BLE jammer to test different setup

[astella]

To the kind attention of Nordic support team, 

I m developping for a nRF52833 device using two different versions of custom boards. According to antenna measurements and diagrams they seem to be very similar, but one of these custom board seems to be prone to random disconnections when in the same noisy environment and same firmware. The problem is that we did collect these observations based on non reproducible random noise. Does it make sense to build a BLE jammer at about minus  30dB 2.4Ghz harmonics and position it at a certain distance from the two custom boards to see if they react always in the the same way or we can sistematically trigger disconnections in one of them only? Is there a way you could propose as an empirical formal test to deal a situation like this so to justify further hardware investigations?

Thank you for your experience and suggestions if possible, 

Best regards

[ketiljo]

You need a very wide band transmitter since BLE is using the whole 2400 to 2483.5 MHz band. Not a realistic way to test it. Maybe use several WLAN links instead? 

21 kHz frequency offset is only 9 ppm and well below the maximum 40 ppm offset that BLE can tolerate, so that is not the problem. Have you tried using Wireshark as a sniffer to find out why it disconnects?

[astella]

Hi ketiljo, thank you very much for your inputs.

We have been successfully using https://www.nordicsemi.com/Software-and-tools/Development-Tools/nRF-Sniffer-for-Bluetooth-LE#infotabs, unfortunately after enabling the authenticated BLE connection with passkey, we lost the chance to be using this great tool, as after LL_ENC_REQ, LL_ENC_RSP, LL_START_ENC_REQ the sniffer is not able to decode what is going on. Please correct me if I am wrong, I d be more than happy to use the sniffer in this case too. We have enabled in fw the chance to start a BLE not authenticated connection, but it only works in test mode, that is different from normal fw working. 
Yes, we could use several WLAN links, we have just to figure out how to do that in a reproducible way. Could be another way to just use another nRF52 and let its transmission sweep the frequency range using radio test sdk routines? Maybe in this case we should have to figure out how to fine tuning the output power. So to get a noise level that is sufficient to affect more only one of the setups under test.  

[astella]

I'm just testing using sniffer_nrf52840dk_nrf52840_7cc811f.hex and wireshark 2.6.6 and I'm not able to decode the ble connection, even providing the passkey to wireshark, before entering it in the device

[ketiljo]

Have you looked at this guide? https://infocenter.nordicsemi.com/topic/ug_sniffer_ble/UG/sniffer_ble/action_paired.html