1/4" from the top and then drill in from the side to get a clean hole from the pressure chamber out to where the pressure traducer will be. I did pretty good for hand tools, but I missed the mark by about 1/6" of an inch. In the picture you can see the gold of the drill bit showing that the holes overlap.

    2.jpg

    Next I chased the outer hole with a 7/16 drill bit and then ran my 1/4 npt tap through. My tap was too long so I had to shorten it after each run, one time use now.

    3.jpg

    Here is the pressure transducer installed, I used Teflon tape on the threads prior to assembly.

    4.jpg

    Before re-assembly I needed to measure the inside of the cylinder so I can do conversion calculations from pressure t0 lbs. I took several measurements and they all ranged from 45.09mm to 45.12mm

    5.jpg

    Finally here is the finished assembled jack, I haven't had a chance to test yet because I need to put the oil back in and bleed it. when I re-assembled it I moved the jack handle to the side, that will give me a good location to mount my string pot and a guide for the pressure transducer wires.

    6.jpg

    Here are my calculations for the conversion from PSI to LBS in the arduino.

    2018-03-01 11_57_37-Greenshot.jpg
    • Like Like x 4
    • Creative Creative x 1
  9. RandomCoder

    RandomCoder Active Member Gold Contributor

    Joined:
    Feb 19, 2017
    Messages:
    185
    Occupation:
    Control Engineer
    Location:
    UK
    Balance:
    1,626Coins
    Ratings:
    +138 / 0 / -0
    @Hoddem that's a quality job, you did well to line the holes up so well. But what made you choose to use a fluid pressure transducer instead of a piezoelectric transducer? The latter would have been more simple to install.
  10. Hoddem

    Hoddem Well-Known Member

    Joined:
    Sep 27, 2016
    Messages:
    358
    Occupation:
    Engineer
    Location:
    Minnesota, USA
    Balance:
    3,047Coins
    Ratings:
    +599 / 2 / -0
    That’s a good question, like most projects this one evolved rather quickly in my head. The original plan was to use a hydraulic pressure gauge and just tap into the jack. No electronics were to be involved and I planned on just converting the psi reading to lbs in my head. Once I had that idea I couldn’t get it out of my head and it kept growing until I ended up with a display and a pressure transducer ("Scope Creep"). I really liked the idea of taking a cheapo bottle jack and converting it to read out in lbs. in fact that was actually the part of the project that got me to pull the trigger.

    I did look at using a load cell, but the cost from my quick search was pretty high.

    FYI, to get an idea of how my mind works.
    I needed a storage shed so I did some googling for ideas, I came across some simple plans and set out to build a shed. Before I even marked the location I decided I needed to add solar power, not long after that I decided the door needed to be automatic. Pretty soon I was building a shed so that I had a reason to build a solar powered shed door ("Scope Creep").

    Hopefully you guys can appreciate this



    I use the term "scope creep" all the time at work but I just googled it, sounds like every single DIY project I have ever done. :)

    Scope creep (also called requirement creep, function creep, or kitchen sink syndrome) in project management refers to changes, continuous or uncontrolled growth in a project's scope, at any point after the project begins. This can occur when the scope of a project is not properly defined, documented, or controlled.
    • Informative Informative x 1
  11. Hoddem

    Hoddem Well-Known Member

    Joined:
    Sep 27, 2016
    Messages:
    358
    Occupation:
    Engineer
    Location:
    Minnesota, USA
    Balance:
    3,047Coins
    Ratings:
    +599 / 2 / -0
    Tested the jack and all seems good, It looks like the jack naturally has some leakage and when under pressure it slowly creeps back.

    Here is a test with only the pressure transducer hooked up.



    The arduino updates all of the values and logs every .1 seconds, each quadrant of the screen is updated every 4th cycle. That's why the max pressure seems to lag behind the current pressure and values jump a little bit. Turns out the screen is pretty slow so if I try to update everything at once it takes almost half a second, in order keep my logging cycles up I had to split the screen updates into quadrants.

    I had my multi-meter hooked up and the values are tracking according to the datasheet, I will need to come up with a way to sanity check the values.

    I did a second test and tested the logging function. Here are the two graphs from that run, one is using the sample number as the X axis and the other is using the time as the X-axis. Once I get the string pot hooked up the values will be plotted using travel as the x axis.

    2018-03-01 20_45_00-TEST18.csv - Excel.jpg
    2018-03-01 20_44_31-TEST18.csv - Excel.jpg


    Things to do
    Build a 3d printed string pot (design is done)
    Destroy stuff!
    • Like Like x 4
  12. RandomCoder

    RandomCoder Active Member Gold Contributor

    Joined:
    Feb 19, 2017
    Messages:
    185
    Occupation:
    Control Engineer
    Location:
    UK
    Balance:
    1,626Coins
    Ratings:
    +138 / 0 / -0
    For a sanity check you could use a set of bathroom scales. You would need to make a small wooden frame so that the scales can sit on top of the Jack and apply pressure against the scales.

