1990 Honda Timing Belt Info
I put this webpage together after fixing the timing belt on my 1990 Honda Civic Si. I heard all the horror stories about the total engine destruction when a belt goes and how it can cost you thousands of dollars. My experience is quite different from that and only cost the amount for towing and the timing belt. The info that I found out can easily be transferred to other models and years of Hondas.
The Timing Belt SNAPS!
I was driving at 120 km/hr (70 mph) when the engine died, there wasn't a sound - nothing. It was as if someone had turned off the key. I looked down at the tach and it was showing 0 rpm (first clue) which didn't make any sense to me at the time. I immediately put the clutch in and the tranny in neutral and coasted to a stop on the shoulder. BTW, putting your car in neutral is good practice when something goes wrong and you don't know what it is.
Troubleshooting
After a few phone calls and a ride back to town, my Dad and I grabbed some tools and headed back for some troubleshooting. The engine turned over very fast and was making an odd soft pinging sound. No spark, so we took off the distributer cap and took a look. The high voltage connector on the ignition coil was very corroded and looked like a suitable candidate for the problem. I replaced the coil with a new one (more on this fiasco farther down!) and turned the engine over. No spark still.
I had just replaced the ignitor (ignition control module) about 4 months ago and I thought that maybe that was the problem. I asked my Dad to turn the engine over with the distributer cap off and right away saw that the rotor wasn't turning - this meant timing belt!
Cam Timing Mis-information!
After having the car towed home to my garage, I purchased a new timing belt and put it on. I ran into two major problems while changing the timing belt. The first was the lack of documentation available for the 1.6L Si engine and the second: getting the damned crankshaft bolt off.
1.6L vs. 1.5L Cam Timing
The 1.6L cam timing marks are located at roughly 7:00 position on the cam gear. There is a small white arrow on the lower timing cover and a small line on the cam gear. A red circle points it out on the picture.
Red circle shows correct cam timing marks
The manuals that I had indicated to use 1.5L timing marks which are the yellow circled UP letters and the yellow highlighted horizontal lines. On a 1.5L engine, the horizontal lines are to be lined up with the top of the cylinder head. When I tried to line up the horizontal lines on my 1.6L, I couldn't get the horizontal lines to line up. I was always one tooth off which just happens to be 18 degrees (more info on timing farther down). Thanks to the newsgroup alt.autos.honda for the correct info on the timing marks.
Checking for bent valves!
Once the timing belt was installed, I removed the sparkplugs and ran a dry compression test - 125 psi in all cylinders. This indicates that the valves are seating properly and aren't bent! I had quotes of $900 for a rebuilt head. So much for the horror stories of interference engines!
Note: At the bottom of this page is information on compression testing
Removing the crankshaft bolt- Oh the HORROR!
The problem is to stop the crankshaft from turning. Here's a bunch of methods that don't work:
- Jam a screwdriver in the flywheel teeth - screwdriver slips
- Use the starter motor to loosen the bolt once your wrench is secure (strips flywheel teeth)
- Put the car in gear and emergency brake on - your complete drivetrain play is actually quite springy
- Have a friend step on the brakes while you do the above.
- Use an air hammer _ probably damage something in the engine, drivetrain and your ears with all that pounding
- Heat bolt with a blowtorch for temperature cycling - probably end up melting rubber seals like the crankshaft seal and harmonic damper.
- Try to secure crankshaft pulley to part of the suspension which moves up and down for some reason ...
What worked for me:
After taking off the left front wheel (driver's side) and plastic shrouding:
On the front pulley there were 6 holes. I found a 2 inch long bolt that fit into one hole, skipped a hole and put another 2 inch long bolt in it. Just over 1 1/2" of the bolts stuck out from the pulley.
Looking at crankshaft pulley from driver's wheel well - showing bolts in holes
I placed a 2 foot long 1" dia steel bar (whatever was handy) between the bolts up against the pulley and the frame so that when the pulley turned, one bolt pressed on the top of the bar and one pressed on the bottom. The bar was long enough that the other end touched the frame. When the pulley was turned, the bar was jammed between the 2 bolts and the frame preventing rotation of the pulley. (this took 3 hours to figure out!)
