I have a 1991 900s naturally aspirated with a 2.1. When you try to start it, it spins over so evenly that there can't possibly be any compression on any cylinder. It won't start, and the compression comes up eventually, but by then the plugs are fuel fouled. If I walk away and try to start it the next day, it starts normally, sometimes running rough for a minute or two. Once it smooths out it won't miss a beat and drives great with normal power. It can be shut down and re-started just fine for that day as long as it doesn't remain shut down for more than a few hours. The first few times, I'd clean plugs and get it going. Or I can unplug injector connectors until it sounds like an engine with compression. Compression on all cylinders was approx 200 psi when checked, but I checked it during a time when the problem wasn't present. Lately, I've left it alone so as to track the nature of it. This pattern is very predictable and it is repeatable:

1)Car won't start, sounds like there's no compression on any cylinder. Keep trying till it starts to sputter, but the plugs are fouled. Do nothing, walk away.

2)Try the next day. Car cranks like it has compression and starts easily. Runs rough a minute and smoothes out. Runs great for as much Saab enjoyment as I want for that day.

3)Go out the next day and no compression again. Leave it alone.

4)Try it the next day, she fires up and drives great.

My first thought was that the fuel leaked down, washing the cylinders. I've removed the rail, observed, saw nothing. No fuel present in regulator vacuum hose either. That would explain poor compression, but I'm tellin ya it sounds like the "all valves bent" kind of no compression. Anyway, I tried my best to catch some fuel leakage in the act at various times and temperatures and nothing found. I'm pretty confident the problem is not fuel leakage.

My next idea was that the timing chain tensioner ratchet failed leaving oil pressure as its only source of tension. Sounds like a good explanation. Say you shut it down, and the cam lobes want to relax in their happy place. The valve spring pressure against the lobes may coerce them to rotate slightly, and in doing so, their desire to rest where they want might overcome the failed tensioner, pushing it back. Maybe even far enough that when I try to restart it, it acts like a severely worn chain that can't drive the cams in proper time. I don't think there'd ever be enough slack to actually jump a tooth in this situation though. When you install a chain, you can't quite put it on with just the slack afforded by a compressed tensioner, you have to have a sprocket unbolted, or a guide removed. Anyway I pulled off the valve cover and put the engine to tdc when the no compression condition was present. Everything lined up perfectly and the chain was tight all sections. I removed the tensioner and found no mechanical fault with it. I even installed a different tensioner and the problem is exactly as it was previously.

How can an engine like this loose compression on all cylinders, and be fine later without doing anything to it? Are the lifters on these engines so soft they won't open a valve the least bit if they're drained? Would every one of them bleed out at the same time? Is there a check valve somewhere required to hold the oil up there? And if that were the case how could they hold pressure overnight when I walk away from the failure to start? The oil is definitely getting up in the head. everything was well oiled when I had the valve cover off. I can't imagine sticking valves or lifters because it runs so great once it gets going. There'd have to be at least 4 doing it by the way it sounds when cranking. Same goes for the theory of warped oval shaped valve seats. You'd think if 4 or more were bad, I'd never get them all to seat well enough to run great. I'd like to know if there's anything specific to look for before I tear into the head.