Satellite Internet forum
Service Providers >> Satellite Connection in Africa >> QPSK BER value on Linkstar modem ?

Message started by quitro on Jun 2nd, 2007 at 7:33pm

Title: QPSK BER value on Linkstar modem ?
Post by quitro on Jun 2nd, 2007 at 7:33pm
hi, can anyone tell me what is the best qpskber value on linkstar to find, on W3A using a andrew 1.2mt,3w buc and rg6u 20 mt cable?
How good the QPSKBER? I have got always on linkstar 0.0003200155. is it good?

Post by Forum Admin on Jun 3rd, 2007 at 12:24am
If you make repeated measurements I would expect each measurement to be slightly different.  A constant value of 0.0003200155 suggests that the measurement is stuck.  Repeat the command. see

A figure of 0.0003 is acceptable, provided it is the best value you can obtain.  Try increasing the elevation by 1/2 a turn on the nut and repeating the measurement several times. Continue in elevation and azimuth till you are certain you are in the exact centre.  This is the only way to make sure that the transmit beam, which is narrower than the receive beam, is properly centered. Having centred the beam pointing you may be able to optimise the polarisation yourself.  Turn the feed well away from optimum and mark the angle.  Measure the BER, e.g. 0.01, and get a good average.  Write it down.  Turn the feed on the opposite side of optimum and adjust for exactly the same degraded performance.  Once you have that, mark the second angle.   Make a paper strip and wrap it around where the two mark are and divide the distance in half.  Adjust the feed rotation to the centre with an accuracy of less than 1mm.  With the correct polarisation the cross polar interference will be minimised.

The best possible BER depends on the satellite downlink beam contour pattern and your dish size.  0.0003 to 0.00003 is normal. What is important is that you have used it to optimise the pointing and that it is significantly less than 0.001  Aim for 0.0001 or better if you can !

Don't worry that your data will have errors, the forward error correction system will easily correct 1 error in every 1000 bits.  What you are seeing is an estimate of the raw BER, prior to FEC being applied.   There is an uncorrected error count if you are interested.  This is useless for pointing since it is always zero except for severely degraded conditions or the odd click.

Best regards, Eric.

Post by quitro on Jun 4th, 2007 at 9:04pm
Hi Eric.
Thank you very much for this helpfull tipp.
best regards

Powered by YaBB 2.5.2!
YaBB Forum Software © 2000-. All Rights Reserved.