New smoother ramps for PW3 type JS2 cams

Just got this data from my cam grinder. Trying to make some smoother opening and closing ramps for the PW3 type cam (JS2).

The earlier design is blue, new design is red.



Acceleration graph below.



This is just the design stage. I don't know how much this will reduce valve bounce for the PW3 type but its a step in the right direction. This isn't a dramatic change. More ramp could be added, but at some point it would increase the duration and make the cam a little more peaky.

Lore has it the PW3 is the jackknife of cams, does good across power band, but wonder about your peaky caution of the new profile, especially for those trying to get the most out of it, ie: staying on a boil by close ratio to hit peak as much as can. A data point that is beyond my need to know but toss out to you, Engine Masters recent contest winner was loosing low end power that cuts points even though had best top power, so they backed off valve lash some to prevent the detonation at 2500-3000 hoping it did not lower top end much, it worked for them with tolerated low rpm detonation and still made almost as much top end 6500 rpm power. i wonder if this has something to do with the large lash listed for the biggest Norton cams.

Steve - Most riders use their bikes in a variety of situations. If the cam spot is harsh and the power band is narrow, you might well be able to use a close ratio box to keep the motor on the boil. However if you are competing, you must eventually lose out to the guys who have smooth power delivery when the road is twisty. It might be lovely to fantasise about having loads of top end power, but 'torque wins races'. However I believe you can engineer to get good midrange power almost regardless of the cam duration, if you are prepared to sacrifice the top end.

The change in the ramps does not make the cam peakier or change the power band. It does not even change the duration when measured at .040" lift. The ramp would have to start a lot higher - say at about .1" or more and change the slope in a more dramatic and significant way. You would have to add a few more degrees in duration measured at .040" lift or higher. In fact - if you look carefully you will see that the ramp dips down more smoothly and there is actually less lift and duration near closing. But any addition in low end torque would be too small to measure.

Ok thanx for entertaining me on nit picking details. Comnoz spintron and back up reports with PW3 history show it did great as designed for the rpm limits of the day, so why else refine PW3 further unless for serious WOT pressing beyond normal red line which implies rest of engine as expensive to tolerant above 8grand. I did not comprehend the peaky remark with such tiny changes. I do not know the market or desire for this further refinement.

Alan in my case I use both smooth and jerky styles per sane intentions or surprise crisis enough now can switch between them on Peel on purpose but MUST be totally on purpose exactly as expected per power band reserves response for tire temp matching that section of surface just then or SPLAT. One my favorite semi scary tactic is to kind of crash ahead of time by exact amount of throttle snap to spin tire a certain amount that re aligns bike w/o risk of front wash out then hook up next instant-length of distance from edge hazards that jerks bike up right better at same time it allowing a harder hook up on still mashed down wider patch of next torque hit out of there. If engine does not chrp,squeech rear out as expected then will fly off path at a tangent like cutting string to a ball swung around your head. This has nil relation to drifting or flat track where they are still counter steering. My dream machine is one that fouls frame about 70* then have torque enough to wheelie front against breath taking centrifugal force up sideways for rather sharper acceleration into decreasing radii. Time will tell if smooth hp constant traction or jerky digital torque traction is more efficient. I know which is more fun.

Hi Jim

Where did you get the PW 3 numbers ? I ask because I have an Excel file of the measurements of my PW 3 and they look a little different in the quietening curve and ramp to yours. I do see the two steps visible in your plot but there is more of a quietening curve visible. I guess we dont know exactly where the tappet clearance would be taken up but suspect the first step at -200 on your curve would not be seen by the valve.

I can send you my excel file if you are interested.

I have been following your discussions with Snotzo becaue Im still working toward a 500 race engine but until I leave Eastern Europe its still just a plan.

John

johnm said:

Hi Jim

Where did you get the PW 3 numbers ? I ask because I have an Excel file of the measurements of my PW 3 and they look a little different in the quietening curve and ramp to yours. I do see the two steps visible in your plot but there is more of a quietening curve visible. I guess we dont know exactly where the tappet clearance would be taken up but suspect the first step at -200 on your curve would not be seen by the valve.

I can send you my excel file if you are interested.

I have been following your discussions with Snotzo becaue Im still working toward a 500 race engine but until I leave Eastern Europe its still just a plan.

