Make New Entry, Make Followup Entry
User name Kai
Log entry time 07:33:23 on November10,2009
Entry number 298733
keyword=weird synch signalAfter tuning the tager, I went to check the parity data. The width of the Blumi asym was fine which means all of them are around 500~600 ppm. The up and down stream bcm asym agree with each other. But something went wrong in sync(see attached left). I checked it in details and found that there are two problems. First is right arm has no "gate" sync reading in both Happex scaler and Pvdis'. Second is the gate sync of scaler11_28 in Pvdis left arm is different from all other "gate sync". I went to check the sync signal in the counting house side, and found signals are all good. I repluged them back, and started a new run, and everyting looks normal again. I can understand why the right arm signal get back, which might be because some interminent connection previously and has gone after I replug them. However, I still don't understand while it can solve the problem of scaler11_28 which is one of the 4 "gate sync" on right arm. I did some calculation on it, and find it will be either 3ms or 5ms that could cause all the counting diff in that run so that it cannot be MPS' (GMN) fault. In addition, there is a spared (copied) channel, which is scaler13_12, that has the same input as scaler11_28, whose result is consistent with all other "gate sync". I hope it was not because there were jitters on the MPS veto of scaler11. Further, I print out the raw scaler cnt of both 103.8K clk and all "ungate sync", and the result of left and right is consistent with each other very well. Lucky enough, it is also scaler11 that have the 103.8K clk input. So I believe the left arm and right arm was still synchronized well, and maybe it is because some bad connection along the path of "gate sync" which can cause either no reading or some reflection in some channel, and after I unplug and replug it back, it recovered itself....... (I'm still trying to convince myself...) However, the point is everyting works fine now, and I will definitely take an eye on it in future.
FIGURE 1
FIGURE 2
