Timing belt replacement

Hi all.
Just wondering if we really need the "special Porsche Belt tension tool" Wouldn't it be possible for someone (anyone) who is lately doing a belt tension with the proper gauge  to do one with a generic belt tester like the ones at this web address http://hmc-international.com/ <http://hmc-international.com/>  , as well so we can find out the tension in real world figures, not Porsche figures.

these gauges are about $9 American. And would save everybody the hassle of hiring/buying a gauge just to check the tension on their cars once in a while. I would see this a great public service to anyone in the know... or currently doing a belt job.

Or is this simply too much to ask????


I have recently installed a new timing belt. I borrowed the official tool (from the GB club) to set the tension. I then measured the tension, so that I can re-set it in the future if necessary. This is what I did:

With the right hand side (facing front) timing cover removed, and all others in place, the lower belt run that is exposed is from the tensioner pulley to the right hand camshaft. Push down on this section until it touches the center section cover. I measured the force at 15kg and the movement at 12mm. You will need to cut a piece of wood to allow you to get a straight, vertical push on the belt, and something to measure the force (I used bathroom scales).

This method produced very reproducible results, even after turning the engine over. All measurements were done with the engine in the timing position, i.e. TDC no1 firing.

Smiffy (1980 "S")
Hi Jes,

Last time I set the timing belt tension I had a play with the tool as compared the readings with those I got from using a depth dial indicator I have  ( similar to a normal dial indicator but with a depth mic base) there were interesting correlation's but I was a bit suspect of the Porsche tool ( from the dealer a real worry as you could see where it had been dropped and damaged) I though try it again next time and get some results this will be soon and I'll see if I can gain any good info.

Its really a matter checking belt defection etc. but the load  needs to de applied correctly it may be a matter of checking the reading on the Porsche tool to a conventional belt tensioning tool. Something I will look into next time I re-tension the belt after run in ( if I get time as I can generally get access to the tool).

I found some info about the RDK and which points to bridge but where did I store it in the computer I'll find it and post it ( or I'll go back to the wiring diag as it was easy if I

My own experience with an intermittent t-belt light left me wanting a way to get two identical measurements in a row.  I would welcome a belt tensioning tool or method that was easier to use and provided consistent readings.  As a non-engineer I would think a good indication of belt tension would be the number of degrees you could twist the belt with the engine at TDC.

Kevin Berez (kaberez@mindspring.com)


I think the tool is not a really easy thing to use and I test my belt quite a few times to ensure repeatability. I was watching  some "Porsche trained " people and I though it was a bit hit and miss.

I don't like this tool or the method but while I agree twist or deflection for a given load could be used I thought I might contact the belt manufacture and get some info ( Gates was a supplier for Porsche some years ago as the original parts had a Gates logo and Part number in a Porsche box) I used to buy direct from Gates so I thought I'd start there.

I' keep you posted but it all takes time.

Dear Bora:
To the best of my knowledge, all 928 timing belts need to be retensioned about 1500 miles (2500 km) after installation. The reason is: new belts will "stretch" a bit, and retensioning removes that stretch. YMMV.

On a personal note; This past July, the tensioner warning light came on my 90GT. I searched the service records to find that the belt was 2.5 years old and there was no record of retensioning. Fortunately, another PCA sharkster (Pete L.) had a belt tensioning tool and offered to show me how to use it. Upon inspection we found the belt to be in good condition, but its tension was 4.0.  We retensioned to 5.1. So far, all is well.

Merry motoring. ~Ed~

I don't know about your mechanic having delusions, but a new timing belt needs to be retensioned between 2000 and 2500 miles per the Porsche Maintenance Checklist for the 928, although Kim Crumb recommends between 500 and 1000. The belt should be retensioned every 15,000 miles thereafter and changed between 45,000 and 60,000 miles. I don't have the maintenance manuals so I don't know what they say.
Good Luck,

Ted Childs

When he stops swearing, direct him to the top of Page 15-14 in Volume I of the 928 shop manual, which says "Make sure new drive belts are retightened after driving car approx. 1,000 km."

That page is dated July 1983, and, as several others have pointed out, the factory now recommends a greater distance before retightening, but it has always been the case that  retightening after a little while has been what the factory (and good sense) direct.

Your (former?) mechanic's vehement-but-dangerously-incorrect statement is just the kind of claptrap that 928 owners used to be vulnerable to before the 928 community organized.

