RF300 Range Issue

[trexmsuee]

Hi: I've got a serial link application using 900 Mhz RF300's (firmware 2.4.17). I've got several sensor "nodes" ucasting a 1Hz data packet back to a single "master" receive bridge node at a base station. Serial baud rate is 38400.

In the field, our effective communication range appears seriously degraded. Using 1/4 wave 0dbi dipoles (w/ ground plane) on the nodes, and a high-gain 5dbi omnibus at the master receiver, we can only get about .7 miles LOS! Using a yagi at the reciever I can do a little better (obviously).

I've done exact same test in previous years with the RF100's at 2.4 GHz, with the integrated F antennas in the nodes and achieved a range close to the datasheet rating. Something must be seriously out of whack...

Is there any possibility that the RF300 firmware I'm using DOES NOT default the txPwr to maximum??? Today, I was not setting it to txPwr(17) in code... I believe that there is a newer firmware release, but I can't find it on the forum? Any other ideas?

Thanks for the assistance,

[jcwoltz]

I do not know much about this issues, but I do know one thing about the RF300's. Have you tried setting NV parameter 53 Maximum loyalty to 0?

a quote for the snap reference manual:

ID 53 – Maximum Loyalty
This parameter, expressed in milliseconds, is valid only for the FHSS (frequency-hopping) firmware
for the Si100x (including the RF300). It will be ignored on all other platforms.

After transmitting or receiving on a particular frequency, a node will wait for a signal on the next
expected frequency for the duration of the loyalty period before it begins scanning all frequencies for
additional communications.
If a node has transmitted or received a message and its loyalty period has not expired before it
transmits its next message, it will transmit with a shorter preamble, expecting that the receiving node
is listening on the appropriate channel within its own loyalty period. The shorter preamble allows for
faster communications, at the risk of packets being missed by nodes that are not currently “loyal” to an
expected channel.

If you adjust the loyalty period, all nodes in the network should be set to the same value. Setting the
value to 0 means that no node will ever expect a loyalty period: all broadcasts will begin with a full
preamble, and all radios will scan all frequencies for transmissions rather than expecting a
transmission on any particular channel. The default value is 185.

I don't know if that will help you, but it might.

[IbarraE]

I'm in the process of investigating an issue dealing with this. I'll report back what I find. In addition to this, I've noticed that the 900 MHz channel space typically has more noise on it than 2.4GHz. If you have access to a spectrum analyzer you may benefit from picking a channel with a lower noise floor.

In the mean time, I'd recommend using uni-cast RPCs rather than multicast and increasing the number of retries.

[trexmsuee]

@jcwoltz - I've done some reading on this today, and it seems worth a try. I'll change, and report back what affect I've observed.

@IbarraE - We do have a spectrum analyzer, and that crossed my mind (but I didn't have it at our field test location). The "built-in" spectrograph indicated some activity on channel 4 (the default), with only two or three nodes active I thought it was fine. We were also in a very rural location, but there is most certainly cell phone activity that may cross near this band.

Thanks very much for the replies,

[vks_01]

I have had a similar issue with the RF300. Just tried installing the sensors with a half wave antenna. I found I dropped out at about 100-150meters. I was not LOS, but fairly close (some tall grass/crops between the 'master' and slave'). I havent had a chance to troubleshoot yet.

While in the field, I did a crude test polling link quality and it seemed like when the signal strength dropped below -90dBm, I started dropping packets very quickly, and lost connectivity completely within few additional meters. Has anyone else come across this issue?

[Jheath]

One thing that can have a real impact on 900 MHz operation is the proximity to ground. As a test, try lifting your device off the ground by a few more feet.

[PhillipH]

I tried using the RF300’s when they first came out in February. Range was a real problem so I have given up on them for now. I haven’t tried changing any of the parameters but would expect the default ones to be as the same as used to give the range in the specs of the device. Went back to using RF100/200 until the next version of portal / firmware gets released ( seems to be always coming out next week ) in the hope this will magically fix the range issues.

Setup - for testing I had 2 Engines set up. One is a master that simply sends a multicast message out every 3 seconds. The other is a slave that simply flashes a LED when this message is received. I also have a power on LED that flashes every 3 seconds as well. To check range, its fairly basic in that I walk watching the LEDS. If the sequence changes i.e. the Receive light starts to miss, then I deem the signal as getting unreliable. Engine / Portal versions are 2.4.17

Range
Was talking directly to one of the Synapse guys at one stage about it and he did go as far as testing the range himself but only within his building so didn’t end up with anything comparable to the los tests we have been carrying out.
With regards to the comments on height, I did my raise transmitter height to 4ft ( as per the synapse test setup from what I understand ) from 2.5 ft. Slightly better trace route figures, but didn't appear to help range at all. Also going to head height i.e 5 ft 8. Noticed first miss at approx. 50m. Signal reception ( no Rx LED flashes ) disappeared completely at about 210m.