Looking from the top at the crankshaft pulley (top of pix) and the bar jammed against the frame (bottom of pix)
Accessing the balancer bolt
Two 10" long quality 1/2" drive extensions resting on a spare jackstand
After breaking 3 cheap 1/2" drive extender bars, I bought two 10 inch long 1/2" drive extender bars (lifetime warranty) so that I could put a breaker bar outside of the fender. I set a spare car jack so that the breaker bar end of the extenders rested at the same height as the crankshaft bolt so that it would keep all the torque in the right direction and not twist out of place while turning. 5 feet of pipe on the breaker bar and a hefty pull and the bolt turned. Didn't sound like a bolt turning, sounded like something major broke but it was the bolt turning.
To tighten to 120 ftlb of torque, just reverse the steel bar so it jams in the other direction.
What worked for David and his 1998 Honda
I've received a lot of emails mentioning that later year Hondas don't have a balancer with holes in it so they can't jam the balancer from turning. David G. emailed me with his experience:
I was about to my wits end over that bolt. Then I found your page, shelled out 40 bucks for a breaker bar, extensions and some impact sockets, (I had an 8 foot pipe lying around) and cracked that guy loose with thundering snap. I was on the end of the pipe and had pulled it about 100 degrees when it broke loose. I really thought the breaker bar gave out.Mine is a 1998, and in case you have time you could mention on the site that it has no bolt holes in the balacer, so I was able to freeze the engine with a deep socket jammed in the flywheel where the shifter cable goes in. There is a little access panel there which came off with two bolts.
I put it back on with an equally powerfull twist in the other direction.
Turning Engine over the easy way!
Turn engine using spare bar on wheel lugs
The standard method of turning the engine to align the timing is to screw the crankshaft bolt in and use it. Instead, if it is a manual tranny, put the car in 4th or 5th gear and use a steel bar to turn the right side wheel using two of the wheel studs. Kind of the reverse of how the pulley was jammed. The wheel is off anyways and it is very easy to turn.
Another way to turn the engine over if the wheels are on the car and its a manual transmission is to put the car in 4th gear and roll it ahead or back a little bit. Make sure the parking brake is off otherwise its a bitch to move...
No Spark Fiasco!
After I got the timing belt in and verified that the valves were not bent. I fired the engine up. It started on the first crank. I drove around the block and noticed that I had no power in low rpms - this indicated ignition timing.
Timing 18 degrees off?
I checked the timing with the ignition timing jumper set and found that I was 18 degrees off. Unlike traditional distributers with vacuum advance, the computer controls the ignition timing. To disable the "vacuum advance", you jumper a connector located next to the ECU (computer). The ECU is located on the toe board of the passenger floorboard.
I tried adjusting the timing but there was only about 5 degrees of adjustment available from the distributer's slotted mounting holes. That's when the engine died with no spark. So far, it took one day to change the timing belt.
Timing that is 18 degrees off, indicates that the timing belt is off by one tooth - again thanks to the guys that frequent alt.autos.honda for that information. Got the right timing marks (discussed earlier) reset the belt and nothing still no spark.
Fixed one problem and then another shows up!
Anyways, to make a long story short, 10 days later, having worked 3-4 hours per day, I found that the new ignition coil had bit the biscuit. During those 10 days, I checked every electrical component and all the wiring of the car, including changing the ECU, rotor, cap, sparkplugs and ignitor 4-5 times. I knew all 10 pages of the wiring diagrams inside out - nothing made sense cause everything worked when tested or replaced. This was particularly frustrating as my field is teaching electronics and troubleshooting courses! Finally, I came across a Technical Service bulleting that described how to check an ignitor and coil on the vehicle.
It said to use a dwell meter, well my dwell meter didn't like the electronic ignition so I used a digital logic probe set for CMOS mode which is used to check microprocessors. I checked the output of the ECU while cranking and it indicated that there were pulses. I checked the output of the ignitor and it indicated that there were pulses going to the coil. This meant that the NEW coil that I had been relying on for the past 10 days as being good, had died! I replaced the coil for the second time and the car started perfectly.
