Reliability of a 2.4GHz link in an industrial environment?
Datarate: max 250kbps, min 100kbps
Distance: about 20-40 meters, line-of-sight not guaranteed
Environment: light to heavy industrial environemnt. Ranging from big open fields to sub surface mining environment.
Application is not _critical_, but still the reliability requirements are high
How would a simple 2.4GHz radio link without any spread spectrum technique work do you think? This link would then be based on any of the currently available 2.4GHz chip out there (eg NordicSemiconductors nRF24L01+).
Is the ISM-band around 800MHz significantly better?
Would a Bluetooth-link do the job best? Is it worth it when you concider BT is more complex and costly.
WLAN is probably overkill.
Perhaps there are other techniques that I have not concidered?
All input is helpful here!
If you are using a computer then what you describe sounds almost like a wifi network. However your datarate is slower than the slowest datarate I've seen on my laptop with a weak signal.
You state a 'subsurface mining environment'. Does this mean you might have 40 meters of rock intervening? That could be your biggest obstacle. Find out if any particular wavelength has been found to work best through soil and rock.
You probably want to use a protocol that automatically drops to a lower bit rate in order to achieve reliable communication. (This is pretty much standard with computer wifi.)
Industrial equipment is liable to send out rf spikes and other interference.
You probably want to use a protocol that echoes each frame of data back to the transmitting instrument, in order to verify correct reception. If data doesn't match then the protocal should order a re-send, until matching data is echoed back. Don't know what degree of tenacity of this capability is built into computer wifi.
The non-line of sight condition means that sometimes the information is coming from reflectoins off of other objects. So you will have a known poor multipath environment. You have low data rates and fsk modulation, so the problems will be that:
1) sometimes you will just have a very poor signal level received since nothing is getting through
2) sometimes you will have a poor signal level because you are receiving 2 equal amplitude but out of phase signals that completely block out the 100 KHz wide bandwidth
Not much you can do about item #1. You want to transmit with as much power as the regulations (non spread spectrum) will allow in your country. You also want as omnidirectional an antenna as possible (so you do not have any reception nulls as the unit is rotated).
Item #2 you can do something about. You can send the information at least twice, once at frequency one, and once at a different frequency two. Also you can have two receive antennas and two receivers on one end of the link, and simultaneously receive from both of them. Then you just pick the receiver that has the strongest RSSI (or no packet errors).
The Nordic NRF905 would be good for the 868 and 915 bands depending on the country you use it in
The company I work for have used the 2.4 and 868 chips
We found the NRF905 more reliable at the 40 metre range
The NRF905 worked really well for us until we had a several reels that were intermittent
It turned out that even though they were from different batches the bonding wires were breaking off inside the chip
To be fair to NORDIC they replaced the reels FOC.
But reworking the PCB's was quite expensive
The last batch we had seemed to be more reliable.
Using the NORDIC parts it easy to do simple channel hop if the channel is busy
"Subsurface mining environment" just means that the system will potentially be used in an under ground area (mine). Communication surface<->underground through soil/bedrock is not what I meant
If realised, this setup will be used to link som process data values to a display. (Therefore not critical, but still reliability is very important)
Yes, the idea of implementing some kind of a very simple frequency hop algorithm has struck me (just change channel according to a predefined sequence if transmission fails for several packets).
If this project is realised, some kind of communication protocol must be designed anyway...
But, that's for later...
--> Right now the most important thing is to choose frequency band and technique. It will be a little hard to change this later in the project. So, you can say that what I'm really looking for is theoretical support for choosing a technique/frequency band for my project. Also, the kind of hands-on experience that SteveNun is describing is also very interesting
Side note:
Manufacturing problem can happend to anyone. I have not heard about NS having problems before, but I will keep this information in mind. Thanks SteveNun
One thing to think about is the countries the product can be legally used in
2.4GHz is a practical solution if the product is to be used world wide
868/915 on the same chip covers EEC & USA etc
That can shorten your options
Yes, I know there are potential issues here. I have not checked this up yet, but thanks for pointing it out!
Just to wrap things up: the project was canceled and this thread can be considered closed.