Peer to peer Simpliciti data lost

Transmitting every 4ms in SimpliciTI is a rather tight timing budget.  Note that if there is any failure of the CCA algorithm, the random back off period is betwen 1ms and 16ms.  If you get a single back off, you have probably blown your timing budget completely.  You might try using the forced form of transmit but this may not qualify for production in the ETSI deployment theater.  Any chance you can combine the data into larger packets and send them at greater time intervals?

I'm a little unclear about you description of the channels.  Since you cannot receive but one radio at a time, wouldn't it be easier to simply use a single channel and time multiplex all the radios together?  If you are suggesting you want to communicate all 4 radios at the same time on 4 different channels, you will need 4 receiving radios as well, is this what your system is?  Either way, communicating every 4ms is a tight timing budget for SimpliciTI.

Jim Noxon

Hi Jim thank you for quick reply, is correct I have 4 peers transmitter and receivers, each peer is working in different frequency channel and different network configuration in order to avoid interference between peers.

the application works very well when I only one of the peers is turned on, I meant the system doesn't lose any data, but if other of the peers is turned on both of them los a few data packages, but when the four peers are working all of them lose many information packages, am not clear why they lose information if the channels are separated between them.

There is a WiFi network operating nearby of my network but I can see in my spectrum analyzer that the WiFi power transmission is less than the signals of my application, besides even if I use a channel where the WiFi has its max power the peer has no problem for exchange data.  So it seems the problem only happen when more than one peer is working.

I cannot explain why this is happening.  What I can do is offer some advice in debugging the problem.  I would start by writing some code to check the RSSI values and see if the energy from one channel is interfering with another.  If all is in good working order, the RSSI value on the radio not receiving the data will be pretty low.  You might want to work with the receiver bandwidth to optimize this a bit.  Also, if there is any leakage of one channel to the next, you will want to make sure the sync words are different on each channel otherwise you could be seeing false syncs from other channels.  I guess, what I'm saying is we need to make sure the channel separation is providing the expected isolation.

Next, I would toggle some GPIO pins when transmissions occur and when receptions are valid.  Then see if there is any correlation between the lost packets and when transmissions occur.  Are these times when transmissions occur simultaneously or is there any other pattern to the failure?

Jim Noxon

Hi Jim,

I'm still working in this problem, now I'm using the packet sniffer to watch the simpliciti data, but I have a question, in the packet sniffer software we can chose the channel we want to receive, and I chose the 0x0b channel (2405MHz) which is the channel one in my application, but in some cases as you can see in the image below, I receive information of transmitter 1 and 2, not always happen but it occurs when I'm working with several transmitters and when the application lose data.

We can watch in the payload 2 header bytes, 0x3131 is the ID 1 and 0x3232 is my ID 2.

Could it be a bug in the packet sniffer or could it be that the CC2530 are changing the channel and making interference between them?

I found that this problem seems to be spurious transmission of the transmitters devices, because the problem increase if I put closer the transmitters from the sniffer, so I think that my problem could be the emission of spurious signals from the transmitters, but still happening even if I decrease the TX power on the signals.

So I think it is possible to reduce it by an incoming RF filter tuning, do you know how can I change the bandwidth of the incoming RF filter in the CC2530, I meant what registers should I change?

Regards

Hello Jim

I was making test on this problem, using an spectral histogram analysis I can see that when there is only one transmitter working the transmission is ok, as shows in the image below…

The power on the transmission is constant over the time, this is the transmitter 1 and is located in the channel 11 of the CC2530 (2405MHz) at this moment there is not any data lose and the application works ok.

Then I add a the second transmitter which is located on channel 16 (2425MHz), as you can see on the image below, the transmission seems to be paused, on time intervals, but the problem is that appear some spurious transmissions or signals interference in many channels which seems to be making noise in the other channels, marked in the orange rectangles on the image.

I suppose that the CC2530 is doing this by some hardware function because the application is not using the frequency allocation function of SimpliciTI. At this moment still working well tha application, but if I add one more transmitter there are many data lost in the application.

So I wish to know if the CC2530 is doing these spurious signals by a hardware function, or could it be a calibration problem?

How can I ensure that the CC2530 is working well? I meant, is it any register to modify or something that disable or reduce the noise in the other channels?

Also I use the CC2500+MSP430 for this application, and as you can see on the image below there is no noise on the other channels, in the image below you can see the 4 transmitters working together at the same time with any lost information, but I need more power on the signals for my application, because of that I need to use the CC2530+CC2591 device.

Are the CC2591's all using the same power supply?  The additional energy at the other bands indicates there must be some modulation occurring between the two radios and the only way we can figure this would be happening would be if the power supplies are not sufficiently isolated.  Can you take a look at the spectrum on the power supplies when running both radios together and singly?  You may want to look at several of the decoupling caps.

As for the blanking between transmissions, one theory is the RSSI measures in-band frequencies but the RF front end will amplify wide band energy as the receiver bandwidth limitation is applied after the IF stage.  Thus even if the RSSI is indicating low energy in the selected band, the AGC will want to increase the gain of the RF front end which could be saturating due to other channels if the received energy in those channels is large as might be happening on a lab bench with the radios closely spaced, especially if the CC2591's are enabled.  Can you lower the output power to a point where the received signals are in the -80dB range and see if you still get drop out issues?  Any chance you can run this experiment without the CC2591's?  I know this sounds like a channel separation issue, but keep in mind the separation capability of the part assumes the amplifier is not saturated for one reason or another.

Jim Noxon

Each transmitter has its own power supply, but the RX are sharing the same power supply, because they're connected by USB ports to a computer.

I also have made the test without the CC2591 already and the spectrum and the drop issues still happening, so if I use a lower power on the transmitter hasn't any positive effect on the problem.

I'll also tried to use the channels chossing the ones that seems to have less interference issues and it seems to be working well, but is not the best way to solve the problem, because there are many RF interference in many channels beside the ones I'm using now.

I will make more tests

Can you describe your test setup more completely?  Include information about your data transmitted like data rate, packet rate, power settings, radio positioning, etc.

There just isn't any history of any issues like this with this part.  There's got to be something I'm missing.  If you are concerned about IP, you can start a private conversation with me from my profile page.

Jim Noxon