Guide pulse direction appears to be reversed, only at declination of 33

[Jeff Nibler]

PHD2 has been working fine, but at specifically declination 33, it appears PHD may be pulsing in the wrong direction?

Mount is an AP110GTOAE with an OAG and rotator. Calibration has been run. 

Guiding works fine at all declinations other than Dec 33. 

In the star image viewer of PHD2, if I use PHD manual guide buttons while looping 1 second exposures, and I click the North button, the stars move South relative to the green target circles (this is simply the results of guide calibration - no issue here). At Dec 33, each automated guiding pulse appears to be made in the wrong direction. Each pulse moves the star further away. I see repeated North pulses, and the star repeatedly moves downward away from the green target circles. 

If I click the North button on the Astro-Physics APCC software, the star in the star view section of PHD2 also moves South, confirming the mount controls direction match PHD2s North and South buttons and the star motion matches

Do nothing else aside from slew to Dec 30 and start guiding - guiding works perfectly. Slew immediately to Dec 35 and start guiding, everything works perfectly. No other changes. Slew back to Dec 33, same behavior. 

Tonight this behavior is 100% consistent at Dec 33. Other nights, it's sporadic. I have never, once, seen this behavior at any other dec, over the course of 2+ years I've run this specific telescope/PHD2 instance.

https://openphdguiding.org/logs/dl/PHD2_logs_4Mea.zip

If it helps, I have diagnostic image saving turned on and can upload the images. 

Thank you,

Jeff