Need some RF Wireless advice

Go To Last Post
18 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I need to be able to reliably send potentially large amounts of data, in packet form over a ~100m max. range. I'm not as concerned about speed as reliability. I have been looking at the Hope RF12 from FutureLec and the nRF2401A from Sparkfun.
Is anyone familiar with these devices and would either suit my purpose or is there something else that I should be looking at.
I also need an inexpensive solution as I will be using quite a few of these devices in a master/slave application.

Thanks for any advice,
Mike

Happy Trails,

Mike

JaxCoder.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The Nordic devices probably won't work reliably at 100m without high-gain antennas. They are intended for short-range applications. The nRF2401A is obsolete, BTW. The nRF24L01+ is much better and is easier to use.

Leon Heller G1HSM

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Leon,

Thanks for you feedback, a point in the right direction is always a help.
I've been struggling over this for some time after purchasing a cheap RX/TX on ebay and quickly realized that they would not work for what I needed.

Happy Trails,

Mike

JaxCoder.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In addition to 100m range, how many walls are there between the devices?

If you made it to Spark Fun's site then you no doubt also saw the XBee series.

I found the XBee chip antenna units to be convenient, but somewhat position sensitive, (i.e. not omni-directional). I was rather disappointed in the performance of the 2mW chip antenna units within and around my house!

Note that the 50 mW units, with an external antenna connector, and a well positioned and aimed directional antenna would give much better performance, at a significantly higher cost.

The added benefit of XBee of course is that you can set up a network with intermediary relay stations to get the data from the most distant sites back to the cental unit. Perhaps Steve will comment further...

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The plan is to have a base unit near a window (preferably) or just outside the window and satellite units around the customers property, we are restricting usage to front yard. Only other restrictions will be the usual shrubs, trees, etc. so line-of-sight might work if we strategically place the units but ideally we want the customer to be able to just go out and drop them and not worry about anything other then the distance restriction which if given a 100m limit and we limit them to say 60m we should be safe. (Crosses fingers)

Happy Trails,

Mike

JaxCoder.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

We use Yishi ys-c20ua modules. They're I think US$23 and have been brilliant so far. They don't do anything fancy, it's just rx/tx power and gnd. They're GFSK so they work really well and since we're running at 433MHz the range is like 500M in a straight line.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks but I couldn't find any place to buy them or even get a price??

Happy Trails,

Mike

JaxCoder.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Use Digi XBee Pro (60mW), series 1, with the U.FL antenna connector option. Then we can do the math on what kind of antennas are needed. A mile line of sight is doable with modest antennas.

Apologies to Atmel.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Steve,

Thanks for your input.
I hate to be a PITA but I am on a real tight shoestring budget, basically unemployed and trying to present this as a proof-of-concept to get funding.
So can I ask a couple of questions:

Is the Nordic device line-of-sight also? didn't see anything mentioned in the literature. I really would like to stay away from line-of-sight if there is any way around it. And would it not suit our purpose?

If the answer to the above is not favorable can the 2mW XBee unit Series 2.5 be used instead? It has 120m range and would be sufficient for our purpose.

Thank you for your help, I am very inexperienced with RF devices and appreciate any help given.
Mike

Happy Trails,

Mike

JaxCoder.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Most wireless devices, if they specify a range, will generally specify "line of sight". That does NOT mean that it won't work if you do not have a line of sight situation.

It is just that not line of sight adds many more uncertainties that you cannot control. It is not as simple as saying "well, if it is good for 200m line of sight, then it will do 50m, not line of sight". What gets you is an "issue" called "multipath". The signal reflects off of various surfaces. These signals arrive at the receiver with varying strengths (depending on the size and orientation of the reflecting object) and with varying time delays (depending on the length of the path each one takes). These signals are all received and combine in the receiver according to their strength and apparent relative phase angle. You can have a receiver at one place and get an excellent not line of sight signal, and move a meter away, and get a really crappy signal. You have NO control over this; it is entirely due to the "physics" of the signals.

In the real world, multipath is often time-varying. It changes as reflecting objects change. This means things such as vehicles and people. For this reason, a strong error detection and retry mechanism can get you past a lot of it.

