You are here

transceiver duty cycle

6 posts / 0 new
Last post
kc1bhd
transceiver duty cycle

It seems that the average power consumption of a device running AREDN firmware consumes more power than running the stock AirOS. Is this due in part to how often the unit id's itself?

KG6JEI
Identification is very

Identification is very infrequent (but within regulations)

It is important to rember these are both radios and computers, both use more power the more they are used. 

For a starter we do send a lot more packets out the RF port than a stock AirOS (relates to linking and other items) this will cause more consumption in the RF stack.

After that we run a LOT more software on the node related to networking, OLSRD for example is constantley recalculating "how do I get from here to there" this adds load to the CPU that does not exist in AirOS.  After that we also have security rules in place that every packet has to check through that takes CPU as well. 

I could keep naming items that are different drom a stock AirOS but really it boils down to "we are doing more and that takes more power" 

AE6XE
AE6XE's picture
The fcc id is once every 5

The fcc id is once every 5 mins and would not be a factor in the power consumption.   This is probably related to the difference between the protocols in use.  AREDN is running 802.11 ad-hoc (or IBSS) mode with OSLR protocol on top.   it should simply boil down that AirOS was transmitting less and computing less than AREDN for the comparison.   But what difference in power consumption are we talking about and how do you have AirOS configured?

KA9Q
I suspect the biggest factor

I suspect the biggest factor is the use of 802.11 ad-hoc mode. Every node transmits a beacon every 100 ms, and it does so at the lowest data rate (meaning the highest transmitted energy per bit). In access point (base station, managed) mode, only the base station transmits beacons.

Ad hoc nodes are also incapable of sleeping their radios to conserve power. In access point mode, the base station announces a schedule for polling clients for inbound traffic and/or delivering any outbound traffic to the client. Clients can then turn off their radios until the next scheduled transaction (the access point must always remain on). This is an optional feature; I don't know whether AirOS actually uses it.

OLSRD "hello" packets are another possible factor as they are also sent at the lowest data rate. But they are only generated every few seconds, so I suspect they're not as important as the 10 Hz beacons and lack of sleep mode.

Some experiments would seem to be in order.

AE6XE
AE6XE's picture
The max power consumption in

The max power consumption in the vendor specs for a Rocket M5 is 8W, a Rocket M2/M3/M9 is 6.5W.    It would be very helpful if someone has time to accurately measure the power consumed of an AREDN node.    That would be great information to help everyone more accurately design systems in the future.   In my case, I run 3 Rockets M2/M3/M5 + 2 ipcams with a ~10 year old 110W solar power 24x7.    I'm not entirely sure how much additional head room I have should I want to further expand :).   If I add a rasbPi to actually measure all this detail on my life system, will I send it over the edge...

KC1BHD,  are there any data points you can share on measured power consumption?

Joe AE6XE

kc1bhd
power

Sure do. i have a number of units and (in the process of making equipment labels), i wanted the actual power used by each type. here goes:

bullet M2 = 3.7w
bullet M5 = 3.1w
loco M2 = 2.5w
nano M2 = 3.7w
rocket M2  3.7w

here is the info in my labels:

Freq: 2.397 GHz          

IP: 10.44.216.175

Channel: -2

Max Power: 2.5w

Bandwidth: 10 MHz

FW: AREDN-3.15.1.0b02

Xmit pwr: 23 dBm

From: http://aredn.org

Ant. gain:  8  dBi

Info: http://m3sh.net

 

Jim kc1bhd

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer