Digital Degree wheel for TDC?

[GaryR]

I like the idea behind the precision of the Stromski degree wheel but isn't there something else out there that doesn't require a big contraption installed into the flywheel side crank? A fancy printed degree wheel is a bit better than a file mark on the flywheel, but still depends on the interpretation of the case split. Would think there is an apparatus that goes on the front pulley that could do it all digitally?

[cgarr]

What about a dial indicator in the plug hole?

[rsscotty]

In the link below, there is a digital setup that NineMeister uses ....see picture in post #29.

Camshaft Timing... (or how to gain 20 hp...) - Page 2 - Rennlist Discussion Forums

[GaryR]

Originally Posted by rsscotty

In the link below, there is a digital setup that NineMeister uses ....see picture in post #29.

Camshaft Timing... (or how to gain 20 hp...) - Page 2 - Rennlist Discussion Forums

That makes the Stomski setup look simple..

I think a degree wheel combined with a digital dial indicator to find actual TDC would be about as accurate as you can get without all the other contraptions...

[304065]

If you look closely at what ninemeister has set up, there is a linear encoder on the rocker, and a rotary encoder coupled to the crankshaft with a belt. Not sure why he uses a belt vs. fixturing the rotary encoder directly to the crank, I'll have to go on rennlist and ask him. (A belt and pair of sprockets have tolerance too)

Anyway the advantage of that setup is you can crank the engine through 720 degrees and map out all of the valve timing events. See how he says, "I set valve timing events not cam timing" that implies he is looking directly at what is happening at each individual valve vs. timing the #1 intake vs the #4 intake and calling it cool.

The advantage of a digital setup vs. a degree wheel and pointer is not necessarily level of precision (degree wheels are pretty big) it's the ability to measure continuously and have a lot of data points so you can analyze the data more easily. No reason why you couldn't mount a degree wheel to the flywheel, set up a fixed pointer, set up your Mitutoyo indicator on each lobe and meticulously crank the engine over a few times, noting the valve lift every degree. . . would take a LONG time to gather all the data points to a resolution that matters (half a degree). Now I suppose you don't need to measure precisely all the way around the circle. . .

A few questions fall out of this.

1) How do you get the cam timing accurate enough side-to-side to notice the difference on the dyno? There is the early nut style, the later bolt style, and the JB racing adjustable cam sprockets (available in both steel and Aluminum).

I don't know anyone who has used the JBR sprockets, maybe they can chime in here. What I do know is what I've experienced, which I'm sure many, many others have-- you get the cam in the ballpark, then rotate it opposite the direction of the nut so when you tighten the nut down and the cam moves it's right on the spec. You then measure (made easier with a pair of Mitutoyo digital indicators that can be zeroed anywhere) one side against the other to be sure they are the same.

This is not a procedure that lends itself to repeatable accuracy without a great many trials.

2) Once you know all twelve valve events, then what? Ninemeister alludes to what he does in the thread-- you try to split the difference side to side, trying to optimize for the most important determinant of performance.

3) Which is what? What is the most important valve timing event that determines performance? Consensus seems to be that it's the intake valve closing. But of course they are all important.

4) How do you make changes given that the cam lobes don't move? Can you change the valve clearance by a tiny amount that a) doesn't affect the point of having valve clearance in the first place but that 2) starts timing a fraction earlier (less clearance) or later (more clearance). Or do you not get a timing effect within the range of adjustment you have (say, 0.01mm?) I don't know the answer I am just looking at what seemingly is the only adjustable point in the system besides the cam sprocket.

Do you gents remember when Levon Pentacost was here (briefly) and took the time to educate us about how he does combustion analysis? I seem to remember that he said he also measured all the valve timing events and all the ignition timing and tried to optimize for the errors. In the NASCAR/ F1 world where every little bit counts and dynos are commonplace, spending the time for this kind of precise setup would seem to make a big difference.

Which also means that it is within the capabilities of our group-- no secret parts-- just tools and informed experience. But sometimes the experts don't broadcast their expertise on the Internet

This has all got me thinking-- I'm going to see if Ninemeister will join the discussion.

[Ninemeister]

Originally Posted by 304065

If you look closely at what ninemeister has set up, there is a linear encoder on the rocker, and a rotary encoder coupled to the crankshaft with a belt. Not sure why he uses a belt vs. fixturing the rotary encoder directly to the crank, I'll have to go on rennlist and ask him. (A belt and pair of sprockets have tolerance too)

I use a belt because with a continuously reading computer set up you need to have a method of smoothly turning the crankshaft around through the entire cycle and a little bit more (~750 degrees) whilst the system samples the lift. Using the bolt head on the end of the crank is the easiest method, hence a pulley and toothed belt are an easy solution (especially when they are from a 964 power steering drive).

Originally Posted by 304065

Anyway the advantage of that setup is you can crank the engine through 720 degrees and map out all of the valve timing events. See how he says, "I set valve timing events not cam timing" that implies he is looking directly at what is happening at each individual valve vs. timing the #1 intake vs the #4 intake and calling it cool.

The big problem with the 911 cams is the fixed relationship between the intake and exhaust lobes, so how do you know that your cam grinder (be that Porsche or whoever else) has got it perfectly correct? The truth is that no pair of 911 camshafts that I have ever measured are truly accurate, neither to the design specification nor to each other, side to side, so the name of the game is to decide what constitutes the closest timing to perfect that you can realistically achieve. Any experienced race engine builder will know this and will hence work out what he considers to be the optimum timing for the individual cam timing events (IVO, IVC, EVO, EVC) and then dial in the cam to his choice. This is all I am doing.