Dear Dave:
I have only serviced one tensioner, so I hope those with more experience will correct any mis-information mentioned below.

There is oil in the tensioner, but it's in there to act as a damper. It is not internally fed to the tensioner, nor will it cause the tensioner to
automatically take up any stretching of the belt. Instead, there are two access holes for the oil. The lower one is the inlet, and the upper is the outlet. As I understand it, gear oil is the recommended fluid, which will behave as a slow damper. YMMV.

Merry motoring. ~Ed~

> <snip>
> You mention "be sure to at least do it as described in the Chilton's guide". As I have never seen this (probably not published in the UK) can you tell me what it says.
> Smiffy, 1980 "S"

Sure, this is what is sometimes referred to as the 90 degree method. When the belt is at the correct tension you should be able to twist a "relaxed section of the belt" 90 degrees. It is, of course, not as accurate as the very expensive and sensitive tensioner tool. The trick is to
know how hard to twist the belt! The original tool, from which this technique was derived, is Special Tool 9131, which I guess is no longer available, although I would love to have one. It's use is pictured on pages 15-15 to 15-18 of the factory service manual.

I believe you can order a copy of the Chilton's Guide from 928 International or maybe from one of the other Big Three. It was prepared for the earlier models including yours.
Daniel Shapiro
'82 928 5 speed 226k


Anyway, I hope it is helpful. The reason I got my 83S-auto for only $4k was that the PO found that the power steering hose was leaking (spraying the side of the engine) and when he went to the Porsche dealer they told him $284 for the hose only!  He realized he was not yet able to handle that kind of expense, so he dumped it. But the useful info is that I got a power steering hose from my local NAPA for $65! I was worried when we ordered it that the replacement hose would not have the restrictor in the middle of the rubber part (yes I sawed the line apart to see what was in there!), but the NAPA part was spot-on. A perfect fit. I did have to weld up a 17mm crow-foot wrench to get the line connection on the back of the pump. Whatta BITCH!

Another note:  In my T-belt replacement job, I found that the little roller that mounts right on the water pump was a little noisy. I tried to press the bearing out, but I took too much force so I just pried out the shields on both sides of the bearing and carefully cleaned it with multiple baths of lacquer thinner. Then a gentle and SLOW spin with the air hose to get rid of any little particles, re-grease it, pop the shields back in and presto, it is smooth and quiet.  This little roller is about one-half the size of the crank sprocket so when the engine is turning 4K, it is turning 8K!

I took the belt tensioner all apart. The o-ring felt about as hard as industrial diamond, so I chipped it out. NAPA had a 1/16 X 1.125 o-ring that is close enough to 1.5 X 28mm to work OK. The dust cover, for which I find no separate part number I replaced with a piece of nice leather from an old pair of boots I keep around for such things. And the leather will outlast the original rubber boot anyway.  This tensioner thing is interesting. There are some cavities in the front of the engine that match cavities on the back of the tensioner. The oil you pour into the fill hole while waiting for the excess to bleed out the bleeder hole fills the body of the tensioner as well as these funny cavities on the back. I guess the purpose is to conduct the engine heat intimately to the bimetal disks that maintain the belt tension. Sure seems like a complicated system!


  1. Adam, I just got done with that job. I got a rebuilt water pump from Devek and got back something like $100 for the old pump, so I wouldn't just throw the old one away. The rebuilt one was done in Germany, (not rebuilt at Mert's Two-Car Garage Water Pump Rebuilders of south Texas), so I figure it's a pretty good one.
  2. Main seal replacement: good.
  3. Take the belt tensioner apart and replace the o-ring (selected from the stock at NAPA auto parts) in the bottom of the cylinder around the piston and also replace the rubber boot (purchased from Devek) at the end where the thing pushes on the tensioner roller. This all costs $17 versus a new tensioner for over $500! Seal the tensioner to the block with gasket sealer, and refill with 90 weight gear oil per the manual.
  4. Take the tensioner roller and the other smaller roller above it apart, clean out old grease and put in new grease.
  5. 5. Replace the belt guide that goes below the sprocket on the crankshaft, part# 928 105 067 00. Porsche says you got to replace the plastic cover that goes behind it too (part# 928 105 147 05), but I just cut two little teeny pieces out of the old one and saved $45. It's easy. The new guide has a ball bearing roller rather than just smooth plastic. It is held in place by two little teeny Jesus clips* and washers on two pins coming out of the front of the engine block.
  6. Rent the belt tension tester from Devek or 928 Specialists. Set tension at 4.5 on the tester.  You must use the tensioner. Did you hear me? I said you MUST use the tensioner.
  7. Recheck tension after 2000 miles or 3000 km.

