Shop OBEX P1 Docs P2 Docs Learn Events
How would you power this solenoid. - Page 2 — Parallax Forums

How would you power this solenoid.

2»

Comments

  • T ChapT Chap Posts: 4,198
    edited 2018-10-27 02:47
    The device TonyP12 linked with built in mosfet is pretty convenient.
    539 x 342 - 36K
  • jmgjmg Posts: 15,140
    edited 2018-10-27 05:51
    T Chap wrote: »
    The device TonyP12 linked with built in mosfet is pretty convenient.

    Yes, but lacks current sense, and is quite niche-price expensive.

    I see new LED driver devices like ALT80802 get close.
    Low cost, wide Vin, 200mV I sense, but lowest Fosc may be on the high side for an iron-cored load.


    Better may be a hysteresis driver, like AL8860 & AL8861- with those, the frequency auto-adjusts to sweep the current the 85mV<->115mV hyst limits. 20mH is about 3kHz at 666mA regulation.
    I see these also have an analog-mode dimming option, where 0.4~2.5V is scaled threshold setting, so the peak current can reduce in holding mode too. AL8861 adds a FET current MAX (fault) sense too.

    16c in volumes, is quite cheap (40V, 1.5A) - shows the economies of scale.
  • T ChapT Chap Posts: 4,198
    edited 2018-10-27 13:55
    In this application what is the goal of current sensing or constant current. Is it to avoid exceeding the .666 because of excess heat? Or to find the minimum current the device will stay ON at and thus achieving the minimum ON current? The way I understand it if you build a constant current supply and can set it for both initial on and low power on, then there is no need for sensing on a device. Shouldn't I be able to take a simple 12V switching supply and connect a mosfet on the low side, set PWM from a Prop pin at 100% and put a meter in series with the high side, check the current. Then set the PWM for the holding force and read the current and set the PWM accordingly for both full on and holding force and achieve the same thing as adding a device with sensing or adding a supply with constant current. This method requires a Prop pin, a logic level mosfet, a series resistor, a pull down resistor at the mosfet.
  • Since the hold current is generally less than the pull in current, you could use two FETS or triacs (if ac solenoid) with one of them direct to the solenoid and the other through a resistance. Then it is a matter of energizing the both devices to pull in the solenoid and then turning off the device with direct connect leaving the device feeding the solenoid through the resistor on as a holding circuit. If power drops or both devices are turned off, solenoid drops out.
  • With a wall wart 1amp switching supply. Meter in series with the + wire to a board with a logic level mosfet FQB30N06L. PASM PWM object with photo showing cycles High, Cycles Period.

    Full on 704 mA
    90% 265 mA. More than adequate to hold
    85% 122 mA. Just below min holding force
    80% 040 mA. No movement of pin at all

    At 90% I would hardly call it warm to touch.
  • kwinnkwinn Posts: 8,697
    T Chap wrote: »
    With a wall wart 1amp switching supply. Meter in series with the + wire to a board with a logic level mosfet FQB30N06L. PASM PWM object with photo showing cycles High, Cycles Period.

    Full on 704 mA
    90% 265 mA. More than adequate to hold
    85% 122 mA. Just below min holding force
    80% 040 mA. No movement of pin at all

    At 90% I would hardly call it warm to touch.

    With those numbers I would put in a 1.5A rated power supply and PWM the solenoid at 90%. Might be tempted to add an optosensor to be sure the solenoid is in the up position if I had an I/O pin available on the Prop and the customer was in the boonies. That would allow for automatically adjusting the PWM and warn me there might be a problem.
  • Yes there is a switch so the PWM loop can always check the switch and adjust the duty.
  • jmgjmg Posts: 15,140
    T Chap wrote: »
    In this application what is the goal of current sensing or constant current..

    If you have a well regulated main supply, current drive it is less important.
    One benefit of current drive, is the Amp-Turns is what sets the force, whilst the heating in the coil can raise the resistance, and with fixed voltage that lowers the current.
    If you PWM drive, the coil heating is modest, and the total power you have here is not large, so you can go simpler.


