FInally I found some time and made some testing of code related
to Dominant Cycle calculation using different methods suggestef by Ehler
and I found some strange behaviours. I enclose two screenshots one for
test signals (chirp and teethsaw) and real time data. Dominant cycle detectors are
as followed:

1 - From the book 'Cybernetic Analysis for Stocks and Futures' by John Ehlers
2 - From book 'Cycle Analysis for Traders' by John Ehlers autororrelation peridogram
3 - From book 'Cycle Analysis for Traders' by John Ehlers combined BandPass output
4 - TASC0308 (BandPass filter bank)
5 - From book 'Cycle Analysis for Traders' by John Ehlers DFT

from screenshoots we can see tht all detectors are OK with chirp signal

as per trends (297 bars long) is clear that some of them settle on completelly wrong period
so nr 3 on 23 bars and nr 5 on 13 bars. Nr 5 has some strange notches down from expected
period of 50 bars.

As per real data and comparison of DC measured all of them show different DC. Nr4 for example
shows the same DC value even market switches from trend mode to cycle node (01/22 - 05/22). Nr 3
shows similar behaviour.

So look like a bit of BS... unfortunatelly..at least at the moment

Krzysztof