sonett wrote:
OK, back on topic. I removed the exhaust manifold and reduced the secondaries from 30'' to 18'', the results were mixed and it didn't have the effect i was expecting, top end power was increased, but i lost some torque low down.
We took the revs a little bit further this time and at around 6700rpm the power is still not tailing off

, on the road i have taken it to 7000 rpm for a short while and there is no valve bounce
I've dug out the valve train weights I originally collected and run the valve bounce calculations again. There is small margin of error because I don't know the exact fitted length the heads were actually assembled at since I didn't do that so I've assumed the optimal situation where the valve springs have been properly shimmed up to be just away from going coil bound (by about 1.5mm) at full lift. I'm also aware that you didn't actually do any of that so the engines might not meet the targets below.
Also I've had to estimate the spring rate of the standard springs from their dimensions using a program of my own devising since I don't have a measured spring rate for them on my spring tester. I do have the exact figures for my own uprated springs.
Anyway the results I get are as follows.
1) Stock valve train on stock springs - 7000 rpm
2) Using my lighter inlet valves (82g instead of 94g) which is how the 44m valve head actually is but still on stock springs - 7200 rpm
3) As above but using my uprated springs - 7600 rpm
4) Adding the lighter cam followers (69g instead of 95g) but still on stock springs - 7600 rpm
5) As above but on my uprated springs which is how the 46mm valve head is - 8000 rpm
So what revs do we actually need? Well I'm astonished that with the modified exhaust manifold the power still hasn't started to drop at 6700 rpm given that it first peaks at about 6200 rpm on most of the power runs. The head must be flowing a godawful amount of air to be able to keep supplying the cylinders that high up on a not particularly long duration cam. I'm beginning to wish I'd actually flow tested both heads now.
It's also an established fact that to get the best acceleration it's necessary to rev any engine about 500 rpm above the point where the power starts to drop which means at least 7200 rpm and we still don't know what the power curve actually does above 6700. It might hold on even higher.
So it's looking like even the 44m valve head needs the uprated springs to let it run out to 7600 rpm. The 46mm valve head should be ok with its 8000 rpm limit but anyway there's not much else we can do short of finding even stronger springs.
It goes to show how important light weight valve train components are though and how badly Saab got things wrong originally. The stock 42mm inlet valve didn't need to be anywhere near 94g with proper design if I can make a perfectly reliable 44mm one at 82g. The stock cam follower is also ridiculously overweight and could easily have been as light as the ones I used in the 46mm valve head. Together these two items give an extra 600 rpm which means if Saab had done so too the stock valve springs could have been much softer and therefore less power sapping and causing less wear.
Running the equations, to retain the same 7000 stock rpm limit, which in fact is far higher than the stock engine needs, but with properly designed light valves and followers the stock springs could have been 15% softer than they actually are. With a 6500 rpm limit which would have been ample for the stock engine the springs could have been nearly 30% softer than they actually are.