my attempt to measure the stock 2.5 turbo roller cam

[acannell]

I'd like to get a setup going where I can analyze cam geometry and valvetrain response out-of-vehicle. Will probably come in handy for any valvetrain upgrades/troubleshooting/experiments that come along.

I only have stock cams on hand, so first up is the 2.5 turbo stock roller cam.

The center hub of an adjustable cam sprocket is bolted to the cam.

That hub then gets centered on the precision rotary table.

A stop arm is bolted to the head and clamped in a vise.

The rotary table is bolted to the table.

At this time the cam can be turned with 30 arc second precision arbitrarily. Measurements can be made on the valves, lobes, or anything else, with ease.

The hydraulic lifter was made solid using shims, so zero valve lash. (described in another thread)

The setup was highly repeatable. I could turn the cam 360 degrees and come back to the same 0.050" lift spot and the rotary table dial read within 10 arc seconds of the previous full rotation. Also, I verified the setup is not twisting excessively by indicating off the head gasket surface during the highest torque part of the rotation and verifying there was no significant movement.

Looks like the exhaust valve on the stock 2.5 turbo roller cam specs are:

Duration from 0.050" open to peak valve lift = 50 degrees
Peak valve lift = 0.414"
Duration from peak valve lift to 0.050" from closed = 51 degrees

The peak valve lift was easily discernable within a fraction of a degree on the rotary table, using a 0.001" indicator.

So total duration at 0.050" valve lift is 101 degrees, and peakvalve lift is 0.414". I'd say the error in degrees is +/- 1 degree, and for the peak lift, the error is +/- 0.005".

Does this sound right? The FSM has very different numbers but doesn't seem to specify the lift its measured at.

[acannell]

http://i242.photobucket.com/albums/ff197/acannell/20141203_143522_zpsjkwybxxl.jpg (http://s242.photobucket.com/user/acannell/media/20141203_143522_zpsjkwybxxl.jpg.html)


http://i242.photobucket.com/albums/ff197/acannell/20141203_143541_zpsjdbpuq15.jpg (http://s242.photobucket.com/user/acannell/media/20141203_143541_zpsjdbpuq15.jpg.html)
http://i242.photobucket.com/albums/ff197/acannell/20141203_161423_zpsd3rm6qoc.jpg
http://i242.photobucket.com/albums/ff197/acannell/20141203_150709_zpsdmiun5tn.jpg

http://i242.photobucket.com/albums/ff197/acannell/20141203_150034_zpsebpj3tn7.jpg (http://s242.photobucket.com/user/acannell/media/20141203_150034_zpsebpj3tn7.jpg.html)

[acannell]

http://i242.photobucket.com/albums/ff197/acannell/20141203_150059_zpsa8vqyzps.jpg (http://s242.photobucket.com/user/acannell/media/20141203_150059_zpsa8vqyzps.jpg.html)


http://i242.photobucket.com/albums/ff197/acannell/20141203_150102_zpstkquq3k3.jpg (http://s242.photobucket.com/user/acannell/media/20141203_150102_zpstkquq3k3.jpg.html)

[shackwrrr]

Most cam specs are based in crank degrees. That 101 degrees would actually be 202 degrees.

[acannell]

Most cam specs are based in crank degrees. That 101 degrees would actually be 202 degrees.

makes sense..thats pretty close to FSM then if we assume the FSM is at a lower lift than 0.050"

okay I measured LSA and its 114 (cam degrees)

and I measured peak exhaust lobe lift at 0.242"

So that would make the follower ratio .414/.242 = approximately 1.71

http://i242.photobucket.com/albums/ff197/acannell/20141203_163041_zps3lxhdbqc.jpg

[zin]

You'll want to get the .006" durations too, that's pretty much industry standard and can be used to get a feel for how "fast acting" a cam is relative to others.

For example, a lot of stock Chrysler V-8 cams look like they are pretty high duration when one looks at the .006" numbers, but taken with the .050" numbers (and compared to another cam with similar "advertised" duration), one finds that the stock cam is VERY mild and is actually using most of that "extra" duration at .006" lift to slowly lift the valve and gently place it back on the seat.

