I went in for a 5k major, it cost XXXX more and here's what happened"

When I find myself in times of trouble
Mother Mary comes to me
Speaking words of wisdom
Let it be

Let it be, let it be, let it be, let it be
Whisper words of wisdom
Let it be

 

So, Mitch, all fooling around aside, you have not commented on how much the timing can change if you just do a belt swap. We could have a discussion on this if you want. Or, now much does timing change due to belt stretch as the belt ages, if it does indeed stretch.
Image Unavailable, Please Login

 

That's damn interesting.

 

Got nothing to do this morning so WTF. Referring the figure below, the belt is tensioned by the tensioner moving in and out. So, and slack on the tension side of the belt,between points a and b, will automatically be taken up. The question then becomes what happens if the belt length between a and b changes, for whatever reason, belt swap, belt stretch, baring any mechanical issues. With crank shaft locked for reference, if the length of the belt increases the cam pulleys will rotate counter clockwise and the cams will be retarded. If the belt length decreases the pulleys rotate clockwise and the cams are advanced. So the question how much do the cams rotate? The manufacture states that the tolerance in belt length is +/- 0.75 mm. Typical quality control would require that the tolerance be 3 sigma, or 3 standard deviations from the design belt length of a large sample. 99.7% of all belts would be below the 3 sigma limit. The remaining 0.3% would be rejects. 95% would fall into the group less the 2 sigma variation, or +/- 0.5 mm, and 68% would be better than 1 sigma variation of +/- 0.25 mm.

Additionally, the segment of the belt between points a and b is approximately 1/3 of the total length. So, evenly distributed, the tolerance on the a-b segment would be at worst +/- 0.25mm and it could be as small as +/- 0.08333mm or less 68% of the time, assuming the 3 sigma rule applies.

All that aside, going with the worst case, of +/- 0.25 mm variation in length from a-b for two different belts, the maximum change in belt length between an old belt, when it was new, and a new belt would be +/- 0.5mm.

To find the impact on cam timing all that is needed to determine what fraction of the cam pulley circumference the change in belt length is and relate it to the angle of rotation. That's simple,
Angle = delta length /(Pi x PD) x 360, where PD is the pitch diameter of the pulley. I don't recall the exact pitch diameter of the 355 pulley but assuming it's around 100 mm yields a change in cam timing of +/- 0.57 degrees worst case. Based on the QC discussion above it would be less in most cases.

If the belt were to stretch by 0.25" (6.35 mm) which seems unlikely, with 1/3 of that (2.117mm) over the a-b segment cam timing would be retarded by 2.43 degrees. But FWIW, I ran the belts on my 308 for 25 years. When I compared the length of the old belts to what would be the design length, one belt was 1 mm over and one was 0.8 mm over. Assuming that both these belts started life on the short side of the design length at the tolerance limit, one belt stretched a maximum of 1.55 mm and the other 1.75 mm, or 0.517 to 0.58mm over the tension segment of the belt, about the same you get in the worst case between belt swaps.

So, if you swap belts and mark the belts and pulleys, carefully transfer the marks to the new belts, and install them correctly, you are looking at a +/- 0.6 degree change in cam timing from the last time new belts were installed, worst case. In fact, if you look at the picture you can see that whoever did the work also marked the location of the positing pin on the pulleys and cam. By doing this you could remove the cams, take off the pulleys, do what ever work is required, reassemble the cam/pulleys, put them back in the heads and put new belts on and still retain correct timing. So the reality is that unless something major is done to the engine, or you suspect a belt jumped a tooth there is no need to time the cams once they are correctly timed. If you know that the cams were timed to with in +/- 0.5 degrees at some point, the worst they can ever be if the position pins are never changed is on the order of +/- 1 degree.

Now, one last comment. I looked of the thermal coefficient of expansion for aluminum. Assuming the distance between cam pulley and cam drive pulley is about a foot, and that the engine block heats up about 100*F when running, thermal expansion will change the distance between those pulleys by about 0.4mm which means the tensioner has to retract a little to allow the belt segment a-b to increase in length by a similar amount. But that means the cams will advance by the angular amount consistent with a 0.4m increase in the belt length, about 0.5 degrees.

So, in summary, the only way cam timing can change, other than by jumping a tooth, is if the length of the belt changes or if the distance between the drive and cam pulleys changes. The former is limited by belt tolerances/belt stretch and the latter by thermal expansion, provided the engine is mechanically sound.

Open to any and all discussion.

[edit] Oh, and as James pointed out, I'd bet that the errors in finding TDC are easily +/-0.5 degrees. So just depending on what tech does the work come make as big or bigger difference than swapping belts.

