Hi Dan
From your first para I assume that you had DEC guiding in the Brain turned OFF but you had trained PHD to guide in RA.
If this is the case then your graph is perfectly normal. It shows RA being guided well but as you are not guiding in DEC the red line is drifting up
in a steady manner. This is because your mount is not exactly polar aligned.
To use PHD effectively you first need to polar align your mount at least
approximately as if the DEC drift say up as in your case is excessive, PHD has a hard time continually bringing the rapid drift back to zero (ie guiding in DEC). This appears to be the symptom of your original problem.
I usually set the DEC drift so that is is slightly biased in one direction (say up as in your case). PHD then has no difficult in correcting this small drift and keeping the DEC guiding near the zero line. To reduce your current large drift you may use PHD to trim up your polar alignment.
Firstly chose a guide star fairly high up on the celestial equator. The move your mount's azimuth knobs say half a turn, noting the direction of movement. This will shift the guide star on your PHD screen out of PHD's guiding range so you will have to stop guiding and reselect the star.
Then resume guiding and note the magnitude of the DEC drift. If it is smaller then you have moved the knobs in the right direction, if not, do this again reversing the movement of the knobs. In this way you could then achieve the condition where the red DEC line is oscillating about the zero line. This would indicate near perfect polar alignment. However, as above, I like to introduce a small drift from the zero line so that when
guiding, PHD DEC guiding does normally not have to cross the backlash zone.
Having done, go into Brain and set the DEC guiding to auto and then retrain PHD for both the RA and DEC directions. I usually set the steps so that there is about 10 or so training steps in each RA direction. After this is done commence guiding and you should now find that you may guide normally in both RA and DEC.
BUT, if you do the above and you still find that the red DEC line is continuously drifting above the zero line this could indicated that your
guiding PC is not communicating with your mount in DEC. How do you check this? You can go into the Shoestring site at the following url
http://www.store.shoestringastronomy.com/downloads.htm
and download the software utility GPINT Check or GPUSB Check whichever suits your PC/mount interface hardware. Run this small utility and it will present you with a screen showing RA and DEC direction buttons. These may be selected with your mouse. Select the appropriate connection protocol from the top left hand list (same as you use for PHD). Set up your mount etc at any time (ie daytime and no need to mount scopes ) and select as normal and go to a guide star and the mount head will move to its position and then hit enter to confirm the completion of guide star position. Then on your Synscan paddle select Utility Function and Show Position. Hit enter in this location to show the actual coordinates of RA and DEC.You should see the DEC steady and the RA moving slowly as the mount tracks the "star". Select the utility RA and DEC movement squares on your PC screen with your mouse. If you hit DEC up and then DOWN you should see the DEC coordinates of your mount move up and down. Similarly RA + will move the RA coordinate change faster and RA minus will show the opposite change in RA direction. If these functions work it shows that your PC is communicating properly with your mount.
Long winded Dan but see how the above goes and let me know.
Cheers Peter
I can see my RA guiding is quite good for my image scale. 
