Main Site Documentation

From Lab to Prod?


#1

Hello guys,
I own a FEZ Spider, and I’ve been a sleeper on this forum for a while now ;), still didn’t have the change to crack open my Spider and play with it, work related.
While away from the Spider I’ve thinking about a project that i would like to realize and instead of waiting to get home and play with the hardware i figured i may take advantage of the free time and write a white paper about how i would realize such a project in the most effective way ( stability, performance and economics), so i have most of my ideas down and how to do the whole process.

Now the only thing that is driving me nuts and i can’t quite figure out is as follow:

lets assume i want to read a sensor Ex: Temp and a PIR or Photolight sensor. in the lab all i would need to do is connect it to the proper socket and do the readings just fine. but in Production naturally the sensors are not going to be next to the Board, let’s say the Board will be in a central location, and each sensor will be placed at a distant location; let us assume between 20 to 100 meters… (Not an Bird eye View), cables looping around walls (cement walls) so on and so forth.

So is there a simple way to amplify the sensor’s readings so it would make it to the central unit without interference and intact? i mean how would one communicate with a sensor far from the station:

so far I’ve come up with the following:
Using XBee: cost prohibitive because each sensor’s price would go up. (yes i can put certain sensors in one location and use one Xbee) but still expensive.
Using Radio frequency: not sure about performance and stability of this solution. especially with interference.
Some other methods where using cables Ethernet (Cat-5e or 6e) cables would be best… how could one do this…?

To make this easy let us assume one of the most basic Project one could do at home: HOME AUTOMATION…

so i have a Spider near a Router and i have a PIR and a Light Sensor OUTSIDE, PIR for motion detection of course and Light sensor for light control based on the daylight. now how would those communicate with my spider without the above mentioned methods or using the above methods but without sacrificing Quality and increasing the per sensor budget?

Thank you.
Jay.


#2

Simply, you can’t.

It’s or wired, or wireless, it’s as simple as that.

There are cheap alternatives, but do expect a sensor-node to cost at least 10 bucks.


#3

A standard PIR detector (alarm system type) has all the conditioning circuitry built in. The unit will run off of 12VDC or 24VDC and usually provides a a set of relay contacts to signal when something/somebody is detected. This type of unit would be the simplest and most cost effective to use.

For something analog, like temperature or light level, you will likely be better off using a simple temperature/light sensing IC that has a built in communication bus or using a tiny uC to read a raw sensor and communicate back to the FEZ board via a wired serial COM link.


#4

The cheapest i can come up with:

Atmega328 5$
NRF24L01+ 3$
Various components (Vreg, caps, xtal, pcb) 1$
Your sensors and psu/batt holder to be added

So this node is able to connect to pir’s, temp sensors, ldr’s, … and send it’s values over RF to a central FEZ board that also has a NRF24L01.

Hope this gives you an idea


#5

Thank you guys for the quick replies.
Ok so i guess RF is the way to go… for now…
the only problem i have against the RF is not knowing up front how well it would perform at certain location, meaning we can have it all working at the lab, but may run into troubles at a client location for example. a building with a lots of interference for example or thick walls and so on.

so i always would like to have a backup plan in such situations, that’s the reason behind my question, know what’s possible at what cost.

@ EriSan500 (Eric)
thank for the part list i will look into those and check their specs…

I’m sure I’m not the only with these kinds of situations, but wouldn’t it be a good idea to have a WIKI for this sort of things.

especially on ways a board can communicate with a sensor that is not near by, but in the same location.

thank you guys.


#6

Why not use the electrical wiring? Or, in most buildings, there is already wiring for computer networks.
The internet is full of stuff like that.


#7

@ Lurch,
yes i was in the same line of thinking hoping to use the Cat5 network cables but i do not think you can just plug a sensor pin onto that cable and expect the value to travel 100 meter and make intact to the station…that’s why i was asking if there is a way to profile the data and send it down the wire…

as for using House Electrical wire i think that’s gonna cost more than the RF.

EriSan500 (Eric):
i check into the the proposed chips and noticed NRF24L01+ uses 2.4GH range which is prone to interference from WIFI… so that’s a no go…

i looked into the Arduino world trying to find out if anyone came up with a good way of transmitting data from sensors to base station either wired or wireless and came up short. i kind of wonder why not much have been done in this area. except for the Xbee solution… maybe I’m missing something.
and that’s why i think a WIKI for this kind of stuff will be a good contribution to the community…

thanks.


#8

You can probably find some inspiration in industrial automation.

At the low end it is just on/off signals like the PIR relay. This can reliably travel very far on cheap cable.

For analog signals you’ll find 4-20mA current loops. The idea here is to transmit an analog value by changing the current in a closed loop. It transmits a true value over distances without the effects of voltage drop. There are lots of sensors for this.

Going digital the choice here is RS422. This is a reliable bus also with plenty sensors. Cheap sensors on twisted pair cable.

CAN should be able to transport quite a distance.

Up the stack you’ll find Ethernet. Fantastic except that sensor nodes now get pricey. Lantronix make beautiful little Ethernet modules.

On the wireless side you get sensors with ISM band radios. Pretty reliable.

Then you’ll get Zigbee and then Wifi.

Why not mix it up? You can use relays and current loops to a remote station, then maybe rs422 OR wifi on the long haul to the central station. It is good to understand the strengths and weaknesses of each technology.


#9

Well, there’s also the RFM12B at 868Mhz, but probably is prone to interference from your garage remote 8)


#10

Thank you again for the information…

@ realiser
you gave me some ideas to try, but I’ll do more digging, and like you said “It is good to understand the strengths and weaknesses of each technology.” and that is exactly what I’m trying to do, find out what all my options are… and evaluate where and how each can be used in the most effective way.

do you have any sample sensor in this line… i would like to learn more about this.

What am i looking for here, i mean how do i find a sensor that does this. a sample link to a sensor would be a great start.

Thank you again for your help.
Jay.


#11

Look at these

http://www.globalw.com/products/we700.html

Read wikipedia to see how they work.

I use the Radiometrix modules on 433Mhz up to 1km. I think 2.4Ghz spread spectrum will be more reliable.


#12

Looks like you found enough.

[quote]yes i was in the same line of thinking hoping to use the Cat5 network cables but i do not think you can just plug a sensor pin onto that cable[/quote]The idea was to use the network with a Panda to run the sensor and send the data to another Panda collecting/using the data.