Image Unavailable, Please Login

 

Academic BS.

"A man's got to know his limitations."

Try designing a test rig rotating at 600 RPM with 2' arm with a clear Plexiglas block on the end of it. The Plexiglas has a serpentine passage through it with cross section 1" x 1" and ribbed walls top and bottom. Fluid is to be pumped through that passage, while the rig is rotating. The fluid temperature must be held constant to within +/- 2*C. A laser must be focused at points in space where the focal point will be inside the channel at various angular positions across the channel at the instant the Plexiglas block flies by, to measure the local fluid velocity. The angular position must be accurate to better that 0.02 degrees. And did I mention that the laser must go straight through the Plexiglas with no diffraction at the fluid/Plexiglas interface?

Welcome to the real world. We did it back in '94.

 

Just tell me where I'm wrong. Once you have pined the cam pulleys to the cams how does timing change other than by differences in belt length if there aren't mechanical issues? The only other means is by altering the distance between the cam and drive pulley centers. How does that happen other than thermally if there are no mechanical issues? We are talking about dimensional changes on the order of mms. The angular relationship between distance (pulley circumference) and degrees is straight forward, circumference/360 =distance per degree rotation. As Dave Rocks said, it's about tolerances, where they apply and their magnitude, and unless something changed mechanically, which would be indicative of a significant problem, the only thing that changes with a belt swap is the deviation of the belt length from the design length, within the belt tolerances. You could argue that cam wear effects timing. Sure, but if the cam wear was enough to effect timing, you couldn't properly time the cam anyway.

So post something constructive that we can discuss. What's incorrect? What has been over looked? I told you, I'm open to a productive exchange. I can except a difference of opinion. For example, if the old belt was time so that the cams were advance 1/2 a degree and the new belt length is such that it would further advance the timing by 1/2 degree, resulting is a full 1 degree of advance, I can accept and respect that you may feel 1 degree off is not good enough. I would disagree, but accept that as a difference of opinion.

 

Exactly, Grant. You can look at it as a change in the relative difference between cams or with reference to the drive pulley. On the 5-8 bank the change in timing of the exhaust cam will be slightly greater that that for the intake. On the 1-4 bank the change for the intake will be slightly more than the exhaust. This is simply because the path from drive pulley to cam pulley in question, on the tension side, is greater for one of the cams.

That would be about the limit. Maybe 0.6. Depends on the actually cam pulley diameter.

But there is another issue and Dave Rocks deserves credit for planting this seed when he talked about tolerances adding up. As was discussed, TDC is not a function of rod length. But it can be effected by bearing clearances. Here is what I mean. At TDC the piston position is insensitive to crank angle. For example, if we assume zero clearances, the piston in a 355 would move down from the TDC position only 0.0075mm if the crank were to move off TDC by 1 degree. In fact, it would move down just a little over 0.1 mm if the crank moved off TDC by 4 degrees. Next consider what the min and max tolerances are for the crank journal bearing clearances. Someone can correct me if I'm in error, but from the numbers listed in the work shop manual the clearance for the connecting rod to crank journal bearing is from 0.05 to 0.1 mm (0.002 to 0.004 in). (Example: Minimum connecting rod bore is 47.129mm, Max bearing thickness is 1.721mm, max crank journal diameter is 43.637. Clearance = (47.129 - 2 x 1.721) - 43.637 = 0.05mm.)

If you are using the Ferrari method to find TDC you place a dial gage through the spark plug hole, touching the piston and turn the crank, raising the piston in the cylinder until the dial gage stops moving. Of course, it won't stop moving until you get to TDC and you want know it stopped moving until you go past TDC and the gage starts to move again as the piston starts moving down. But here is the rub. Due to friction in the cylinder, as the piston moves up the crank journal will be pushed as tight as possible against the top of the rod bearing. Then as you move past TDC, friction will hold the piston in the cylinder while the crank rotates until the crank journal is pushing down against the bottom of the rod bearing. Due to the insensitivity of the piston movement to crank angle at TDC, this method of finding TDC can easily lead to an uncertainty of a couple of degrees due to this bearing slop. A dead spot, as I think Dave called it. There will a range of crank position where the dial gage doesn't move. A skilled tech can minimize this by going back and forth and averaging things out but getting exactly on TDC using this method will always result in some uncertainty and judgment of the tech. So it wouldn't surprise me to find the if the timing were checked on a given engine by a number of tecks there would be a range of that could easily span +/- 1 degree or more. Using the piston stop method can reduce this uncertainty but even then there is a limit which is a function of the divisions of the degree wheel used. And certainly (are you reading this Mitch, FBB?) experience plays a role.