Outback ignition fault

[Daniel33]

Hi everyone. 

 

I am a newby here, but have been trying to get onto the site for months. Each time it came up in a google search, the return timed out and I couldn't get a page open. So, hopefully, I can get back on to read the responses to my question, but just a heads up that it might take some time to respond or even read any answers.

 

So, here is the question.

 

I have a 2006 Outback that I bought for my son. It had blown the head gasket on the left hand bank some time ago, and also had a frozen thermostat which locked all the heat into the left hand bank. The previous owner was not very savvie, and used some silver seal to attemp a fix.

 

I have no idea how long he drove it for with the locked themostat, but it had pushed the valve guides down on the exhaust stems, and they were quite carboned up so I would say it was a while. 

 

Anyway, I bought a 2004 wreck for spares, and after getting the '06 apart, found the block to have softened, so took the block and left hand head off the '04, changed the camshaft over, re-seated the valves, and put the entire thing back together, using the '06 cam and crank gears. 

 

I also had to advance the valve timing, as I was a little over-zealous with the sander when I cleaned the heads. I had just done the head gaskets on my daughter's '07 Forester, which had a performance exhaust, and had the valve timing 6 teeth advanced, and it runs just fine!!! 

 

This '06 was running fine before the head work, but now shows a very strange fault.

 

It is firing the coil on numbers 2 and 4 at around twice the rate it should be. 

 

I can see this with a timing light. 1 and 3 are sparking in sync with the engine sounds, but 2 and 4 seem to be operating at about double the speed.

 

The only thing I did that I probably shouldn't have, was that I broke off one of the teeth from the crank wheel when I was removing it. I have glued it back in place with some epoxy metal filler. Can't remember the name of it off hand, but I tried to get it to move towards a strong magnet while it was wet, and it didn't move, so I assumed it was non-magnetic...(I know JBWeld ismagnetic).

 

It is running very rough at idle, pops and farts when cold, stutters and hesitates down low, but launches at around 3000 rpm as it should.

 

We have checked the coil pack connection, and re-fit the main loom plug a few times, as well as checked the leads to make sure they are all pushed onto the plugs properly.

 

What could be causing the high spark frequency???

 

How would it even be possible?? I have not put a scope on the crank sender, to check for a frequency match with the power leads to the coils, but that is next if no-one has any solutions. It is the only thing that could be causing the issue, as the cam only has four lobes to trip the sender.

 

What causes the coil to stop firing??? Is it ECU magic?? Why would it be both 2 and 4??

 

The initial test drive was done without a front AFR sensor on the vehicle, and it was moved a few times without either the O2 or AFR sensor connected, but I don't think this would have killed the ECU.

 

Any help will be smiled upon. 

 

Thanks.

 

 

[montana tom]

I suspect the crank gear with the broken / repaired tooth. Also the left side cam gear must match the ecu.  Crank gear and cam gear must stay with the ecu. 

[Daniel33]

Thanks Montana Tom. I did keep the crank and cam gear together with the 06 ECU. It might be the repaired tooth, I agree. I might have to accept I need to source another one or attempt another repair, this time with some non- metal repair epoxy resin. I just don't understand how it is increasing the spark frequency so high. I read on another forum of an ignitor for the coil pack, and that it could be somehow at fault. It lives somewhere on the firewall but can find no sign of such a thing. Ever heard of this? Could it be the problem? I am in Australia, if that makes any difference.

[montana tom]

I would source another crank gear rather than trying another repair.   Does the comp show any codes ?  In Australia you are offered different options than we are limited to in the US .  Twin turbos ... diesel engines ...  they even have slightly different names, but they are still all subarus under the sheet metal. I don't  think the outbacks here used the  coil igniter at all,(i could be wrong)  I remember them on 90's impreza's . Many others out there with more knowledge than me about electrical issues ,hopefully one of them can help you.      

[Daniel33]

No codes are coming up relating to the ignition. I am getting codes for the AFR sensor, but I know that is faulty. I will have a look locally. We are in the bush a bit, but you can get anything in the post nowadays. 

 

Thanks for the input.

[Fairtax4me]

You'd be better off welding the broken tooth back on (Theoretically at least). Replace the crank gear.

A break in a magnet creates a different magnetic field. Same thing applies to the broken tooth on the crank gear. The magnetic field of that tooth is weaker now and the ECU is reading a gap in the AC wave coming from the crank sensor. That gap is messing with the ECUs calculation of the crankshaft angle and speed of rotation, and throwing off the timing of the ignition command, or skipping a spark command.

The coil fires in pairs. So two plugs get spark at the same time. 1-2, and 3-4. So on a normally commanded ignition cycle you'll see each plug fire twice as often as it should.

 

Igniter on the 06 is built into the ECU. Not on the firewall. Pre2000 cars had a separate igniter on the firewall.

 

You advanced the valve timing? How far?

6 teeth is 45° advanced. Works out to about a 22.5° change in valve timing compared to crank/piston position. That would mess with the spark timing enough at idle to create a misfire.

[Daniel33]

You'd be better off welding the broken tooth back on (Theoretically at least). Replace the crank gear.