John

Click to expand...

Every cam grinder does things a little different. There are variations in each copy and I don't know which is original. Which version do you think has the most quieting ramp? The measurements I use are just the cam lobe measurements - not the valve.

Please send the data you have to:
Jim@jsmotorsport.com
and we'll make some comparisons.

Hi Jim

I sent you the file. Measured off the push rod with the head off the bike. I measured all the lobes and unlike many cams they are identical on both cylinders. Many of the 4 S cams about are not the same on both sides. See Dynodaves site. This is a PW 3 cam on a Dommie core bought from Mike Hemmings about 10 years ago.

John

On behalf of Peter Williams:

Cam design
I am the designer of the PW3 camshaft for the Norton twin engine. I have read some of the contributions to the forum and hope I can add a little understanding of the workings of the valve train (valve, rocker, push-rod, follower, cam) but I should refer you to “Cam Design Handbook” by H. Rothbart for a much deeper understanding (.pdf accessible on line).
The foundation of cam design is to understand of Newton’s Laws of Motion and in particular the second one which is, in fact, intuitive. Force = Mass x Acceleration; the larger the mass or the acceleration or need, the more force you must give.
It is also a waste of time to only look at a lift curve of the valve (or cam follower) motion; you have to be able to look at the velocity curve and, most important of all, the acceleration curve. If you have the kit to measure lift for each degree of cam rotation, as seems to be the case from the graph on the forum site, there is a simple way to create the velocity curve by subtracting each lift value from the following (or preceding one) to get the velocity per degree. Do it again with the velocity values to get the acceleration per degree diagram. (This ‘difference’ method is practical mathematical differentiation.
0.026105 0.025659 0.00089
0.051764 0.024773 0.00131
0.076537 0.023463 0.00171
0.1 0.021752 0.00208
0.121753 0.019669 0.00242
0.141422 0.017249 0.00271
0.158671 0.014534 0.00296
0.173205 0.011571 0.00316
0.184776 0.008409 0.00331
0.193185 0.005104 0.00339
0.198289 0.001711
0.2

There are a number of parameters within which the designer of a cam has to work.
1/. For reasons of cost the flat faced cam follower is retained for the Norton (as opposed to a radius face)
2/. Generally speaking, the ramp is simply a “take-up” of clearance between the cam and the follower which is necessary to allow for the thermal distortion of the metals of the engine and inevitable imprecisions of manufacture. The ramp is important to mechanical noise but relatively unimportant to performance.
3/. Picture the cam and follower as the cam is about to start lifting the follower. There is a contact line of the curved surface of the cam on the flat surface of the follower. As the cam rotates the line wipes across the flat face of the follower as the cam lifts and accelerates it. The distance of the contact line from its original position on the base circle is called the eccentricity and is obviouslyrestricted by the width of the follower, and more explicably, by the cylindrical radius of the Norton follower, but maximum velocity of the follower and valve, etc is at the maximum eccentricity.
4/. The first proper phase of valve movement is the acceleration imparted by the cam. The timing (i.e. the crank angle when lift begins) is crucial to the filling of the cylinder with new air assisted by the gas dynamics during the closing of the exhaust (valve overlap) and so is the cam angle during which acceleration occurs.
5/. The Norton twin has push-rods and rockers. These not only have mass which adds to the forces and stresses involved in accelerating them but the stresses cause strain. When I calculated the force involved at top engine rpm and then statically applied the force statically to the valve train, I measured about 0.010 inch deflection. This means that the valve train is like a spring. So when the force is removed the valve train returns to its relaxed original shape. This does not mean the valve is bouncing and out of control, but it could be if the acceleration is too great (or to use the seemingly in vogue word ‘aggressive’).
This is why I kept the acceleration as low as possible and much lower than on the Norton ‘S’ series which I designed in John Player Norton racing days. The result is the valve staying in control at much higher engine speed than in those days. The highest I ever revved the JPN in the ‘70s was 7500rpm - in anger. The PW3 allows 8000rpm before ‘going over the top’ of power when the valves start to bounce (watched on the dynamometer with a stroboscope). I believe it to be pointless to run the Norton engine any faster than 8000rpm when I has run out of power and breakage is risked. There is more controllable torque and power available in a higher gear. That’s what I did, anyway…
6/. After maximum velocity is reached the valve train is under as much control of the valve spring as by the cam. The spring has to decelerate and stop the valve at full lift before pulling the rest of the valve train and the valve back towards the seat. However, the cam profile during this phase on the PW3 is dictated by a mathematical sin curve suited to the springs available twenty-five years ago. At maximum velocity the cam takes over control to lower the valve back on its seat. The PW3 is unsophisticated in this respect having symmetrical opening and closing flanks.
The duration of crank angle over which the valve is open is also dictated by the cam and the duration given by the PW3 is quite short in order for the valve to close as early as possible. This gives good compression of the new charge of air and fuel which results in the very good engine torque which was not as good in my racing days.
A properly designed, modern cam profile is designed mathematically and made by CAD/CAM without the use of ‘masters’ which went out of use thirty years ago. The cam profiles on the PW3 camshaft were designed mathematically and the cams are chilled cast iron as used by the automotive industry where ever translating flat faced followers are used. It seem to work quite well and I think can only be bettered by a new cam profile paired with radius surface followers, which will cost much more.