    have fun/
    83s a/t


> -----Original Message-----
> From: Mark Rosenfield [mailto:chefmark@pacifier.com]
> Sent: Friday, March 03, 2000 12:02 AM
> To: 928
> Subject: [928] Belt Tension Light
> I am in the process of fixing the toothed belt warning light on my 89 S-4. The previous owner had it disconnected because it was sending faulty signals. The microfiche doesn't give me anything to go by  so I am soliciting information from the group. I have examined the system to the lead on the belt tensioner. Can anyone tell me what part on the tensioner is the switching mechanism for the warning light? A part number would be very helpful as well as any information regarding part replacement.
> Thanks in advance,
> Mark Rosenfield
> 89 S-4, 122k
mounted on the carrier for the tension roller is a simple copper strip which connects to the pushrod on the end of the tensioner housing---the "warning system" is looking for a constant ground--if the tensioner arm bounces away from the tensioner usually under hard acceleration/deceleration the light flickers on---the timing belt is pulled around by the toothed gear on the crankshaft along the way it turns the oil pump gear, turns the drivers side cam, the smooth back side spins the water pump, it pulls the passenger cam and finally goes around the tensioner pulley before it gets back to the crankshaft gear--in this system the tensioner is there to take-up the slack on the no-load side---like a bicycle chain there is a tight side and a no-load side. One important note the tensioner is not hydraulic, it does not self adjust, it is not supplied with engine oil. It is a simple mechanical tensioner stacked inside the housing are a number of slightly cup shaped bi-metallic washers these washers when heated change shape to loosen the belt to compensate for the thermal expansion of the aluminum block. As the design progressed the tensioner was filled with 90 weight gear oil, to better transfer heat to the washers, the next step was to drill two holes on the back of the housing and put a gasket between the block and the housing allowing oil to run into the webbing for better heat transfer-There are no oil holes in the block- You must still fill the tensioner when changing the timing belt--engine oil for later models- the last improvement was to insert a baffle with a controlled orifice like a shock absorber inside the tensioner--this was to help limit the bounce of a loose belt when the engine is cold- Back to the "belt warning" system--it is very unreliable-and can only warn of a possible loose belt--it cannot tell when the belt is worn, too tight, has cracks or teeth have broken off--it is good to have it working but do not depend on it saving your engine.

The question of the belt tension tool appears somewhat regularly on this list.   Anyone doing their own engine work will think of alternatives to that expensive tool.    I will share my little escapade from 18 months ago.

Original tool is now manufactured by one maker in Germany.   It's profile (size) is important because it fits in the required area for tension measurement.   It's also important that it measures the belt between the crank and camshaft gear on the passenger (US) side.  Other areas of the belt are either too difficult to get to or will not give reliable/repeatable measurement.   I rented the proper tool and purchased two 'Kriket' tools from NAPA (with different ranges).  Correct tool has rollers that fit over and under the belt to take the reading.   Procedure requires pressing down on the tool until it is pressed against the air pump bracket (going off memory here).   All of this is important because it reliably positions the tool and uses a fixed surface - the mounting bracket - as a reference for every measurement.

My plan was to take measurement with the 'correct' tool and use a Kriket to find a reference.   If you use the location described above (belt from crank to cam gear), there is little actual belt available.   Most importantly, the ribs face up.   This pretty much cancels using the Kriket because it works by pushing down until enough belt deflection occurs to overcome the Kriket's little 'snap' at which you then read the scale on the tool.   After many variations and attempts, I could not get reliable, repeatable results and had a heck of a time manipulating the position of the tool.

Another alternative is the old 'squeeze' tool for measurement.   It works pretty much like the factory Porsche tool, but it's actuated by squeezing down on the actuator with your palm while holding the tool in place with two fingers.   Because of it's size and geometry, measurement must be taken at the top of the belt - the part exiting the cam gear on passenger side. There is no room to fit the tool in the proper place on the lower belt.

