Range test results:
- On the top of PSA with the default hostap driver
- SMC card works at 140 feet above the ground
- On the roof of RTH, with default kernel (without roofnet patch)
- with hostap driver, SMC card shows similar result.
- with Orinoco driver, Orinoco card works at 160 feet even on the ground even with smaller antenna!
Also, Orinoco card shows better RTT ( < 2ms ) between two stargates than that( >2ms ) of SMC card
Experiment in front of RTH building (on McClintock Ave) with both of SMC and Orinoco cards and smaller antenna.
- Over the ground (2~3 ft above the ground), both of SMC and Orinoco cards work more than 600 ft.
- Orinoco card works a little longer range (about 50 ft more) than SMC.
- On the ground, we have not measure the maximum distance, but
- Between 300ft distance, Orinoco card works, but SMC does not.
- During the test,
- SMC card showed almost 0% of loss rate.
- Orinoco card showed around 2% of loss rate.
- As Om's suggestion, we checked the signal strength of both card using Wi-Spy that is a radio spectrum analyzing tool.
- The signal strength that both cards show is not quite different from each other.
- One of the big differences of today's test environment from previous one is the height of the testing environment.
- Currently we cannot get benefit of high power cards
- One thing we can suspect is the hostap driver.
Daniel's experiment
At distances of 100ft, 200ft, and 300ft in an open space using all possible combinations of cards and antennas. UDP packet stream to 1 minute at 0.5 Mbps (3841 1024-byte packets sent in 60 seconds). Surprisingly, in ALL cases, not a single packet was dropped!
On RTH roof:
We used continuous 'ping' (as fast as possible), and say 'working' only if the ping gets through near 100%.
1. Works for sure at 100ft (approx). 2. Original SMC antenna works 100ft when above ground, (when a person is holding the box above ground) but does not work on ground. 3. Big antenna works 100ft regardless of whether it is on the ground or not. 4. If above ground, no performance difference experienced between different antennas at 100ft. 5. Barely works at 120ft (approx). 6. Big antenna seems to be better, and works most of the time at 120ft, but not confident enough to say "sure". 7. Actually, the only 'definite' benefit that we get from the big antennas seemed to be the position of the antenna. (on ground, near wall, etc) Couldn't see clear benefit in terms of distance. 8. Changing the tx power didn't have much effect on distance. Couldn't feel much difference between 'default power' and 'max power'. Alway, the cutoff point was around 110~120ft.
Om's test:
After installing the windows driver for SMC card on KiYoung's and my laptop, we went to the roof, put the two laptops on the two ends of the roof, intalled the antenna, and blasted 535 pkts (pkt size=100 bytes) from my laptop to KiYoung's laptop every second for 64 seconds. The packet loss rate was 1.5%. We did similar experiment in the hallway on the 4th floor and got similar results. The results are confirmed in the reverse direction as well. Our conclusion is, with laptops, and in windows, we can get pretty reliable connection from one end of the building to the other end. It remains to be seen if it is Stargate (hardware) or Stargate driver (or other software issues) causing dramatic loss in yesterday's test.
On RTH-roof.
The setup was, - RTH roof - Orinoco cards - with Antennas (two types) - box *on the ground* - and use "ping" with "-f" option to test.At steady state (when we are not moving, touching, etc) 130ft ( < 1% loss) 140ft ( ~ 1% loss) 150ft ( ~ 3% loss) The losses increase drastically but temporarily when we move the box, touch the antenna, or move around. But it stabilizes soon after the environment stabilizes.