Measuring Compression on a Honda
To measure compression, you need a compression gauge which has a hose that screws in place of the sparkplug into the engine. Normally, you just crank the engine over about 5-7 times and read the compression off the gauge. There's a couple of things that you need to do different for a fuel injected Honda.
- Disable the fuel pump - Before you do a compression check on a fuel injected Honda, you need to disable the fuel pump by removing the fuel pump fuse. If you don't, raw fuel will be injected into the cylinders washing the oil off the walls. This results in the oil rings not sealing and low compression readings. In the worst case, the rings will not seal and you'll measure no compression as if you have broken rings or a blown head gasket.
- Foot top the floor - When cranking the engine, the gas pedal must be pushed all the way to the floor or the throttle must be rigged fully open. This allows air to be sucked into the cylinder and compressed. Otherwise low compression readings will occur.
Compression test readings
When performing a compression test, higher values are better. My 1990 HOnda 1.6l measured 125 psi which is pretty close to the specs. Some high performance engines like VTECs may read 170 psi. The important point to note is that the compression reading should be very close for each cylinder within 10 % of each other. If there is a wide spread than that indicates that there is a problem with the bottom end and a leak down test (discussed further on this page) should be preformed.
Wet and dry - sorry not bathing suits!
There are two compression checks made: dry and wet. The two tests are identical except that before the wet test, a couple of shots of oil are squirted into the cylinders. You can also do static (engine not running) or dynamic testing (engine is running) Here's a very good step by step description on how to perform a compression test.
Curing a flooded fuel injected engine
If you suspect that your fuel injected engine is flooded. Push the gas pedal to the floor while starting. The ECU (electronic control unit) will detect that the throttle is more than 2/3 of the way open and shut off the fuel while starting for a period of time (I believe 15 seconds) to clear the flooded condition. Then it will apply fuel to start the engine.
Leak down testing
Leak down testing is a great way of determining the condition of the bottom end of the engine. The basic principle is to pressurize a cylinder and measure the rate in percentage that air leaks out. Each cylinder is tested one at a time at top dead center during the compression stroke where the piston is at the top of its stroke and the valves are closed.
The leak down test checks the sealing of the head gasket, valves and rings. All cylinders will leak and a good cylinder will leak about 10%. While a well worn cylinder will leak about 20%. A leak down test will indicate which area of the engine is problematic.
While performing a leak down test, remove all sparkplugs and pressurize one cylinder. If a high leak percentage is noted then listen for air escaping in the following areas:
- Adjacent cylinders leaking air through the sparkplug holes - this indicates a blown head gasket between the two cylinders.
- Dipstick hole leaking air - a little is normal, but a lot indicates bad rings
- Intake manifold leaking air - this indicates that the intake valve is not sealing or is bent or burnt
- Exhaust pipe - this indicates that the exhaust valve is not sealing or is bent or burnt
- Coolant as observed in radiator cap hole has bubbles - indicates that a there could be a blown head gasket between cylinder and water jacket or cracked head or block. Usually a cloud of white smoke comes out of the exhaust when the engine is run.
Here's the VFAQ (VTEC frequently asked questions) about leak down testing with nice pictures.
Finding top dead center (TDC) for a cylinder
I found that the easiest way to find top dead center for a cylinder is to put a long drinking straw into the spark plug hole. I've read quite a few articles and websites that say to use a metal or wooden rod. I use a long drinking straw because their soft and can't damage the piston or valves.
I rotate the engine by hand (see previous picture just before "Spark Fiasco" section) and when the piston is at TDC, the straw will have the least amount located in the cylinder. You will see the straw move up as it approaches TDC and then move down as it passes TDC.
An alternative way to find exactly TDC if you're engine is a V6 or V8 and the sparkplugs come off the side is to use a manometer as described on my DIY Kustom - How to find TDC webpage.