Manchester serial decoding does not work for the output signal from my TSOP33436 IR receiver (RC5 protocol). Instead of positioning the first bit at the first transition (which is from +5V to 0) it does so at the second transition. The results are wrong packet length (13 bit instread of 14) and wrong data.
I have raised this for investigation by the team, it looks as if we may not be detecting the first edge as the start bit so end up being 1/2 a bit out throughout.
I think the start flank is missing in this signal.
Every example I have found has a halfway signal when idle and has a edge half the amplitude as a start of package timing trigger. Also if you count the full bits then you see the first bit should have a staring edge before the first edge.
Something I have added in red in the below scope image.
I have also exported to sigrok and gave the same inconsistend timing result.
Perhaps our problem is that Manchester encoded signals look different on a data line (no DC component) than on the output of an IR receiver (idle = 5V).
This is actively being worked on by the team as part of a new Manchester based decoder we are introducing, so should be resolved in an upcoming release of the Early Access software
In the latest release of PicoScope 7 Early Access 7.1.17 we have added a new option on the Manchester Decoder setup panel. This allows you to specify the number of edges to ignore at the start of a packet. For normal Manchester this is set to 1, but for signals such as this RC5 set this to 0.
The download is available from within the software, or from the downloads tab at the top of the page.