I never faced any guiding issue with the AP1100 so far since years (I have the serial #10) and GTO CP3
But as far as I understand, it is very unlikely that the issue comes from an defective USB cable nor the RS232 cable if everything works well except the AD guiding.
I would rather suspect a PHD or AP Ascom driver wrong setting.
How is PHD calibration ? -> calibrate once on the ecuador close to the meridian at the begening of your observing session and then slew to the target.
Of course if you have a fixed setup and don't remove your guiding camera, you only have to calibrate PHD once forever.
So first of all review PHD calibration data. A good calibration is really a keep point to get a good guiding.
The verify the AD agressivity and the AD MnMo :
AD agressivity should be higher or equal 50%. 70% is a good starting value for average conditions.
MnMo in the threshold under which PHD will not correct the guiding error : It should be set to a value corresponding to about half of the main camera pixel size.
But the unit value is pixels of the guide cam, so you have to do so basic maths to calculate it from the main cam pixel sampling.
For instance if your main cam has a sampling of 1" per pixel and you guide cam a sampling of 3". You would like to set the MnMo to 0,5".
So 0.5" is equal to 1/6 of the guide cam sampling -> so you would set the AD MnMo to 1/6 = 0.17
Concerning the DEC, I use to set the MnMo to the double of AD MnMo and agressivity to about 30% -> this is to avoid guiding in 2 directions in DEC which is the second keypoint to get a good guiding.
In the AP driver, verify the guiding speed and maybe the PEM state. try disabling the PEM if enabled.
Example of what I am curently geting right now form a remote EQ8 in Chile (I am in France) : the EQ8 is not as good as the AP1100 in AD tracking (from far) but the tuning principle remains the same.
Look at the graph : you can see that the DEC stays at one side of the axis and then on the target you can also see that the hits are located only at te uper half and well grouped in DEC at least.
The main limit of this EQ8 is the AD
(the star is a little bit saturated and has a triangle shape, but the guiding is good enough for the moment....until we replace the EQ8 by an AP one day)