Mixing RFM69W with RFM69HW

[jgilbert]

Moteino users --

I have a fleet of about 8 nodes running 433mhrz RFM69Ws in a 3 story townhouse and am noticing that the nodes two floors up from the gateway are very inconsistent, and sometimes go hours without getting a packet through. I was wondering if its possible to mix the 69W and the 69HW in the same network. Surprisingly I cannot find a single confirmation of anyone actually doing this although folks have said it is theoretically possible.

My plan would be to replace the gateway and the upstairs nodes with the 69HW and leave everything else the same (e.g. the 69W nodes would remain on the network.) Any advice?

Another option is to create a separate 915mHz network. I can't really find anything definitive if 915 is better range/lower interference indoors. I live in a fairly dense residential area (Cambridge MA) so plenty of interference here.  Would I be better off with the 915 mHz?

My average receive RSSIs are between -84 - -90, varying day by day (have been logging this for several million packets since last week.)

Thanks,

Jeremy

[Felix]

Of course you can. They are the exact same semtech chip, except one is implemented with a PA_boost (HW) on the PCB, one is not (W). You will only need to load the right settings in the sketch so that the registers are setup correctly.
You are probably not setting it right or missing some small detail. But all sketches have all the necessary settings, some may be commented out. You need to be careful with the IS_HW setting. That only needs to be enabled for HW radios.
A -84-90 dBm RSSI is pretty good for so much structure in between.
The 868-915mhz is what I recommend to most people that start out. It is very good for indoors and outdoors. Interference can be a problem but I would not expect much at 433mhz.

[juan3211]

Hi, sorry for opening again this post, but I think that it is the good one for me to ask, as the subject is exactly my question.

I have bought several RFM69W and RFM69HW.

I know, that if they have the same frecuency, they can be mixed.

But what about the better option to work them as RX/gateway or TX/node ?

Two examples:

• you have only two nodes. Is is better Gateway RFM69HW and Node RFM69W? Or is it the same as the node also will send the ack payload ? Or will be the same to have "1 x RFM69HW and 1 x RFM69W" or "2 x RFM69W" ? (about energy it will be better the second option, but I don't know if the range is equal or not)
• If you have two RFM69HW and two RFM69W: is this better: Gateway RFM69HW and the most far node also RFM69HW, others nodes as RFM69W

Hope your answer and thanks a lot.

[TomWS]

If your communications protocol is two way, ie, sendWithRetry() which requires at least an ACK from the far end, then your range will be limited by the weakest radio in the link.  So, in general:

Gateways would be HW since they have to reach the farthest node in your network.
Nodes can be W if they are close to the gateway, but should be HW if they are the farthest nodes and required by range.

If your duty cycle is low at each node, then, in general, your battery life is not significantly affected by using an HW and this is a safe design choice in most cases.  Using ATC to dial back power to a reliable value makes this practical.  The exception is if you are powered by a very weak battery, such as a coin cell.  I would never use an HW radio with a coin cell unless a very large cap was in the circuit to handle the current surge.

Tom

[juan3211]

If your communications protocol is two way, ie, sendWithRetry() which requires at least an ACK from the far end, then your range will be limited by the weakest radio in the link.  So, in general:

Gateways would be HW since they have to reach the farthest node in your network.
Nodes can be W if they are close to the gateway, but should be HW if they are the farthest nodes and required by range.

If your duty cycle is low at each node, then, in general, your battery life is not significantly affected by using an HW and this is a safe design choice in most cases.  Using ATC to dial back power to a reliable value makes this practical.  The exception is if you are powered by a very weak battery, such as a coin cell.  I would never use an HW radio with a coin cell unless a very large cap was in the circuit to handle the current surge.

Tom

Thanks a lot @TomWS.

Sorry about delay, but I have to check my settings as I don't recieve any suscription email.

My feelings were as you says, but I need to confirm them. Thanks a lot.

I understand that HW can surge up to 130-150mA.

Regars,

[TomWS]

I understand that HW can surge up to 130-150mA.

This 'surge' only occurs when you are transmitting at full power and only lasts as long as the transmit time (a few 10s of Milliseconds), which is why duty cycle matters.  If you're only transmitting once every 10 minutes, then the duty cycle is ~20ms tx/600000ms cycle time for an average TX power of ~5uA (plus quiescent/sleeping current).

Tom

[juan3211]

This 'surge' only occurs when you are transmitting at full power and only lasts as long as the transmit time (a few 10s of Milliseconds), which is why duty cycle matters.  If you're only transmitting once every 10 minutes, then the duty cycle is ~20ms tx/600000ms cycle time for an average TX power of ~5uA (plus quiescent/sleeping current).

Tom

Yes, I know. I was talking about what you say about coin cells.

Thanks a lot for your help.

I am a bit upset about NRF24 and its inestability and/or poor range, so I want to check RFM69.  Thanks a lot again