    T Chap wrote: »
    Full on 704 mA
    90% 265 mA. More than adequate to hold
    85% 122 mA. Just below min holding force
    80% 040 mA. No movement of pin at all
    Something seems a little awry, as a 10% reduction in on time, should not drop current to 37% ?
    Because of i^2*R effects, if you can hold fine with 40%, that's dropped to 16% of full-power, a significant saving, and things would be more reliable.
  • I’m not sure how you would explain the drop at 90% duty. Unless there is some particular error in the method of inserting the fluke in series with the the +12v line. But I am watching the signal on a scope at the gate pin and it is 90%.
  • kwinnkwinn Posts: 8,697
    jmg wrote: »
    T Chap wrote: »
    In this application what is the goal of current sensing or constant current..

    If you have a well regulated main supply, current drive it is less important.
    One benefit of current drive, is the Amp-Turns is what sets the force, whilst the heating in the coil can raise the resistance, and with fixed voltage that lowers the current.
    If you PWM drive, the coil heating is modest, and the total power you have here is not large, so you can go simpler.


    T Chap wrote: »
    Full on 704 mA
    90% 265 mA. More than adequate to hold
    85% 122 mA. Just below min holding force
    80% 040 mA. No movement of pin at all
    Something seems a little awry, as a 10% reduction in on time, should not drop current to 37% ?
    Because of i^2*R effects, if you can hold fine with 40%, that's dropped to 16% of full-power, a significant saving, and things would be more reliable.

    There may be some other effect at work as well but I suspect this non-linearity of current vs PWM is due to the inductance and resistance of the coil along with the diode across the coil.
  • T ChapT Chap Posts: 4,198
    edited 2018-10-28 04:01
    The only thing I have across the coil at the moment is a varistor plus the diode in the mosfet.

    https://www.digikey.com/products/en?keywords=F6213-nd

    https://www.digikey.com/products/en?keywords=FQB30N06LTMCT-ND
  • jmgjmg Posts: 15,140
    T Chap wrote: »
    The only thing I have across the coil at the moment is a varistor plus the diode in the mosfet.

    Oh.. You should also have a catch diode, that gives the inductor flyback current a path, (Diode conducts when MOSFET turns off, and inductor terminal tries to flyback above V+)
    Diodes needs to be fast recovery and rated to Inductor current, ie > 704mA
    1A might do it, or 2A ? Can be Schottky or fast recovery.

  • jmg wrote: »
    T Chap wrote: »
    In this application what is the goal of current sensing or constant current..

    If you have a well regulated main supply, current drive it is less important.
    One benefit of current drive, is the Amp-Turns is what sets the force, whilst the heating in the coil can raise the resistance, and with fixed voltage that lowers the current.
    If you PWM drive, the coil heating is modest, and the total power you have here is not large, so you can go simpler.


    T Chap wrote: »
    Full on 704 mA
    90% 265 mA. More than adequate to hold
    85% 122 mA. Just below min holding force
    80% 040 mA. No movement of pin at all
    Something seems a little awry, as a 10% reduction in on time, should not drop current to 37% ?
    Because of i^2*R effects, if you can hold fine with 40%, that's dropped to 16% of full-power, a significant saving, and things would be more reliable.

    I may be wrong on this. These values will seem awry if you are using a meter rather than a scope reading the drop across the coil and some math to figure the actual current though the coils resistance to see the current per pulse. Also, the impedance of the coil at the chosen frequency may further alter what the meter sees.The meter is probably giving an average value while the real current should not change for a fixed voltage and coil resisrance. Varying the frequency will probably do the same thing to the meter readings.
  • kwinn wrote: »
    jmg wrote: »
    T Chap wrote: »
    In this application what is the goal of current sensing or constant current..

    If you have a well regulated main supply, current drive it is less important.
    One benefit of current drive, is the Amp-Turns is what sets the force, whilst the heating in the coil can raise the resistance, and with fixed voltage that lowers the current.
    If you PWM drive, the coil heating is modest, and the total power you have here is not large, so you can go simpler.


    T Chap wrote: »
    Full on 704 mA
    90% 265 mA. More than adequate to hold
    85% 122 mA. Just below min holding force
    80% 040 mA. No movement of pin at all
    Something seems a little awry, as a 10% reduction in on time, should not drop current to 37% ?
    Because of i^2*R effects, if you can hold fine with 40%, that's dropped to 16% of full-power, a significant saving, and things would be more reliable.

