9

I'm trying to measure the activation time of a pull-type solenoid.

When the solenoid doesn't have a load connected to it, you can measure its activation time through the current it consumes. At first, the solenoid will consume current until the magnetic field is strong enough to move the armature, at which point the current consumption decreases until the armature has reached its holding position. At that point the current will rise a bit more until it stabilizes.

However, if there is a load attached to the armature, the curve of current consumption never dips, which means that you can't quite tell at what point the armature is fully pulled in.

I'm wondering what technique could be used or is standard for measuring the time it takes the armature to reach the end position with a load.

I've thought of making the load close a circuit so you could measure the time between activation of the solenoid and the closing of the measuring circuit but I'm wondering if there are other approaches which I haven't thought of.

(p.s. if this is off-topic let me know and I'll remove the question)

Edit: Preferably the solution can be embedded in a machine with up to 1000 solenoids. This means that the added hardware (cost for example) needs to be somewhat realistic (i.e. no 200$ camera per solenoid ;) )

Florian Humblot
  • 191
  • 1
  • 4
  • 16
  • 1
    High speed video camera? – user57037 Oct 21 '19 at 07:19
  • 1
    Why are you need to do this? Are you trying to characterise the solenoids or detect when the desired position in a machine has been achieved? – Transistor Oct 21 '19 at 12:52
  • @Transistor I'm trying to determine which (if any) solenoids are failing to meet their hard real-time deadline. This can be done while the machine is in a 'test mode' which lets me activate the solenoids whenever I need to. – Florian Humblot Oct 21 '19 at 12:54
  • 1
    About how long does it take when it is loaded? Is it 1ms, or 10ms or 50ms? I can't help but feel that the current waveform will undergo some change when the solenoid stops moving. If you look at the current waveform with a current probe maybe you can figure it out with some experimentation. If so, then sampling the current waveform and doing some type of processing on it might help figure it out. – user57037 Oct 22 '19 at 00:07
  • 1
    Does the solenoid coil inductance change somewhat linearly with position? If so, then any technique which allows you to measure the instantaneous inductance should also allow you to figure out the position. Is this a fruitful branch of thought? – user57037 Oct 22 '19 at 00:17
  • 1
    @mkeith The deadline with load is 25ms. The average unloaded time is between 15 and 20ms. I have been measuring the current consumption, which dips when the armature reaches it's final position, however, when there is a load, this dip does not appear. I am unsure how exactly to measure the inductance, but I'll look into that! – Florian Humblot Oct 22 '19 at 06:02

8 Answers8

9

I'm wondering if there are other approaches which I haven't thought of.

Idea:

If you superimpose a small high frequency AC signal on top of the DC that drives the solenoid, you can infer the inductance change from the inactive to activated position. When the armature is activated, the coil inductance is usually (and measurably) higher.

The high frequency signal needs to be high enough so that one cycle provides enough time resolution but not so high that eddy current losses swamp that signal amplitude change. Feed the signal from a signal generator via an impedance that is nominally the reactance of the coil to obtain optimized signal amplitude change.

Your signal amplitude should be chosen so that its nominal amplitude barely affects normal DC operation. You might also need to DC feed the solenoid via a small impedance so that the power supply used for activating the solenoid doesn't significantly attenuate the AC signal.

A bit of experimenting might be required.

Andy aka
  • 434,556
  • 28
  • 351
  • 777
  • What about just PWM'ing the solenoid and looking at ripple current slope? Is that practical at all? Shouldn't slope vary predictably with position? – user57037 Oct 22 '19 at 00:21
2

If you don't need very high accuracy/precision, then the voltage across the solenoid can start a timer and a limit switch can stop it. For more accuracy, a small vane entering an optical interrupter can be the stop signal.

AnalogKid
  • 16,865
  • 1
  • 13
  • 25
1

You can measure its deceleration if your OK with pulsing it with a much higher initial voltage so the inductance is not the dominant factor.

From your existing tests you know roughly how much current vs how much time to move the armature to the end of travel, well you can equally give a much higher current for a proportionally smaller period of time e.g. a higher voltage pulse

This pulse will accelerate the armature, and even when the pulse is over, the armature will be in motion towards its end of travel,

As it moves through the magnetic field it will be generating a voltage you can measure, when it stops as it reaches its end you will see this voltage drop off. the sharpness of this drop off tells you if it was a soft or a hard stop,

A soft stop could be that you didn't quite travel the full distance, so it will somewhat resemble half a sine wave, travelling at a good speed, slowing down and then moving backwards, while a hard stop will look much more like a square wave, where its moving at a good speed, and bounces suddenly off the end.

Reroute
  • 4,377
  • 6
  • 14
1

A capacitive (or inductive) bridge could be used to measure (in a non-contact manner) when the armature approached the end of travel. Or simply use the change to de-tune an RF oscillator.

Spehro Pefhany
  • 376,485
  • 21
  • 320
  • 842
  • I'm sorry but I'm not quite sure what a capacitive bridge or inductive bridge is supposed to be or how it works, couldn't find a lot of information on it by googling, could you expand a bit on that? – Florian Humblot Sep 12 '19 at 05:35
1

One more - a microphone attached to listen to the start and stop of the mechanical movement.

Dirk Bruere
  • 13,425
  • 9
  • 53
  • 111
1

Design a differential amplifier and observe the current with it, with a careful design and analyzing the resulting pattern, you should be able to figure out the moment of impact.

Ayhan
  • 1,194
  • 9
  • 17
  • 1
    What do you expect the user to see? What change in the current waveform do you expect when the solenoid actuator stops moving? – Transistor Oct 21 '19 at 12:48
1

I'm not sure if the machine you're observing already exists or has already been designed, but...

While not exactly an answer to the question, a coil energization speed-up circuit will improve pull-in time and potentially bring any of your solenoids that haven't outright failed into the desired speed range. This approach is commonly used in industrial protective relaying gear where the fastest possible coil energization, and thus the speed of actuation, is of critical importance.

Bort
  • 1,084
  • 9
  • 18
0

I don't know if there's a proper way to do it, but I would first study what changes in current happens in all the process of the movement of one of the solenoids.

This is pretty easy to do with an oscilloscope and a low value resistor in series with the solenoid. Using two probes in the oscilloscope (or using a differential probe) you measure the voltage drop in the resistor in all the movement process.

After you have this, probably there's a current drop after the movement stops. Something like this can be measure with a low-cost microcontroller with a low value resistor again, and you can program it to detect the pattern you observed in the solenoid you studied.

Josemi Rodríguez
  • 34
  • 4