I could not make any other method work to my satisfaction.    I believe that for 16 valve engines, the 'finger' method of twisting the belt is probably good enough.    Not so sure I'd trust it for 32 valves.    Why take the chance?   I decided to rent the tool (the Big Three all provide this service) whenever needed.   It's entirely possible I may have missed something or should have tried another variation - so by all means see what you can do and be sure to post your results!

JP Rodkey
79 euro
Hi guys (and gals),
I'm in need of some desperate help.  I'm taking apart my front end to do the t-belt, water-pump, pulleys and so forth.  I release tension from the belt in preparation for removal, and then slide the belt off the sprockets.

I look at the passenger side cam sprocket only to notice in horror that the sprocket has moved 4 teeth counter-clockwise.

The belt must have still has some tension on the top end and pull the sprocket when set free.

Oh Sh#t!!!!!!!!!!!
I'm at TDC, can I simply rotate the sprocket clock-wise 4 teeth and line it back up with the mark?!?  Seems to feel tight, I'm guessing this is just the valve spring I'm feeling?

Please advise if I'm F#CKED!
89' 5 speed and first time t-belt do-er.



Not to worry.  You've called it correctly...the valve springs and cam position move the p-side sprocket as soon as the belt is removed.
This must therefore be addressed during reinstallation.  17mm box wrench in the hands of an able assistant will allow you to proceed with the proper "threading" of the new belt.

Best regards,
David Lloyd

Hi everyone
Well maybe I know the answer to my long and boring threads on variation of tension in the timing belt I found when running the engine without covers that there was CONSIDERABLE harmonic vibration in the belt at most rpm's This did not seem right Looking at the tensioner adjuster wheel with a strobe light ( i used different plug wires to see it in different positions ) I noticed that the belt appeared to part company with the idler as it rotated. Took the whole thing out and checked bearings, seemed fine, but greased them up anyway! So , I then put it back and reset tension with finger thumb method . OK  this is not a good method ( I know everyone knows this) as it is difficult to calibrate one finger/thumb unit. But I did find I could tension the belt past where it was before and still get a 90 degree twist. What I did was to tension till I could not quite get 90 and then run back a notch or two.
Result: Tension in all positions. Now I realize when the belt is too loose it allows the tensioner to come to its stops in travel ( spring loaded I assume) and then there is NO tension in the belt and it is completely loose. Clearly this is way too looses as it ceases to drive the water pump

As the PO had a new T belt fitted just before I bought it and I now have 2000 miles it clearly needed a bit more tension.

Running the engine now, there is no visible harmonic vibration and I can still do the 90 degree twist reasonably comfortably.

1) I'll rent / beg / borrow a tensioner tool asap.

2) Too loose causes significant harmonic vibration in the belt with probably excessive belt and cover wear as the tensioner cannot extend enough to take up the slack in some engine positions.

3) Too tight can cause serious wear problems as well, though I really don't see how this could directly cause drivers side camshaft breakages unless there is a marginal oil feed to one of the bearings and the extra load causes it to pass its limit ????

Makes sense?
Any comments?


I agree with your methodology.  Personally, I use 1.5 Fingthumb Units (FU's?) of tension. Can't get too close to a 90 deg. twist - maybe 45 degrees.  My book says to use SAE 90 oil in tensioner, but I don't know what year yours is.  Does it have two fittings?  If so, one is the fill port and the other is the bleed hole.  I really can't imagine breaking a camshaft from too much tension, but maybe so.  Mine would just make noise if I got it too tight.  Used to race Nissan V-6 engines - 850 hp at 8500 rpm.  We used what I think was a larger belt than standard and tensioned it so tight I couldn't believe it.  No broken camshafts or belts or problems of any kind.  I am not convinced of the necessity for a tension gauge...

Gary - living on borrowed time - Casey
86 5 sp


Too tight caused excessive cam sprocket wear on my 88.  When I got the car, the dealership had simply been cranking down the tensioner adjusting bolt to extinguish the light (t-belt warning) for the PO.  Don't let anyone do this.  It had been doing this for a year or so with him.  I knew the guy and the dealer and several other techs that had seen the car.

When we took the front off to change the belt and re-build the tensioner, we noticed considerable wear in the sprocket faces, looked like 1/16" grooves where the belt ran.  Not good, lucky it didn't break a camshaft.  I have heard of broken sprockets too, I guess several things could cause this but "too tight" would sure be one of them.  Went ahead and replaced all wear parts, water pump, etc. and replaced the front seal.

