NRF52840 Coded Phy Sensitivity with soft device 6.0.0-6.alpha

RE: NRF52840 Coded Phy Sensitivity with soft device 6.0.0-6.alpha

Susheel Nuguru — Thu, 05 Apr 2018 11:42:57 GMT

The documentation was updated today, Engineering B chips are supported by SDK15 and S140v6.0 production release. Link to compatibility matrix is here

RE: NRF52840 Coded Phy Sensitivity with soft device 6.0.0-6.alpha

MartinBL — Wed, 21 Mar 2018 09:55:30 GMT

Unfortunately we only recommend sticking to the Compatibility matrix and use development SW on development HW.

RE: NRF52840 Coded Phy Sensitivity with soft device 6.0.0-6.alpha

matsublox — Tue, 20 Mar 2018 13:03:51 GMT

What about the combination of production SW and Rev B chip. Will that be appropriate to do range testing with?

RE: NRF52840 Coded Phy Sensitivity with soft device 6.0.0-6.alpha

MartinBL — Tue, 20 Mar 2018 09:25:36 GMT

Hi,

Alpha -6 is not strictly compatible with Engineering B silicon (refer to the compatibility matrix).

The alpha -6 includes a workaround for erratum [164] RADIO: Low selectivity in long range mode (which isn't actually necessary on Engineering B since it is fixed in silicon).
The alpha -6 lacks a workaround for erratum [191] RADIO: High packet error rate in BLE Long Range mode which will degrade the range performance.

So I wouldn't recommend doing long range tests on Engineering B with alpha -6. Production versions of both the IC and an accompanying Softdevice are just around the corner.

RE: NRF52840 Coded Phy Sensitivity with soft device 6.0.0-6.alpha

matsublox — Thu, 15 Mar 2018 13:42:28 GMT

Picking up this thread again ... We have Rev B devices and would like to do some range measurements for an application where the range is very important. We are using your range measurement example. This means that we want to have an optimal setup for this.

When is the Softdevice with support for Rev B sensitivity fix available?

If we use the Rev B devices and -6 revision of the Softdevice do we still loose about 6 dBm in sensitivity?

RE: NRF52840 Coded Phy Sensitivity with soft device 6.0.0-6.alpha

MartinBL — Thu, 14 Dec 2017 15:06:45 GMT

A second thing to mention is that although we initially stated that the 840 would be capable of +9dBm output power we have had to reduce that value to +8dBm. This is documented in the OPS 0.5.1. Have you measure the actual output power?

RE: NRF52840 Coded Phy Sensitivity with soft device 6.0.0-6.alpha

MartinBL — Thu, 14 Dec 2017 15:06:29 GMT

Ok, sounds like a solid setup.

What chip revision do you have? I'm not sure if rev B has reached the market yet, so probably Rev A?

There is a Product anomaly in IC Rev A that degrades the sensitivity. A workaround for this anomaly was introduced in S140 v6.0.0-6.alpha. However, although the workaround is included in -6-alpha it hasn't fixed the product anomaly entirely. We have corrected the anomaly in Rev B, but a we have yet to release a Softdevice optimized for this new revision.

RE: NRF52840 Coded Phy Sensitivity with soft device 6.0.0-6.alpha

Christian — Wed, 13 Dec 2017 10:56:48 GMT

With the digital attenuator set to 0dB, the whole setup has an attenuation of 8db, I used a transmission power of 9dbm, but I was only able to set the attenuator to around 97 db until the connection of the boards dropped, and the output of the boards showed -96 dbm rssi which seems logical for this setup, but I was not able to reach the -103 dbm.

Thanks in Advance

Christian

RE: NRF52840 Coded Phy Sensitivity with soft device 6.0.0-6.alpha

Christian — Wed, 13 Dec 2017 10:54:56 GMT

Hi Martin,

I used the att_mtu_throughput example and modified it to start RSSI measurement after connection establishment (by using sd_ble_gap_rssi_start(m_conn_handle, 0,0);). The RSSI is then printed by adding the following lines to the event handler:

case BLE_GAP_EVT_RSSI_CHANGED: { NRF_LOG_INFO("RSSI: %i",p_ble_evt->evt.gap_evt.params.rssi_changed.rssi); } break;

For the measurement I used two nRF52840 boards, each one of them in a Rohde & Schwarz shielding box to avoid external interferences. The boards are connected to the output of the boxes via SMA Cable and an SWF to SMA Adapter on the J1 Port of the Boards. The boxes itself are connected via a digital attenuator to guarantee a controllable channel.

RE: NRF52840 Coded Phy Sensitivity with soft device 6.0.0-6.alpha

MartinBL — Fri, 08 Dec 2017 09:41:38 GMT

Hi,

How do you measure this?