Peter Williams
In the original you find three graphs I could not copy into this, see http://www.andover-norton.co.uk/PW3%20Cam%20Design.pdf

Joe Seifert/Andover Norton

Peter/Joe,

Thanks for this insight into the design of the PW3, which I am using in my 920 with good results. The obvious question being, do you fancy designing a modern cam, radiused follower, & spring package. It seems to me there is a reasonable market for such a set up.

Martyn.

John
Thanks for the data. Everything is pretty close but both graphs I'm working with at the top of this thread show slightly smoother ramps than the PW3 you measured. There could be some measuring variance between your measurements and mine and this is hard to avoid. The closing ramps shown in the graph at the top of this thread appear to be a little smoother and more so on the latest cam with the modern ramps. The closing ramps are important because that's where the valve bounce occurs.

Matchless
The graph at the beginning of this thread shows my version of the PW3 cam that is already re-configured for radiused BSA lifters. I'm wanting to make the closing ramps a little more modern and smoother with the latest version.

Joe/Peter
There's some discrepancy in the RPM/valve bounce levels that different people are reporting. I've been prompted by a customer to give a little safety margin.

Peter William's explanation re the rationale behind his PW3 cam should suffice to satisfy all who have an interest in understanding the why's and wherefore's of this outstanding design.

The book mentioned, 'Cam Design Handbook' by H. Rothbarts, carries very comprehensive explanations, with the appropriate mathematical equasions, of many types of cam forms. Unfortunately only a small number are applicable to the valve train mechanisms of i.c engines. A much more useful book is Don Hubbard's 'Camshaft Reference Handbook', unfortunately very hard to find at present, but worth it's weight in gold for the practical common sense it contains. More recent articles on various aspects of cam design are to be found at http://www.profblairandassociates.com, and are available as free downloads.

A problem I am continually running up against is the inability of the camshaft grinder to grind ramps that are true to the original design. Why this should be so I do not know, but it is the subject of continuing investigation on my part. In this respect I do not doubt that the PW3 cam is any different. I do not have Peter Williams actual design details, but from careful measurement of a newly ground profile, anomalies exist which suggest that the ramp is not a faithful replication of his original design.

That these anomalies exist is indisputable. Whether their presence is detrimental in any way might be difficult to determine, but common sense would suggest that it would be better all round if they could be avoided.

ZFD said:

On behalf of Peter Williams:

Joe Seifert/Andover Norton

Click to expand...

Thank you both very much.

The reason I think I might need smoother ramps is because higher RPM levels are being reached with the JS lightweight pistons and longer rods etc. I recently had a report from Bernd (see his "My new Commando race bike" thread) who put his 850 racer on the dyno and the guy who was running the dyno revved it up to 8400 RPM till the power dropped. I didn't think an 850 could rev that high. There was no physical evidence of valve bounce with the JS2 cam (PW3 type but for BSA lifters) and conical beehive springs. But recent spintron tests suggest that valve bounce would be happening. So either the valve bounce is of no concern or I can make the ramp changes previously described without changing the character of the power band.

Snotzo
I shot the lift/duration per degree specs of the new cam to you in a PM.