cam sprocket alignment
 

I cannot get repeatable readings for the cam sprocket distance. The main issue is that the case has about a 0.2mm step at the parting line of the 2 halves. So obviously if I use just 1 case half, I get a rather small span on which to lay the straightedge, and based on that span my readings vary. If I use both case halves, I can get almost any reading I like based on where I set the straightedge.

Certainly neither method is repeatable within 0.25mm, though if I pick the exact same spot for the straightedge, I am usually within 0.05mm, so the repeatability issue is not with the operator or his equipment.

I am curious if anyone else has experienced this, and what you might have done about it. You'd think the surface would have been machined with the 2 halves clamped together, but obviously this was not the case (haha pun intended ...)

My solution has been to just use the same number of shims the factory did. The camshaft position is set by the chain housing, and since I have not switched cases or chain housings, I see no reason why the cam sprocket alignment should have changed.

Of the engines I have redone, the factory shims have always resulted in good cam sprocket alignment especially if the cam towers are not changed.

I'm no expert, my 1 cent.

When I redid my top end I spent way too many hours using the Stomski tool and straight edge/etc. along with a spreadsheet to measure what shims I should reinstall. After all that the shims originally installed were correct.

I've often wondered is +/- 0.003"? over a distance of say ~1 foot really going to make any difference? And it is on the tension side, the slack (feeding) side is not constrained very well as is comes off the "floating" tensioner arm. Seems like the feed side would be the important one.

I think you're just fine putting the shims back the way they were originally installed, as long as the parallelism measurement checks out reasonably well.

I agree that the parallelism is a proper thing to check. However it's not absolutely critical that it be dead nuts on. I found on my '87 3.2L rebuild that the 4-5-6 cam sprocket had been installed backwards by somebody. The parallel measurement was way off (can't recall by memory) and the cam sprocket showed some side wear on it, which obviously should not come as a surprise. But my point is that the engine showed no ill effects in terms of noise or massive wear on the sprocket.

I only did the top end so I don't know how the intermediate shaft sprocket or the shaft bearings look. Engine continues to run fine to this day, 15K miles later.

Would you mind explaining in more detail? I was thinking of buying the Stomski tool to do my first engine, but if you are saying that it was troublesome to use, then I would look for a different way to do this.

Was the tool difficult to use? Why did you have to spend so many hours analyzing the data?

Thanks

3 shims are the "Norm" but not against any known law to do other wise...

I was very careful to keep all of my gears, sprockets, shims, etc. together when I did my tear-down. Only last night when I re-installed the cams did I notice that the right and left side both had 3 shims each while the reference material calls for four on the right side (1987 3.2)

Is it possible it came from the factory like that? I saw no obvious signs of the engine ever having been opened up before. Has anyone run into this?

I'm now searching for a caliper to borrow, but given the difficulties some have posted about being able to get conclusive, repeatable measurements, I don't want to create an issue where there is none.

GK

I use a simple straight edge (not even a precision one) and have been able to get results I feel are accurate and repeatable. It's nothing more than a long ruler (literally, it is a long measuring ruler) that is reasonably straight. I hold it on the intermediate shaft bore and take my measurements accordingly. Seems to have worked OK for me over the years.

KTL - I planned to use a simple straightedge (a long carpenter's square) across the front surface of the engine case, but don't you need a precise depth from the straightedge to the IMS, and to the sprockets? They've got to be within .25 mm, correct? I'm looking for a caliper that has a good depth gauge to get that measurement.

Also - assuming that the shims turn against each other, sort of like a bearing surface, I lubed them with assembly lube. Was I correct in doing so? The reference books say nothing about this.

Thanks,
GK

Yep, use a good pair of calipers with a built in depth gauge. I have a nice Starett caliper that I use.

No worries about lubing the shims. The shims get splashed with oil anyway during engine operation.

A carpenter's idea of precision and an engine builder's idea of precision are in two different leagues. I once tried to determine how flat a cylinder head was using a 36" carpenter's square. By my measurements, it seemed the head was warped by about .020".....it was not, the machine shop declared. The carpenter's square is just not that precise.

