Data Digitizer Optimization Setting

Question from Clint and Bill...

Is adjustment of offset into data digitizer a calibration concern? Currently, the plan is to measure the level at the Correlator input filter, and adjust the digitizer input level with the BBC output step attenuator under M&C control (I think). See The basic question is what is the optimum power level into the digitizer during the observe and calibration cycle, and how does the system software know both the optimum level and the actual level. This is related to the issue of the TP digitizer dynamic range requirements.

The digitizer input must be optimized based on either the correlator input or the total power results. I am told the correlator input is preferred. I am also told without this optimization, you risk having a signal outside the 3-bit range of the digitizer which would cause the digitizer to output all 1's or all 0's.

  • Has this question been answered by the answer to the TP digitizer dynamic range issue?

  • From where did the suggestion that the correlator input level for optimization is best come?

-- JeffMangum - 23 Jul 2004

Comment by BillBrundage - 23 Jul 2004:

The 2-4 GHz baseband data digitizer (DG, 4 GS/sec)) optimization and the total power dynamic range are related. The baseband total power digitizer input level = k * input level to the baseband data digitizer, where k is a fixed ratio. The total power digitizer can/will have an input (and output) dynamic range (Vinmax / Vinmin) of 20:1 (11:1 operating range * 1.8 = 20:1 for setting tolerances, per JeffMangum's TWiki on total power digitizer dynamic range).

However, the 4 GHz digitizer is only three bits, so its input power must stay within a limited range of approximately 8:1. The worst case operating scenario has its input power range of 11:1 -> 20:1 for setting tolerances. Thus total power gain setting for the worst case could result in data digitizer DG outputs of all zeros on Tsysmin, all ones on Tsysmax and zero correlator data.

Thus it appears that some combinations of observing band and calibration temperatures would result in no correlator image data.

It may be possible to change step attenuator (system gain) settings for each condition of Tsysmin and Tsysmax in order keep the digitizer near optimum in each condition, but doing this probably complicates amplitude calibrataion and (especially?) phase calibration.

The monitor control software must set the gains and thus the power level into the DG with reference to either the corresponding total power level or to the corresponding state counters in the digital filters of the correlator. State counter statistics directly determine the optimum level into the DG, but use of a total power detector requires calibrating its relation to its corresponding DG's optimum level, which in turn must be determined by the corresponding state counters. Such calibration of the total power / DG level relation could vary over time and would require re-cal with every change-out of a TP detector, or of a TP digitizer, or of a DG.

-- BillBrundage - 23 August 2004

Resolution of this question will require a CRE (soon).
Topic revision: r4 - 2004-08-23, BillBrundage
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback