Fullauto Cylinder Head

I disagree that complex systems are not applicable to old Nortons. Most of us have learned that when developing a race bike, it is usually better to vary only one thing at a time, then test the result of the change. If you think of a race bike as a system, there are probably about 6 critical variables which must be optimised to achieve best function on various race circuits. If you change the power characteristics of the motor, trick the frame handling, change the tyres. You change the whole scenario. There is a limit to how much the rider can adjust to the machine - failure to adjust is usually displayed in poor lap times. On this forum, there are a lot of bragging rights attached to dyno figures. With test equipment there are always two critical factors - accuracy and precision of measurement - and you still need to be certain that what you are measuring is relevant to the performance outcome.
Personally I don't accept that flow benches reveal what configuration will give best cylinder filling, because I don't believe similar circumstances occur on the bench in comparison with what happens when a motor is running. When you get good figures on a flow bench, don't you then make an ASSUMPTION that the head will work well in practice ?
I'd be interested to know if any Dynotec dynamometer has been calibrated by a national measurement laboratory for the repeatability it's torque figures.
As I've said previously - most Commandos have an upper rev limit of about 7000RPM before they self-destruct. Going higher in revs safely to get more power, is extremely expensive. So the way to go is improve the torque characteristic and the gearbox. Then adjust the handling to suit. The biggest boost in acceleration that I achieved was when I replaced the standard gearbox with one with close ratios.

There is a problem with torque - if you don't load the bike up, you never know how much torque you have. I run very high overall gearing and use the close ratios to keep the crank spinning well up in the usable rev range. Doing that with the 4 speed close box creates a problem with first gear during clutch starts to races. However once the bike is rolling, it is very fast. It is a very deceptive situation. You might believe the bike is accelerating as fast as it can until you raise the overall gearing and make the motor pull harder. I think it is the inertia effect of the heavy crank. The thing you have to overcome to go fast on a race bike, is the head-wind. Once that heavy crank is wound up, there is no stopping it. So I don't believe lightening the crank is a good way to go.

acotrel said:

Personally I don't accept that flow benches reveal what configuration will give best cylinder filling, because I don't believe similar circumstances occur on the bench in comparison with what happens when a motor is running.

Click to expand...

I would say there's a vast body of knowledge and experiences that would suggest differently. The flow bench is just one tool. The work (or use of data) needs to be tempered with knowledge and good judgement. Naturally if you totally hog out a port you will certainly get better valve flow coefficients but it will not perform well. In the case of the FullAuto head I have solid faith that it was done with good knowledge and experience.

acotrel said:

When you get good figures on a flow bench, don't you then make an ASSUMPTION that the head will work well in practice ?

Click to expand...

Not necessarily. First you need to define "good figures"; what I mean is there's more to it than just CFM. Furthermore, if you modify a port on a new head for a given engine to something you know has worked well before on a given engine, I don't think that expecting the new head to work well is an assumption any more than driving over a bridge that is properly designed - you have the expectation that it will perform in a certain way.

Rohan said:

If the dyno of a FA head on a std bike shows north of 45 hp, thats already better.
Regardless of how knackered the old one was.
Enquiring minds, enquiring minds...

Click to expand...

Dances with Shrapnel said:

Well then, it appears that you are suggesting no need for a baseline dyno pull and that a FullAuto head dyno pull showing anything over 45 hp would be attributed to a FullAuto head.

Again, your charming habit of totally reversing/twisting what I actually said to not what I said at all - and then firing it back at me. You should go into politics...

If you were really involved in this stuff, you'd know the packages and possibly the folks.
We notice you didn't pm me or LAB for details, as I offered.
I don't wish to quote other (independent) folks on a forum where they are not involved, nor paraphrase them.

