Inquiries on linear actuator

[villager]

I'm doing a prototype project on automated car parking. It's basically a structure of three cube-shaped parking slots that can move up and down only (one degree of freedom). While I've been looking for a simple electromechanical solution that can do the stated objective, I've stumbled upon linear actuators. Given that I'm a novice in the area of motors and motor control, I'm a bit confused about the variants of these linear actuators and how to electronically control each one.

From my research, I've known that there are two types of electromechanical linear actuators. The first one is based on brushed DC motor and the second one is based on stepper motor. I don't know much about the difference between the two apart from the fact that stepper motor has the capability to rotate in microsteps giving the ability for position control. I want to know which type would be better and how each one can be controlled to extend and retract the shaft (reverse direction). All I've suggested is an H-bridge to reverse the power supply polarity, but I don't know if this holds true to stepper motor, too.

I've come to know that the stepper motor linear actuator, due to its microstepping nature, would be better for my application because I would be able to control the stopping position of each vacant parking slot, which would be on ground level, without the need to use an external sensor like a Hall Effect sensor. 

I would like to read your answers and suggestions if you please.

[ajb]

If this is truly meant to be a prototype or proof of concept for a real car parking system, then you should pick the actuator that most closely resembles the actuator to be used in the full-scale system, at least in terms of control and positioning capability.  In particular, you shouldn't rely on the precise positioning ability of a stepper motor unless you plan to use some bloody great big stepper motors to raise and lower a bunch of cars. 

More likely, you'd use hydraulic actuators to do the lifting, which don't have any sort of inherent positioning capability (although they do automatically stop at each end of their travel, which is nice), so you'd need to add positioning switches to determine when you're at each target position.  To command movement, you'd drive one solenoid to raise the apparatus, or another to lower it.  You can emulate this quite closely by using a DC-motor-driven linear actuator controlled by a simple relay H-bridge.  You will of course need to add limit switches for the ends of the actuator's physical travel if your actuator doesn't have them built in.

If your goal for this project doesn't really require accurate simulation of the full-scale system, then of course you can use whatever mechanism is easiest/cheapest/closest to hand. 

As far as controlling stepper motor and brushed DC motor based linear actuators, the principles will be exactly the same as driving any stepper or brushed DC motor.  Any stepper or brushed DC motor tutorial will be directly applicable to a linear actuator driven by either type of motor (unless the actuator has drive circuitry built in, in which case refer to the manufacturer's documentation).

[villager]

So I understand that motor-based linear actuators can't be used in a full-scale system, is this right? from my research, I've come to know that they're not used in elevator applications contrary to a linear hydraulic actuator, but I'm not sure of this conclusion. the prototype is supposed to be of a real parking system, so I suggested to use two motor-driven linear actuators holding the parking structure from the center distance of the right and left sides in order to balance the structure (something like this: ), leaving the front and back sides for car entry and exit points.

How do you evaluate this solution? and how applicable is it for a real system? if motor-based actuators aren't applicable to reality, I can suggest that, for a real unit, we would replace the actuators with hydraulic linear actuators. Both provide the same functionality (linear motion around one axis), but the control method/circuitry would need some modification alongside the software running on the microcontroller (yes, we use a mcu for decision making).

As for position control, again, I've two options:

1. Attach a magnet to each one of three parking slots and install a hall effect sensor at the ground level of the structure. Whenever the hall sensor detects the magnetic field of the magnet, it would send an interrupt signal to the mcu. At this point, the mcu will decide to stop the motion of the actuator.   

2. Using a stepper motor based linear actuator, we calculate the number of steps the motor has to rotate until the vacant parking plot reaches to ground level. This can be done in software, but the thing will depend on the weight that the actuators are currently lifting. It may not be as accurate and easy to implement as the first option. What do you think?

Illuminate me if you please.

 

[SeanB]

In any case you choose it will have to pass the safety standards of a goods elevator. This includes being power fail safe, actuator fail safe and to have positive braking. Thus you are either going to have a regular cable system or chain with a motor, brake and gearbox driving a wheel, along with a counterweight, or a hydraulic system with a pump, fluid reservoir and a set of lock valves to hold the car in position.

As well you will need guides, and whatever you choose the standard safety design consideration is that it must not fail at a 12 times overload, and must still operate with a 2 times normal max load in it.

Sensing as well must have at a minimum a sensor per level, along with a mechanical overrun switch placed at close to max and min travel, along with an emergency power disconnect ( which shorts the power rails to trip a breaker or blow the main fuses in the supply) just before hitting the main bump stops or emergency dampers.

Your controller must have documented code, and must not fail if any inputs are wrong, but must detect external input faults and fail safe to not operating the motor. It must also have protection from faulty code, which typically is a watchdog timer that when it times out resets the system and locks the motor drivers into a non operating state. Controller inputs and outputs must meet regular industrial input and output protection ratings, and must survive connection to mains rails for long periods without failing. This typically means opto isolated inputs and relay outputs. Power supplies as well must be designed for long period operation with the ability to survive all the transients that come in on the mains, including massive overvoltage events and lightning strikes. thus typically you find still a mains transformer feeding a bridge rectifier, along with a self resetting thermal overload in the transformer. Motors must have current limiting and overheat protection to protect them. Motor breakers and physical latches must have a safety link that opens on failure so that any fault there results in no movement. this circuit must be a physical break of the motor's drive signal, not a software only operation.

