Optimising data comms using ERF,XRF and SRF devices

Perhaps one directly for Technical, but all response appreciated.

I am constructing a "personal area network" (PAN) using a combination of ERFs (PICAXE firmware) and XRFs and SRFs. All working most effectively within a short but reasonable distance. I am sending 6 byte packages through the system to control various devices, with a reply-back protocol to ensure the transmission has succeeded. I am using PANIDs and repeaters. Most of the nodes are connected to 18M2+s but some may have other PICAXEs such as 40X2 later. The throughput of data is low; 6 bytes in, pause for a second, 6 byte reply. Low frequency - once or twice an hour at maximum.

The PICAXE Editor 6 has splendid and effective wizards for setting up these transceivers and my question relates to the optimum settings (Using the interface board that I constructed according to Hippy's directives http://www.picaxeforum.co.uk/showthread.php?26188-Configuring-ERF-with-Wizards-and-Ciseco-Explorer-Plus). My prime concern is achieve maximum range and maximum reliability. Therefore:

1). In the Data Transfer Tab; what would be the optimum data rate to set? Default is 250K. Would a reduction increase range and/or reliability?
2). Is there any advantage to altering the packet length from the default 12 to the standard 6 bytes that I am using?
3). Is there any advantage to altering the TX/RX or RX/TX delays in this application?
4). The default serial port rate is 9600 baud - Obviously for this low data rate I can work with a lower rate, is there any point in changing this?

After a venture into X-Bee, these devices are so comparatively easy and effective, but a little info on optimising the settings would be much appreciated.

Circuit said:

1). In the Data Transfer Tab; what would be the optimum data rate to set? Default is 250K.Would a reduction increase range and/or reliability?

Click to expand...

In theory reducing data should result in an increase in range or reliability at the same range.

However, the practice may be different depending on how the comms have been implemented in the RF device itself.

To be sure, its one of those things that you have to actually test.

For the payload size, 6 bytes, you could use a very low data rate indeed.

For a UHF device, in the UHF band at the license excempt limit of 10mW, devices of this type should be capable of 12kM or so for direct line of sight.

However you also need to appreciate that range will vary hugely depending on the environment. In an urban environment, no direct line of sight, the range might drop to 100M or less.

Changing the packet length or TX/RX and RX/TX times can reduce data latency, the time it takes for data to be received after being sent, but probably does not matter in the setup you have.

For a fixed packet size you should only need to alter packet length, and I wouldn't then expect any gain from also altering TX/RX and RX/TX times.

I also can't see much to be gained in altering from 9600 baud.

Thanks for that, Srnet & Hippy.

I guess that Hippy is suggesting that I should just leave things on the default settings; perhaps that is the best idea.

srnet said:

In theory reducing data should result in an increase in range or reliability at the same range.

However, the practice may be different depending on how the comms have been implemented in the RF device itself.

To be sure, its one of those things that you have to actually test.

Click to expand...

Yes, I thought that might be the case; range I can certainly test but it is tedious. Reliability is more of a beast to measure and validate statistically. Oh Bother! Given your "sole-worn boots of experience" with radio, Srnet, would you just stick with the default values or would you set the data rate lower? I think I read somewhere that reducing the data rate with these devices did not improve the range because of a firmware issue, but for me reliability is really the key need.

srnet said:

However you also need to appreciate that range will vary hugely depending on the environment. In an urban environment, no direct line of sight, the range might drop to 100M or less.

Click to expand...

Oh, how right you are! My garden is 100 m long and mainly wooded; I cannot get one of these devices to reach from the house to the outbuildings at the end. Fortunately Ciseco document the antenna connection on the ERF module with instructions for replacing the chip antenna with a quarter wave one; vastly improved performance! Anyway, XRF repeaters with appropriate PANID settings seem to do the business. The same applies to the house which is rather solidly built with double-brick internal walls; no plasterboard/studs and twenty metres from one side to the other leads to less than ideal RF conditions. Trying to get through eight layers of rather dense house bricks is rather challenging. (I am rather reminded of our exchange on autonomous flight and just wondering, with a chuckle, if you are going to suggest that I just pop down to B&Q and buy a couple of hundred metres of cable instead! But seriously, RF is the way I want to go.)

I guess that "radio guru Stan" down in NZ is likely to glance at this thread - any ideas Stan? How would you set up the various attributes in the ERF units for best performance to send 6 bytes back and forth occasionally?

Anyone else with experience in altering the settings in the PICAXE ERF wizards?

Circuit said:

I guess that Hippy is suggesting that I should just leave things on the default settings; perhaps that is the best idea.

Click to expand...

Pretty much but I wouldn't rule out altering Data Transfer Rate to see if it improves things. The way I would test that is to start with 250K and move TX and RX apart until reliability drops. Then try other settings to see if reliability improves.

Circuit said:

I think I read somewhere that reducing the data rate with these devices did not improve the range because of a firmware issue, but for me reliability is really the key need.

Click to expand...

I have never tried altering the data rate beyond testing that it can be changed and it probably would be useful if someone did some tests and measured the results for range and/or reliability.

There's an RSSI reading which can be extracted from the receivers using an AT command which might be useful and, for later firmware, there may be some test modes which bounce RSSI readings back and forth which could make things easier to measure.

I've fiddled with all manner of settings more out of curiosity than need.
The only things I've found that effect range are:-
Frequency - has a large effect.
Power level - speaks for itself.
Baud - no noticeable effect.
Horizontal vs vertical mounting. - depends on location and surroundings.
Data rate - not tried yet.