I got myself a digital caliper with depth gauge last night - probably not the best quality one available, but choices are limited when you go to local retailers. I took measurements of arbitrary things with it, and found that it was repeatable, and since I was looking for distances of objects that were relative to each other, I felt that it would do the job.

I used a carpenter's square after holding the edge down against my granite counter, and it was flat all the way across - no daylight. If nothing else, I proved that my square and my counter were both equally straight, or equally warped.

I measured the depth to the IMS and the right and left gears. Bentley's diagram on this was most helpful to me as it lists the expected delta between the measurements. I took the IMS depth, added the delta for each side, and created a range for each that took into account the .5mm margin for error.

The tough part was getting repeatable measurements to these points, holding the square in position with one hand while holding the caliper squarely against it, and sliding the moving part of the caliper with the other. After some practice, I got a range of measurements for each reference point that were within +/- .3mm - I used the IMS measurement that was in the middle of that range as my reference point.

After doing the simple math, I was satisfied that the left and right were within spec, so having 3 shims on the right was intentional, whether done at the factory or by another wrench.

This killed a solid hour of my evening Porsche-time, but it was an hour well-spent. One less thing to worry about when getting ready to hit the ignition key...

GK

I have used the Stromski kit on the last several engines built and as stated in an earlier post, if taken apart and kept together the measurements are usually less than the 0.25mm variance. My next Stomski tool will be the DigiDix for setting cam timing. As our hobby becomes more expensive it is nice to have tools that really work.

This was one of the handful of things on my rebuild that really was not fun...the other being the cam timing. Anyway, I wondered how critical the extreme measurements were on this then I recalled KTL's story about someone installing the sprocket backwards...with minimal damage. If I ever do another rebuild this is one of those things I'm not going to lose sleep over. I'll save that for the cam timing, and case sealing.

Indeed, and keeping in mind that the tolerances here are pretty broad, Porschey-wise, since the adjustments are in .5mm increments. My value-priced calipers measured them out to be between .49 and .51, BTW...

Good job. Sounds like the same hand-eye precision measuring coordination trick i've used for the job.

It's one of those things when you're done, you say "I could really use another set of hands...." Often times when I employ the wifey for those second set of hands, I end up regretting it....... :p

I drove my self nuts using a modified carpenter's square (short leg cut off). The square was fine as a straightedge, but you really need four hands this way. My first effort led to me making a scatter diagram and finally picking one in the middle. Normally, the IMS protrusion from the case stands a bit proud of the two machined edges for the motor mount which are farther out. Very easy to rock the straight edge so it is pressed against two supports on one side, or two on the other. What you want is to have it resting only on the two sides of the IMS case protrusion.

I really like the Stomski tool. I had purchased a set of depth micrometers (actually, one mike and a bunch of extensions), but found that my digital dial indicator was long enough, and stable enough, even without the nifty base plate I bought for it. And the first time I used the Stomski I found I was already right on. I've got 4 motors (3 for one race car, one in the SC), so I go through this often enough to be happy I paid the $ for the Stomski (which is much like the tool the factory shows in its manuals).

One of my race motors was made from two different case halves (one motor blew up one side, the other the other, I guess), but the machinist who did the work got it to work out just fine. No issues with that. However, on my first rebuild of the motor (I bought it with about 25 hours on it), and getting to setting parallelism, I found that the outer (motor mount) protrusions stood proud, at least on one side, of the IMS seal plate area. Aaaargh. Eventually I used the die grinder so that the straight edge sat true on the IMS stuff, and I could get a feeler gauge under it at the motor mount area.

But without a tool, I think it is very very hard to get a relatively thin edged straight edge to work well. It wobbles up and down, as well as the problem of not being sure what two points side to side it is resting on. Though a helper can make up for all that.,

[QUOTE=KTL;7413609]...I found on my '87 3.2L rebuild that the 4-5-6 cam sprocket had been installed backwards by somebody. ..QUOTE]
You know that the 1-3 side has the spocket facing out, while the 4-6 side has the cam sprocket facing in, right? It is supposed to be that way !

