Distance from coordinator to end_device at association step is very low.

It's 3 months since the above post, and I am also seeing this problem using ZStack Pro on MSP430 with 2520/2591.

Does Zigbee require a stronger signal to allow a join?

I can send large test packets to a joined node at great distance, but have to bring the node 2-3 times closer to the coord to get it to join.

-Rick

Update:  I just got a reply from TI pointing me to this code:

nwk_globals.c:
// Minimum lqi value that is required for association
byte gMIN_TREE_LINK_COST = MIN_LQI_COST_3;

zmac_internal.h:
// LQI to Cost mapping
#define MIN_LQI_COST_1  12     //24
#define MIN_LQI_COST_2  9      //20
#define MIN_LQI_COST_3  6      //16
#define MIN_LQI_COST_4  4      //12
#define MIN_LQI_COST_5  2      //8
#define MIN_LQI_COST_6  1      //4
#define MIN_LQI_COST_7  0      //0

Hi Launix,

Have you solve this problem of yours? Thanks Cheers [:)]

Lbaerd

Hello,

yes, we managed to correct this issue. We saw a problem in the energy

measurment for the end_devices.

Launix.

Thanks for looking into this, Double 0, this problem's potential severity had me a bit concerned, though I have not yet experienced it myself (still waiting for PCBs).  Your explanation and solution seems credible and well thought out.  I hope it works out for us all.  Thanks!

Thanks alot for your recommendation. This sound viable. I will be able to try out later today and keep you guys posted. [:D]

HI Double 0,

I works!!!! Now the end device could join network at considerable distance (20m) in my environment. I guess it could be better. It's such a great help from you. Although I can't quite figure out all these constant, I suppose it's the calculation of the LQI.

 Cheers[<:o)]

I am able to go till 60 meters with these settings, and devices are getting associated in Line of Sight. But my problem is, i am able to send and recieve normal (not beacon) data at a distance of 400 meter, then why this is the problem for association. Is there some settings need to be done for LQI.

Hi All,

I found something mysterious in Z-Stack, they are calculating link quality using RSSI value but in CC2520 datasheet it is mention that it is having disadvantage (section 20.6. It is advised to use correlation value in place of RSSI. I used the same in calculation of link quality and got good amount of range. Now i want to go in open air and see the range, right now i got 60 meters with these values which i got previous settings as posted previously. But now i am not using those values in spite i have changed only to correlation from RSSI. I will keep you updated what range i got for association.

That would be great. Pls keep us updated once you have any good news

kudos... i got 130 meter till yet.. now again i ve to go in some other aream to find out how much extra i can get..

Hi Varun,

These are some good ideas, but I would caution against simply using just the correlation value since around 85 for the corr value means that the receiver is at its sensitivity limit (PER = 1%). This is with NO interference. The max it ever goes up to is about 107, so there is only a very small "useable" range for the corr value if you are using it for the stack LQI measurement. The intent of the stack LQI is to not give an accurate assessment of range. Corr value has a limited value for range measurements.

Also, if there is a strong interferer at +- 5 or 10 MHz you'll see a sharp change where if there is a strong interference of let's say -28 dBm, at 10MHz away, a corr value of around 93 means PER = 1%, whereas if it is at 5 MHz away, the PER shoots up to 100% and therefore the corr value is 0. In both these cases, the wanted signal is at -85 dBm.

Something clever might be to combine RSSI + corr value to ge the best of both worlds, as suggested in the datasheets.

Regards,

-- "Double O"

Hi Double 0,

You are right, i have also gone through some of the papers, they also suggest use combination of RSSI and corr value to calculate LQI. But they don't mention how to compute it. So, i used only corr value, if performance is good, then i guess i should stick with it. As my PER is coming to 0.45% @ 400m. So ideally association should be successful at that level, as radio is able to recieve packets at that distance.

If you can suggest something, that would be great.

Regards,

Varun

Hello,

I'm using the TIMAC with the CC2430 db.

the CC2430 has a temperature sensor?

I would like to ask if i can mesure the temperatire by this sensor then send the value to the coodinator?

Thank you

Bs

Oussama

Hi Radek,

1. Are you sure about your PCB (especially the RF part) design, wrong or ill design may lead
   to very poor RF TX/RX performance.
2. I suggest you to check if CC2591 configured for maximum dB. 

1. I will check it again, but all rules was carrefull check before we made this boards. I remember some lenght from CC2530 to CC2591 and next to chip antenna. This was very important rule I think so.

2. In code I check settings and I think that all is ok. I have HGM on 1. Have You got some example/test project? 

New: I was reminded of one thing, we use to test four boards on two I get ~60meters from board to sniffer on my laptop, on second two boards I got ~400meters.

I think that two of them are broken, mayby CC2591 don't work correctly but on other two mayby work good.

Radek Bazwisko said:

1. I will check it again, but all rules was carrefull check before we made this boards. I remember some lenght from CC2530 to CC2591 and next to chip antenna. This was very important rule I think so.

The stack up is very important too in these cases (not only tracks' length and width)

I ask my college, he look at this project again and board structure looks ok.

Mayby in code you change some more settings, that I don't have.

Hi,

If you say that two boards works fine and the other two are not,

while the same FW loaded on both sets, then something tells

me to go and look in/on the PCBs.

Igor, yes only two boards works fine, but as you can imagine I use them to tests, other to are under my desk and wait for corrections.3

Then now we have two boards working fine they have been getting "Beacon request" from range 300meters to sniffer,

but they coundn't connect to each other (two boards) using zigbee protocol on about 30meters.

These recommended parameters:

#define  MAC_RADIO_RECEIVER_SENSITIVITY_DBM      -98 /* dBm */ // orig  -95

#define  MAC_RADIO_RECEIVER_SATURATION_DBM       -4  /* dBm */ // orig 10

Should they be the same for a CC2520 platform and CC2520/CC2591 Platform?

At this point just comparing the CC2520EM and CC2520-CC2591EM.

Thanks

Hi,

I made two more boards independently with CC2530F256 and CC2591. I used those settings which Double 0 wrote. Now i get ~110meters.

But question is that on this settings, the Zigbee Logic will be maintained?

And second thing is that i use CC2530 with CC2591, shouldn't I have a better range?

We are looking at solutions for this, because it does not meet our expectations.

Greetings.