Same with camshafts, several things can cause failure, but don't write off "too tight" as a possibility.

It may be interesting for some to note that again in the past 6 months I have seen times of intermittent t-belt warnings.  We knew the belt was right, we check it often, especially with this "phenoma".  Finally Mo took it personally and went nuts.  Took the computer out and replaced it with a known good one.  No luck.  Wires had continuity, all seemed OK.  Were about to tear into the pod and look around when Mo decided to completely inspect the tensioner mechanism.  He found that the 3" long metal "sensor" that breaks the ground to trigger the warning lamp, had a fracture (crack) in it near the 90 degree turn the metal takes near the attachment points.

A crack.  Not a complete break, but still hanging on.  This explained the intermittent nature of the warning.  Sometimes it worked for a month or two as it should, other times it would display the warning for a month.  Came on about 3 or 4 minutes after driving (normal warning mode for a problem in the circuit).

Now after replacing the arm, no problems.  Tip:  Carefully inspect all parts and connections (even for little cracks) as I believe even with a 60% crack in the metal, it got enough juice to satisfy the warning system.  At 70% it didn't.

Weird, huh?

Marc White

There is a lot of variation in loading on the timing belt, which leads to flutter and jerking of the belt. This jerking and fluttering can cause the "switch" in the tensioner to lose ground contact, sending a signal to the warning system computer that the belt is loose. The tensioner oil appears to serve two main functions - damping these fluctuations, and transferring heat from the engine block to the bi-metallic washers that serve as the spring in the tensioner so that they compensate for engine expansion due to heat. Failure of either of these functions could lead to false belt tension warnings.

To fill the tensioner on a later car with two nipples:
Get a trigger-type pump oil can. Fill it with 90 wt oil. (One tech sheet from Porsche says to use engine oil, but I prefer to go with the heavier 90 wt called out in the shop manual - I think that it will give better dampening.) Get a foot-long piece of clear plastic tubing that fits on both the spout of the oil can and the nipple of the tensioner.
Remove the rubber caps and open both bleed nipples. Attach the tube and oil can to the inner-most (upper) nipple. Pump oil slowly (that's the only way it will go in!) into the tensioner until you get a solid stream from the outer-most (lower) nipple. Close both nipples and re-install the rubber caps.
Check the engine below the tensioner for oil leakage for the next week or two.

If you add oil to the tensioner and it promptly leaks out, you need a tensioner rebuild. The rubber cap gets hard and brittle with heat and age.
Wally Plumley
928 Specialists


The sensor is nothing but a strip of copper attached to the tensioner arm, when the tensioner pin is touching the arm it grounds the circuit and the central warning brain is happy -- if the contact strip starts to break the resistance can be too high and reads as no ground as does a loose wire . This system is so primitive it should never be relied upon but should not be ignored .

Jim Bailey


Although I'm mainly a lurker on the post, Mr. Kempf was kind enough to allow me to try his version of the Porsche T-Belt indicator as a beta user.  I've owned a '79 for over 15 years, and have replaced the belt three times now.  The last time, the tension was set using his new indicator.  As I mentioned to Jay,  I took and made a very fine scribe line so as to see if I could actually see and possibly graph the belt stretching over a period of time.  The reason I even contemplated this exercise, was that I felt that I could get enough repeatability between readings so as to be worthwhile, along with my curiosity.
Alas, I decided to maintain tension, rather than to attempt to  chart an ever loosening belt.  The belt relaxed very little over the next 5000 miles and after the third adjustment hasn't moved for over 7000 miles.  Actually the third adjustment probably wasn't necessary, and probably wouldn't have been noticed except for my scribe line.  I  recently acquired a '89 s4, but so far have not taken the time to run a base line tension check.  I have no qualms in recommending the new tool.  Not being one attempting to impress others by how much I can spend for 'Porsche' tools, how can you miss with this one.

Keeping the rubber side down


There was a question on how to calibrate or check the "Kempf" Timingbelt tool. A simple answer is: attach a string to the assembly to simulate the timingbelt resistance to twist and attach 5 lbs (2.2Kg) of weight to the string. That should make the indicator stay right at full scale of the "calibration window".