F18Rep

Interesting discussion. We might learn a little by comparing the new Gates belt to one of the originals with some simple testing. If we put them onto a jig similar to what is seen in this photo we could confirm there were differences. Perhaps a difference factor could help in finding a recommended tension value for the new belt.

PS... This was from an some early experiments with measuring tension for a 944 tool...Bruce

ptuomov

What happens between the crankshaft sprocket and the driver-side camshaft sprocket matters for absolute timing. But only what happens between the two camshaft sprockets matters for the relative timing of driver-side and passenger-side camshafts.

I think we can resolve the question about the relevance of thermal expansion with a little bit of math. Assume that the thermal expansion coefficient of the block material is 13 e-6 / F and that of the belt is approximately zero. The distance between the sprockets is about 24 inches. The block heats in use by about 150F. Putting it all together means that thermal expansion pulls the two cam sprockets apart by 0.05", give or take.

Is 0.05" significant? I think it is. The circumference of the sprocket is 2 pi r. The impact of 0.05" expansion in degrees is therefore (360 * 0.05") / (2 pi r). I haven't measured the sprocket ever, but suppose that the radius is 2.5 inches. If so, the thermal expansion advances the passenger-side cam relative to the driver-side cam by approximately 1 degree. In order to have the cams of a hot engine to have equal timing across banks, one would therefore have to retard the passenger-side cam by one degree relative to the driver side cam when timing a cold engine.

I took SWAGs at all the numbers, so this is not accurate. The result is in the ball park, though. Also, coincidentally, the result is very close to by how much the factory install instructions retard the passenger-side cam.

928Myles

Instead of making up a jig like in the photo above, would it not be possible to do the following:
Install and correctly tension a std timing belt . If I am correct we are looking for longitudinal tension to be the same between both types of belt so if we hang a 10kg weight off the std belt and measure delection we have a reference for this. Doing this where the belt tension is measured is probably the best place for this to happen.
If we then install the racing belt and tension it to the same deflection using the weight, could we not then measure the belt tension with the kemph or porsche tool in the normal way to get a comparison?

Just a thought,
Myles

jon928se

There are two issues. (I mentioned bending of the belt in an earlier post but I think it is a negligible issue)
Tension

Tension is assessed by deflectiing the belt a known amount (thus inducing more tension) and measuring the force required to achieve that initial plus additional tension.

A belt that is less stretchy due to a larger cross sectional area and or having a higher modulus of elasticity will require more force to deflect it the fixed distance.

It simplest to put some arbitary numbers to it for explanation purposes I'll use T as my tension unit.

Let's assume a stock belt which we've already set at the correct tension sits at 90T, and the act of deflecting it the fixed amount the gauge uses adds 10T and the gauge indicates the desired result .

A)
Consider a belt that has twice the modulus. deflecting it the fixed amount would add an additional 20T. Setting the tension using the gauge would mean that once the gauge was removed the belt would sit at 80T - a bit on the slack side.

B)
Then consider a belt that is 10 times the modulus. deflecting it the fixed amount would add 100T. Setting the tension using the gauge would mean that once the gauge was removed the belt would have zero tension (0 T). Not good.

There would be no slack in the belt at TDC when stationary. However as soon as the engine was running the additonal tension on driven side of the belt caused by the resistance of the camshafts pushing valves down would stretch that part of the belt thus increasing it's overall length. This could result in sufficient slack in the belt on the undriven side which could allow the belt to jump teeth on the cam pulley.

Compare that with the stock belt. the driven side would stretch more, but there should still be some tension (albeit reduced) in the undriven side as the belt was stretched to start with.

The problem we have is that we don't know how the GRB compares to stock. We don't know the relationship between the static tension and the additional tension when measuring the stock belt. Nor do we know the increase in tension on the driven side of the belt when the engine is running.

Assuming my initial assumption of the gauge adding 10% to the static tension when measuring is in the ballpark then my gut reaction would be that a belt with twice the modulus (half as stretchy) is not going to cause issues.

Who knows?

borland

That's really not necessary. All you need to do, is measure the additional contribution of the belt stiffness added toward the tool measuring the belt tension.

Both tools gage the belt tension, by measuring two components of belt tension:

1. Belt tension before tool applied, plus
2. additional belt tension due to tool being applied as the measurement is taken.

The tool is calibrated, so that when the 1 and 2 are measured as correct, 1 will be correct after the tool is removed. Both tools (Kempf and 9201) operate by adding additional tension, by distorting the belt from its rest position.

The force required by the tool, to add additional belt tension, must overcome any stiffness (counter force) in the timing belt. This counter force can be measured with the belt under zero tension.

