Long data post processing

The E.COSY experiment uses extensive phase cycling and in an effort to reduce the long acquisition times the P.E.COSY and P.COSY techniques have been developed. However these techniques require excessive post-processing of the data and consequently only the E.COSY experiment will be illustrated in Check it 5.4.1.6. For a comprehensive review of the other techniques the reader is referred to the literature [5.34,5.35]. 

Autocorrelation (see Chapter 2) can be performed on raw time series data collected by MCS cards such as those described earlier, although there are a number of inefficiencies associated with post-processing autocorrelations (see Chapter 2), not least that long data acquisitions have to be performed and the analysis has to be carried out before one can tell if the experiment has been at all successfiil. Hardware digital correlators [46] (e.g. the ALV-5000 series, ALV GmbH, Germany) can take the digital output from APD modules and direcdy perform an auto- or cross-correlation and display the result in real time. The method of operation of one particular hardware correlator was covered in detail in Chapter 2 Section 2.4.2. 

Vernier and Ko, in a 1990 patent(d ), claim that a secondary post processing heat treatment may be used to accelerate and perhaps stabilize the aging behavior and note that the untreated, partially carbonized fiber will continue to increase in resistivity, albeit slightly, for very long periods. This has become an area somewhat in dispute, as our experience with contemporary fibers indicates otherwise. Figure 7 is linear plot of room temperature resistance as a function of time for two samples of the above-described set. This data indicates that the room temperature resistance for these... 

It is questionable whether this approach would be any better than the postdetection data manipulation techniques described in Section 4.3. The trade-off is between complexity of instrumental design and operation, programming complexity and post-measurement data processing time. As long as the instrument response could be correctly and precisely modelled in software, the balance would probably lie on the post-analysis processing side. 

With these icons and the foundational data we are ready to map out the process that we are studying. An example of a very short process map is shown in Chart 8.2. Most maps will be much larger because the number of steps in a process can be quite extensive. Often we use butcher paper that is about 3 feet wide and as much as 20 feet long to graph out the complete value stream. We post the paper on a wall and tape the value stream icons and Process Baseline tables on the paper, and then draw the appropriate connecting lines. 

---