Some thoughts based on my experience helping others with a MK FC with similar spike issue.
The same setup between those copters are large frame with big TM U-x motors and TM ESCs. Those high power motors and ESCs generate lots of noise and for the MK case, I2C comms of the MK is interfered. In some cases, the FC simply can't talk to the I2C to PWM convertor. The partial working solution was to replace the 12s HV BEC with a dedicate FC battery and re-route the wirings. Ideally true opto isolators should be used(there are no opto isolators in the so called opto escs). But I haven't seen any available products for that purpose now.
For the AQ, I can only make assumptions based on info in this thread that there may be some sorts of interference in your setup. Based on the AQ6 about page, it has two dc-dc converters and emi hardened IMUs, but there are no such things for M4. As the schematics is close sourced, I can only guess somehow the emi hardening eliminated the interference so your AQ6 flies well. But the M4, hmm... it's another thing.
Suggestions
Try to add dedicate battery w/ BEC to power your avionics. And make sure the ground of esc and ground of pwm output are the ONLY ground connection. Ground loops (
https://en.wikipedia.org/wiki/Ground_loop_(electricity)) will cause problems. You can then do a constrained ground test(e.g. props reversely installed and frame secured) to check if the spikes are gone.
Regards,
Ye