Third fiasco with range tests, reason - cascade of LNA .
Based on calculations from www.ti.com/lit/an/swra046a/swra046a.pdf, particularly on formula from page 23 path loss = 20*log(4Pi/Lambda)+10*N*log(distance) with 1mwt transmitter and distance == 625m, value N == 2.2 matches loss of signal (2 is free space path loss, N==2.2 is close to it). Vancouver's rain does not influence on that value (on heavy rain it should be 2dB per 1km). Then attempt to work with 3.5 km and 3 cascades 10dB LNA brought fiasco. Assuming N is same 2.2 (over Burrard Inlet) gives 3 LNA combine value less then 12dB. From another hand 3 LNA somehow works, range from 625m definitely improved to 1050m (can not be measured more in the same environment, same N value). That mean losses btw LNA or LNA and switches needs to be re calculated and impedance matching should be adjusted. Do make sure that problem was in LNA instead of antennas both ground station's and cubesat's helical antennas was measured for a HPBW (Half Power BW) on a school's football field. 16 turns antenna showed 25 degree (half power on -85dBm sensitivity with distance for 1mWt is 149m) theoretical value is 26.5, and helical 3 turns showed 61 degree matched theoretical value.
Positive part of the tests – was found spot over 33th street, half block from Royal Canadian Mounted Police down hill to Oak st., top of the hill Granville st.(1100m), then again down-and-top of the hill on Vine st. or Larch st. (1000m). On 33th no cables over that distance and top 50m on Granville allows to calculate (http://www.tapr.org/ve3jf.dcc97.html) diffraction loses over single obstacle modeled as a knife edge (the loss of non-LOS path) it gives loss 32 dB. Sliding position from Larch st. allows to adjust loss in range of 3 dB. That makes that place practical to test 1WT (30bBm) transmitter (instead of 200km range LOS Baker Mnt - Grouse Mnt).
test 1 Schematics:
Path loss calculation with different formulas:
Some video will be posted.