[Oberon] FPGA - nRF24L01 `RPI Net' server

Jörg joerg.straube at iaeth.ch
Wed May 30 11:23:15 CEST 2018


Hi Thomas

ARC and SCC.SendTries try to solve the same issue but on different levels:
- ARC retransmits the 32 bytes on chip level.
- SendTries repeats sending those 32 bytes on driver (SCC) level.

With your change of ARC=15 and SendTries (most probably still at 50), SubSnd tries to send those 32 bytes 750 times until it gives up.
750 is perhaps a little high!

Try to count how many retransmits on chip level and how many retries on driver level are needed until the packets went thru.

br
Jörg

-----Original Message-----
From: Oberon [mailto:oberon-bounces at lists.inf.ethz.ch] On Behalf Of Tomas Kral
Sent: Wednesday, May 30, 2018 9:30 AM
To: oberon at lists.inf.ethz.ch
Subject: Re: [Oberon] FPGA - nRF24L01 `RPI Net' server

On Wed, 30 May 2018 07:24:39 +0200
Jörg <joerg.straube at iaeth.ch> wrote:

> In ideal conditions without any retransmits Net.SendFiles would need
> 1.5 seconds to transmit your NRF.Mod (6158 bytes)

Good, this is an ideal target, I may one day get close to.

> As you need about four times as long, there must be many retransmits 
> going on under the hood. Did you move the channel out of the way of 
> possible Wifi in your environment? Try setting the channel to 
> something bigger than 86 —> WriteReg1(RFCH, 86);

True, I have quite a few retransmits, with or without decoupling caps.
Also tried various other power supplies. Will try to play with channels once again.

> rPi and your Oberon board is roughly 1 meter, try set the power level 
> to -18 dBm. —> WriteReg1(RFSETUP, 0);

In fact 10cm, that was my first discovery, with full power, 0dBm, 30% ping was the usual standard, now I get 98-99%. Also `SCC.SubSnd()'
delay has some influence on it.

I have added these lines to `SCC.Start()', 

WriteReg1(4, 0*010H + 15); (*SETUP_RETR delay + retry count*) WriteReg1(6, 03H); (*RF_SETUP <= 1Mb rate, 0dBm wide range, -12dBm short reach*)

Will try -18dBm again, last time noticeable improvement was the stepdown from 0dBm to -12dBm. 

--
Tomas Kral <thomas.kral at email.cz>
--
Oberon at lists.inf.ethz.ch mailing list for ETH Oberon and related systems https://lists.inf.ethz.ch/mailman/listinfo/oberon



More information about the Oberon mailing list