TBASERDN040210

February 10, 2004

Q-band receiver in the Equipment Room, cold.  Using lab 83620 as LO1,
set at 9250MHz.  Connected IF of L1 and L2 to OR 1 and 3 respectively,
by coax.  These tests are after mixers in chanels L1 and L2 were
swapped, and microphonic LO multiplier driving L2 and R2 was replaced.

L1 -> OR1 -> CM1 -> SF1 -> ACS J1 (bank 1, rcvr 1)
	     cM2 -> SF2 -> ACS J3 (bank 2, rcvr 1)

L2 -> OR3 -> CM5 -> SF5 -> ACS J2 (bank 1, rcvr 2)
	     CM6 -> SF6 -> ACS J4 (bank 2, rcvr 2)

T=120s, integration=10s.  ACS in two bank, two sampler/bank mode.

Scans 1-6		  Rm Temp load on FE.

Scans 7-21		  LN2 load on FE.

Scans 22-23		  T->300s; still LN2.


Changed to R1, R2 instead of L1, L2.  T -> 120s.


Scans 24-29		  300K load on FE.

Scans 30-43		  LN2 load.

Scans 44-45		  T->300s. still LN2.


Analysis:

Note: All the Ta/Tsys plots are multiplied by 100 so the vertical
scale is in percent.  Unless noted, the traces are offset by 0.5 for
clarity.

Plots in this directory show typical baselines for sequences of 120s
scans on the LN2 load.  Note that unlike the data taken on February 3,
all four channels are similar, within a factor of two.  Either the
replacement of the microphonic LO multiplier, or some unknown change,
appears to have improved the L2 and R2 baselines.

However, I calculated the standard deviations over "typical" 200MHz
portions of the baselines, and these ranged from 0.05% to 0.07%.  The
radiometer equation (for BW = 800MHz/2048) reaches this range for 10
second integrations.  Hence, the spectral sensitivity still seems to
be limited by baseline structure.