At the suggestion on the Synapse guy I also tested using the traceroute in Portal rather that just my LED or no LED tests.

About 6m away through a wall.
USB Stick -to Unit 1 - 68 to 80%
Unit 1 back - 68 to 80%
USB Stick -to Unit 2 - 59 to 70%
Unit 1 back - 59 to 70%

Unit 2 positioned outside 10m away on road.
USB Stick -to Unit 2 - 51 to 58%
Unit 2 back - 51 to 58%
21 to 262ms round trip

Unit 1 positioned outside 100m down road from Unit 2. (USB Stick about 90m to Unit 1 but through a couple of walls )
USB Stick -to Unit 2 - 23 to 27%
Unit 2 back - 23 to 27%
19 to 121ms round trip - some timeouts start showing up in event log.

Unit 1 positioned outside 150m down road from Unit 2. (USB Stick about 140m to Unit 1 but through a couple of walls )
USB Stick -to Unit 2 - 25 to 33%
Unit 2 back - 25 to 33%
19 to 32ms round trip - no timeouts showing up

Unit 1 positioned outside 200m down road from Unit 2. (USB Stick about 190m to Unit 1 but through a couple of walls )
USB Stick -to Unit 2 - no response - timeouts in event log Unit 2 back - no response - timeouts in event log Dissappears Off Node list when click in Node views and pinging doesn't bring back Ocassional LED Rx Flash.

Portals Channel Analyzer.
Looking into my range issues I started to use the channel analyser to check what else was happening. With all modules off I get noise below -90dbm pretty much across all channels, but I presume this is normal. At about 912Mhz ( channel 25 ) I get a red square at about -90dbm which I presume is probably a cellphone frequency. Using the above setup with the units sitting about 3 feet away from a USB mounted engine I see blue squares start to come up right throughout the channel range when I turn the engines on, though with higher peaks on the top channels . I would have thought I should only see them on the channel range specified. Signal strength for these are between -40 to
-70 dbm. Did not get any reason why this looks like it does i.e. why it appears to be operating outside the channels specified.

Antenna
Tried 3 different ones, including the ones being used by the guy at synapse and the recommended ones.

So probably like you – pretty disappointed in these and wouldn’t recommend them ( but still holding out hope of the fact they should be able to work to specs. )


Phillip

[Frank]

We have the same trouble with RF100

The maximum range is 260m, free sight, good weather, same polarisation of the antenna....
My testsetup was like Phillip, with the difference, that i send a rpc command from a master to a slave and increment a couter. This counter was send back to the master and compared with a counter inside the master and view on the 7-sement-display...
4 test were carried out, with different antenna, RF100PD6 and RF100PC6...

[RickS]

RF300PC1 very poor range

We have a situation where we are in the installation process of monitoring temperatures in a very large distribution center using close to 2 dozen of the RF300PC1 (antenna built onto the PCBA). The range is VERY POOR as we might get less than 100' of range. We are having to use the RF300PD1 with SMA 1/4 wave Linx antenna to pick up most of the modules. It does have good range. We have the latest firmware installed and ID53 set to "0" as has been suggested but see no difference.

[gvoce]

Rick how high off of the ground are the RF300 modules? We have verified that due to the 900Mhz wave length as compared to 2.4Ghz that the RF300 devices need to be higher off of the ground to get the same range as you would if using the 2.4Ghz.

[RickS]

Thanks for your response gvoce. The modules are placed about 5 feet off the ground on commercial shelving in a very large distribution center-much like Home Depot. Right now we have 15 modules that are spaced about 70 feet apart.

[gvoce]

For test purposes can you raise the modules say twice that high? Also can you limit the number of nodes to say three until we get a good "network"?

Can I ask why you chose 900Mhz?

I have a customer that has deployed 20 modules throughout a manufacturing facility but is using 2.4Ghz with great success.

[gvoce]

What external antenna are you using?

[RickS]

All 15 of the modules are RF300PC1 (antenna on PCB). However, we are having to use a RF300PD1 with a 1/4 wave Linx to pick them up with the snapstick. The remaining 15 or so modules we are hoping to get installed tomorrow along with the E10. Unfortuneately, there are 3 floors and so the ceiling height is limited to about 8 feet on each floor, making mounting them on the ceiling difficult. The reason so many is that a major restaurant wants their food temps. monitored is certain key locations so we have temperature sensors in each module and are relying on them to be repeaters which they can't do right now. Therefore, we have to relay on the PD1 which is mounted at the very top of the highest part of the facility overseeing these 2 major sections. We intend to buy 3 more RF300PD1 with 1/4 wave Linx antennas to help cover the rest of the installation. The facility is 1 million sq. ft.

I certainly appreciate your help and guidance.

[RickS]

To check the network, for some reason "Traceroute" was working fine but has quit working (greyed out like no longer available). Is there a way we can get that back so we can actually see the communications between modules?

Thanks!