Tx power not at 0 dBm

6 posts / 0 new
Last post
wisilica
Offline
Last seen: 1 day 9 hours ago
Joined: 2015-03-17 08:16
Tx power not at 0 dBm

Hi,

We are working on DA 14585 chipset using SDK 6.0.2. We are implementing a code in which the data is getting advertised every one second. While observing the RF signal using spectrum analyzer, it is observed that, instead of having peak rssi value at 0dBm, the values were observed to fluctuate from -12-120 dBm.
Please suggest the possible reasons for the same.

Thanks
Wisilica

Device: 
MT_dialog
Offline
Last seen: 1 day 7 hours ago
Staff
Joined: 2015-06-08 11:34
Hi wisilica,

Hi wisilica,

I am not sure about the setup that you are using in order to measure the 0db RSSI, are you using antennas between the analyzer and the 580 or you are using a direct connection between the two devices in order to perform the measurements ? Also what exactly are you trying to measure with the spectrum analyser and while advertising. When advertising the device will jump between the 3 advertising channels emmiting for a short period, due to that i dont think that the spectrum analyser will give back a value near to 0db rssi measurement. The way to test this would be with the prod_test tool and instruct the device to continusly transmit.

Thanks MT_dialog

wisilica
Offline
Last seen: 1 day 9 hours ago
Joined: 2015-03-17 08:16
Hi,

Hi,

We are not using antennas and are directly connecting 585 device to spectrum analyzer. We had observed a huge rssi variation while the packets sent by 585 device were captured by another scanning device. So we were analyzing whether the device actually sends packets at 0dbm tx power or less. We are configuring the central frequency of the spectrum analyzer to each of the 3 channel frequencies and then checking the results.

Thanks wisilica

MT_dialog
Offline
Last seen: 1 day 7 hours ago
Staff
Joined: 2015-06-08 11:34
Hi wisilica,

Hi wisilica,

Regarding performing the measurements with another central device please try to do that in a shielded enviroment in order to avoid noise that will corrupt your measurements. 

Regarding your setup, power readings of a BLE RF signal on the spectrum analyser are hard, because the signal is modulated and the signal is hopping from channel to channel, perhaps you have used a narrow RBW setting on the tool and that reduces the chances of capturing a hopping signal. The RF signal will be on a channel for a short time only before it hops to the next channel, and easily missed by an analyser having a narrow filter sweeping over the full BLE frequency band. Max hold setting and measuring for many sweeps makes chances bigger that a hopping signal is captured in full power.

In order to capture/measure all three advertising channels in one go, its best to use a big RBW (e.g. 3MHz) and put the analyser in max hold or peak mode and let it run for many sweeps (multiple minutes), but even then it might be hard to get 0 dbm readings. The easiest way to take the measurements is to use a CW signal, un-modulated or constant TX signal (not-packeted). Those signals can be generated via the prod_test fw. So please use the prod_test tool in order to measure the TX power of the device.

Thanks MT_dialog

wisilica
Offline
Last seen: 1 day 9 hours ago
Joined: 2015-03-17 08:16
Hi,

Hi,

Thank you very much for your prompt reply. By executing the prod_test code in the board, and while sending unmodulated carrier wave using the prod_test.exe tool, we were able to send the carrier in 3 frequencies - 2402, 2426, 2480 Mhz. These waves were analysed using spectrum analyser by direct coupling method. The entire setup is shielded as well. We are using max hold configuration too. The images of the spectrum analyser output are attached herewith. Still, we are observing discrepancies in the output TX power. Instead of having the max TX power at 0dBm, the spectrum analyser output shows variation from - 1.4 to - 2.5 dBm. We have also observed that, the TX power in all the 3 channels are different.
Kindly look into this, and suggest the reasons for the same.

Thanks
Wisilica

MT_dialog
Offline
Last seen: 1 day 7 hours ago
Staff
Joined: 2015-06-08 11:34
Hi wisilica,

Hi wisilica,

In the measurements that you have you have to consider the RF cables losses as well, also the used bandwidth of 100KHz is to low, usually 1MHz is used to measure the output power. Additionally the typical TX output power of the 585 is -1 dbm the maximum is 0 dbm (you can check the datasheet for this).

Regarding the variation that you get in the different frequencies, a nominal power variation over channels of 0.8 dB is not unusual, your measurements fit the expectations.

Thanks MT_dialog