    I can relate to"scope creep" as my desire to build a motion sim has resulted in me needing to build a workshop. But this required the garden completing (chance to play with a mini-digger which I highly recommend if you ever get an opportunity!). Then the 'workshop' turned into a second-hand prefab garage too big for the garden but ideally sized after removing the garage door and re-arranging the panels. Then the roof needed upgrading etc etc... I guess you know how the rest goes! hug:
  13. Hoddem

    Hoddem Well-Known Member

    Joined:
    Sep 27, 2016
    Messages:
    358
    Occupation:
    Engineer
    Location:
    Minnesota, USA
    Balance:
    3,047Coins
    Ratings:
    +599 / 2 / -0
    @RandomCoder Good to know I'm not alone ;)

    I finished up the assembly and electronics.

    Battery pack mounted, I had to ditch the nice battery pack with a switch and cover because it was just a hair too large. luckily I had this one in my pile of stuff.

    IMG_6917.jpg

    Electronics and box are mounted to the front of the display with wires running in from the back. I also added a little toggle switch to turn whole system on.

    IMG_6914.jpg

    My custom designed string pot mounted and calibrated.

    IMG_6915.jpg




    IMG_6916.jpg

    So far everything works as expected.

    A couple of issues that I am working on are
    1. Running the arduino off battery gives me different ADC values than using USB power from the computer. I need to measure the 5v supply when using each source to see what the difference is and compensate for it.
    2. The return springs are counter-acting the jack so I need to add distance based compensation to overcome that.
    3. There are losses and creep within the bottle jack due to the pump valve and friction. I have a floor scale that reads up to 400lbs, so I am going to try to "calibrate" the system, but that is a relatively small portion of the scale.

    In hindsight as @RandomCoder pointed out a more common scale method may have been a better direction. I have been looking at that option a little bit more, but there are complications associated with that method also.
    • Like Like x 3
  14. RandomCoder

    RandomCoder Active Member Gold Contributor

    Joined:
    Feb 19, 2017
    Messages:
    185
    Occupation:
    Control Engineer
    Location:
    UK
    Balance:
    1,626Coins
    Ratings:
    +138 / 0 / -0
    Wow :thumbs
    That's some serious engineering from what I initially thought you had in mind. Excellent work! You sure don't do things by half measure. ;)
  15. Hoddem

    Hoddem Well-Known Member

    Joined:
    Sep 27, 2016
    Messages:
    358
    Occupation:
    Engineer
    Location:
    Minnesota, USA
    Balance:
    3,047Coins
    Ratings:
    +599 / 2 / -0
    I spent a little time playing with my floor scale and trying to compensate for all of the variables I have run into, but I didn't have much success. My floor scale only runs for a short time before it turns off so I can't really do any extensive testing. I made some corrections and found I was reading about 35lbs heavier on the jack then the scale (limited testing). I was able to implimen't a spring compensation that for the most part works well, but once I start loading the jack the other forces start to take over and so far I cant accurately map them. Long story short I think I will have to buy some load cells and start down that path. I found a couple of 2000kg load cells on amazon for around $40 and with a HX711 amplifier I will be in for under $100 not including building a frame.

    On a side note, I decided to destroy a part anyways knowing that the values are off by around 35lbs (again limited testing). This is half of a 3 way hose connector that I printed up for a friend. It is printed at 0.1mm layer heights and printed solid.



    The load cells are from china so 4-6 week wait for them If I do pull the trigger.
    • Like Like x 4
  16. Hoddem

    Hoddem Well-Known Member

    Joined:
    Sep 27, 2016
    Messages:
    358
    Occupation:
    Engineer
    Location:
    Minnesota, USA
    Balance:
    3,047Coins
    Ratings:
    +599 / 2 / -0
    So after a long hiatus I am finally getting back to this project. I ended up admitting defeat with the bottle jack and switched gears to good old load cells. I picked up a couple of 2,000KG load cells and some HX711 ACD converters. My plan is to build a stage that will sit in the jack frame, The touch screen interface will be the same just utilizing the load cells instead of the pressure sensor.

    upload_2019-1-3_14-23-29.png

    I also picked up screw drive jack, my thinking is that the bottle jack will still exhibit the pulsing output where as the screw drive bottle jack should be a much more steady load.

    upload_2019-1-3_14-30-39.jpeg