    There may be some other effect at work as well but I suspect this non-linearity of current vs PWM is due to the inductance and resistance of the coil along with the diode across the coil.

    When the plunger advances, is there any armature in the coil or is it just air?
  • T ChapT Chap Posts: 4,198
    edited 2018-10-31 21:07
    I am experimenting with a magnetic clutch for the first time as part of the experiment with the original solenoid discussed before. On my driver board that uses the logic level mosfet from the Prop it was designed with a 34V varistor as I had seen that used in a design as a shunt to take any excess voltages to GND. Obviously it's not working as in the photo 34V Varistor the spikes are present. Maybe I screwed this up. I put in a 2A 40V Schottky

    https://www.digikey.com/products/en?keywords=1655-1519-1

    The photo labeled No Diode Spike shows a spike with NO diode on the first pulse, and the diode connected on the second pulse. The spike is 76V. The regular pulse is 24V. The other photo 34V Varistor shows the varistor in place of the diode. After further reading on the subject of varistor as shunt on solenoids/coils it is commonly warned that they will stop working at some point. This is metering the coil so the collapse of the coil is when the pulse goes high in the photo. The pulse is .5second on and .5 off.
    819 x 486 - 125K
    819 x 486 - 127K
  • jmgjmg Posts: 15,140
    T Chap wrote: »
    On my driver board that uses the logic level mosfet from the Prop it was designed with a 34V varistor as I had seen that used in a design as a shunt to take any excess voltages to GND. Obviously it's not working as in the photo 34V Varistor the spikes are present. Maybe I screwed this up. I put in a 2A 40V Schottky
    If there is no change on adding/removing the varistor, either it is not working, or it has a breakdown above the FET. The limiting you can see, will be the fet avalanch.

    FETs do have and avalanche rating, but that should be rarely pushed.
    A flyback catch diode is the correct solution, if you intend to PWM driven any inductor.

    Rare instances where you might choose a solution like a high voltage zener, would be a relay, or solenoid valve, where you seek to speed the reaction time. (mostly, there a few ms do not matter)
    With any inductor, the lower the clamp voltage, the longer the i takes to fall.

  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2018-10-31 21:22
    A varistor is fine as a safety device, but the more they operate, the more they degrade, so they are not designed for this type of circuit. Your diode is the way to go and they are fast.

    BTW, be aware that if your PWM frequency is too high in relation to the solenoid's inductance then it will not respond proportionally to the duty cycle since the reactance will be too high. If you go too low then the PWM may even be audible so I would expect a a few hundred Hertz would be a good start. Also, due to the current flowing in the catch diode when you release the solenoid, which will also flow through the coil, this will tend to maintain the magnetic field (reversed) longer and slow the release of the plunger. If fast release is desired then you should use a reversed zener diode in series with the catch diode so that you will still get a spike but it will be clamped at a tolerable level but give you a faster plunger release.
  • Thanks guys. I wonder if there is a test ie sweeping a frequency range to determine an ideal rate based on the inductance/reactance factors you described to get best performance.
  • potatoheadpotatohead Posts: 10,253
    edited 2018-11-01 01:07
    I did something like that when working with some coils. They were out of tattoo guns of all things. What I did was hook up a magnetometer and then sweep it up across the frequencies from about 10 hz to 5khz.

    The ramp-up to real, significant induction was surprisingly quick.

    Reactance was a significant factor at 1khz.

    My meter was slow, so I would just step 50 hertz at a time, pulse a few different ways I knew I needed, capture the numbers, move on.
  • Lenze makes a good clutch/brake combo.
  • T ChapT Chap Posts: 4,198
    edited 2018-11-01 05:17
    I assume what potatohead is saying is that he put a meter at a fixed position on the coil, then ran various frequencies and duty into the coil and noted the magnetic behavior. Then better the magnetic behavior means the better performance ie ideal frequency. The frequency range closest to reactance I assume means bad magnetic performance. So you want to avoid the reactance.

    Edit. I’m looking at magnimeters I see a number IPhone apps. Turns out I can stick a magnet near the phone and get a reading.
Sign In or Register to comment.