I would venture the guess that Ken knows that. Backward on the right side would have the sprocket facing out.

Once you realize the relationship between the cams and the cylinder banks, with left forward, and right back, and that the cam wheels are to be installed to minimize this offset, you won't make the mistake. Of course, Wayne's or Bruce Anderson's books make this very clear with diagrams and such.

I am interested to hear that the misinstalled gear actually worked pretty well. A mechanic/machinist buddy once told me that if you were two shims off you could tell while hand rotating the engine with the covers off that things weren't meshing right. I've tried looking at how parallel the links are, especially on the idler. But not to the point of being systematic about it, so I can't say as this is a valid test or not. I thought I saw something, but not sure.

Thanks for all the replies! I ended up finding a feeler gage that matched the mismatch of the case halves & taping it to the case where it meets up with the engine mount. This made it fairly easy to make the measurements. I'm using a 24" Starrett straightedge (they're 1/4" wide, so they are easy to keep square) and a 6" HF digital caliper. I hold the caliper so the edge of it rests against the straightedge (so looking along the caliper, the caliper and the straightedge make a "T", not a sandwich) - this minimizes measuring error if I'm not holding everything perfectly square.

Ended up with 3 shims per side. I t appears that the chain has about 0.75 mm slack sliding side - side (fore-aft relative to the car) on my cam sprockets, so that would also be an indicator of the degree of precision required - if the IS sprockets have a similar amount of slack, one could be off +/- 1.5mm without the chain needing to run uneven.... not suggesting one should try that, but it does indicate the +/- 0.25mm is a bit excessively anal.

There was a comment re. cam timing being difficult - that's one thing I've never understood. Cam timing is one thing that always seemed a breeze to me - takes me 20 - 30 minutes for both sides. Everything else I do seems to take double or triple what the Dempsey books say it should, just not cam timing... I always have been an oddball I guess :)

I think I have got cam timing on the older, big nut, cams down. But with the bolt end cams I no longer was able to rotate the cam but not the gear (pin out), which is what makes doing the older cams pretty easy once you get the hang of it.

With the bolt style, eventually I could get the pin in so the cam was close to where I wanted it. Then I stuck a punch end through a suitable hole in the gear and used that against an inner tooth to nudge the cam so the dial indicator got closer to where I wanted it, and I could tighten the bolt some, enough to keep the dial indicator there, and then remove the punch, insert the gear holder tool, tighten more, and finally rotate the engine to see if it was close enough to go to final torque and move on.

But those who have worked much longer and more with this style cam end must have come up with something better? There is so much slop that I don't see how one can use just the pin to get where you want to be unless you are very lucky. Or so patient you can do something I'm not sure I understand, which is to skip teeth and take advantage of the fact that (I guess) the holes on the gear aren't symetrical with the teeth.

I have the later style cams. I put assembly lube & motor oil on the gear where it sits on the cam. Chain tension is just provided by the stock tensioner - adding force just makes everything harder to do, it does not seem to alter the results.

To make the adjustment, I plug the pin with sparkplug into one of the holes, no cam bolt at all; I don't bother unscrewing the sparkplug either, it just sticks out of the cam gear. Once the dial indicator reads what I'd like, I pull the pin & rotate the crank. If the cam rotates along, it tends to do so in one direction but not the other - so I overcorrect and then approach the timing from the other side, where the cam seems to stay put.

Once it's all set, I torque the bolt & re-check - it rarely moves more than 0.002 or 0.003 ".

YMMV

I stopped the cam-rotation issue by sticking an awl into one of the other holes in the gear, then using that to ever-so-slightly "wedge" the gear-to-gear alignment to get the indexing pin into the next hole over, followed by a re-test.

I think the issue is caused by rocker pressure on the cam lobe, causing it to rotate - particularly when only one rocker is installed, and there are no other rockers holding the cam in place, or perhaps even pushing another lobe in the opposite direction. When I practiced this process during disassembly (described here: http://forums.pelicanparts.com/porsche-911-technical-forum/729856-87-drop-top-drop-n-top-end.html#post7252404) I did so with all the rockers still installed, and there was no rotation of the cam when I jogged the crank backwards a few degrees. That was easy. This time was a little harder.
GK

GK - I am with you. That's what I think also.

