Frequency hopping experiments on the NRF24

[Martin]

This article, dating from 2015, https://forum.mysensors.org/topic/1153/ ... e-emerging, says that most of us have probably never used a genuine nRF24L01+. I don't think the situation has improved since then.

There are "good fakes" which use a little more power, and are a little less sensitive than genuine chips, and then there are the "bad fakes" that are missing features or don't work at all.

If you buy from AliExpress or Ebay, it's a lottery - you might get good fakes or bad ones, and it can vary from one order to the next from the same supplier, or even within the same order.

But what can we do about it? Most likely not much, as there probably isn't a reputable supplier that makes and sells the item we want. It would be nice buy an RF-Nano with all genuine chips (the nRF chip is likely not the only fake component on these boards), but probably no company is making such an item. The only way to be reasonably sure would be to design and make our own RF-Nano boards and order all the components only from very reputable suppliers. A lot of time and effort, and the resulting boards would likely cost about ten times what we're currently paying - even if we manufacture them in fairly large (hundreds) quantities.

[Mike_K]

Hi Phil

I have 4x rf-nano, two which Tobe kindly sent which were dated 1937BU and two I bought recently, one of each of your dates. And I've got the same problem with the "1913CL", it works OK at higher baud, but not at 250k. I ran your sketch and it looks like its NOT the plus, what are yours reporting? Googling this problem, there appear to be hundreds of reports about old non-plus chips having fake "+" marking put on them as well as completely fake chips.

Most people wouldn't spot the difference as they'd select the fastest supported speed thinking that faster=better and not realise it's a trade-off between speed and range.

As Martin says there's only one reliable way forward - design your own SMD circuit board! I've done a quick price for 20x "rf-pro-mini boards" assembled by JLCPCB and you could get them made for a similar price to the Keywish rf-nano, but knowing (hoping?) that as JLCPCB uses components supplied by LCSC they'd be genuine as they are listed as an approved reseller. I priced them based on an ATMega328P with 8MHz xtal, 3.3V regulator + caps, the RFN24L01+ and 16MHz xtal and a circuit board large enough to accommodate your servo headers and an ICSP connector, but no USB or USB-serial converter. Time to get KiKad or EasyEDA fired up!

Mike

[RichardT]

Here's a bit of my exchange with Keywish:

“Dear Friend, Nice day! our tech group already find the reason, this issue is caused by the supplier， could you please check the amount of the shipping fees toreturn the rf nano, and we send the working rf nano to you.”

So it seems as though they know about the issue. Just checked that the replacements all have the 1913CL marking so ties up with everyone else's findings. Yet to hear from Keywish re my latest report of the duff replacements. The only consolation is that the boards do work at 1 mbps. Anyone done any range tests at that datarate with antenna based tx? Could be ok for my boats but annoying nonetheless.

[Tobe]

How do I use the sketch to check my RF-24/RF-NANOS?

[Phil_G]

Sorry Tobe I should have attached it rather than listed it. Attached below.
Load the sketch into your RF-Nano via the USB & bootloader, when its finished loading open the serial monitor which will need setting to 115200 baud. Opening the serial monitor will cause a reset and after a second or so the result should pop up. During this second its important that you cross your fingers

isplus.ino

(917 Bytes) Downloaded 159 times

I'll configure my OS Cougar for 1mbps and when it stops raining I'll repeat the same range check with one of the non-plus RF-Nanos, I'm curious to see how degraded it will be - but there's always boats and cars to fall back on

For sure I couldnt solder the NRF chip, way too tiny for my bleary-eyes & banana-fingers

Meanwhile, back at the ranch....

I have previously bought several of these items from Keywish and they have been perfect, however the most recent ones do not identify as "NRF24L01+" but "NRF24L01" and do not support 250k bitrate. This makes them unsuitable for my application as I need the 'PLUS' version. These are advertised as PLUS versions and are marked as such, but physically this appears not to be the case. At over £5 each these are expensive chips. The NRF24 library reports the chip as NOT the plus version and does not support 250kbps. The bad ones have the date marking "1913CL" and the good ones "1937BU". Please advise.

Dear Friend,Nice day! I am terrible sorry about this issue and I already rerport it to our tech group, this problem is caused by the supplier, and we already solve this problem, do you accept the solution that you turn these 5 RF Nano to us, and we resend 5 working RF nano to you?
we bear the shipping fees on both sides. Sorry again.

Thank you, will the replacements definitely be '+' versions capable of 250kbps? How do we proceed with the return please?

Cheers
Phil

[Tobe]

Thank You Phil, will have more than one finger crossed.

[Tobe]

I'm confused, this stuff is not really my cup of tea but my inquisitive mind is some time getting me ...
How is the relationship range versus Bps rate? So if I should go to 1/2 Bps rate from 250 Bps I would have even more range? Is the "strength" in each pulse at 1mBps getting so low so actually the output power gets extremely low so there is not enough energy for more range? Would that be a matter of how we modulate the signal?
2.4 gHz as frequency should reach the line of sight with just a few mW.

[Martin]

It's to do with the rate information can be transmitted over a noisy signal path (and any real-world signal will be affected by noise to some degree). Think of two people speaking over a very bad telephone line or radio link - if they're unable to communicate effectively, one thing that helps is to speak slower - or switch to something even slower than speaking, such as Morse code, assuming they both know Morse.

Technically it's covered by the Shannon-Hartley theorem which provides a formula for the speed you can pass data through a link with a given signal-to-noise ratio (S/N). One thing which affects (S/N) is the path length (range) because the signal falls off with distance according to the inverse square law.

[Tobe]

Thank You Martin, you actually make remember something I learned over 40 years ago and also to realize how much I have forgotten over the years! I would imagine that noise is increasing drastically due of "radio Waves Pollution not the least"
As an old ham I sure know that "slow" Morse code wins over noise and distance which are the grounds for QRP where many chase the world with mW....that is also history for me due of my hearing losses unfortunately.
So theoretically we could have different range depending where we are located with the same equipment as I guess noise is the variable here?

[Phil_G]

This I didnt expect!
Keywish have sent me a pack of replacement RF-Nanos even before receiving my returned ones ( I only posted them back to China the other day)!
The replacements pass the test just fine, ie they are plusses and run 250kbps as they should.
Brilliant, well pleased

AND

I just had a Paypal notification - they refunded the postage!!!

AND

they've halved the price back to £2.79 again