With zero tension on the belt, it takes very little effort to distort the belt from a static position. With the Kempf tool, it is easy to see how much twist is needed (less than 90 degrees of twist). The large spring on the Kempf tool provides a clue as to how little counter force is presented by the belt's stiffness.

The more sensitive 9201 tool deflects the belt much less, so the counter force due to belt stiffness will be much less than with the Kempf tool.

mark kibort

interesting. I thought it was due to the expansion distances or angle because the difference of position comared to the cam pulleys, oil pump etc made the distance of belt different .

then, what is the reason for the difference in valve timing from side to side. some cars have .2mm vs .4mm differences?? (2mm vs 1.6mm or 1.8mm for some based on the distance the cam lifter moves with 20degrees crank rotation after TDC)

PorKen

The distance to the crank isn't as the crow flies. The 5-8 side is twice as far as the 1-4 side along the belt, and has more block expansion affecting it.

The block expands, but the belt cannot get significantly longer which advances the timing. The WSM has the 5-8 side set approx. 2° retard (opens later) to compensate.

Belt modulus does change the timing up to 2° at different points of rotation (EG. between #1 and #6 TDC.) depending on how many valves are being pulled open. I've seen this with Conti and Porsche belts. I'm curious to see if there is less of a difference with the new belt.


Different cam profiles. IE. more or less lift at the same point.

ptuomov

How did you measure this? Have you pinpointed the exact source of the difference? Is it the camshaft lobes being different between cylinders, camshaft twisting, worn cam chain sprockets, stretching cam chain, worn timing belt sprockets, or stretching belt?

If belt stretching would be the problem that the factory was addressing with the camshaft install instructions, they'd have advanced, not retarded, the passenger side camshafts. From that, I've concluded that what they are trying to address with the cam install instructions is not belt stretching. The most logical explanation is thermal expansion of the block and heads.

GregBBRD

I'm hoping you are just confused.

M. Requin

All of this is sort of interesting, but there is a better wasy (I think I vaguly hinted at this in an earlier post). From the Gates catalog:

Belts, like strings, vibrate at a particular natural frequency based on mass and span length. Gates unique Sonic Tension Meter simply converts this frequency into a measurement of tension.

Here’s how it works:First, enter belt mass constant, belt width and span length into meter using built-in keypad. Next, hold meter sensor to belt span, then lightly strum belt to make it vibrate. Press “measure” button to obtain reading and the meter quickly converts vibrations

into belt tension. Readings are displayed on a liquid-crystal screen.

* Belt mass constants are listed on a data card that comes with the Sonic Tension Meter."

I've been using the .wav file that used to be on Jager's site for an audible check on tension for a couple of years now, and I've confirmed it with the Kempf tool. It does take a somewhat practiced ear, which is why Gates' electronic version is much better.

PorKen

32V'r. Momentary belt stretch at different points based on the force necessary to turn the cams, stock tensioner. Direct measurement using the 32V'r at #1 TDC is different than at #6.

Agreed. 32V'r measurement shows approx. 2 degrees advance from cold to hot at #1 TDC on cyls 1-4 (right). Negligible change on 5-8 (left).

PorKen

Guilty re. 1-4 vs. 5-8. Typing too fast without finishing coffee.

1-4 (right, US passenger) side is retarded vs. 5-8 (left) to compensate for thermal expansion. I consider the 5-8 side to be the base timing, where you are saying it's advanced vs. the 1-4 side. Six or half dozen.

There is the base tension and then there are momentary overtension events caused by the load exerted by the cams when turned by hand or acceleration when running.

ptuomov

If the standard Gates belt has the properly pretensioned vibrating frequency of h1, does a properly pretensioned Gates racing belt have a vibrating frequency h2 = h1*sqrt(m1/m2), where m1 and m2 are the standard and racing belt masses?

I am going to not worry about this in practice. This for two reasons. First, I think people overestimate the importance of precisely optimal pretension on timing belts and underestimate the importance of having the system mechanically in good working order (matching belt and sprockets, tensioner working, everything aligned correctly, etc.). Second, since I my low-compression engine now has custom cam profiles, higher-load and non-linear beehive springs, Gates racing timing belt, and Porkensioner, it's so far from stock that I am just going to throw my hands up and hope that Porkensioner somehow gives me just the right slack-side tension.

Could this be just the slack between the belt and sprocket teeth? When running, those conditions should stabilize.

M. Requin

Although I am interested in the physical theory of vibrating strings, in this app, who cares? My point was simply that there is a better, in that it is more accurate and easier to use, tool on the market for what I think is an indisputably important maintenance check.

ptuomov

Sure, I'd buy the tool and just use it if I had otherwise stock motor and a new belt.

Here's my question to you: What would you do if you'd have new springs, new cam profiles, new racing belt, and Porkensioner?