dcraid I would think that your vertical integration strategy might make it easier to add PEC support for your upcoming AM5 mount.
If you look at the AM5's documentation, you will find that each cycle of the error curve is not precisely the same as the previous cycles. This will make PEC harder to implement. The data may need to be collected for a longer super cycle, assuming such a super cycle exists.
Just in the three cycles that are visible on the published curve at ZWO's site (what they documented as "single" cycle, but actually shows 3 cycles :-), you will already see deviations from one another.
Each cycle for the AM5 should be about 430.82 seconds (the length of a sidereal day [86164.1 seconds seconds] divided by 200, assuming ZWO is also copying even that from the RST-135 mount, too).
The "full periodic error" curve shows very gross low frequency drifts (with peak-to-peak error changing over the hour angles). That slower drift could be autoguided away -- however, you end up having to autoguide anyway.
Chen
