Some questions on nRF52811 documentation

RE: Some questions on nRF52811 documentation

Armin_B — Wed, 04 Sep 2019 16:34:15 GMT

did you ever try to test CTE feature (continuous tone extension) at transmit?

RE: Some questions on nRF52811 documentation

Armin_B — Wed, 04 Sep 2019 16:25:14 GMT

thank you for fast response.

good luck

RE: Some questions on nRF52811 documentation

Dmitry — Wed, 04 Sep 2019 07:33:11 GMT

Hi,

we're developing our own board, and our firmware is ran on bare metal 52811, without SD and libraries (own protocol, not BLE-compatible)

RE: Some questions on nRF52811 documentation

Armin_B — Tue, 03 Sep 2019 21:29:47 GMT

Hi Dmitry

how did get a nrf52811 board to test? and which library you used to configure the DFE and I/Q sampling?

Thanks

RE: Some questions on nRF52811 documentation

Simonr — Tue, 09 Apr 2019 13:42:30 GMT

The range of I/Q samples depends on the AGC during the packet. It could be that the signal is so strong in this case that the gain is lowered a lot. Difficult to say from the description. One could separate the two devices and see if the I/Q sample range increases.
The settings are very different to what we typically use for the BLE specification and I got confused about the switching slot/state. For the BLE spec typically SWITCHSPACING = SAMPLESPACING and then there is no sampling in switching slots. In general there is sampling in the Reference period and in the "Switching period" (i.e. switching and sample slots). We should probably have called it "Switching state" to align with the DFESTATUS register description. The 1us skip is expected, as it skips like that to ensure the switching period starts at the same place even with different sample settings.
The documentation is a work in progress, and we will take your notes into account! For now, keep this value at 0 or 1, as these are the recommended values. Thanks for your input.

Best regards,

Simon

RE: Some questions on nRF52811 documentation

Dmitry — Mon, 08 Apr 2019 17:51:48 GMT

Finally got a nRF52811, IQ sampling is awesome ) Still have no transmitter with antenna switch, so now playing with TX pattern.

My DFE configuration:

NRF_RADIO->DFEMODE = 2;
NRF_RADIO->DFECTRL1 = (16 << 0) | (0 << 7) | (2 << 8) | (6 << 12) | (0 << 15) | (6 << 16) | (0 << 24);
NRF_RADIO->DFECTRL2 = (0 << 0) | (0 << 16);

Here what I got with pattern F0 F0 FF F0 FF FF FF FF 00 FF ...

And again questions:

1) Range of I/Q values is about -150..+150, isn't it too low for 12-bit ADC? I use 10056 devkit as transmitter at 2 cm distance, TX power is 0 dbm, ADC backoff gain is 0.

2) I see there is 1-usec skip after 64 samples of reference period (8 usec). After that, a wave is continuous, so receiver does sampling even in switching slots? ok, it's very good for our case, just would like to be sure that it won't change in the next chip revision )

3) Still can't get TSAMPLEOFFSET. I figured out that changing it affects start of reference period (not start of sampling after it), and 0 corresponds something about 16 usec after ADDRESS event. Maybe it worth to update a documentation to make this item more clear?

Thanks!

RE: Some questions on nRF52811 documentation

Dmitry — Fri, 15 Mar 2019 14:31:33 GMT

Thanks Simon!

RE: Some questions on nRF52811 documentation

Simonr — Fri, 15 Mar 2019 13:17:30 GMT

Hi Dmitry

Sorry for the late reply, but it took some digging and asking around to find all the answers.

1. There is no IQ sampling during switching slots.

2. TSAMPLEOFFSET is one of the two start offsets for direction finding, and is the offset value in number of 16MHz clock cycles after the reference period (see figure 71 in product specification).

3. Energy detection is featured in the IEEE 802.15.4 operation, yes.

4. There is a correlator which can only operate on ADDR0. It cannot work on 8 addresses in parallel. Therefore only ADDR0 will be demodulated with a 3dB gain while the others will not. Putting addresses in ADDR[1..7] does not degrade ADDR0 sensitivity. So sensitivity shouldn't drop when only filling it with 2 addresses.

Best regards,

Simon