The difficulty we had back then, when this stuff was all in its infancy, was that it was terribly difficult MATHEMATICALLY to distinguish very slight differences in engines where it made rather large differences in performance. One of the much touted examples of this was the BSA B33 and a Goldstar. With quite similar designs (same family, obviously), its 'only' a matter of cams carb compression, and chamber/port shapes between them. Not to mention the whole physical package. But if you tweak a B33 to Goldstar specs ( which of course is not really possible, they are different enough) it doesn't/can't go like one.
The engine simulation stuff I've followed, and using these 2 as a test, doesn't do particularly well at sorting this out either. And if it tells you a B33 WILL go as a Goldie would, what does that say ??

And, as said earlier, a computer can't do anything that a pencil and a supply of envelopes and time can't do.
They just do it an eck of a lot faster. (I say this because some seem to think they possess magical powers).
As long as they are fed the right info, and someone who knew what they were doing programmed it.
Which is different to data being entered into it, since you seem to have twisted this.

With all your big words, have YOU actually used an engine sim package to any commercial purpose.
I'm at a disadvantage not knowing a name or any of your background.
Or if you are the Motor Teamleiter at Merc ....

Click to expand...

What's with this secret society where one must PM someone else. Certainly, there are differing opinions out there based on real world experience or lack thereof.

Good that you dabbled in something with a B33 or Goldstars but this dates what you experienced which is relevant when discussing computer assisted design. The point of computers is not only that they can compute faster but the reduced time required allows efficient integration of many different numerical models of phenomena in an IC engine; it can quickly integrate and act on large sets of data and render very good information with the right code and understanding. Things are no longer in their infancy so wake up and have a look around. Where are those "Enquiring minds, enquiring minds..."

Using an engine sim system in "any commercial purpose" is another red herring here. I've cited my work and involvement and successes and have a low-cost system at my home office. You have established you have no pertinent experience with modern engine simulations and fail to understand the fundamentals of numerical modelling (ancient or modern), yet here you are espousing why it is so difficult to do or so costly to do. If you had at least tried, there would be something to work with and there is a chance you would have had some sense of context.

'And, as said earlier, a computer can't do anything that a pencil and a supply of envelopes and time can't do.
They just do it an eck of a lot faster. (I say this because some seem to think they possess magical powers).
As long as they are fed the right info, and someone who knew what they were doing programmed it.
Which is different to data being entered into it, since you seem to have twisted this.'


Pretty difficult to do the data manipulation required in rationalising the matrices in pattern recognition, with a pencil and paper. You might be able to deduce an algorithm, however the rest would be impracticable. If you look at a bike as a system and rate all the variables, it would probably be relatively easy to establish what patterns exist. ( I believe the system is bimodal for road racers). The problem would be collection of performance data. What I am suggesting is computerising something which humans are very good at - PATTERN RECOGNITION.
I regard computers as a tool to be used, they don't THINK intuitively.

Isn't engine simulation a matter of going in the reverse direction to pattern recognition ? However how do you know in which direction to send your computer if you haven't previously analysed the patterns ?

Dances with Shrapnel said:

Good that you dabbled in something with a B33 or Goldstars .

Click to expand...

You are good at being obnoxious with every single word, so we keep it simple for you.
You seem unaware you are repeating near everything I say ?

That was only the beginning - and a very convenient sample to test this stuff on, to test how reliable it is.
If it gets the B33 faster/stronger than the goldie, you know what that proves !

Current things.

These heads just happen to produce approx the same hp.
Try getting your sim package to prove that...

One plan was to manxify a featherbed ES2, using data from the head on the left.
Unfortunately the big C caught up with the guy that was doing the casting.
Which has rather stymied things....
This is the shortstroke rod for it - pre polishing...

acotrel said:

Isn't engine simulation a matter of going in the reverse direction to pattern recognition ? However how do you know in which direction to send your computer if you haven't previously analysed the patterns ?

Click to expand...

The ole scattergun approach !
Just send it haring off with iterations in all directions.... !

Presumably, you'd hope that whoever programmed up any particular package would have had sufficient knowledge to keep it within bounds, and know the value inherent in changing various parameters. Getting 2" inlet valves into 125cc cylinders might be a fruitless quest say....

