Project pitch: DIY dosing pump with 3d printed peristaltic pump head, R-Pi controlled

Discussion in 'Do It Yourself (DIY)' started by Mandelstam, Dec 26, 2017.

  1. Mandelstam

    Mandelstam Well-Known Member

    Joined:
    Oct 29, 2017
    Messages:
    565
    Likes Received:
    888
    Location:
    Malmö, Sweden
    Couple of weeks ago I got an idea in my head that I've been playing around with. The idea was to make a dosing pump that's Arduino/Raspberry Pi controlled. I just like the idea to be able to customize it completely and have total control over it. And maybe be able to save some money as well.

    My plan is to design and then 3d-print the peristaltic pump heads and mount them on stepper motors, either Nema 14 or Nema 17. I have a 3d printer at work that I can use. For rollers the idea is to use standard steel ball bearings to achieve a good low friction movement against the hose inside the head.

    Printing your own pump heads gives the possibility to use different hose thicknesses depending on the amount of flow needed. Using a thinner hose gives a higher precision on smaller volumes. Thicker hose means quicker dosing of larger amounts. For example, a hose with id 4mm holds 3.7ml of fluid per 10cm. A hose with id 1mm holds just 0.25ml per 10cm. With a stepper motor that would give the possibility to dose even very small amounts equally spread over a whole day in theory. Using micro stepping, a stepper motor is virtually silent as well so it won't be a nuisance if it's running for longer periods of time.

    I just did a quick run down of costs for all the materials needed and for a 6 head dosing pump (as an example) the cost would probably be a bit under $200. Including pumps, controller (R-Pi + drivers), housing. That's about double of a Jebao of the same size and under a fourth of what three DOS pumps would cost you. Quality wise it would be a lot closer to a DOS than a Jebao.

    Before I run any further down the rabbit hole with this project I just wanted to run it past all of you. Anyone done a similar project? Any pitfalls you can see? Any functionalities that you would add or that you are missing from commercial dosing pumps? My own experiences with dosing pumps are very limited at the moment (=zero :rolleyes:)..
     

  2. glb

    glb Valuable Member R2R Supporter Reef Squad R2R Excellence Award Partner Member 2018

    Joined:
    May 12, 2013
    Messages:
    6,091
    Likes Received:
    1,853
    Location:
    Miami
    Sounds fascinating! I’m not sure how much money you’ll save though. How many pumps would you make for this cost?
     
  3. Mandelstam

    Mandelstam Well-Known Member

    Joined:
    Oct 29, 2017
    Messages:
    565
    Likes Received:
    888
    Location:
    Malmö, Sweden
    There are some costs that would be pretty much the same whatever the amount of pumps would be, like the controller and the housing. My example was for a 6 header dosing pump, so 6 pumps with one controller for about $200. Every pump is about $20 or so, so if you make 3 header dosing pump it would be around $140. Rough estimate.
     
  4. Shameless_Dood

    Shameless_Dood Active Member Build Thread Contributor

    Joined:
    Oct 10, 2016
    Messages:
    175
    Likes Received:
    153
    I've looked at doing something similar to this. I think the trick will be getting consistent dosing out of it. Especially if you are making your own.

    A quick search of eBay and I found dosing heads with motors for 7.99. God knows what kind of accuracy they'd get.
     
  5. Mandelstam

    Mandelstam Well-Known Member

    Joined:
    Oct 29, 2017
    Messages:
    565
    Likes Received:
    888
    Location:
    Malmö, Sweden
    Yea that's why I've been thinking on making my own heads and supply them with stepper motors. Stepper motors are way more accurate for this kind of thing and if you have your own controller it's easy enough to calibrate them if it's necessary. Most cheaper dosing pumps use regular small dc motors. With them accuracy is much harder.
     
  6. Ranjib

    Ranjib Valuable Member R2R Supporter R2R Excellence Award Build Thread Contributor

    Joined:
    Apr 16, 2016
    Messages:
    2,620
    Likes Received:
    3,685
    Location:
    Foster City, Bay Area
    for reef-pi I am going with cheaper DC motor based dosing pumps for the beginning, then I intend to switch to stepper motor based dosing pump. I am not too worried about the dosing volume, as long as I dial it very slow and monitor things, that will be enough to keep things safe. My only advice will be to get the end to end thing working first, then optimize individual bits. Which means getting housing , drivers etc done first. For the first run I'll recommend going with some pre-made stepper motor based dosing pumps, that way you can focus on 3d printing housing and getting the electronics + code sorted first. Then focus on only one the stepper motor head.

    I use pca9685 for pwm , which gives me 16 channels, enough for LED, dosing pumps and DC power head based wave makers. There are other cheaper options as well (pca9685 breakout board will cost 12$). The workload is small enough to be powered by a pi zero.

    Godspeed
     
  7. Mandelstam

    Mandelstam Well-Known Member

    Joined:
    Oct 29, 2017
    Messages:
    565
    Likes Received:
    888
    Location:
    Malmö, Sweden
    Some good advice Ranjib! Thank you! And yes, I'll probably just start with a cheap stepper motor pump and sort out the other hardware and coding before moving along. I have to learn Python too so it will probably take me a while to sort that thing out.
     
