Wireless solution needed, suggestions please...

Hi Folks,

I am in need of a wireless link for my project. My project will accumulate about 1gb of data, but I can only physically get to it once a month. The link needs to be able to transfer a whole 1gb (from an SD), preferably fairly quickly. I mean like an hour or two at most. It needs to have a 50m+ range. (Not necessarily LOS). THe SD is interfaced to an LPC ARM, and the arm needs to pull the data off the card and send it to the RF TX. This bit I can do, but I don’t know enough about the wireless products to help me choose. Obviously, x bee and so forth have a limit of 115600 I think, which is waaaaaaaaaay too slow. I need a robust link in a small size.

I was thinking of a cellular module, as that way I can pull the data off whenever and wherever. However, from what I can see, this is another USART only device.

Any kind of RF link considered!

I definitely like the nRF24L01 chips from Nordic. While they’re not trivial to interface, they can transmit up to 2 Mbps (to keep in spec, about the best possible data throughput is around 1.75 Mbps). I have examples on my website (see the signature below) in C for the 24L01 with an LPC2148, which is basically an LPC2138 with USB.

Awesome, thank you for the quick response. I will have a look at the products later, can I assume they meet the requirements as stated above?

You should have no problem getting 50+ meters of range, especially out of the RP-SMA version, if you could get LOS. But without having LOS, you might have a hard time. The RP-SMA could also be interfaced to a coaxial cable to put the antenna in a place that you could get closer to LOS. It’s also possible to add an external power amp to the chips to get more range. A higher gain antenna is generally a better solution than a power amplifier, though.

aaronharper1984:
Hi Folks,

I am in need of a wireless link for my project. My project will accumulate about 1gb of data, but I can only physically get to it once a month.

Any kind of RF link considered!

No doubt you considered using a removable media storage card (like SD).

as you look at wireless - beware yield versus air link bit rate. Yield is 40 to 80% of the air link rate - depending on data payload overhead bytes, TDD delays, media access delays, forward error correction bits, probability of retransmission for error correction, and so on.

Example: WiFi 11g at best is 54Mbps air link and rarely more than 22Mbps yield at the IP layer. About 15% of this is from IP, the rest is duplexing, CSMA, ACK delays in 802.11 and other causes.

Stevech, I have no idea what you just said :shock: Can you put it in laymans terms for me?

Brennen, I don’t have the option of a repositioned antenna, but I could put a better/bigger one on my board. Getting LOS is going to be nigh on impossible most of the time. This is a wildlife tracking / monitoring project, with a mobile ’ target’. Similar to the cat-collar projects, except for badgers and other endangered animals. Getting close is the best I’m going to be able to do. The best time to retrieve data is when the badger is in the sett, which is gonna be 2-3 feet underground, bit like a rabbit warren. With that in mind, could your idea still work? A spinoff project wil be a ’ nursery cam’ where the unit is left in the sett before the sprogs are there. Hence the high data usage. Therefore it needs decent tx power. I can put a silly amount of batteries on site, so power usage is not really that much of an issue. Getting the signal and hence data, in a short amount of time to minimise the intrusion is critical.

I know there may be a better method of doing this whole project, but this is what I’ve been tasked with.

Also, on your site, the pdf Tutorial 0 doesn’t work??

And another question to you. My target uC is the 2148, so that is a right stroke of luck. However, I don’t know the pin connections for the nordic to the 2148. Can you help?

Thankyou for your excellent help!

What stevech was driving at is that the actual amount of data that you are able to send from one node to the other is generally 40% to 80% of the quoted air link rate. The 20% to 60% that you lose is generally overhead due to the protocols that allow the data to go to the node(s) you want it to, while minimizing collisions and (in certain protocols) automatically retransmitting data if it is not received properly or at all.

One great advantage you would have with the 24L01 is the fact that it has extremely low power consumption. Since you will be strapping your device on animals (at least as I understand), you’re obviously going to need to hook it up to a battery. As far as the range goes, you’re probably going to be best off to use the RP-SMA version of the board and get as high of a gain antenna as you can find that can fit your particular animal. You would also want a high gain antenna for your main unit that you would carry around.

As far as how to hook up the hardware, that should be in Tutorial 1. And the broken link was actually because of the fact that I recently switched my hosting from Windows to Linux, and there was a conflict with cases between the file link and the file name (Windows doesn’t care about cases but Linux does). It should be fixed now.

aaronharper1984:
Stevech, I have no idea what you just said :shock: Can you put it in laymans terms for me?

Bit = smallest unit of data. A bit has to binary values: 0 or 1.

bps = bits per second, or data rate.

Yield (data rate yield) is the useful data rate to YOUR program, after overhead for protocol and error detection and possibly error correction.

If you don’t know these basics of wireless, (and we just skimmed the surface), then use low cost Xbee series 1 modules with their serial port emulation configuration. All you have to do is send AT commands like a Hayes modem, to set the source and destination addresses then just send and receive data as if it were a serial port. But even this is oversimplified.

In wireless data, a robust protocol is needed. It isn’t trivial. So if you don’t want to go into that learning deep-dive, then go the plug-and-play module with the protocols provided within the module’s microprocessor and all you have is a simple serial port interface (logic level, not RS232).

Maybe use a very simple/small wireless link for the collar unit, and station a more complex unit nearby which can receive the signals as they happen and log the data. Either walk up to the base unit and swap SD cards, or use some other high speed transfer mechanism like a USB cable or WiFi.

Once again thank you for your help.

Stevech, I’m afraid I do not know the basics of RF and wireless. My chosen ’ field’ has been sensor systems and data processing, not sending that data. I’va always been able to do just what macegr has suggested.

So my apologies for my ignorance. I’ve gone down the ask for help route because I understood RF to be a bit of a black art. I have until February 09 to have a working prototype of this project, and I did not feel I would have the time to design and build the rest of the project if I spent 6 months learning the do’s and dont’s of RF. What I need to try and do is, as stated first, find a “solution” that I can tailor, but not necessarily write/build from scratch.Which I hope brings me neatly back to Brennens idea.

Note: I think I now understand yield versus air link bit rate!!!

thefunman:
Try the TRW-24G from www.csjelectronics.com. They have 280M range and are low cost under $15 and very easy to use in C code. I have bought and used many of these from them…

Spec sheet says TX Pout is (about) 0 to 1 mW. OK for really short range operations. And no built-in full-up MAC software.