Yeah, that is indeed confusing. It has been a while since I read that document (I am more familiar with the RFM73 clone), but note that sentence states configuration registers, while 8.3.1 talks about command and status registers.
My best guess is that the limitation applies to a small subset of the registers, probably only the CONFIG register. Or it is a sentence that was left in the datasheet totally by accident. Note that it contradicts the 3d paragraph of the introduction.
I have used this chip briefly (mainly to test whether it is compatible with the RFM73 - it is), but I never had to put the chip in standby to do reading or writing of the packets.
Yes, you can have multiple units receive the same broadcast (or multicast) message, if they are each configured to use the same receive address that the transmitters is sending to. (And in range of course).
Remember that each can have up to 6 configured receive addresses, so one of the ways to use the device is to configure one receive address which is unique to a given device, and one receive address which is shared with other devices (eg: all receivers, so some subset of them). The transmitter can send to a single receiver by using its unique address, or to all of the receivers in the group by using the shared address. ("Multicast" may be a somewhat better description in that it goes to multiple receivers in this scenario, but it need not go to all receivers - you can control the subset of receivers using a given shared address).
Of course, if you are sending a multicast message, it must not use ACK - because multiple receivers would try to send their ACK packets at about the same time, which would collide. (Also: since the hardware defined packet structure does not have a "source" address, the way that Enhanced ShockBurst ACK works is that the receiver sends the ACK packet to its own address, and the transmitter must be configured to receive at the address it's sending to. Thus in multicast, a given receiver's address - to which it will send an ACK - is also the shared receive address for the other receivers). So ACK must be off for multicast (as your code says).
There is not quite enough information to answer why your particular setup causes only one receiver to accept the packet. Suppose you have one transmitter T, and two receivers R1 and R2 (both using the same address). R1 is the one which "always wins". Is this the situation:
(R2 not operating): R1 reliably receives from T
(R1 not operating): R2 reliably receives from T
(both R1, R2 operating): only R1 receives from T, R2 does not
That is, adding a R1 to the second situation causes R2 to stop receiving from T. Can you verify this by starting the T to R2 test in a loop and then after a while also powering up R1, at which point R2 stops receiving; then power down R1 and see if R2 starts receiving again?
If so I'd suspect that some kind of ACK problem is involved, or there is some problem with R1 and R2 being too close with some kind of RF or antenna intererence, or they are on the same power supply and there is some interference between them through the power line.
The nRF24 family is a bit sensitive about power; sometimes you need to put a filter cap across the power input to the module; eg: a 10 uF electrolitic (observing + and -). I've had what I thought were software or config problems clear right up when I improved the nRF24L01+'s power supply.
Best Answer
No. The chip can take a maximum of 3.6V as its supply. And since its supply current is so low (<15mA) you can use one of those SOT-23 voltage regulators, so no reason not to.
You must keep its supply voltage within its supply voltage range. You can apply a higher voltage at its digital inputs, i.e. connect a 5V CMOS output to one of the chip's inputs directly.