  8. Ranjib

    Ranjib Valuable Member R2R Supporter R2R Excellence Award Build Thread Contributor

    Joined:
    Apr 16, 2016
    Messages:
    2,620
    Likes Received:
    3,685
    Location:
    Foster City, Bay Area
    You are welcome sir :) . Adafruit has an excellent tutorial on pca9685 and they are all in python. Let me know if you need a hand with the coding.
    When I started with reef-pi I didn't know whole lot about electronics either... the best part about the hobby is it gives us ample time :)
     
  9. Mandelstam

    Mandelstam Well-Known Member

    Joined:
    Oct 29, 2017
    Messages:
    565
    Likes Received:
    888
    Location:
    Malmö, Sweden
    Awesome! I'll start a real project thread when it's a bit more hands on. I will probably need help sooner or later lol. My girlfriend studies to become a software engineer but refuses to learn Python for some reason. Well I don't blame her as long as she's in school and have enough of other languages to learn.

    We are hopefully getting a laser cutter at work and if we are I'm probably going to design and cut the finished housing with that out of acrylic.
     
  10. Stigigemla

    Stigigemla Active Member

    Joined:
    Mar 20, 2015
    Messages:
    208
    Likes Received:
    108
    Location:
    sweden
    I dont think its a good idder to 3d print your pump heads. It is small margins to get tube to work good and thermoplastics often will change measurements if it is under pressure. Buy heads made of nylon and rely on tjem instead.
     
  11. Mandelstam

    Mandelstam Well-Known Member

    Joined:
    Oct 29, 2017
    Messages:
    565
    Likes Received:
    888
    Location:
    Malmö, Sweden
    What pressure do you mean? I can't really see any real pressure working in a peristaltic pump.
     
  12. T Carey

    T Carey Member

    Joined:
    Apr 28, 2017
    Messages:
    19
    Likes Received:
    13
    Interesting project. Go to WWW.POLOLU.COM for motor control boards that are cheap and work well. Those should interface with a PI easily. Their stepper motor controller boards are great. I use them in my mico CNC build.

    Good luck. The only thing I would be concerned about is the bearings on the tubing compression rollers. A little Delrin drilled out for a shaft should do it.

    I think the pressure the previous poster was taking about is the pressure between the roller and the sidewall that squeezes the tube. In the parked position there will be constant pressure on the sidewall at that spot. Perhaps printing with ABS would be a good choice.
     
  13. Mandelstam

    Mandelstam Well-Known Member

    Joined:
    Oct 29, 2017
    Messages:
    565
    Likes Received:
    888
    Location:
    Malmö, Sweden
    Thanks! I've checked out Pololu before and for this project too. Great site. :)

    What's your concern about the bearings? My reasoning behind them is that they don't have much friction so will roll gently against the hose but also that as I don't have a machine shop I'm very limited to what other rollers I can fabricate. Delrin rollers would probably work fine but I need to drill them pretty precise so they are symmetrical. Printed rollers will have too high friction to roll smoothly. As the pumps run dry and never get into contact with any fluids I'm thinking that normal steel bearings should be ok. But I'll gladly listen to any concerns! Nothing is set in stone!

    For the pressure, I must admit I've never played around with a peristaltic pump before so I have no hands on experience, but as the hose is super soft it doesn't take a lot of pressure to compress it. And my idea was to leave 2x the wall thickness of the hose as distance between the roller and the wall. So the hose will be closed off but the silicone itself not compressed. But I hear you, the printer we have at work uses PLA. Maybe just beef the walls up so they are a bit sturdier could work.
     
  14. Stigigemla

    Stigigemla Active Member

    Joined:
    Mar 20, 2015
    Messages:
    208
    Likes Received:
    108
    Location:
    sweden
    The roller do press the hose against the wall of the house so the hose dont let water through. To get it tight you need some amount of pressure.
    Silicone tubing is not to recommend. All quality pumps has a hose made of santopren, novopren or similar.
     
    Mandelstam likes this.
  15. Mandelstam

    Mandelstam Well-Known Member

    Joined:
    Oct 29, 2017
    Messages:
    565
    Likes Received:
    888
    Location:
    Malmö, Sweden
    Ok, then I get what you mean! Problem is that I haven't really been able to find any cheap heads without some random little dc motor already attached. Thanks for the heads up about the 3d printer plastic. I'll keep an eye on it and see how it behaves.

    Never heard of santoprene before but googled it and yes, its properties look pretty good. But only comparison between silicone and santoprene I could find was:

    Soft ventouse cups are becoming increasingly popular for assisted vaginal delivery. We report the results of a prospective random allocation controlled trial comparing two such cups. The results suggest that the silicone rubber cup is more likely to achieve delivery than a similarly designed cup made from santoprene [odds ratio 4.65, 95 per cent confidence interval 1.3 to 18.8]. Consequently we recommend the silicone cup for delivery when a soft ventouse cup is indicated.
    :D:D

    Why don't you recommend silicone for use in this case? Mechanical properties? Fatigue? I won't be dosing any strong acids or anything like that so those santoprene properties I don't think would be of any real value in this case.
     
  16. Stigigemla

    Stigigemla Active Member

    Joined:
    Mar 20, 2015
    Messages:
    208
    Likes Received:
    108
    Location:
    sweden
    For one of the popular pump heads the life of silicone hose is estimted to about 600 liters and santoprene to about 3000 hours.
    I have read about a few silicone hose breaks the first year for tanks about 120 gallons.
     
    Mandelstam and Ranjib like this.
  17. Mandelstam

    Mandelstam Well-Known Member

    Joined:
    Oct 29, 2017
    Messages:
    565
    Likes Received:
    888
    Location:
    Malmö, Sweden
    Found some additional comparisons between different hose materials: http://www.gradko.com/tubing/products/
    Clearaprene sounds really interesting for this purpose.

    Silicone:

    Santoprene:
    Clearaprene:
     
Draft saved Draft deleted

Share This Page

Loading...