Friday, September 19, 2008

Improving Signal Strength - The Easy Way

Well I found a simpler way to increase the signal strength. Just get a long USB cable and keep the device outside the window. In my case, the window is facing the nearest tower (around 2kms away). I did happen to notice one thing though. Although AT+CSQ indicated 31 (the maximum value), the signal strength indicator in the MS-Windows dialer indicates 3 bars, and sometimes 4.

Friday, September 5, 2008

Improving signal strength of EvDO/CDMA

I found a simple technique to increase the signal strength/reception quality of my EvDO connection from BSNL. The first requirement for this is to be able to get a numerical measure of signal strength which is more reliable than the signal bars indicated on the connection manager. Since I use Linux, I really have no way to observe the signal bars. So instead I wrote a script in Python that gives me the signal strength (power) in dB. I'll put it up when I feel it looks good enough to not be embarrased. Anyway, one could always use the "at+csq" command to get the signal strength. For more information look at one of my posts below.
CDMA in India uses the 824-849MHz band for uplink (mobile to tower/base station), and the 869-894 Mhz band for downlink (tower/base station to mobile). So the ideal center frequency is is 850MHz. Frequency (F) and wavelength (L) of electromagnetic signals are related as follows:
L=c/F
where c is the speed of electromagnetic signals or light, which is 29,979,245,800 centimeters per second. So the wavelength in centimeters is approximately 35.3cm. At L/4 (quarter wavelength), this works out to approximately 8.8cm. So one should just keep a plate or sheet of metal roughly at this distance from the antenna of the EvDO modem. In my case I first adjusted the device to provide maximum gain and then place the sheet.
The advantage of placing the sheet at quarter wavelength is that it need not be too big in size. Note that the antenna should be perpendicular to the modem and the sheet parallel to the antenna. The sheet could also be grounded if required. Using this sheet gave me an increase of 12dB in the signal strength from -67dB to -55dB. Use of the metallic sheet will most likely cause the reception to be more directional than earlier. So one could further adjust the orientation of the entire setup to increase the signal strength.

Those who are really desperate to boost reception could try out a quarter dipole antenna design at http://www.mattstuff.nq.nu/antennas/quarter-wave.html. Note however that the design is for 435MHz. Length of the dipole will have to be modified for 850MHz.

ps: I have also come across people using CDs. Did attempt using them, but did not get a good enough gain. There is also no simple way to ground CDs. I am assuming that the value returned by "at+csq" query is reasonably accurate.

Monday, September 1, 2008

KPPP confused about kbps and KBps

It has been a long time since I used kppp (used to during those dial-up (pre-2004) days). Looks like kppp statistics is confused about kbps (kilo bits per sec.) and KBps (kilo bytes per sec.). Shows KBps as kbps in the logs. Confirmed using ifstat that it is actually KBps and not kbps. Seemed that I had observed this similar problem during the dial-up days. Could someone maintaining kppp please explain why this has not been fixed for so many years.