Wednesday, July 10, 2013

Fuel injection works, clutch doesn't

It's been a good news, bad news couple of days working on the Volvo.  The good news is that the fuel injection issues seem to be resolved.  I double checked all the wiring for the various solenoids and all checked out.  But when switching on the ignition, the fuel pump would run continuously when it should just come on for a second.  I had picked up a spare ECU and decided to give it a try to see if it made any difference, and it didn't.   That's good news because it means my ECU is OK, but it's bad because I still have the problem.  Next I took a closer look at the connector that attaches to the ECU and noticed a few of the contacts seemed more recessed than on the old connector.  I'm not sure if I mentioned that it's a new wiring harness for the fuel injection, so this connector is new too.

I picked at the contacts and found a few that were binding a bit in the connector shell, and I bent a few a little bit to give better contact force.  After that I reattached the connector to my original ECU and presto, it works.  The fuel pump now always runs for just a second when you first switch on the ignition.  And rotating the throttle gives the tell-tale 20 clicks of the injectors.

A while ago I played around with another adjustment which seems to have made a big difference and is worth considering if you have a car with the D-jet fuel injection.  The throttle switch does a couple of things.  It has an idle contact that tells the system that the throttle is completely closed and the engine is idling.  Then there are a series of contacts at different rotation points through the full travel of the throttle.  Each of these contacts causes an extra pulse of the injectors to give a little kick when you press on the gas.  There is also an internal contact to enable the pulses only when you are pressing the gas down, and disable them when you let the gas up.

There are procedures to adjust the throttle switch to be sure the idle contact is active when the car is idling, and that it releases as soon as you crack the throttle.  In theory, shortly after the idle contact releases the pulses should start.  If they are delayed then the throttle flap is open letting air in and there is no corresponding  pulse of fuel so the engine stammers rather than jumps.  It turns out this spacing can be adjusted, and it appears to drift over years of use.  After 40 years, there was a noticeable lag before the first injection pulse.  Inside the throttle switch there is a screw that holds the idle switch contact in place.  By loosening the screw the switch position can be nudged around to adjust the gap between the idle switch and first pulse.  It took some playing around to get it right, but I was able to take out the slack and now the injectors start pulsing just after the throttle comes off the stop.

With that side track aside, let's get back to the engine.  You may recall that I suspected a leak around the injector seals for the #1 injector.  It didn't go in easily when I first assembled it, and the hissing sounds I was getting earlier suggested the seal wasn't right.  So I pulled off part of the fuel rail, removed that injector, checked the seals, lubed them up, and reinstalled it.  This time is went if as expected.

With the fuel system back together, all that remained was to static time the distributor, put on the cap and wires, and fire it up. With static timing set to 10 BTDC and wires connected, I gave it a crank.  Cluck, cluck, cluck, zoom!  Off she fired.  It felt a little sluggish but ran.  Turned it off, did a quick check for leaks, then started again.  There was a reasonable amount of smoke burning off the exhaust manifold and header pipe, but that's normal after a rebuild, especially when there has been a lot of cleaning.  The sluggish operation suggested retarded timing so I hooked up the light and sure enough it was off by about 10 degrees.  With that adjusted it seemed pretty good, and the engine sounds nice and tight and smooth.

Next off for a little test drive.  I hopped in and pushed in the clutch and it didn't feel right - like it was bottoming out.  I then tried to get the car into gear and it wouldn't go.  Crap.  I figured all it needed was an adjustment to the cable so up on the lift she went.  I took all the slack out of the cable - more than I normally would - and lowered it down for another try.  It was a little better, but not right.  I was able to start the car in gear and get it out of the shop, then restarting in 1st I took a little spin around the yard, but the clutch clearly isn't right.  Double crap!

Back up on the lift she goes for a closer look.  After a better check, it's clear that the release fork is bottoming out on the bell housing before the clutch is completely released.  Something isn't right.  After a bunch of research I found that there are two different clutches that were used on the car, and they use different release bearings.  One is 36mm tall and the other is 42mm.  I was able to assemble a stack of washers 6mm thick and, tied to a string to prevent loosing them, I tested to see how things would be with another 6mm of release bearing and it looks like that will do it.  But it still doesn't make sense.  From what I can tell the 36mm release bearing goes with a Sachs clutch.  That's what came out of the car.  The 42mm bearing is for a Beck Arnley Clutch, but I'm pretty sure that's not what I have, and it's not what VP Auto Parts says the sold me.

The other possibility is that the release fork is bent, but it shows no signs that I can detect.  The way it's formed I think would result in cracks if it were bent, but you never know.  It's also possible that it's the wrong fork.

The bottom line is that the trans needs to come back out so this can be sorted out.  I've arranged for help weekend after next, and in the mean time will stock up on parts to address whatever we find.  I'll get a 42mm release bearing since I'm pretty sure that will resolve the problem, even if it doesn't explain it.  And I'll get a new release fork to have something to check against in case the existing one is either bent or the wrong part for the car.

In the mean time I'll work on a few loose ends, including running the engine for a while to be sure temperatures are right and that there are no leaks or other issues.

2 comments:

Anonymous said...

Read your latest post about your clutch woes. You may have already done this; however, check the pins on the release fork that engage the release bearing. They tend to develop flat spots. My recollection is that your car is really high mileage so they may be quite worn

Peter Hayden said...

The fork is definitely a suspect. It appears fine with no cracks, however the contact points for the release bearing are shoes, not just round pins. All the forks I see offered for sale today have pins. One possibility is that some time in the past the fork was replaced with an incorrect fork, so I've got a new one on order.

There are two other clues that might turn out to be part of this story. First is that the clutch cable-to-fork attachment has a rubber bumper and retaining washer which is not part of the normal 140 kit. Looking into it, the bumper/washer arrangement is from a 240. Was this done just to soften the clutch action, or to address a fork travel problem? Either way, it's a sign of past hackery which might be a contributor to today's problem.

The second clue is that the clutch peddle has a stop bar that limits how far up the peddle rises when released. Normally the peddle should be flush with the brake peddle. The bar on my car was bent, allowing the peddle to rise up noticeably higher than the brake. Someone would have had to intentionally pull up hard on the peddle to make such an "adjustment", and further suggests past issues with the clutch release mechanism.

Although this is still a mystery, I suspect that once I have the clutch and fork back out of the car along with a replacement fork that all these clues will fall into place to create a clear picture of what's wrong. One thing I know for sure is that I'll be installing just the bell housing to test the clutch before I put the transmission back in.