Mike

[acannell]

okay when I do both intake and exhaust ill grab the .006" duration data

[acannell]

hmm

exhaust .006" valve lift duration appears to be 120 cam degrees (240 crank)

that would put it more in line with an older T1 roller cam, not the 2.5 T1 cam, correct?

is there some way I can positively identify the cam? it has a number printed on the end of it but it doesnt seem to match up to any part number: "70100416".

I'm actually not sure where I got this cam. Its possible it came with the head on the 90 daytona but its also possible that head came from who knows where before I got the car.

Or is it that the factory spec is at 0.006" tappet lift and not valve lift? If we say its at tappet lift then the valve lift would be something greater than that, and the duration shorter. Lets say the follower ratio goes down to 1.5, then the valve lift becomes 0.009" at 0.006" tappet lift. And if I measure that I get a 117 cam duration (234 crank).

[Vigo]

That's a nice setup.

Just goes to show what can be accomplished when someone cares enough to try. This is a lot more than i saw from all the people pissing and moaning in the other thread.

I assume you would be willing to measure other cams if people were to lend them to you? I dont have any aftermarket cams but this could create a go-to database for real cam numbers.

[acannell]

That's a nice setup.

Just goes to show what can be accomplished when someone cares enough to try. This is a lot more than i saw from all the people pissing and moaning in the other thread.

I assume you would be willing to measure other cams if people were to lend them to you? I dont have any aftermarket cams but this could create a go-to database for real cam numbers.

Thanks, just to be clear, theres nothing expensive or difficult about that setup, or at least, that needs to be.

The rotary table in the pictures is way overkill. Thats a 6" rotary table that also tilts and has specified 10 arc second precision. Thats 1/360th of a degree lol. I'd say for measuring cams more like 1/2 a degree is probably good enough, as long as its repeatable to say 1/2 of that. So any cheapo chinese rotary table would probably do.

I did buy a chinese 3" rotary table so I can make a fixture to do this sort of thing in-vehicle. I'm not sure if it will be too difficult to turn the little handle on that or not, we shall see. But it was only about $60 shipped.

You want a worm gear so that the spring force cant turn the table backwards. Pretty much every rotary table has worm gears, and they have precise degree indications, so its a perfect match. Low backlash also helps here and so a rotary table and worm gear drive make sense since they usually have a fraction of a degree of backlash and/or its adjustable to be very small. The table I was using has 1/120th degree backlash, I'm not sure about the 3" one in the mail.

You need to make sure the cam is concentric with the rotary table. You can dial it in using your dial indicator and tapping around till its true. A way to bolt your cam sprocket (or whatever you will use) to the table also needs to be whipped up..but theres a million ways to do that so no big deal.

A 10$ harbor freight dial indicator and a mag base for whatever they're selling them for should do the trick. Scrounge around and you can get the whole setup for super cheap at garage sales or what have you..all very common ingredients. Or buy it all new and I'd say you could get the whole she bang running for $100 (rotary table, dial indicator, mag base, misc hardware). Dont forget to make your lifter solid and set to zero lash as I described in another thread.

For cam lobe measurements, its also important to make sure the dial indicator is square to the cam axis in 2 dimensions and intersects the cam axis. You can approximate this by mounting the indicator on a vise as a makeshift line axis, approx. tangent to the cam base circle, then approx the dial indicator to be square, and move the vice jaw until you find the peak.

I dont think all the "eyeball" errors add up to much, but there are more precise methods to do all of this (as far as strategy not $$$) and I'll be getting into those when I make my fixture.

If you are just measuring valve lift you can nail it good enough just eyeballing the indicator parallel to the valve axis and perp to the valve face. I say good enough eyeballing because crane and lunati both seem to accept the idea of eyeballing the dial indicator to align it for cam degreeing it, and this is the same type of measurement. A 7 degree tilt in the indicator only adds up to .001" error at typical distance.

I will probably end up measuring every cam I get my hands on. So are Shadow and 4-Lbodies. Combined that will make for multiple datasets for certain popular cams which should really nail the lid shut on not knowing the specs.