With the old style/big nut cams, you had plenty of leverage on the cam to resist the spring pressure of the #1 or #4 intake. Once set, you didn't care. And installing just one rocker (or an exhaust also, so you can check valve clearances) makes sense if you might need to pull the heads off again.

But with the bolt style cam, I'm thinking a guy ought to figure out which other cam/rocker/valve would be pushing the opposite way. The valve you use for setting timing is going to be pushing open, so it will want to move clockwise. Ought to be one of the other five which will be pushing counterclockwise. Less hassle than installing them all. The one time I rechecked timing with them all installed, I was pleasantly surprised to find the cam didn't want to turn all on its own due to spring pressure.

As a newbie to this process, the many posts on the subject were very helpful, but this video really de-mystified it for me:
N Fulljames cam timing.wmv - YouTube

It was kinda fun but I'm glad I'm done. It's really good to see the covers going on, the whole engine being closed up, and I'll be putting tins on shortly.

Yes I know that. I just didn't get into the detail of which way it was supposed to be facing, dished part facing in or out, because I couldn't recall from memory.

I found the bolt-end cams to be pesky when the sprocket would not slip on the cam snout with the pin removed. I used to attempt setting the cam to a position that anticipates that inability to rotate it with the pin out and nothing to grab onto. It was a frustrating vicious cycle of back and forth fiddling that didn't allow me to precisely set the cam where I wanted it.

I think it was flat6pac (Bruce) or Mike Bruns who suggested to insert the bolt with the big washer removed and snug it in place by counterholding the sprocket with the special tool and then bottoming the bolt head on the end of the cam.

No need to gorilla tighten it. Just mildly tight enough to allow you to rotate the cam via the bolt head w/out the bolt loosening. Kind of like double-nutting. Then you can use the bolt head to rotate the cam forward or back to set your desired lift spec.

Kevin

I tried that, but must have done something wrong. I can see how it would work for getting the pin in a likely hole.

But not for dealingwith the wiggle room at a given pin setting. Because at that point you need the big washer on and to snug it down to hold the reading. And loosening the bolt moves the cam, and the holder doesn't help because it grabs the gear, and not the cam. So spring pressure can let the cam move until only the pin holds it. So if that is not optimum, you haven't gotten anywhere.

Porsche seemed to think that using the pin to do all the locating before torquing was satisfactory. But they weren't thinking of guys like us who want the two sides to be almost exactly equal, or who are fixated on being within a couple of thousandths of an inch.

Nex time I do this (which, alas, is sooner than I wanted), I'll have to look into the "add a second rocker" notion to deal with spring pressure.

Very true about dialing in the setting when you've already got your pin selected, and you're just "getting it to within a gnat's ass" within the pin hole slop.

Solve all your precision problems and get a set of Mike Bruns' JB Racing adjustable cam sprockets. They look like the bee's knees (never understood that expression....)

JB Racing - Porsche Engine Components

Just gotta decide if $400 is worth it to you to get it dead nuts perfect the way you want it. :D

You are right - I have hankered after those. I've got a VW Beetle with a somewhat similar setup (much less expensive). That would cut through all of this.

Porsche, on the other hand, now requires you to use a sort of jig on the newer water cooled 6s (maybe started with 993s?) which plugs into back end of the cam. That sets the cam, set the crank to TDC, tighten, move on. Maybe they needed something like that with vario-cam and all that?

Sorry for late reply.
The tools were fine. I took time to develop a method and a spreadsheet for taking the measurements and checking my math. I tend to dwell on a task until I understand what I'm doing which can take a while and I wasn't in a rush, was taking my time and enjoying the method I came up with. If you can take the measurements just a few times and be confident in your math, then it doesn't need to take much time.

