Shop OBEX P1 Docs P2 Docs Learn Events
PUMP circuit needs a bit of outside the box thinking. — Parallax Forums

PUMP circuit needs a bit of outside the box thinking.

BitsBits Posts: 414
edited 2012-12-20 16:23 in General Discussion
I am trying to fix a production PCB that has the circuit listed below. It was a circuit that was not suppose to be populated and now the boss wants to try to use it, so naturally its my job to make it work and do so without cutting / hacking the PCB too much.

My concern is "noise" as this PCB will have to pass CE testing. I can actually hear the system making a high pitched buzz so even if it passes CE this would drive people nuts.
My gripe is that I don't have good control over the pump. This pump draws liquid through a 6 foot tube and the flow rate needs to be adjustable.

I am using the PWM object with these settings
Pump = "PWM object"
PC = a number from  0 - 100

Pump.Setperiod(60)
Pump.SetDuty(PC)

When PC is < 40 the system works lousy. PC numbers Ranging 40 - 100 work fine except that since there is no feedback PC of 40 at time can not overcome the pumps inertia.

** The mosfet is actually an IRLL2705

I am considering a transistors, placing a RC circuit after prop to make it an analog voltage (no longer a PWM circuit).

Any other ideas? Perhaps Code mods will help?
Thanks

Comments

  • Don MDon M Posts: 1,653
    edited 2012-12-18 14:23
    The "buzz" you hear is the PWM frequency "playing" through the motor. To get around it you'll need to change the frequency. I'm guessing you hear it more when the motor is running slower?
  • BitsBits Posts: 414
    edited 2012-12-18 14:27
    Don,

    You are correct as I tweak the period the sounds change. Currently 60 is the least-detectable by ear for some reason.
  • ElectricAyeElectricAye Posts: 4,561
    edited 2012-12-18 14:34
    Do you have a "bleeder" resistor from the Prop pin to ground? With mosfets, I find that sometimes helps when I'm switching things quickly. Just a thought.
  • BitsBits Posts: 414
    edited 2012-12-18 14:36
    No I have not installed a pull down resistor yet only because the propeller might be doing that internally. Otherwise its an easy patch and ill try it here in a bit.
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2012-12-18 14:47
    Hi Bits;

    1.The IRLL2705 in the SOT-223 package can't be operated in the linear mode. It can only dissipate a maximum of 2.1W with ideal board layout. and 1W with more common layouts. Your just going to have to use PWM.

    2. The noise is probably the cycle frequency. You might want to raise it to 20KHz or more so you can't hear it.

    3. Since this is a commercial product you are operating the IRLL2705 off speck. They suggest a minimum of 4.0V gate drive. There are MOSFETs that are designed to be driven by 3.3V.

    Using a scope test to see if the drain voltage is fully on and off. With 3.3V gate drive it's possible the MOSFET is self oscillating.

    Duane J
  • ElectricAyeElectricAye Posts: 4,561
    edited 2012-12-18 14:56
    Also, when it comes to the control of your pump, keep in mind possible mechanical effects. For example, you could get siphoning effects going on one way or another if you don't have check valves and there are significant height differences, etc. Also, control could depend on what sort of pump it is, whether or not it has a rotor with large inertias, if it's centrifugal, what sort of response times are required, etc.
  • BitsBits Posts: 414
    edited 2012-12-18 14:57
    ElectricAye
    Pull down made little difference i suspect made it worse because the gate voltage drops just a little.

    Duane
    Good point. I am shopping for a better mosfet. Ill set up the scope now and take a peek.
  • BitsBits Posts: 414
    edited 2012-12-18 15:15
    ElectricAye,

    What do you mean response times?Ergh on what system? The pumps flow rate or the Mosfet, circuit?
  • BitsBits Posts: 414
    edited 2012-12-18 15:24
    Duane
    Where do you see "
    They suggest a minimum of 4.5V gate" I see the data sheet showing 1 - 2 V on the IRLL2705. I wanna learn darn it!
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2012-12-18 15:42
    Hi Bits;
    Bits wrote: »
    Duane
    Where do you see "
    They suggest a minimum of 4.5V gate" I see the data sheet showing 1 - 2 V on the IRLL2705. I wanna learn darn it!
    The spec shows guaranteed on resistance for 4.5V and 10V.
    The threshold is somewhere between 1V to 2V at 250uA of current.
    Lets say it is 2V then there is only 1.3V more to turn it on. Not a lot.
    Now, you might get away with this but it would not be the "Engineering Way".

    Take a look at this one:
    FDT439N
    This has a threshold of .4V to 1V and a guaranteed on resistance of .058Ω @ 2.5V.
    Its actually cheaper to boot!

    Duane J
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2012-12-18 15:44
    That's "VGS(th)", the bare minimum, kind of a throw-away spec.
    Look at Fig. 3 on page 3, ID vs VGS
  • tonyp12tonyp12 Posts: 1,951
    edited 2012-12-18 16:15
    attachment.php?attachmentid=97989&d=1355875696

    If you are using a prop pin with its 3.1 V out, you're just getting in the lower zone of acceptable.

