atben/atusb real life signal strength test

Werner Almesberger werner at
Thu Jan 20 19:09:12 EST 2011

I did a performance test of the latest atben/atusb version, testing
the quality of two-way communication at various locations in my

The test setup consisted of a Ben with atben, running on battery, and
atusb plugged into my PC at the front side, only few cm above the
floor. I tested at 2425 MHz, with maximum transmit strength (3 dBm or
2 mW), and maximum-sized packets (127 bytes).

Command on the PC:  atrf-txrx -p 3 -P 1
Command on the Ben: atrf-txrx -p -P

My apartment is a duplex, and I tested at both floors. Here's the
approximate layout of the upper floor:

The blue dot is the PC with atusb. Distances are in cm. The living
room is underneath the bedroom and the office. There's also a terrace
on top of it all. This is a high-rise building with steel reinforced
concrete on the outer walls and hollow bricks in the interior walls.

Communication inside the office was perfect. I also had a good signal
in the bedroom, unless I put myself in the path between door and Ben,
with my body close (< 20 cm) to the Ben. As I approached the stairs,
communication became more unreliable and failed completely at certain

Going even further away, communication became unusable. Inside the
bathroom, the two stations would hardly notice each other, let alone
communicate. It is of course somewhat galling not to have wireless
while on the throne, so I guess I'll have to put another "access
point" somewhere nearby.

Interestingly, I also had very good reception in the entire living
room. Nothing made it to the terrace, which suggests that there's
some metal sheet for isolation.

I also tested Atmel's non-standard 2 Mbps high-speed data rate. It
worked perfectly in the office and was still acceptable in the bedroom
and in the living room. Beyond that, it decayed more rapidly than with
the normal data rate of 250 kbps.

All this confirms that an atben/atusb combo can now achieve the "same
and adjacent room" coverage I was hoping for.

- Werner

More information about the discussion mailing list