A series of tests were done using the spectrometer in 12.5 MHz bandwidth mode, with 4 frequency bands, hence 8 samplers being used. The path connections are summarized in the following table:

Receiver/Beam	KU/B2 or K/B4
Opt Rcvr	OR6				OR8
Freq Band	1	2	3	4	1	2	3	4
pol	LCP	LCP	LCP	LCP	RCP	RCP	RCP	RCP
CM#	9	10	11	12	13	14	15	16
ACS port	J25	J26	J33	J34	J29	J30	J37	J38

Thus, the paths using CM9,10,11,and 12 are using all old hardware; CM 13, 14, and 15 use the new optical receiver module but the same old converter modules; and CM16 uses all new modules.

For each setup, the IF Rack and Spectrometer were balanced as well as possible using the astrid Balance command or the balance buttons in Cleo. In each case, the attenuator settings in the IF Rack and Converter Rack were recorded.

In looking through the following results, keep in mind that when the converter module attenuator value is zero or 31.8 that means that balance was NOT achieved.

Several setups were tested:

test		IF Rack Attens		Conv Module Attens (db)
	IF Target	S6(db)	S8 (db)	CM9	CM10	CM11	CM12	CM13	CM14	CM15	CM16
A.	Freqs: 12000, 13000, 14000, 15000; IF filter: All Pass
	1	21	21	0	0	0	0	1.2	1.1	0	4
B.	Freqs: 15000, 13000, 14000, 12000; IF filter: All Pass (bands swapped)
	1	22	21	0	0	0	0	0	1.1	0	11.6
C.	Freqs: 14000, 14300, 14700, 15000; IF filter: 3/1280
	1	18	16	0	0	0	0	2	7.2	5.3	10.8
	9	8	6	7.6	9.5	8.5	5.1	10	16.6	14.8	20.1
D.	Freqs: 14000, 14100, 14200, 14300; IF filter: 3/320
	1	11	9	2.5	5.5	4.8	2.3	6.3	12.3	10.1	16.7
E.	Freqs: 14000, 14020, 14040, 14060; IF filter: 3/80
	1	6	8	8	11.1	9.7	8.1	13.8	18.3	16.8	23.3
F.	Freqs: 14000, 14002, 14004, 14008; IF filter: 3/20
	1	5	2	14.6	15.8	16	14.3	19.1	23.8	22.5	29.8
	(change to K-band/B4 :)
G.	Freqs: 24000, 24020, 24040, 24060; IF filter: 6/80
	1	5	6	6.5	8.8	7.8	6.1	9.6	13.2	12	19.2
H.	Freqs: 24000, 24100, 24200, 24300; IF filter: 6/320
	1	12	13	1.1	2.8	0.7	0	4	8.1	7.1	12.6
I.	Freqs: 24000, 24300, 24700, 25200; IF filter: 6/1280
	1	17	18	0	0	0	0	0	2	0	3.2
	3	12	13	0	2.8	0	0	4.6	8	4.6	8.2
J.	Freqs: 22500, 23500, 25000, 26000; IF filter: All Pass
	1	25	24	0	0	31.8	31.8	0	0	0	0
	3	19	19	0	1.1	0	31.8	0	4.6	0	4.2
	9	15	14	0.8	4.5	0	0	3.1	8.8	4.1	8.8
K.	Freqs: 26000, 25000, 23500, 22500; IF filter: All Pass
	1	25	24	31.8	31.8	0	0	31.8	0	0	3.7
	9	15	14	0	0	2.5	0.6	0	4.1	7.5	12.5
	(change to KU/B2 :)
L.	Freqs: 15000, 15002, 15004, 15008; IF filter: 3/20
	1	5	4	14.3	17	16.2	14.6	19.2	23.5	22.5	29.8
M.	Freqs: 12100, 12102, 12104, 12108; IF filter: 3/20
	1	3	3	14.3	17	16.2	14.6	19.8	23.8	27.5	30.1
	3	0	0	18.1	20	19.7	18	23.1	27.8	26.1	31.8
	0.5	6	7	11.6	13.3	13	11.5	15.3	20	19.5	26.3

Comments

There were only two cases in which balance could not be achieved using the new hardware (CM16). These were case J with IF target=1, and case M with IF target=3.

Cases J and K used a very wide spacing of frequencies, with the frequencies running 22.5, 23.5, 25, and 26 GHz for case J, and the frequencies running in the opposite direction for case K. We can achieve balance for CM16 in case K, but not in case J. The difference is apparently due to slope across the band. We are not very far from the balance condition, because we can achieve balance in case J by increasing the IF target level to 3 volts.

The narrow band case, with the IF filter = 20 MHz, is tested in F, L, and M. In these cases, the attenuation in CM16 required to achieve balance is near the upper limit of 31.8 db. We can balance in case M by decreasing the power level in the IF Rack from 1 to 0.5 volts.