Despites the appearances and the recent “problem” with some of my code, I’m still working on coding
My goal with this project is to use it as a replacement for the “time system” (I don’t know the english word) that employees are using in my business. But, as I’m doing it, it could serve other purposes, like monitoring some machines in the factory, when equiped with appropriate sensors…
The principle is simple :
- each board acts as a TCP client which connects to a Windows TCP server
- each time something happens on the board (here a RFID tag to read), it sends the data to the server that will deal with it
Of course, at this stage, this is more a proof of concept thant a real industrial project. Mainly because I still don’t know how to deal with network since the FEZ Connect shield is discontinued.
This is why I’ve used my Moxa shield on the Panda : this module can act natively as a TCP client (among many other things) and connects to the Panda via serial TTL. This is an “old” shield I’ve made for something else but it’s now useful You can find more info here : GHI Electronics – Where Hardware Meets Software
I will try to replace this module with this one : http://www.moxa.com/product/MiiNePort_E3.htm It doesn’t need soldering and can be placed “where I want” in the future device.
Also, the use of such a module simplifies the code in the FEZ board as I don’t have to deal with TCP but only serial communication.
For the test, I’ve used a Panda II with a Connect shield and coded the TCP client (using a modified version of this code : http://www.tinyclr.com/codeshare/entry/198 ). On this Panda II, I’ve connected the Seeed RFID /TTL reader.
A third board is used in this test, which is a Domino (also discontinued, no luck ) with a “special” Connect shield. Here again, I’ve had to deal with TCP in code. As I had only one Seeed RFID reader, I’ve put a Phidget RFID on the Domino USB host, so that all the test would “look like” the real thing with many RFID reader on many boards in the differents workshops.
All in all, it does work pretty well. TCP code is small enough, RFID reading is ok (as shown on the Windows TCP server capture).
Now, the biggest part is only starting : I have to create a set of commands/responses so that the TCP server can talk to its clients and so that they understand each others I already have such an experience with a (commercial) software I’ve done for astronomical cameras, so it won’t be that hard, I think. The hardest part mainly consists in choosing the right commands and their answers.
Also, how it will be delt with on the server is another great part. There will of course be some database and nice windows, along with a connection to an existing software we already have.
The example shown here only deals with RFID but you can imagine that I can/will connect some sensors if I want to use it to monitor some machines in-house. Only the commands/answers would need to be changed (and the server part, of course). But the core wouldn’t differ that much.
About the network, it will use wired ethernet because of interferences and distances. And also because I have anticipated such things and I’ve put many switches at strategic places in the factory
Of course, I still miss one important thing : which hardware will I use ? It have to think at it a little more. The Moxa board (not the shield) is expensive and will only be used for testing purposes.
Once I get something really presentable, I will come back with code. Right now, it’s more alpha stage, only to see it can work. Of course, the code will be shared and useable by everybody (except one person, for those who know…). Maybe the only part that I won’t be able to share is the interface to our existing software, I don’t know yet. But this is not a critical part of the project for everybody else than me.
If you have any question or remarks, do not hesitate do ask or tell. I will be please to answer.
Thank you for reading.