Hi Jean. I think your overall guiding results could be improved significantly if you avoid some operational mistakes you're making. Before getting into that, the Dec guiding issues you're talking about seem most pronounced when you were pointing far to the north at Dec = 78. The behavior looks pretty asymmetric, suggesting the scope might be out of balance in Dec. I suggest you start by returning the scope to that position and checking to see if it's well-balanced in Dec. In both guiding sessions, it looks like you may have some static resistance in Dec that's making it hard for the guiding to reverse direction - there is a pattern of very sluggish response followed by over-shoot. It's shown here, Dec in green:

At point #1, you can see there's a delayed response getting the mount to move south, probably some combination of backlash and stiction (static resistance). Then the resistance is finally overcome and the mount over-shoots (point #2). The mount is not in good shape to do bi-directional Dec guiding at this point so you should either make some mechanical improvements to the Dec axis or consider doing uni-directional Dec guiding.
Getting back to the operational mistakes, here are the things I see that are holding you back:
1. You're trying to use 1-second guide exposures without doing multi-star guiding, not a good idea. You should start using the auto-find button to choose stars and you should leave multi-star guiding enabled.
2. You got a calibration alert for orthogonality error that you chose to ignore. That's probably because of the backlash you have in Dec. Before doing a calibration, you can clear the backlash manually this way: a) start PHD2 looping on a star-field, b) use the hand-controller to move the scope north ('up' arrow) until you see the stars in the display moving significantly, c) start the calibration. This will need to be done every time you calibrate until you can improve the mount mechanics.
3. I don't see any evidence that you've used the Guiding Assistant. You should do that to set better min-move values and to measure the backlash on your Dec axis.
You are basically having to deal with a balky mount but I think you can get better results with it. To get a systematic view of how the mount performs, you can follow this procedure:
You should also take some time to review at least the best-practices document and probably the trouble-shooting section of the PHD2 manual.
Good luck,
Bruce