update...
I believe that I have resolved the low compression on #4. On my last post on Sept. 13th, I checked the valve clearances and also sprayed some carb cleaner into the cylinder. I took this last step a bit further by filling up the cylinder until cleaner was coming out the spark plug hole with the piston at the top of its compression stroke. Then pressed down as hard as I could on the valve buckets using a ratchet extension to extend the valves as far down as possible into the cylinder and into the carb cleaner to make sure they get soaked. Let this sit for 10-15 minutes, then turned the crank by hand for several complete rotations. Next, I positioned the cam lobes on #4 away from the valve bucket tops and used my large screwdriver to press down on the valve at the edge of the bucket, then quickly moved the screwdriver off the edge allowing the spring to recoil and slap the valve surface against its seat on the head. Did this maybe 10-15 times. Then turned the crank by hand several complete rotations which evacuated most of the remaining cleaner still in the cylinder. Put valve cover back on and cranked engine with starter for 30 seconds to obtain oil pressure (had EFI fuse and coil resistor unplugged). Then re-assembled everything and have been driving the car daily since then..about 10 days.
Idle is perfect, very calm and smooth and turning on the accessories will not effect it poorly like before, and no codes 25 and 26. Prior to all this cylinder cleaning, I was getting the codes every day or two after resetting the ECU since mid- June when the problem first started.
Also, there is an audible change to the engine. I noticed that when under small load, before any boost, I could hear the exhaust leak. It was especially noticeable with pass window rolled down and if I drove by..say...a large fence or wall jut off the side of the road, so the noise from the engine would reflect back at me. My best comparison is that is sounded like a steam locomotive...you know...chug,...chug...chug...chug. It was very very subtle, but I was aware of it. I know what I was hearing was the leaking valve/s on the compression stroke, and no it was not the exhaust manifold, its new and only has 3k miles since I replaced it and all the bolts are tight and torqued correctly to 30 ftlbs.
So, in addition to the perfect idle, no codes, that locomotive sound is completely gone now and under load the engine sound is uniform and relatively quiet.
I think there was carbon build up on the valve seat or valve surface, or somehow a piece of carbon from the piston head came dislodge and became stuck on the valve which was the cause of my leak. I was able to remove or dislodge all or some of it to the point that the valve/s are now seating properly on that one cylinder
I am sure that the problem is related to how I drove the car the day before the problem started. I was on a late night drive for several hours and decided to open it up and drove for about 10 minutes between 90-100mph on the open freeway. I had never driven the car that fast before, and I think doing that dislodged something. I know there was carbon build up in the cylinders because prior to October of 2009 (when I replaced the vale stem seals), they had been leaking for several years before I decided to replace them. The next morning while continuing my journey is when the codes and poor idle started (mid-June) Maybe just a coincidence, but I think this theory is plausible.
Yes, the final test would now be to check compression on that cylinder, but for now, its running better than it has in months, and as long as the codes are not thrown, idle is good, and that sound is gone, I'm happy and confident that my "home-made" sea-foam method has helped.
I'll update if things change and the symptoms return, or if I finally do a compression on the cylinder..but for now, I'm just sick of working on the car and happy to be driving it again, code free!
**update on Jan 18, 2011**
4 months later all is good, ie.. no codes and idle is stable.