Rohan said:

acotrel said:

Isn't engine simulation a matter of going in the reverse direction to pattern recognition ? However how do you know in which direction to send your computer if you haven't previously analysed the patterns ?

Click to expand...

The ole scattergun approach !
Just send it haring off with iterations in all directions.... !

Presumably, you'd hope that whoever programmed up any particular package would have had sufficient knowledge to keep it within bounds, and know the value inherent in changing various parameters. Getting 2" inlet valves into 125cc cylinders might be a fruitless quest say....

Click to expand...

This is a marriage made in heaven! :lol: Carry on.....

Rohan said:

Dances with Shrapnel said:

Good that you dabbled in something with a B33 or Goldstars .

Click to expand...

You are good at being obnoxious with every single word, so we keep it simple for you.

Click to expand...

and

Rohan said:

With all your big words, have YOU actually used an engine sim package to any commercial purpose.

Click to expand...

Simple for me? From your previous post shown above, you have expressed some difficulty with my apparently big words.

Rohan said:

You seem unaware you are repeating near everything I say ?

Click to expand...

I am acutely aware that I am quoting near everything; I do this so that others who read this can see a clear, concise, and stark contrast between your statements and the vast base of knowledge to the contrary as well as your espousing incorrect information and contradictions. You have consistently avoided any effort in trying to understand as this is not what it is about for you, correct? Early on I invited you to take this subject to another new thread but you have not taken that clue. Perhaps you should choose your words wisely; compose a thought or two? I repeatedly see the skiff of your understanding and enlightenment scraping the bottom of the river of your knowledge on the subject matter.

Rohan said:

Current things.

These heads just happen to produce approx the same hp.
Try getting your sim package to prove that...

Click to expand...

We can agree that if you are looking for a very specific "proof", then a dyno is the answer but single comparative dyno pulls will only be a proof of one "state" or "condition" and will not be able to indicate true potential nor differences due to environmental changes. Do you really think an air cooled versus water cooled will have the same behaviors at -20C through plus 40C?

I could model those provided you were able/willing/really interested in providing the necessary data to the necessary level of detail! Furthermore, if you purchase for my use, Virtual 4 Stroke and 4StHead, we can bring it to another level of information and enlightenment. I am aware of two fundamental approaches. If you are interested, PM me and I will explain and put it all in context. I see no point of showcasing the pounding sand in one's ear only to sadly see it dribble out the other.

Rohan said:

One plan was to manxify a featherbed ES2, using data from the head on the left.
Unfortunately the big C caught up with the guy that was doing the casting.
Which has rather stymied things....

Click to expand...

Sorry to hear that. Unfortunately, we are all approaching that time of our lives where friends, alliances, and collaborators are leaving. Sad but hold on to the memories.

Rohan said:

This is the shortstroke rod for it - pre polishing...

Click to expand...

Looks like the metal on top of the big end bearing might be a tad thin.

As for the context of answering questions pertaining to performance of the FullAuto head, I stand behind the approach and I also have no problems with dyno data when reviewed in the limited context.

Having breakfast looking at the contents of box #168, I'm now totally unconcerned about the size of my penis

Hey, Ken; how's it going? I'm impressed with the length of this thread over something as seemingly innocent as kudos for a job well done. Maybe I'll start my own about global warming, or the ozone depletion. Y'know; something not quite so inflammatory...

I agree, it has veered off into a totally irrelevant discussion about Gold Stars and ES2 parts, what has that to do with the fine Fullauto product, which has to be a lot better than anything from the factory!.....there should be a separate section for certain people to argue amongst themselves.....and that con rod looks way too short for anything to do with an ES2 motor, what is the bore and stroke?....just to stir it up a little more!....

