Not an EE, need advice - pump control

Discussion in 'The Projects Forum' started by JC_Biggs, Oct 28, 2013.

  1. JC_Biggs

    Thread Starter New Member

    Oct 28, 2013
    12
    0
    Hey ladies and gentleman! Glad I found what seems to be an awesome resource here!

    So my company has never before done any kind of electronic, or automated designs or products. there honestly has never been a need for it in our field, until recently.

    I have been tasked with designing a hand help dispensing system that uses a pump and a 12v batter pack. (think cordless drill)

    So I looked high and low for an affordable positive displacement pump that would of simply needed a rev counter..to no avail. So I decided on a low pressure high volume turbine pump that will operate on a large voltage margin, and has a decent power draw.

    What I need now, is a way to control that pump to give a set dose. I have used a bench power supply to give timed pulses at given voltages to determine a "table" that will give us the correct amount of fluid dispension. its pretty accurate. (enough for what we are doing)

    Im thinking that I need a 3 value look up table. 2 of the values would be preset by the user (dosage size and fluid viscosity) and the 3rd would be the current charge status of the battery pack.

    The way I envision this working is that when the user squeezes the trigger, the electronics will (via a relay or mosfet) turn on the pump for the preset amount of time, and then add 1 to the bean counter.

    SOunds simple enough to me, but I have no idea about anything electronics. I was hoping maybe you guys could point me in the right direction. We cant really pay a engineering firm 50 grand to design somethign that might not make it to market. Id be perfectly happy with a schematic I can bread board and use for our prototype.

    HELP!!! :D
     
  2. Danm1

    Member

    Jul 19, 2010
    55
    7
    could I ask several questions?
    * How precise does it need to be?
    * About how much is dispensed each time?
    * How fast of a recovery do you need? for example do you need to be able to dispense 10 per second, or is it only one per press?
    * How is a are you showing the dosage amount? Do you want a small LCD display with a couple small buttons? Did you say it needed a counter displayed?

    I'm envisioning a microcontroller based system which precisely controls a linear actuator, all in a hand held device.
     
  3. JC_Biggs

    Thread Starter New Member

    Oct 28, 2013
    12
    0
    Thanks for the reply.

    Precision is within 2% of the user demanded dosage. dosage amounts can vary from 30cc's all they way to 600. I think this is more a function of how close i can get my variables in the lab than anything else. a lookup table with enough resolution should work.

    A dosage cycle is from 2-5 seconds, and will only be activated once for a squeeze of the trigger. Then the user will stage the next product and dose again. so 5 or 6 doses a minute if they are extremely fast and already have multiple products pre-staged.. Its possible they might do 10 on the lowest dosage setting, but at 2 seconds per cycle there is plenty of time there to spare. more than likely though there will be 2or 3 doses a minute

    The idea was to have an LCD display. there would be a single button that allowed you to move to different screens. when you got to the dose and viscosity screens, you would have a rotary dial (potentiometer or something similar) that would allow you to adjust the dose. pressing the button would lock the value in and move to the next screen.

    I had orginally wanted to use a linear servo motor, or solenoid of some kind, we found some of these and bought enough "simple" control to make them work, but the cost of the motors alone was just to high on that kind of device. not to mention we would still have to develop valving and other parts. customers never would of payed the money.

    The only way I can get this done with the cost my company expects is with a pump. since then its basically just 3 parts (pump control battery) I like the pump a little better to that i can seal the housing and it still stay cool.
     
  4. Danm1

    Member

    Jul 19, 2010
    55
    7
    I just took a quick look on the internet, Cole-Parmer has a line of dispensing pumps. Check out the piston, peristaltic and diaphragm pumps. Would these be similar to what you are trying to design?
     
  5. JC_Biggs

    Thread Starter New Member

    Oct 28, 2013
    12
    0
    positive displacement pumps would be ideal. this is the second avenue i pursued after linear motors.

    you set the dose, a simple circuit tells is the correct number of cycles, counts them, and your done. but they are all ridiculously expensive, and the only pumps in our size range are parastolic pumps. which wont work for us.
     
  6. Danm1

    Member

    Jul 19, 2010
    55
    7
    Those others are expensive because it takes a significant design effort to dispense such small amounts so precisely. It's not a trivial thing. Plus all the regulations that are involved with anything medical. There is also the liability factor.

    How much time do you have before you need the design for the prototype?
     
