DCR Through The Analog Filter Rack

Figure 6: Gain ratios versus relative input power for the DCR signals coming from the Analog Filter Rack. $3\sigma$ errors are shown. The black line is the expected result for a linear system response.

Next we decided to repeat the above experiment of changing the IF Rack attenuation with the DCR signals coming from the Analog Filter Rack (AFR). If the same non-linearities are seen as when the DCR signals come from the IF Rack then we can narrow down where the non-linearities enter the system to be between the IF Rack attenuators and the splitter between the IF Rack DCR V/F path and the Optical Fibers.

Table 3: Attenuation settings for optical drivers during TGBT01A_004_08 tests with the DCR signals coming from the Analog Filter Rack.

	IF Rack Attenuation
Scan	OD 2	OD 4	OD 6	OD 8
28	12	12	10	17
29	15	15	13	20
30	18	18	16	23
31	9	9	7	14
32	6	6	4	11

The signals were sent down Optical Drivers 2, 4, 6 and 8. Each signal was split in the Converter Rack into two exact copies. These eight signals were then feed to the DCR and the spectrometer at the same time. Splitting the signals in this way will allow us to determine if there are any non-linearities after the Converter Rack splitter.

Table 4: IF Paths for the Spectrometer and the DCR_ACF tests.

Receiver	Optical	DCR_IF	Converter	DCR_AFR	Spectrometer
Polarization	Driver	Sampler	Module/Filter	Sampler	Sampler
XL:1	2	A 2	1	B 9	J9
YR:1	4	A 4	5	B 11	J13
XL:1	2	-	3	B 10	J17
YR:1	4	-	7	B 12	J21
XL:2	6	A 6	9	B 13	J25
YR:2	8	A 8	13	B 15	J29
XL:2	6	-	11	B 14	J33
YR:2	8	-	15	B 16	J37

The DCR was feed eight signals into ports 9 - 16 in its Rack B input. The spectrometer was feed the same 8 signals and was setup in its A1B1C1D1 mode with 50 MHz bandwidth, 2 samplers per bank, and 9-level sampling (1N2-XX-50-9 modes). The relative data paths are shown in Table 4.

We tracked 3C 274 (Virgo A) and balanced the IF Rack to 3 Volts in each RF power samplers for Optical Driver paths 2, 4, 6 and 8. After the IF Rack was balanced to 3 Volts the Spectrometer was balanced. Again the hi cal was used and data was taken at several different attenuator settings which are listed in Table 3.

The results are plotting in Figure 6. We have again plotted the gain ratio from Equation 5 versus the relative change in input power (i.e. $dB(scan n) - dB(scan 30)$). As can be seen in Figure 6 the gain ratios are not one - indicating that the DCR response is again non-linear. In this case we have not driven the DCR into gain compression. This is likely due to the extra attenuation that was provided in going through the Optical Fiber system, the Converter Rack and the Analog Filter Rack (i.e. the real input power to the DCR was less than before).

From Figure 6 we see that the data match within the error bars for signals split in the Converter Rack that came down Optical Drivers 2, 4, and 6. This suggests that these paths are not adding any dominant non-linearities to the system. Since the lines are separated in the same direction for different the different paths after the Converter Rack split we suggest further tests that to obtain lower error bars to check for smaller non-linearities.

However, From Figure 6 we see that the data that came down Optical Driver 8 do show significant differences. This suggests that some significant non-linearity exists in one of the following components: CM13, CM15, CF13, CF15, DCR B 15, and DCR B 16. We can rule out the LO2 and LO3 mixes as the source of this non-linearity since they are not seen in the data from Optical Driver 6 which were mixed with the same LO2 and LO3 signals.

Figure 7: Gain ratios versus relative input power for the DCR signals coming from the IF Rack and the Analog Filter Rack. $3\sigma$ errors are shown. The black line is the expected result for a linear system response.

In Figure 7 the gain ratios from the DCR through the IF Rack are compared with the DCR through the Analog Filter Rack. For signals going down Optical Drivers 2, 4 and 6 we see that the data agree fairly well. (Recall that the AFR data has more attenuation and does not put the DCR into gain compression.) This suggests that the major contribution to the non-linearities is generated before the split in the Optical Drivers in the IF Rack and after the attenuators in the IF Rack.

However, we see from Figure 7 that the Optical Driver 8 data do not agree very well between the DCR through the IF Rack and the Analog Filter Rack. This suggests that Optical Driver 8 may contribute to the non-linearity of the system. However, from Figure 4 we note that the data from the DCR coming from Optical Driver 8 in the IF Rack does not have the same shape as for the other Optical Drivers. The problem may well lie ahead of the optical fibers.

Figure 8: Comparison of DCR IF Rack non-linearity with DCR AFR non-linearity. $3\sigma$ errors are shown.

If we take the ratio of the gain ratio for data coming down Optical Drivers 2 and 6 with the DCR signals coming from the IF Rack and then the Analog Filter Rack we can check to see if there is any possible non-linearity in the IF system after the IF Rack splitter (i.e. the optical fiber and downstream). These ratios are shown in Figure 8. Although the error bars overlap, it would appear that there could be some non-linearity in and downstream of the optical fibers. This is based on the fact that every data point is less than one for the DCR through the IF Rack and is greater than one for the DCR through the Analog Filter Rack. Further tests are obviously needed.