I suggest many people think of things in terms of simple cause and effect - two dimensional graphs. However when you are working with a number of variables you are probably working with a number of inter-related graphs, so then it comes down to pattern recognition. The Fullauto head on it's own when fitted to a motor with no other changes probably gives some advantage because of different comp.ratio, slightly better port shapes and valve angles, and the fact that the head being replaced might have had leaky valves. However until you optimise the whole system - head, cam, exhaust system and carburation, you probably won't see full advantage. Even then the power characteristics you achieve must suit the bike's handling and gearing.
As I said, I inspected a new Fullauto head and the casting and machining were superb. I would not know how effective the new port shapes and valve angles might be. However it is probably a better starting point than a standard head.

I do know one thing for certain. If you over-port the inlet tract, you can destroy the torque characteristic of the motor and there is no inexpensive way back. That is why I always read Jim Comstock's stuff about gas flow with interest, mainly to get an idea of how far you can go without ruining the head. I have also set myself a maximum rev limit of 7,000 RPM for my 850 motor, mainly because of the bottom end and the piston weight. Even that is probably too high. Once you have over-ported the head, you are probably committed to using higher revs and a different style of riding.

I noticed Rohan's reference to Gold Star BSAs and ES2s. There are two bikes which might be better used for comparison purposes - the 1959 7R AJS and the 350cc Aermacchi Ala D Oro.

Alan, while it doesn't directly ? have much to do with FA heads, there is a very famous story in bench flow tuning.
Jack Williams, Competitions Manager for in the 1950s for AMC (and Peter Williams' father) did a lot of bench flow work on the postwar AJS 7R - a little 350cc ohc single cylinder racer, one of the manx nortons rivals.

He did a lot of flow work on the head, while injecting a thin trace of coloured ink into the centre of the airflow stream. He found, through much trial and error and port design and shapes, that if the ink could be flowed down the port and swirled into the cylinder such that the ink didn't impinge on the inlet manifold walls or cylinder walls, then the engine bmep could be improved (ie more power), quite notably, over a fair amount of the rev range. This resulted in the 7R going from quite average performance with mid 30's hp into the 40's hp bracket - and thus to being quite competitive, such that the 350cc manx wasn't nearly as successful as its bigger brother.

Later again his 7R's went to a 3 valve design = 2 inlet valves - when it began to become more appreciated that larger inlet valves was the key to more performance. Hitherto it had been thought that getting the exhaust out was the hindrance to more performance, so larger exhaust valves had been commonplace.

This is by-the-by, but one clever fellow....
And can't really be modelled in engine simulation software ...

Folks have wondered over the years why Nortons didn't ever really go to notably larger inlet valves ??

acotrel said:

I noticed Rohan's reference to Gold Star BSAs and ES2s. There are two bikes which might be better used for comparison purposes - the 1959 7R AJS and the 350cc Aermacchi Ala D Oro.

Click to expand...

The Goldie and the B33 were perfect for comparisons. Same family of engines.
50% more power from the goldie, so plenty of areas to play with to see where the increases are coming from.
And you can swap cams and pistons and assorted pieces too, for dyno evaluations.
And, see if the engine sim packages can correctly assign the increases.

Heck, you can even put goldstar cams into an M20 (sidevalve WW2 clunker).
Someone in classic racing got one that would do 100+ mph.
(Which is utterly remarkable !!!, they produced ~10 hp, and you need ~34 to do 100.)

I watched with amusement where someone did the V8 engine evaluation - same V8 as mine.
More brute force with improvements than getting optimal configurations though. !
And you can buy all the bits off the shelf, there is a vast supply chain of this stuff out there.
So he had a list of - a new cam and roller cam followers and roller rockers and stroker crank kit and alloy heads and extractor exhausts and 4 barrel carb. And for icing on the cake, a tunnel ram manifold and supercharger and nitrous kit. Taking a mildly tuned 40 hp/litre engine to 1000+ more ponies - and a hefty $$$ shopping list.
Lot of ski boats ran these engines in those tunes, before the chev 454 became more commonly available.

We diverge.
But that sort of performance market is almost catered for in Commandos.
Except you can't buy all of it in one shop.
The FA head being a new addition to the goodies available.
And its mostly a DIY market.
Mick Hemmings, Norvil etc might disagree.