  7. JC_Biggs

    Thread Starter New Member

    Oct 28, 2013
    12
    0
    i already have a prototype assembly using a turbine pump, but now i need the electronics to replace my bench based power supply. id like to have ciruit schematics done and a bread board prototype fabbed up in a few weeks at most. if the proof of concept makes it through the board, then ill move on to pre preduction PCBs
     
  8. Danm1

    Member

    Jul 19, 2010
    55
    7
    OK so you already have a prototype with a turbine pump. Are you able to achieve your 2% precision requirement with this?
     
  9. JC_Biggs

    Thread Starter New Member

    Oct 28, 2013
    12
    0

    "sometimes" i cant dial the power supply in close enough to give an accurate dose across the entire spectrum, when you think about battery voltage dropping you start needing increments of tiem that are in the .1± range, and my power supply only does 1 second increments. I can dial the voltage in to get the correct dose for a given time, but in the real product the time will need to be based off the voltage and not as it is on the bench.

    Im thinking I might need to go with a PWM control. Basically what Im using now is a small fuel pump that is capable of working with PWM. I do beleive though that with a high enough resolution on my timing circuit I can get this turbine to work. I really like this pump because it draws very little power at 0-4psi. Relative to some of the other pumps i have tried. its also 1/4th the price of other pumps in its class and a tenth the cost of any displacement pump i can find.
     
  10. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,769
    969
    What specifications/standards does this device need to meet? UL/CSA/EN/ROHS/medical,etc... You must list them all and anyone who helps must have access to them.. The knowledge to design it right and to a given list of requirements is a good chunk of "engineering costs"
    Who will you hold liable when it fails?
    Who will ensure this design is safe?
    Can you do microcontroller programming? Do you expect that given 100% to you too?
    Can you do the PCB layout given a BOM?
    Would your company donate $5000 to this site for the help given?

    Sourcing the pump parts is the easy part.. The pump control is what you "should" be paying for as its the most difficult.. (easy though for those that know how to do it)

    Why don't you just buy off the shelf.. This kind of stuff is already out there.. Its not reinventing the wheel..

    No offense here.. Just pointing it out..
    There is only so much we can help before it becomes "free engineering".
    Sounds like you expect everything to be done for you and then you will profit from it but the people that did all the work (aka users of this board) get nothing.

    Helping a DIY/hobbiest out is one thing.. But giving "free engineering" support to a company looking to profit is another different beast in my opinion..

    And those that can't do..must buy.
     
  11. Danm1

    Member

    Jul 19, 2010
    55
    7
    mcgyver is right, there is a limit to what can be answered in any forum. For you to say "Id be perfectly happy with a schematic" made me smile. ;)

    In order to provide you with a functioning schematic would take some time. You would be better off asking a simpler question, like maybe:
    " how do I drive my turbine pump?"
    and then provide the technical details of this pump.
     
  12. John P

    AAC Fanatic!

    Oct 14, 2008
    1,632
    224
  13. wayneh

    Expert

    Sep 9, 2010
    12,088
    3,027
    I'm another vote on the "buy" side. Why reinvent the wheel if commercial products exist that can do the job? I've used many different sorts of dispensing pumps in my lab career. Peristaltic pumps with a timed motor, motorized syringe pumps, filling by weight sensors, on an on.

    Thumb through the Cole Parmer and Fisher Scientific catalogs and you'll see what some of the options are.
     
  14. JC_Biggs

    Thread Starter New Member

    Oct 28, 2013
    12
    0
    Look im not here asking for someone to give me something for nothing. this is business, and i understand how it works. What i was asking for was some guidance on how i should approach building this. I am a former avionics technician, and have had several hours of basic electricity training, both in the military and in college. Ive spent alot of time around technology and so im not completely ignorant.

    basically what im asking you folks is "how" in the grand scheme of things should i do it. what parts would i need? I assume you have some kind of chip that references non volatile memory, and then triggers a relay. ..what kind of company should i be looking for to do this cheaply and easily, or hell is there even a product out there that may already do what i want? Ive looked and havent been able to source anything that can do what im looking for, but then again, i dont know the best sources, or exactly what im looking for. Im sorry you mistook my question for seeking a free schematic, but i fully intend on building and programming it myself, with guidance from people who know more than me. Like i said, i work for a company that has NEVER done anything with electronics before, and i really dont know where to begin.

    anyways, this whole idea came from the time i have spend flashing ECUs for the performance bikes I have built. they use Throttle posistion and rpm and output a value for injector pulse width.. Im trying to to basically the same thing, with just one extra value for the Z axis of the table. I thought it might be easy to accomplish.

    so far the only thing i know for sure is the PCB has to have a voltage regulator on it :D




    and John P, there is small cheap pumps available, such as you source, but I have yet to find one that gives the flow we need, in the package size, and design format that fits our product. Like another said, the pump is the easy part. this turbine fits the bill perfectly and from bench testing seems like it will work. just needs more precise control. and with the diplacement pumps id still have to have electroncis to set the dose and count the pump strokes. same boat really. Thanks for the link though.
     