    When I look at that graph I say 4v looks really good if I my current requirements are not max
    If I want spikes of 30amp I would need 5v on that gate.

    Using a 74ABT126 TTL buffer would be nice as to boost the voltage and gate charge a little.
    Or dedicated gate driver IC's, or choose a mosfet that is at least 80% 'on' at 3.1v
    199 x 220 - 11K
    VGS.gif 10.7K
  • ElectricAyeElectricAye Posts: 4,561
    edited 2012-12-18 16:32
    Bits wrote: »
    ElectricAye
    Pull down made little difference i suspect made it worse because the gate voltage drops just a little....

    For bleeders, I usually use about 10K resistors. But you could probably go up to 100K without any problems. I can't say for sure, but I don't think the Prop has internal resistors on its pins, does it?

    Bits wrote: »
    ...What do you mean response times?Ergh on what system? The pumps flow rate or the Mosfet, circuit?

    When it comes to controlling mechanical things driven and controlled by electronic devices, you often must consider the response time of the entire system. Control theory lumps together the electronic and mechanical properties so they can be analyzed as an organic whole. For example, you can have high speed action in your electronics, but it might take several seconds for a centrifugal pump to get up to speed or to stop. And when you try to stop the flow with valves, moving fluid can create water hammer effects, etc. I'm just saying you can't consider only the electronics without regard to the mechanical time lags, etc. Maybe these won't matter to your application, but I'm just speaking in general terms here. I didn't mean to ergh you out about it. :)
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2012-12-18 16:40
    Hi Tonyp12;

    Your not using that graph properly.
    This graph is based on the typical specs.
    It is useful for estimating the characteristics v.s. temperature.
    I would add to the gate voltage in the graph the difference between Vthmax and Vthtypical.
    Some specs give us the typical threshold specs, the IRLL2705 doesn't. So I would add 0.5V to the
    gate voltage in that graph

    Duane J.
  • CircuitsoftCircuitsoft Posts: 1,166
    edited 2012-12-18 17:45
    The PHT8N06 is another option that should handle plenty of current at 3.3v.

    As for trying to make the pump run, you probably need some sort of voltage feedback to figure out how fast it's spinning. It'll take more power to start the pump moving than to keep it moving, and voltage feedback could be used to measure how fast the motor is spinning. It won't be terribly accurate, but it's a starting point.
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2012-12-19 05:16
    Hi Circuitsoft;
    The PHT8N06 is another option that should handle plenty of current at 3.3v.
    That MOSFET won't even turn on at 3.3V date drive. It's specified for a gate drive of 10V.

    Duane J
  • Don MDon M Posts: 1,653
    edited 2012-12-19 05:40
    It'll take more power to start the pump moving than to keep it moving

    Isn't that the premise behind PWM motor control to have max power at starting? You're switching on the max available current to the motor only in a thin slice of time.
  • BitsBits Posts: 414
    edited 2012-12-19 12:59
    Hi Bits;

    1.The IRLL2705 in the SOT-223 package can't be operated in the linear mode. It can only dissipate a maximum of 2.1W with ideal board layout. and 1W with more common layouts. Your just going to have to use PWM.

    2. The noise is probably the cycle frequency. You might want to raise it to 20KHz or more so you can't hear it.

    3. Since this is a commercial product you are operating the IRLL2705 off speck. They suggest a minimum of 4.0V gate drive. There are MOSFETs that are designed to be driven by 3.3V.

    Using a scope test to see if the drain voltage is fully on and off. With 3.3V gate drive it's possible the MOSFET is self oscillating.

    Duane J

    Thanks I see now.
  • CircuitsoftCircuitsoft Posts: 1,166
    edited 2012-12-20 11:11
    Hi Circuitsoft;

    That MOSFET won't even turn on at 3.3V date drive. It's specified for a gate drive of 10V.

    Duane J

    While the gate drive is specified at 10V, it's also specified below. I looked at the voltage range of the gate cutoff voltage, subtracted the average (1.5v) from max(2.0v), and looked at the output current vs drain voltage graph (Fig. 5), Choosing a line around 2.8v (3.3v - 0.5v offset of max vs avg cutoff). Looks like Drain Resistance will be < 1/10 ohm when drawing 1amp through the drain. Halfway between the 2.6v and 3.0v lines on the graph are horizontal at ~4-5 amps, so it really should be fine.
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2012-12-20 12:30
    Hi Circuitsoft;

    What spec are you using?
    I used this one by Phillips PHT8N06.
    The gate threshold max is 4V. and the average is 3V.

    Duane J
  • CircuitsoftCircuitsoft Posts: 1,166
    edited 2012-12-20 16:23
    Ah. I didn't realize how important the L at the end of the part number was. If you search DigiKey for PHT8N06, all results are PHT8N06LT transistors.
Sign In or Register to comment.