SPIRIT1 link budget

[angeletti2]

Good afternoon,

I'm developing a low power radio stack based on SPIRIT1. I encountered a strange behavior in the radio, using the SPIRIT1_Library from ST. I positioned two board within 5cm between each other and using the function "SpiritQiGetRssi" I cannot get more then -20dbm at reception. The module I'm using is SPSGRF-868 (chip antenna version) with 10dbm output power, FSK, 38400 baudrate, 100k channel space, 20k freq dev, 100k bandwidth. I tried with different combination but still the receiving power seems to low to me. I expected at least something near -5dbm.

Anyway, moving the modules 50cm apart further drop the receiving RSSI to -79dbm, and after less than 10m the signal reaches -90dbm.

Any suggestion or possibly a fix?

Regards,

Fabio

[enrico239955]

Hi,

-20 dBm is more than good at 5 cm.

Enrico Migliore

[angeletti2]

I was not specific enough, -20dbm is the best signal ever received between two spsgrf module near each other. The average is -35 dbm at 5 cm. The real problem is that at 10m it is terrible! -90dmn I think is way too low. Am I wrong?

[enrico239955]

Hi,

are you using the module on a custom mother board or an ST's motherboard?

Possible problems:

1 - A power supply problem

The regulator is not able to deliver enough current and the VDD drops when the Spirit 1 draws current.

2 - Module placement

No objects or ground planes can be placed near the module's antenna

3 - Module hardware problem

Your module have hardware problems.

Ciao,

Enrico Migliore

[angeletti2]

I have the module mounted on a custom board. I tried the code also with the nucleo expansion boards with spirit1 and I encountered the same problem. Searching the community I'm not the only one with this kind of issue, is maybe a problem of the module itself or wrong configuration? I'll try the power supply section to check if there are significant voltage drops

[angeletti2]

Here I attached the measurement on the power supply pin wrt ground and also I attach the part of the PCB. The yellow one is ground, similar to other boards that use SPSGRF

Do you think that the problem could be the ground plane?

[enrico239955]

Hi,

I assume here that you measured the 3 V power supply voltage during RF transmissions.

That being said, the 3 V power supply signal seems stable during the RF bursts. Therefore the problem is not the power supply.

I took a look at the second picture you published. It seems that below the antenna there's no ground plane. If that is the case the problem is not the ground plane.

Are the TX and RX channels the same? I mean, are you sure that both the channels have the same frequency?

Ciao,

Enrico

P.S.

Ground planes below antennas are bad. They modify the parameters of the matching netowrk that was carefully designed and tested by ST.

ST designed the matching network in order transform the impedance of the antenna to the complex conjugate of the output impedance of the power amplifier. That means that the imaginary part of the impedance seen by the power amplifier is zero and half of the power is delivered to the antenna.

The other half of the power becomes heat.

[angeletti2]

Hi again! Yes the boards run all the same code (and thus the same channel), since they use the same module also the crystal has the same frequency thus the channels. Now we bought the uFL version of the SPSGRF but nonetheless we expected better performance with the integrated chip antenna. I'll keep anyone interested updated on the tests with the external antennas.

Anyway the measurements on the oscilloscope were made during both rx, sleep and tx state, the radio stack handles them.

Bests,

Fabio

[enrico239955]

Hi,

>> Now we bought the uFL version of the SPSGRF but nonetheless we expected better

>> performance with the integrated chip antenna.

External antennas works much better than chip antennas.

When the antenna is close to the PCB you always face performance degradation in both RX sensitivity and TX power transmited.

The U.FL connector allows you to place the antenna far from the PCB.

The insertion loss of the U.FL connector is usually less than 0.5 dBm.

The coaxial cable, depending on the actual length, may introduce a loss of 1 dBm or less.

Ciao,

Enrico Migliore

[SOMah]

Hi,

We are experiencing the same problem as Fabio. Very similar if not worse performance than what Fabio is quoting. We have tested with uFL connector version too, and we have both not seen any performance increase. Our mechanical constraints prohibit external antennas anyway...

We have followed recommended design considerations (power supply, ground plane etc...) and we also have tested Rx power on the USB SPSGRF dongle and on the NUCLEO devkit (Tx and Rx)...

Can you make available the CE EMC testing report available that gives the EIRP of the module? I see on the 915 version (from the FCCID this is available) that the EIRP is roughly 0dBm and not 5.3dBm as quoted in the datasheet?

Is there any firmware settings (public or withheld...) that your testing has shown to increase output power, or any specific modulation schemes that you have tests to show increase Rx sensitivity and thus increase range?

Thanks

Sam