Long story short, the default values appear to give the best range.
I've never been able to get anything out of RSSI but mine are quite old and I think it was only implemented in newer versions.

Circuit said:

Yes, I thought that might be the case; range I can certainly test but it is tedious. Reliability is more of a beast to measure and validate statistically. Oh Bother! Given your "sole-worn boots of experience" with radio, Srnet, would you just stick with the default values or would you set the data rate lower? I think I read somewhere that reducing the data rate with these devices did not improve the range because of a firmware issue, but for me reliability is really the key need.

Click to expand...

Actually its quite simple to test relaibility, its only a guide you want, not evidence for a Scientific paper.

I would be amazed if distance and/or reliability did not improve if the data rate was lowered from 250kbits to say 5kbits or lower. No guarantees, of course, maybe the chip set was actually optimised for 250kbits, and it really does not like working at slower speed.

A testing program of some sort is very useful. Just set up a transmitter to send the packet\payload say every 2 seconds. Set up a receiver to beep a buzzer, flash an LED when a packet is recieved. Have the reciever count the number of packets received too, and print to serial terminal. That then is the basis of a simple test rig.

Set the TX and RX up a short distance apart, on say plastic tables in the garden. You may need to cut the TX power a bit so the range comes down to small backgarden range, putting the battery powerd TX in a metal box is one way of doing it.

With the TX and RX running, move them apart until the RX stops receiving about 50% of the time.

Run for 5 minutes, how many packets recieved ?

Reprogram for a lower data rate, keep TX and RX in the same place.

Run for 5 minutes, how many packets recieved ?

You can pick holes in the scientific method, but as a indicator of the effect of changes (to data rate etc) its good enough and simple to do.

hippy said:

There's an RSSI reading which can be extracted from the receivers using an AT command which might be useful and, for later firmware, there may be some test modes which bounce RSSI readings back and forth which could make things easier to measure.

Click to expand...

It seems I was perhaps wrong on being able to read an RSSI value - confused it with XBee - but the RSSI bounce test is described here -

http://www.openmicros.org/index.php/articles/84-xrf-basics/146-rssimode

Maybe this is a silly question but have you tried asking on openmicros.org, where you might get a response direct from Ciseco?

Thanks for all the most constructive advice; looks like I need to get the picnic tables out and spend some time in testing after all. Hippy, your reference to the RSSI bounce test looks most interesting. I will try and find some time to implement this; I think that it will help in optimising the antenna locations around the property if nothing else.

If you don't need high speed communication between the devices then go with the lowest over-the-air data rate, 1.2k baud.

Ages I made this little vid to show the difference between 1.2k and 250k in terms of how much radio transmission there actually is with the same amount of data within the same timespan, holding one of those spy bug detectors next to the XRF to visually show when the radio is transmitting:

With the original v1.3 XRF's I was just about able to achieve 3.4km line of sight communication running at 1.2k baud, the subsequent v1.5 XRF's had some changes made which unfortunately reduced their range running at 1.2k baud, but still managed a respectable 1.7km distance.

I guess that "radio guru Stan" down in NZ is likely to glance at this thread - any ideas Stan?

Click to expand...

Modest blush ...are these modules at 433 or 866 MHz ? I'll assume the former -at 433 MHz your 10mW TX (we're allowed a "paint blistering" 25mW here in NZ) may indeed be marginal thru' diverse RF attenuations. Herewith brief propagation/antenna slanted thoughts-

* Propagation can be somewhat of a black art at UHF freqs., since diverse reflections & RF "view corridors" may unexpectedly enhance/attenuate signals. I personally verified this just recently when moving our large digital UHF TV to another room (& well away from it's cabled Yagi rooftop antenna). Dreading the significant recabling chore,on implulse I tried a small portable antenna & found a bizarre but reliable signal sweet spot which gives perfect all channel reception.Yah !

* For your PAN needs I'd hence look to first doing a site audit with a signal strength equiped 433MHz scanner etc. Other band devices may be present too, & these may influence your RX sensitivity. Know your enemy!

* Elevation can be everything! Hence bypass the waist high picnic tables & instead raise the units as high as possible! (Ah-are you nimble enough to install & access elevated setups !?)

* Are you in a region where seasonal vegetation/serious weather/snow loading/animal damage could be an issue? Leaf fall & dry air can improve coverage significantly-even here in temperate NZ I occasionally note improved propagation when nearby deciduous trees are bare.

* At 433 MHz TX antenna are usually banned (but RX ones ARE allowed), so consider a passive reflector? Almost any sizeable metallic sheet/mesh at a suitable distance (traditionally ¼ wavelength) from the unit may assist. You might even be able to use existing metallic building features to good effect -garage siding, flashing, balcony railings,sat. TV dishes etc. Mirrors (silver backed of course) or metallic claddings (alumium foil backed screen maybe?) or metal sided home appliances could have influence indoors.

Footnote: Parabolic dishes only come into their own when several wavelengths across,so at 433 MHz this means a bulky device several metres wide. However the reflector doesn't need a solid surface -skin gaps of 1/10 wavelength (= ~70mm at 433 MHz) are tolerable. Hence quite large but low profile "parabolic" reflectors made from galvanised bird netting may have some merit. EXPERIMENT !

Extra: Be inspired by the higher freq. 3G (~900 MHz or 2GHz in our region) cellular setup below. He's used an aluminium sheet cut to approximate a parabola, with a 3G data dongle positioned at the resulting focal point. I once managed something similar for 2.4GHz WiFi with a "popped out" golf umbrella! Stan.