Originally Posted by RR_NDB
1) They used RC oscillator (frequency dependent on batt. voltage) at TX
2) Peak detector at RX
3) Conventional RX front end BPF (not a very sharp one)
We speculated a long time back about whether the pingers have advanced to crystal oscillators or are still using RC or LC oscillators, and whether the ping rate is digitally divided or runs from a separate oscillator. I don't recall any resolution of that. The required spec is 37.5 kHz ± 1 kHz; I only hope the pingers are really better than that.
Conclusion:
Clearly there is room for improvement for much better ULB systems than the current std.
We can all agree on that.
Originally Posted by rh200
How accurate is the cycle, one could imagine using pulsar detection algorithims. Another words folding the data in time over the the full pulse cycle, the noise should decrease with a root two factor.
We don't know (see above), but if it is stable, this might be done (root(n) noise reduction). I had speculated in my original post on this subject (#951 in the second thread) that something like this might have been done with the data from Emeraude. It was just a guess, I don't know what they actually did.