Worn valve tips , valve action geometry and Mushroom head adjusters.

Measuring rocker action is very painstaking. You have to use a dial gauge and measure the full lift of the valve. Then you have to find the halfway point, then judge if the valve is at 90deg to the rocker!! Simple-NOT HUH! I have monitored our Forum with regarding the Commando problems with valves and geometry and nobody seems to fully understand or has measured it. I have done this on a Reliant engine-got it wrong-decided that Reliant got it right!

It's not easy to get a clear side view with the Dommie head design.  But one thing to help must be to swap out all the springs for some much lighter ones (like Amal carb springs).  Then it will be much easier to watch (and perhaps take pictures, with a degree disc fitted) to see and record all the movements without the cam kicking it past the watch points.  An other issue is - should it fit best half way up the cam movement, or towards the peak of the cam where the forces must be biggest?  

............the whole point of maximum efficiency of your valve gear on ALL 4 stroke engines is that at half lift the rocker and valve need to be at right angles to each other. This not an issue of where you are on the cam. Also David I would not think you have to do all the rockers/valves once you have measured one then we must assume you have the same rocker advantage and valve length in your head. Then of course if you have a higher lift cam the 90Deg point will change again!!
This is effected by so many things-the length of the valve stem the length of the push rods, which in turn are affected by the cam followers which in turn are held in the barrel’s. Then there is the head height/barrel. It seems  that Norton's in their wisdom over the years changed several of these points without full thought of ALL the points, hence the rocker/valve action was compromised (messed up), then another factor ie valve length would be changed, and then of course the wrong valve would be put in with another set of push rods, then the compression was upped by lowering the head so certainly on the Commando we had a mobile disaster. And I do not know if anyone has managed to measure the result, which of course would only be effective on the engine it is measured on.
Nest of worms is a perfect description.

Both inlets will be the same - and both exhausts - so only need to do one of each...

Could you mark (with fine paint pen or  bow pen) across the rocker boss to the head, at each end of the rocker movement?  Then move to the half way point and inspect to see if the tappet is either in line (correct) or out of line (incorrect)?

Marking the rocker boss/head might work but I doubt it will be accurate enough. 
Robert-with regard to the 650SS with high compression, 'maybe' the higher compression corrected a push rod inaccuracy? (or made it worse). This is the problem here it is very hard to measure accurately-best of luck all round on this one, the Commando has had many changes in valves/pushrods/compressions??

Shouldn't that be maximum travel across the valve tip - but in the right place. This will also be less stressful on the valve train as well. Cam lift and rocker ratio are fixed, with the max travel across the valve tip the valve will be opening the maximum amount possible. 

Here are two valve stems from my 850 showing the contact sweep from AN mushroom adjusters, with standard size Maney steel pushrods and Black Diamond inlet valves. Removed and refitted on my RH4 then Fullauto head so two different contact lines and later replaced with +1.5mm Black Diamond inlets.
The bigger valves show the same stem pattern after around 20k miles with very few clearance adjustments, valve train is very stable. Not to the edges with my standard type Web #312 cam and 0.006" valve clearances but very happy with the geometry.

Contact sweep pattern is a useful way to gauge geometry at the end that matters, without all the dial gauge or lighter spring hassle. The 90°method is commonly used on pushrod V8s with hydraulic lifters and stud mounted roller rockers as in the second pic, rather than our solid lifter pushrod rockers on spindles with screwed adjusters.
The V8 mark is at 90° to a Norton because of the roller but they aim for a thin central line for minimal side force on the guides.

Interesting indeed, your head may be OK with new valves but easily checked as is by blacking out the stem with a Sharpie and see what the contact sweep looks like. Go from there.
The valves in my Mk3 when I bought it were horror film rejects.
 

From the late 60's when the 'standard' length pushrod was introduced and used right through the Commando production, the valve length changed at least twice, but strangely the pushrod length stayed the same. The police used a shorter pushrod as did the production racer, maybe they were onto something. My predecessor did an A0 drawing of the complete valve operating components  including the lengths and angles involved to determine the radius needed on the mushroom adjusters, we don't have it here so I assume he kept it. 
It surprises me that the adjuster runs very close to the edge of the valve tip on most engines, why Norton did not adjust this by machining the adjuster position 0.015'' further out from datum surprises me as it would be simple to do and I suspect would make a fair bit of difference.   

After the engineer made such a good job of designing the best mushroom adjuster available, to move the rocker datum 0.015" further out would be a retrograde step!
The difference in most engines is probably due to tolerances in the cylinder height and valve train setup from cam chain to valve seat depth but they run perfectly fine with this.

If you want to get as close as possible to the drawing, it's up to the engine builder to correct these to get optimum central sweep with maximum valve lift for whatever cam is used.
Paul Dunstall mentions in his tuning manual that the Commando valves being longer with shorter pushrods was to give a slight improvement in rocker geometry.

Historical hardware differences are irrelevant in this case, although techniques and advice from enlightened old school and modern tuners is essential. Work with what's in front of you, usually modern equivalents.

Neill, and this is where we must be careful - Commando pushrods? and longer Commando valves? the pushrod lengths used were around since 1964, so long before the Commando and they were used in the late Domi's, the Commando used two different length valves in its life - fact.
Paul is correct in what he implies, but it is a case of it being poorly written, the early Commando valve was the same length as that used in the Atlas and the push rods and rocker used are the length we know are used on the Commando to this day. The problem arose when Norton added the stellite tip to the valve, the calculations were royally screwed up it seems resulting in the longer valve used in the later Commando's with the standard length pushrods, I agree with Paul this it not ideal. They must have known this as they described using shorter pushrods in the Combat engine when it was an optional extra before it was fitted as standard by which time it used the standard length pushrods, the production racer used shorter pushrods and the Police forces also specified a shorter pushrod, along with Paul Dunstall, but Joe public had what 'fitted' and needed to be used up it seems. 
Norton in the day most probably realised, couldn't care less as it was just 0.036'' longer and had ordered 10,000 of them so they were just fitted anyway.
If using the now 'standard' length Commando valves I would always fit the slightly shorter pushrods.
I did not mention moving the rocker datum, I said 'from the rocker datum' also the drilling for the adjuster, when you look at the rocker the hole drilling for the adjuster is taken form a position that looks to be where the adjuster to valve intended contact is made, so the 0.015'' I mention would be form a position on a valve tip and not on the rocker itself. I will do video showing how Norton did it and the drawing, this will help you and others whose to get the system to an optimum position.  
'Work with what's in front of you' - always, and to this add 'do not assume that Norton had it correct' especially as the 70's progressed, lack of money and being in administration ensured bikes were just 'put together' and if it worked sold to an unsuspecting public.