2

For a pet project of mine, I'm interested in finding a tubular solenoid which force is spread linearly over a longer length. The project is based on IA and uses a programmed controller (and sensors) to feed several solenoids (the question is for one solenoid) quickly in order to "grab" something falling vertically, from below, this silently (done) but also a bit more smoothly (the problem). Currently, while the programming seems to work, the test prototype with two solenoids lack control. Basically the solenoids force goes from 0 / 1 [didn't try to control current yet].

I bought and tried from the Net, something like enter image description here

but the force is basically 0 / 1... meaning the force grows exponentially over a short distance.

Ideally, the force would be linearly spread over a length I decide. Thus my question

Preferably with only one solenoid, how to choose/select the core and wires in order to spread the force over a longer length (i.e. not exponential growth)

If 2+ solenoids are necessary please explain how you would arrange them to reach such result (1 is preferable though).

Note: if that helps, the current applied to the solenoid can be precisely controlled from a chip.

Déjà vu
  • 281
  • 5
  • 12
  • 1
    Torque suggests rotary action. It looks more like a linear actuator, in which case you should be referring to the 'force' and not the torque. If so, please edit your question. – Transistor Jan 01 '16 at 13:04
  • Torque is a rotational, wkipedia torque. How linearly do you need? Do you mean a constant force across the entire length of travel/movement, or the force builds linearly from 0 to a maximum? It might be wort giving some information about the application to help people give good answers. – gbulmer Jan 01 '16 at 13:07
  • Question re-edited – Déjà vu Jan 01 '16 at 13:15
  • 1
    I second the thought of using Linear Actuator, It is industrially recognized and works the way you want. I would recommend you to make your own Linear actuator with stepper motors, A lead screw design would be a kick-ass solution to this problem. You can calculate pitch of the lead screw or buy it off market. The mechanical coupling of the stepper motor and the lead screw generally slips if not matched well. But hey there can be some problems to any linear actuator. – ammar.cma Jan 01 '16 at 13:32
  • @ringø: That reads better now, thanks. If I understand correctly you are trying to control the position of the actuator (to move something a controlled distance) rather than control the force applied to the load. Can you tell us (1) Is there a spring returning the solenoid to the extended position? (2) If you power it from a lab power supply and gradually increase the voltage (or current) can you control the position or does it suddenly 'grab' and pull fully in? (3) If not, can you post a short table of stroke vs. voltage? e.g., 0 V -> 0 mm, 1 V -> 1.2 mm, 2 V -> 3.6 mm, etc.? – Transistor Jan 01 '16 at 13:41
  • @transistor 1) no spring - once the object is grabbed, power is down and that's it [it's a prototype!]. 2) actually thanks to a couple of sensors close to the solenoids, speed and position of the object are known ; that decides when and how to control the solenoids 3) there is a 2 ;-) – Déjà vu Jan 01 '16 at 13:56
  • @ringø: actuator = solenoid. All my questions are about the solenoid, not your control system. You haven't answered any of my questions. Read them again. – Transistor Jan 01 '16 at 14:15

3 Answers3

4

Solenoids don't work that way because magnetic fields don't work that way.

To get a constant linear force independent of position, a rack and pinion driven by a motor is a possibility. Some voice coil mechanisms can be pretty constant in force with position too, especially if position feedback can be used to adjust the drive current.

Olin Lathrop
  • 313,258
  • 36
  • 434
  • 925
  • Thank you, and I trust your reasoning. This is also my (novice) conclusion. But I was expecting from you guys, experts, some innovative ideas on how to reach something close to my objective. A motor (tried already) would be 1. slow to start 2. noisy – Déjà vu Jan 01 '16 at 13:18
  • 1
    @ringø If a motor doesn't work (and keep in mind that there are probably linear actuators that are quieter than a solenoid), what about hydraulic/pneumatic actuators. – uint128_t Jan 01 '16 at 19:10
  • @uint128_t Thanks, but quieter than a solenoid seems difficult. Would you have a page about hydraulic/pneumatic actuators? – Déjà vu Jan 02 '16 at 04:09
  • Olin btw trying to go to your site, Chrome says "The site ahead contains harmful programs" – Déjà vu Jan 02 '16 at 04:24
  • @ring: There is something in one of our low level library routines that apparently matches one of the signatures that some of the virus scanners are looking for. Unfortunately, they just to pattern matching and don't look into it more carefully. There is nothing harmful in our programs. – Olin Lathrop Jan 02 '16 at 13:16
  • @OlinLathrop I think there is a way to teach Google the downloadable package is safe – Déjà vu Jan 02 '16 at 13:32
1

The force of a solenoid is a function of both its position and the current. If you have position feedback, you should be able to control the current to achieve nearly constant force over some portion of its travel.

That said, there are linear actuators that don't involve rotary motors with gears or screws. They are basically linear BLDC motors, and they can be very quick. The downside is that they are rather expensive and require a relatively complex controller.

Dave Tweed
  • 172,781
  • 17
  • 234
  • 402
  • Thanks, that seems to be closer to what I need ; have to check the "expensive" aspect of those devices, though... and the noise as well! – Déjà vu Jan 01 '16 at 13:59
1

It is possible to have some intermediate control of push/pull solenoids like this one by using very carefully controlled PWM signals. The easiest way to find the ramping function is probably trial and error.

bigjosh
  • 10,088
  • 32
  • 52