    Last edited: Oct 29, 2013
  15. MaxHeadRoom

    Expert

    Jul 18, 2013
    10,503
    2,367
    Looking at it in another way, is there a possibility of using a liquid dispensing valve for accurate measure?
    The pressure for the liquid could be a small compressor such as used in Automotive applications from the miniature 12vdc under the hood type that supply suspension pressure, also fit-in-your hand type that are used for emergency inflation and are fitted with a small reservoir.
    This way the metering should be exact and not rely on a pump run period to dispense?
    Or something along these lines?
    Max.
     
  16. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,769
    969
  17. MaxHeadRoom

    Expert

    Jul 18, 2013
    10,503
    2,367
    Or another possibility on the same lines instead of a more expensive dispensing valve is to use the miniature compressor to pressurize the liquid and pulse an ordinary miniature solenoid valve for a certain programmable period?
    Max.
     
  18. JC_Biggs

    Thread Starter New Member

    Oct 28, 2013
    12
    0

    That is also a method we considered, but then the end user would need a tank that we supply with our equipment hooked to it. As it is now, they can just use anything that has a tube to supply our pump. which is how our current manual product line works. The other method is to use pressure to fill a chamber, but then you still need mechanical devices to push the fluid out. again, to costly.. so were back to "controlling the pump"


    sorry, our customers are cheap LOL. they are used to buying a $12 product and now we're putting something together for them that is 10-14 times the cost.
     
  19. elec_mech

    Senior Member

    Nov 12, 2008
    1,513
    193
    Welcome to AAC.

    I can appreciate where you're coming from. I'm assuming this is a small R&D project you're using as a proof of concept and if and when you move to production, you'll get said device approved by the appropriate agencies if required.

    While I agree this forum isn't designed to offer "free engineering services", I'm looking at this more as a small project you're getting paid to learn electronics and expand your skill set. I didn't get the impression you're expecting us to gift wrap this for you, more that you are looking for guidance.

    I've learned a great deal myself working for small companies on projects such as this. While there may be a COTS product that does the job, it may not be the right size, have the desired features, etc. Of course, I'm the first to look at something and think I can make it cheaper and I've been told numerous times that probably isn't the case, so you can take my advice with a spoon of salt.

    That said, let's have some fun. :D

    Can you elaborate on your requirements?

    Example:



    Device has LCD screen so user can select/change following:
    • Dosage Size (what controls this, length of time power is applied to pump?)
    • Fluid Viscosity (what controls this, size of pulses to pump?)
    LCD will also display number of times button is pressed.

    When button is pressed, PWM signal will be applied to pump for x time and increment will increase by one.

    Anything else?

    How many different values for the PWM and times do you foresee needing?

    Also, what is the current draw of the pump you're using? I assume it is running from 12VDC?

    You might be able to do this with a PICAXE microcontroller (uC). They are cheap, programming them is simple, and initial investment to get up and running is minimal. They are not fast and thus not ideal where multiple simultaneous operations are needed, but I think you'd be okay for your application if a 1-second pulse is close to meeting your requirements.

    You could also use a uC to handle the bean counting and LCD and add a PWM circuit with a selector knob if you don't have too many timing selections.

    You might be able to skip the uC altogether and use a PWM circuit and a 7-segment display, but we need to know more about how the viscosity and dosage size is selected and controlled. Of course, from a power standpoint, an LCD with uC is the way to go if using a battery.

    Found a few PWM circuits you may be interested in trying. I have not used any myself yet, but these look pretty straightforward. This and this are the same circuit - both discuss the operation and one includes a video. SgtWookie provided an even simplier one here some time back.
     
    JC_Biggs likes this.
  20. JC_Biggs

    Thread Starter New Member

    Oct 28, 2013
    12
    0
    Thanks for that reply, I will save my reply for tonight when i can sit down and answer your questions more throughly. in the meantime, I will look at the links you provided.
     
Loading...