This PEAC algorithm is just a Pisano-type algo, with some added End Around Carry.
As I'm about to test it against other algos, and I have already examined the Pisano/Fibonacci aspect, the EAC side now looks strangely and sadly under-covered. This is an essential sub-system that has deep effects on the behaviour of whatever uses it, so it should be fully characterised.
I have noted before that EAC seems to perform a mod-2^n-1 for very cheap but I suspect there is more to it than that.
When working with an 8-bit byte, values mod 255 will range from 0 to 254, leaving the value 255 "untouched". This is a weakness because this reduces the odds of detecting an error by 1/256th. OTOH if the checksum is 255 this is a clear sign that something very bad has happened.
With end-around-carry, this range is shifted and now, 0 is the "forbidden value"... If the sum is 256, then the byte wraps around to 0 but the carry is added to give 1. Thus 0 is never encountered, unless the system is initialised to 0 and fed with 0s.
This is an important distinction because PEAC allows 0 as a valid value, while Fletcher does not. As mentioned there, mod and EAC are not the same but have enough similarities that EAC could be studied by going through mod.