Walt, if you install a second rocker be sure it is fairly loose so you don't run the valve into the piston while spinning things around! With just the intake, you tend to stay clear of the piston. If you also have a correctly adjusted #1 / #4 exhaust rocker, for instance, it will make contact while doing the adjustment ... you can guess how I know? :)

The pins on my car seem to be accurate within +/- 0.002, which I'm pretty sure is indiscernible from an engine point of view. Based on how the rocker arm geometry (how was the face ground / worn, what angle is the rocker at based on valve seat & face grind) and changes in valve timing caused by expansion of the engine case when it's hot I bet you're getting that level of variation and then some anyway.

If you decide to be picky nonetheless, you could probably rig up some sort of thing from a big (1/2" or 3/4" x 1" long) rare earth magnet and a thin steel plate with a "V" ground in it glued to the magnet, inserted through one of the other rocker arm cavities onto the camshaft. They are surprisingly strong, and would probably hold the cam enough to move it - just be sure you have it attached REALLY well to whatever handle so you can remove it later, if the handle were to come off you would have serious issues getting the magnet off again! I'd drill a 1/4" recess in bar stock & use superglue. We have a few on our fridge, and they're almost impossible to pull off without sliding them off the side. I bought them for my kids to play with, but decided immediately upon trying to separate 2 of them that they are not suitable for kids...

I don't quite follow the interference issue. You are saying that starting with an approximation of the intake overlap (dot on cams up), you might be far enough out on an exhaust that you'd get interference? I'll bear that in mind when I try this balancing act. With the big nut cams, I can get the intake spot on from the get-go (+/- pin slop). But has been harder with the bolt end style. I should be able to at least get a pin in so the overlap is within the range of my pin slop, which should mean that when I find a counterbalancing valve (another intake? or an exhaust, but one which is pretty much as close to closing as the intake is open - hmmm, sounds like the exhaust on the same cylinder at overlap) spring I should be close enough not to have this happen.

I wish my experience showed only 0.002" valve height difference due to pin slop. I am remembering more like 0.025", and I never have gotten as close as 0.002 with just the pin. If I could I'd be a happy camper. Not sure how we have such varying results. Maybe I'll take some spare parts and just measure slop. I probably can do it directly with a fixed inner, a pin, and a dial indicator on a tooth of the outer. Should be able to translate that into degrees, though that won't be exact as to overlap, as that depends on some other things, the cam profile most acutely?

On my engine (motronic has a high dome piston) there is a very narrow range of cam timing where intake & exhaust clearances are adequate. Other engines are probably different!

It sounds odd that you're getting .025" variation - that seems really really high. Valve lift spec for 3.2 motronic at TDC is only 0.049", +/- 0.007" or something like that! Sure you're not remembering variation in mm? The pin seems a pretty tight fit, I cannot imagine more than a degree of slop in there. If there is a bunch of slop for some reason, you could turn a new pin that fits tighter maybe?

It's also possible that your cams measure lift at a point where the change in valve lift per deg crank revolution is much greater than on my 3.2, in which case the tolerances ought to be much larger to make up for it?

I've seen this (pin slop greater than any sensible tolerance) on my 2.7 and SC motors (overlap about 1mm) and race motors (overlaps 5.8 -6.2 range). Dial indicators in Metric are ridiculously more expensive than those in inches, so it was only when electronic ones with a switch for either system became available for reasonable prices that I could use metric. I'd never have worried about 0.002" variation. I've used (for better or worse) 0.005" as my side to side tolerance as long as both readings were within limits for the race motors (clearances so tight that not much room to move timing around), or close to the dart I threw for the SC, given all the settings which work for that cam.

As I recall, the early 911s had an inner which had full holes, while the subsequent motors have half holes (look more like gears, though not). But I've had the slop with both.

There is a system which takes advantage of the different ratios between the crank/IS gearing, and the IS/cam gearing. Since the pins are a vernier system, if I remember this right, if the timing you want does not end up with the valve spring pulling the pin tight against the outer's hole at a reading you like, you can loosen things all up, and move the chain wheel one tooth one way or the other. This will give you a slightly offset selection of holes, and you can continue to do this until you find the magic combination.

But this is kind of tedious. And ultimately it is not the pin which withstands the torques involved. The pin just holds things in place until the big nut or small bolt are fully tightened, at which point it is the friction caused by fastener stretch which locks things in place.