You need door detection, along with detection of objects in the openings, and switches to show both door open, door closed and door locked status, along with having all door sections having a switch, both on the platform and on the outside. Whatever motor drives the door must have protection as well.

[villager]

SeanB

I value your input and your intensive concern on the safety issues, but we're not attempting to make a full-fledged prototype. Minimum safety issues are going to be considered, like heat sinking and such, but not anything like you described. On the elevator making, the thing is that we preferred to use something as easy as motor-driven linear actuators rather than a more complex mechanical system like the cable system. I'm sure that neither me nor my colleagues are good at these mechanical systems since we're coming from an electronics background (yes, it's a senior project by the way).

We want to make a mechanical elevator which can, to a certain degree, emulate the working of a real system using linear actuating. As for the electronics part, that's where we will take care of things more intensively like overheating issues, and that's where we will be expected to answer questions relevant to safety, that is, we're not expected to handle safety concerns regarding the elevator.

 Again, I would like to see answers and suggestions to my inquiries in the previous post, if you please.

[ajb]

If you use a basic DC motor-based linear actuator the only difference between that and a full-scale system using hydraulic actuators will be that with the prototype you'll need limit switches on the actuator to keep the motor from stalling (or worse) at its end of travel.

The stepper motor may be tempting, but consider that steppers are not fool-proof.  Just because you've commanded the motor to move X number of steps doesn't mean it actually has!  What if the driver has failed, or the system is overloaded?  Even without a catastrophic failure an open-loop stepper system can easily lose a step here or there and after a while your elevator stops hitting its marks. 

So no matter what you're going to need some additional sensors to at least give your system a sanity check.  I'd probably do three switches per target position: One precisely on target, one just above, and one just below.  That allows you to detect over- and under-shoot, and if you want to get fancy you can have the system move rapidly to the first limit switch and then slow down until it hits the 'on target' switch.  That way you can improve cycle time without risking overshooting the target.  It also gives redundancy, so that the system knows in which order the switches should hit and can identify a malfunctioning switch if it fails to get an expected switch transition.

[villager]

Thanks for your answers, ajb.

I would like to take the opportunity and ask about another thing not related to linear actuator, but of course related to the prototype. I plan to use a through beam photoelectric sensor for car detection and other things. I researched the subject and found that there are a lot of specifications that have to be taken into consideration when choosing one, like supply voltage and saturation output voltage, npn/pnp switch output..etc.

I still don't know if it would be suitable to get one of these sensors like the one in the figure below and whether they will be suitable to provide an output pulse the oscillates between two states (high and low) that can be fed into an interrupt input of a mcu. The datasheets and specification tables for the sensor, that I took a glance at, show that most of the sensors of this type are supplied by a DC voltage between 10 and 30 volts, so the output would saturate (high level) at a voltage close to the power supply (say 10 v). I see that, in this case, the signal has to go into come circuitry (interface) before it's fed into the mcu so that it produces an output suitable to the mcu's pins (0 to 5 V max).

Would you please illuminate me on this sensor?

[ajb]

Industrial systems aren't generally designed with signals that swing between high and low voltages like you'd expect to feed into a microcontroller.  The convention in industrial systems comes from the days when everything was based on relays and mechanical switches, so you have input 'coils' and output 'contacts'.  Look at the wiring diagrams in this PDF.  The rectangles connected to terminals 2 & 4 in diagrams 1 and 2 represent the input coils of a control system--these could actually be the coils of electromechanical relays, or they could be solid state, either way they are triggered by a voltage applied across them.  A sensor, like the beam receivers in that PDF, has output contacts that are either closed or open.  In the case of the NPN output devices in diagram 1, each of the outputs is actually the collector of an NPN transistor, the emitter of which is tied to L-.  So when the sensor 'closes' an output contact, it turns on the transistor effectively connecting it to L-.  Otherwise the contact is open, ie, effectively not connected to anything.  The PNP devices are the same except the transistor is, of course, PNP, and connected to L+.  So you can read the output of one of those sensors just like you'd read any other open collector output, which can be as simple as a pullup resistor.  But if you want to make your system reliable, you'll need to incorporate some level of protection as well.  A google search should turn up lots of information on protecting the inputs of a microcontroller against all sorts of threats, but the most robust solution would be an isolating relay or optocoupler would do the job.  Or you could just use a PLC, which will have well-protected inputs designed exactly for this type of application.

[villager]

That explains why I see the almost lack of discussion and demonstrations on using such industrial sensors with microcontroller. Anyway, I still prefer to use an mcu for decision making in the design because I don't have experience with PLCs. The problem is that the output level of a typical industrial sensor (e.g. inductive, photoelectric) would be high (I guess that it's because they're designed to directly drive loads) to be fed directly intoa typical mcu with standard +5 V max. To bring the level down to +5 V, I believe that a typical transistor switch as a level shifter would do the job providing a protective interface between the output of the sensor and the input of mcu. How about this solution? would it be more ideal over things like voltage dividers or optocouplers?

[aflex]

For me it will be a more better idea to consult with some linear actuator manufacturers. They will have a clear picture about the actuators. They will certainly let you understand the specific uses of specific actuators and provide the best solution.

I got one of the reference - Venture Mfg. Co. (http://www.venturemfgco.com/) from Dayton, OH. They are experienced and experts.