There are new technologies, generally called "MIMO" (Multiple Input, Multiple Output) that use several antennas at each end. It really reduces the effect of multipath. This is especially used with LTE, one of the touted "replacements" for WiFi. I have not yet seen it applied to 802.15.4 hardware, but there is no big reason to keep it from happening (except, maybe, cost).

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Jim,

Thanks for your explanation, that cleared up a lot of issues I had with line-of-sight.
Although I have not worked with RF I am familiar with RS232 and others so had anticipated that for communications to be reliable I would need to use a packet passing mechanism with a 16 bit CRC which is already in place using UART as a test harness. Additionally I anticipated that a need for some kind of signal strength arrangement would be needed so that once the customer "dropped" the box at a "good" location it would work as designed.

Thanks for your input,
Mike

Happy Trails,

Mike

JaxCoder.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Be aware that signal strength does not tell the whole story. It is certainly a pre-requisite but multipath can confuse that, a whole lot.

CRC is a good way of verifying the quality of the received packet. You will, however, need a mechanism to tell the sender to try again. What is used in some packet systems is a required ACK/NAK. That is, the receiver notifies the sender that a given packet is received adequately. If there is no ACK, then the sender knows that the packet did not even arrive at the sender in a way that the sender can tell who sent it. NAK is used to notify, usually faster, that the packet was received with an error. Sometimes, this is done at a "low level" as part of the protocol (ax.25 is an example) and sometimes it is done from higher levels of the exchange mechanism (tcp/ip is an example).

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Jim,

The ACK/NAK and auto retry are included in the stack I had developed. I want the user to do and be aware of as little as possible. The only information they should be told if the auto retry max. has been exceeded is that the satellite unit needs to be moved to a new location because the base unit cannot talk to it.
I've also dealt with customers enough to know it needs to Homer proof. :)

Again thanks,
Mike

Happy Trails,

Mike

JaxCoder.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Glad to hear that you've considered these issues!

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Would it not be easier, and probably more "budget" to use a pre-paid cell phone connected to your device via a GSM modem? You can buy those at wal-mart for about 40$ here, and all you need to buy after is a time card from your favorite service provider. No range or LoS issue there.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Non-line-of-sight multipath, in my opinion, is a non-issue for low power, short-range, low data rate wireless like 802.15.4 and Nordic and so on. Not so 802.11. Reason: the multipath delay (statistical) is quite short as compared to the period of the symbols.

Non-line-of-sight for such kinds of wireless is, IMO, mostly impacted by sheer attenuation by obstructions in the path. As a well experienced wireless guy, I can intuit the extra path loss for this that goes in a link budget with antenna gain, etc. But often, the black art of RF surprises me. Usually it's worse than expected. But now and then the prediction was too pessimistic. So is the non-line-of-sight world.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mike32217 wrote:
Thanks but I couldn't find any place to buy them or even get a price??

Have a chat to the sales people at http://www.yishi.net.cn/ .

http://www.yishi.net.cn/rf/Produ...

That's the product. They'll make it however you want. We have it with an SMA pigtail and vertical pins to solder straight into the pcb.

It's really a basic and reliable system. If you send something to the USART it will transmit and you'll get exactly what you sent on the recieving end on the recieving USART. Nothing fancy. Output = input.

The only thing is that if you have multiple pairs running in close proximity they will corrupt each other as the reciever doesn't know the difference amongst multiple signals on the same frequency band. I think maybe zigbee modules have hardware checking for that stuff but for this one you'll be limited to selecting from a range of 8 frequency bands if you want to run multiple pairs. Apart from that it's a great system.

EDIT: Just found out they don't pass certification in Australia (and possibly not anywhere else either).

Last Edited: Wed. Nov 10, 2010 - 04:33 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thank y'all for your help and just to let you know I ordered the XBee 2mW Series 2.5 Wire Antenna from Sparkfun. Should be good enough for the proof-of-concept and I'll be able to not only work with RF but see if range and data reliability will be suitable for what I am trying to accomplish.

Thank y'all again,
Mike

Happy Trails,

Mike

JaxCoder.com