Originally Posted by 304065

The advantage of a digital setup vs. a degree wheel and pointer is not necessarily level of precision (degree wheels are pretty big) it's the ability to measure continuously and have a lot of data points so you can analyze the data more easily. No reason why you couldn't mount a degree wheel to the flywheel, set up a fixed pointer, set up your Mitutoyo indicator on each lobe and meticulously crank the engine over a few times, noting the valve lift every degree. . . would take a LONG time to gather all the data points to a resolution that matters (half a degree). Now I suppose you don't need to measure precisely all the way around the circle. . .

Well, not strictly true, precision is important as well. A standard degree wheel will be good for 1 degree, a super large degree wheel will get you to 0.5 degree, the rotary encoder gets me to less than 0.1 degree.

In respect of the data collection, yes that is true that you end up with a much higher sample of measurements in a much shorter time. You also end up with close enough data to measure the rate of change of lift (velocity), the rate of change of velocity (acceleration) and the rate of change of acceleration (jerk). Jerk is responsible for beating your valve train to death, so the way to determine a well made cam is to measure the jerk. You can't do this with a dial gauge.

Originally Posted by 304065

A few questions fall out of this.

1) How do you get the cam timing accurate enough side-to-side to notice the difference on the dyno? There is the early nut style, the later bolt style, and the JB racing adjustable cam sprockets (available in both steel and Aluminum).

I don't know anyone who has used the JBR sprockets, maybe they can chime in here. What I do know is what I've experienced, which I'm sure many, many others have-- you get the cam in the ballpark, then rotate it opposite the direction of the nut so when you tighten the nut down and the cam moves it's right on the spec. You then measure (made easier with a pair of Mitutoyo digital indicators that can be zeroed anywhere) one side against the other to be sure they are the same.

This is not a procedure that lends itself to repeatable accuracy without a great many trials.

The answer is that you spend 10,000 hours building engines and testing them on the dyno.
The stock 911 vernier gear is good enough for most applications.

Originally Posted by 304065

2) Once you know all twelve valve events, then what? Ninemeister alludes to what he does in the thread-- you try to split the difference side to side, trying to optimize for the most important determinant of performance.

3) Which is what? What is the most important valve timing event that determines performance? Consensus seems to be that it's the intake valve closing. But of course they are all important.

4) How do you make changes given that the cam lobes don't move? Can you change the valve clearance by a tiny amount that a) doesn't affect the point of having valve clearance in the first place but that 2) starts timing a fraction earlier (less clearance) or later (more clearance). Or do you not get a timing effect within the range of adjustment you have (say, 0.01mm?) I don't know the answer I am just looking at what seemingly is the only adjustable point in the system besides the cam sprocket.

The answer to these questions depends on how good the cams measure since the decision has to be made on the lift data results, but usually with 911 cams you prioritise the intake valve opening point as the factory method implies.
The second option might be to match the intake lift centrelines left to right, so with a stock cam you might set both to 114 degrees.
The third option is to average the intake & exhaust centrelines left to right - for instance on a typical race cam with 108 centres you might end up with 108.5/108.5 left and 107.5/107.5 right.
As I mentioned, the reason why we have to compromise is simply because the cams are never made accurately enough to be perfect at all four points on both sides of the engine.

Originally Posted by 304065

Do you gents remember when Levon Pentacost was here (briefly) and took the time to educate us about how he does combustion analysis? I seem to remember that he said he also measured all the valve timing events and all the ignition timing and tried to optimize for the errors. In the NASCAR/ F1 world where every little bit counts and dynos are commonplace, spending the time for this kind of precise setup would seem to make a big difference.

I'll give you a phrase which sums up these efforts:
"Normalising the combustion events"
If you measure the individual cylinder torque production of any production based engines you will see a large statistical variation, all of which averages out at the mean torque at the crank. To keep the maths simple, a 300lbft 911 engine makes, in theory, 50lbft per cylinder, however in practise it will probably make anything from 45lbft to 55lbft on each combustion event. If the cam timing differs left to right, that variance might be 43-53lbft on the left and 47-57lbft on the right. Hence logically IF we could get the cam timing perfect on both sides we could immediately see a gain on three cylinders of 4lbft. Secondly, IF we could also ensure perfect air filling, fueling and ignition timing of each cylinder we might be able to see the torque spread improved to within 55-57lbft per event. Do this and our 300lbft 911 engine now makes 336lbft. That's a 12% gain without changing the engine design.....

Originally Posted by 304065

Which also means that it is within the capabilities of our group-- no secret parts-- just tools and informed experience. But sometimes the experts don't broadcast their expertise on the Internet

You are absolutely correct, these wins are within the capabilities of the group with simple tools as well as the digital gear that I use. At bare minimum, when setting cam timing get the left side inlet valve lift at TDC to factory perfect (e.g. 1.26mm on a 964) then set the right side to within 0.02mm of the left. It takes hours, but it is worth the effort.

Originally Posted by 304065

This has all got me thinking-- I'm going to see if Ninemeister will join the discussion.

Thank you, I hope your effort was worthwhile?

[GaryR]

I'm glad I got this going, great discussion!

[cgarr]

Now that this is all figured out let's add some carbs and start some balancing! Lol

[afterburn 549]

The fad is to go digital.........It hard to screw up a wheel,
Simple, keep it simple folks .

[cgarr]

Be interesting to check a hot engine when things are not as stable as on the stand.

[911pcars]

The most accurate method of determining TDC is with a mechanical stop on the piston. At TDC, the crank can rotate several degrees while the piston remains stationary.

Caveat: careful with valve interference if using a modified spark plug stop.



Sherwood

[911pcars]

Since crank pulleys are mass-produced, TDC and other advance marks may not be exact.

However, by measuring with a mechanical piston stop, the true TDC on this engine is indicated by the white chalk mark. FWIW, this was also the same "error" when I installed an aftermarket aluminum pulley.