Pick to light 'prototype'

Projects General Guidance
1

Hello,
I am working now on pick to light project. (System should scan barcode and then light up LEDs assigned to the appropriate container. Detection of hand near right container should cause switching off the light (and of course some alarm for wrong pick).)
I've done some research to start with mock-up of the system. I'd like to use dedicated pick-to-light sensors to make hardware part of the project as simple as possible (sensor has integrated job lights). And here come my questions - how many sensors like this: (E-series, PNP no output) http://info.bannerengineering.com/cs/groups/public/documents/literature/125680.pdf
can I connect with one Arduino? Is Arduino Mega the best choice or is it better to use Arduino Uno and have less sensors per one board(or some external modules)?
There are also some sensors with IO-Link, but I'm not sure how it works - is it possible to integrate them with Arduino somehow?

Anyway, I want to start with 2 or 3 sensors and test how it works. If anyone tried to do something like this already, I'll be grateful for any tips/suggestions.
Oh, and is it necessary to use any pull-up resistors while connecting sensor? As I've read some say yes and others say that input circuit already has required resistors...

Have a nice day :slight_smile:

2

jamespi:

at $130 each ....
the limit is your checkbook.

3

since the cost of the lights does not scare you, I would offer that your project is extremely well funded. (Amazon warehose type money)

I would say that using a PLC would be the most professional way to do go, albeit with another large checkbook.

if you were to pose the question as :
I want to create a 'bin-pick' system. the major points are :

  • green light goes on in the bin as per the shopping list
  • motion sensor in the bin to detect that a hand had picked a part
  • motion sensors in other bins, alert sounds when wrong bin has motion
    ** possible red led in wrong bin and recording of wrong pick on SD card

I would offer that you have a huge budget and a very large project.

to light a hundred bins or so, with some simple LED's and motion detectors is not hard and there are multiple methods of doing this.

the questions you ask are not very pertinent at this preliminary stage of the project.

first, is how big is your pilot project ? how many bins ?
second is how big is your first full size implementation ?
third , how large is your maximum conceivable project ?

we would offer hobby style answers and answers that would use an Arduino.

however, if you make a hobby project and it gets huge, and it has problems, you are the only one on the planet who is familiar with it and you get that phone call that the system is down. at night, for years.....

but, to pursue that line, I would make a modular system.
a single Arduino as a slave on a row of bins, 8/16/32.... bins...
have each arduino part of a larger system.
you can have one Arduino talk to a hundred slaves

Each bin would have a sensor and light array. since this is hobby stuff, I would use simple LED's on a board with a motion sensor on that board. board could have a plug to connect to the slave Arduino controller.

as for modularity, the bin units would be very rugged. the slave unit would be easily accessible.
the master that controlled those slaves would also be accessible.

but, generating the signals would come from some PC that reads that shopping list.

as I said, the pilot program needs to be laid out as an overview
then refined to make sure the functions are right
then all possible features and upgrades are listed and included or excluded. (this is huge btw)
and then a prototype be made
then your pilot project be fabricated,

lots and lots of questions to make sure you do not commit to a path that will be abandoned later for one reason or another.

4

Watching a video, I like the light bar and the motion sensor.
the single light with a push to acknowledge would drastically slow down a worker.

I have to say that this is interesting.
you can scan a bar-code, and the bin would light. fill bins
or have the computer have a shopping list, sorted, so the picker movements are more fluid

5

I'd place a LED and/or buzzer near each container and have a single controller activate the place where the requested item resides, just like your car answers on a car park. Why force the user to visit all places, until he happens to reach the right one?

6

Thanks for your commitment!
You're right, probably I sholud give more details... First stage will be testing arduino with up to 5 sensors, to make sure that 'one module' will be working properly.
Main goal is to prepare system for workstation consisting of about 30 containers (<30cm width/container; I doubt that it can be more than 50 bins). It's assembly station, not a warehouse scale. So I was also thinking about modular system with Arduino to control each row of bins and some master device.

I know that using 'dedicated' sensor is not the most cost effective way (although it was initially accepted by my manager) and i'll consider doing it the right way (with motion sensor and some lights ). :wink:

7

There are lots of ways to do this.
I would think that the starting point is where the data is generated.
where does the list originate ?

a hobby motion sensor is less than $5
a couple LED's also less than $5
a board with some shift registers (both input and output)

a single UNO could have both inputs and outputs on an SPI bus
so you would get a signal that the sensor tripped
then you shut off that light, and turn on the next. repeat.

if a different beam were to be broken, then you send alert.

an output shift register can be daisy changed, so lighting 50 LED's or 50 LED strings is only a few pins.

the input shift register would show which bin was in play and that would signal to shut off that LED and turn on the next.

a port expander [ mcp23017 ]might be better as it can do both input and output 16 pins (8 in, 8 out)

up to 50.... I would consider using one Arduino.
for modularity, it would be less wires, but in the end, may be more work to break into multiple modules.

I was thinking 10's of thousands inside of a 5 acres under a roof, with 3 levels of shelving with 200 bins on a shelf.......

8

Thanks for next helpful post.

Ok, looks like one Arduino with expanders will be enough. For me it's already pretty big project, but in general 50 sensors is nothing unusual. That's why it's always good to ask someone more experienced.

I'll try to put it all together and start working on some 'demo' version with few sensors/lights to reveal (and hopefully solve) potential problems.

9

Getting the first 5 bins to work with all the bits and bobs you want will be a small, but challenging project and will use the most of a single Arduino.

getting the motion sensors to work, the lights to light, red for error or green for correct, and possibly white for lighting.

as for the next 45 bins.....

at the low cost of a Mega on e-bay, you could just use the Mega with it;s huge number of I/O pins for most of the project and maybe a second for the rest.

lots of ways to expand for more bins.

if you get stumped or have questions, let us know.

10

Maybe with a part verification array you could build an X-Y matrix and reduce the number of sensors substantially.