'Goldilocks zone' for RSSI values

[A.P]

Hi!

This might be non-coherent, I apologize in advance as my knowledge in RF is very limited.

We're developing a sort of 'TX-negotiation protocol' between our peripherals and central during which the central asks the peripheral to increase or decrease its TX. 
The goal is to have the peripherals use the lowest TX level which still yields 'acceptable' RSSI values on the central.

So far so good. When considering some scenarios, I began wondering if there are cases where increasing the TX level on the peripherals will yield worse results in terms of connection quality or stability (PER/BER etc).

In other words, in scenarios where conserving energy is not an issue, will increasing the TX power of the peripheral always lead to improvement in the connection quality or Is there a 'Goldilocks zone' for RSSI values on the receiver end which we should aim for?

Thanks!

[psychogenic]

Hi,
I can't answer for Nordic or any specific IC or protocol, but I can think of a few reasons why more TX power might not be ideal in terms of RF.

Even if you're within the legal bounds of the tx-power of the band (presumably some ISM band like 2.4GHz), you are an interferer to other users and protocols, so being a good spectrum neighbour is already a reason in itself.  If you're above the noise, past a certain point you aren't more reliable, just louder.

Then some technical reasons I can think of:

• if you were really close and or really loud, you might actually overload the receiver (meaning causing noise, clipping harmonics, mixing products whatever in the rx);
• you may start hitting compression as you're increasing your output power (which is basically when you start boosting your own noise faster than your own signal);
• multipath: secondary bounces that make it to RX may become more present, or even bounce around more than once if they have more juice.

My guess is the tech (if it's certified) probably forces you within the confines of the regulatory framework and, even if it also keeps you from going into compression, being a good spectrum participant and dealing with multipath are still on your shoulders.  With that in mind, some related experimentation would be my next step.

Hope this helps a bit.

Cheers,

PatD

[A.P]

Hi Pat!

Thank you for the answer!
That is part of the reason we wanted to reduce TX power, when we find it is unnecessary.