I will need to be getting a high performance cam at some point so whatever it is will get measured.

[zin]

Most .006" numbers are coming from push-rod engines, so it's a tappet lift number. In our case, the equivalent would be cam lift, but since we aren't "cam over bucket" style, getting a "cam lift" number would be a little challenging, as it would need to be measured at the point of contact on the "pad" of the "rocker arm" (or the center of the roller for later cams).

If we just measured valve lift, and were consistent in our methods and practices, a useful comparison and inference regarding the cam's potential could be had, the point is to see how "fast acting" one cam is compared to another, and valve lift numbers (with 0 lash) will give that data.

I wonder if this would be a good time to discuss valve train geometry... it's something else virtually no one is addressing as it often would require changing the length of the valve stem/adding lash caps.

In a nut shell, cam changes, particularly lift and especially with regrinds (smaller base circle), the static angle the follower is at with 0 lift will affect the "transmitted lift" to the valve. Proper geometry will have the follower's tip sweep from the far side of the valve stem at 0 lift to the near side at full lift. Well, not necessarily all the way from one side to the other, but it should be symmetrical and sweep from one side to the other.

Set up a cylinder head with some "dummy springs" and rotate the cam while watching the valve tip from the side, with stock parts, it should do pretty much what I've described, but with radically higher lift/duration cam, the angles can get "off" and the follower doesn't push the valve all the way down.

Mike

[acannell]

Most .006" numbers are coming from push-rod engines, so it's a tappet lift number. In our case, the equivalent would be cam lift, but since we aren't "cam over bucket" style, getting a "cam lift" number would be a little challenging, as it would need to be measured at the point of contact on the "pad" of the "rocker arm" (or the center of the roller for later cams).

If we just measured valve lift, and were consistent in our methods and practices, a useful comparison and inference regarding the cam's potential could be had, the point is to see how "fast acting" one cam is compared to another, and valve lift numbers (with 0 lash) will give that data.

I wonder if this would be a good time to discuss valve train geometry... it's something else virtually no one is addressing as it often would require changing the length of the valve stem/adding lash caps.

In a nut shell, cam changes, particularly lift and especially with regrinds (smaller base circle), the static angle the follower is at with 0 lift will affect the "transmitted lift" to the valve. Proper geometry will have the follower's tip sweep from the far side of the valve stem at 0 lift to the near side at full lift. Well, not necessarily all the way from one side to the other, but it should be symmetrical and sweep from one side to the other.

Set up a cylinder head with some "dummy springs" and rotate the cam while watching the valve tip from the side, with stock parts, it should do pretty much what I've described, but with radically higher lift/duration cam, the angles can get "off" and the follower doesn't push the valve all the way down.

Mike

I'm planning on modelling our stock valvetrain components in SW for the purposes of just these sort of discussions and ideas, but once the real geometry in is SW, we can start designing parts around the real world and take all the guesswork away.

The question though is what exactly is the end goal? If not using regrinds and only using new castings ("billet") solves the problem at the cost of $125 extra then I think thats the end of that. But if there is something to be seriously gained by going to the trouble and expense of redesigning followers or valve lengths, etc.. then I'm all for it or at least enhancing the discussion with parametric cad goodies.

So what is there to be gained if we assume new castings will be used instead of regrinds?

[zin]

Regrinds are just an extreme example, the geometry will change whenever the cam changes, it becomes a bigger and bigger inefficiency as the lift and duration increases.

If you'd like, I've got a "purple shaft" cam that's pretty radical (made for NA applications), it would show the change fairly clearly. (And I'd love to know it's actual specs).

Mike

[acannell]

Regrinds are just an extreme example, the geometry will change whenever the cam changes, it becomes a bigger and bigger inefficiency as the lift and duration increases.

If you'd like, I've got a "purple shaft" cam that's pretty radical (made for NA applications), it would show the change fairly clearly. (And I'd love to know it's actual specs).

Mike

okay then when I get the models drawn up it shall be interesting to see whats what

let me get back to you on measuring once i've made my fixture up..