A break in a magnet creates a different magnetic field. Same thing applies to the broken tooth on the crank gear. The magnetic field of that tooth is weaker now and the ECU is reading a gap in the AC wave coming from the crank sensor. That gap is messing with the ECUs calculation of the crankshaft angle and speed of rotation, and throwing off the timing of the ignition command, or skipping a spark command.

The coil fires in pairs. So two plugs get spark at the same time. 1-2, and 3-4. So on a normally commanded ignition cycle you'll see each plug fire twice as often as it should.

 

Igniter on the 06 is built into the ECU. Not on the firewall. Pre2000 cars had a separate igniter on the firewall.

 

You advanced the valve timing? How far?

6 teeth is 45° advanced. Works out to about a 22.5° change in valve timing compared to crank/piston position. That would mess with the spark timing enough at idle to create a misfire.

Hi Fairtax4me.

 

The crank wheel is a cast iron item. It could be welded, but the little knob would simply melt away. Also, I do not have a Tig welder. The heat would probably mess with the metallic domains in the material as well.

 

I am not sure that the wheel is a magnet though, but rather, a lobed metallic inductor. The hall effect sensor, (the crank and cam sensor), "see" the lobes and react, closing a switch, creating a pulse for the ECU. The material it is made from has it's domains evenly spread, so there is no polarisation, which would mess with the sensor. I will test this today. I will be removing it, and doing some tests.

 

The coil pack is showing 2 and 4 at a high frequency. Are you sure you have quoted the correct pairing of 1-2, 3-4? Good info on the paired firing, I hadn't thought of that. I thought the firing order was 1,3,2,4 though. This would put 1 and 3 piston oppositely opposed, and firing together, one on compression, one on exhaust.

 

If I remember correctly, the cams are twice the diameter of the crank wheel, therefore, they turn at half the speed. 6 teeth works out to be around 11.75 degrees if I did the calculation correctly, but I may be slightly off. The forester runs fine with the valves like this, and, as the valves now hit the piston on the standard setting on this engine, I thought I would give it a try. It spins up ok, and has plenty of push. No chance to get a mileage calculation yet though with this other issue.

 

I will do a video of a scope probe and put it on youtube if anyone is interested.

[Fairtax4me]

There are only two coils in the coil pack. Plugs on each coil are fired at the same time. Polarity is opposite on the plugs so the spark on one side actually comes from the ground electrode to the center electrode. Causes funny wear patterns on the plugs and necessitates the use of double platinum plugs for the best wear resistance. Called a waste spark ignition.

 

Firing order is 1-3-2-4

Cylinder order:

3-4

1-2

Front

 

Opposed pistons push away from each other then towards each other. YouTube/Google animations?

 

11.25° would be right if you advanced by 3 teeth.

Cams do run half speed of the crank. Cam sprockets have 48 teeth, timing change would be 7.5° per tooth. 6 teeth = 45°. Divide by two... If it runs fine that's great. Just seems like that would make it less efficient to be that far off.

 

No the crank sprocket isn't magnetized but there is a magnetic field induced in the sprocket by the sensor. I don't think those little teeth make much of a change in the magnetic field of the sensor without having the mass of the rest of the sprocket behind them. The broken tooth will not carry with it as much of the field from the sprocket. Might be wrong. Would be interesting to see the wave on a scope to find out for sure.

[GeneralDisorder]

The crank sprockets are less than $40 retail from the dealer. That's less than a half hour shop labor. Isn't worth even attempting to repair one.

 

You really should check replacement part pricing before you sink buttloads of man hours into repairing cheap parts. No real excuse for not exploring the obvious option of just buying a new one.

 

GD

[Daniel33]

Well, the fault with the ignition was that the repair I did on the crank wheel had failed. The lobe had fallen against the next one, when I was fumbling about with the glue, trying to get it to set into the grooves properly. So, I unpicked the goo, and used some five minute epoxy to fix the tooth to the wheel first, then, re-enforced that repair with more resin against the neighboring teeth.

 

It seems to be holding. It is now idling smoothly, and operating at low revs under load as it should.

 

It is the weekend here, and although the retail price may be low in America, there is no guarantee that is the case here in Australia. I don't pay workshop fees. I do all my own repairs, as best I can.

 

Once I got the engine running properly, I took it for a test drive, and it overheated. The previous owner had put so much silver seal in it, he had blocked the radiator core. So, I had to use the radiator out of the '04 model, but that was an auto, so I had to modify the top mounting brackets, and the air scoop to fit it.

 

It seems to be OK on power, with perhaps a slight backfire upon deceleration, but no sign of any hesitation on the power.

 

Here is a vid of the scope. Turn down the speakers. 

[Gloyale]

In the US you couldn't put an 04 head on an 06 engine.

 

here, the 06 had Active intake valves, and a crowned piston top.

[Daniel33]

Turns out I had the years wrong. It is an '01 block with '04 heads. The car is 2004.

[Daniel33]

Hmmm. My posts aren't sending.....

[Daniel33]

Perhaps it is the use of an equals sign in the text??

 

Unless I can't quote yet.

 

So, this is to reply to Fairtax4me.

 

 

 

Hi Fairtax4me.

 

Thank for the info. You sound like you know what you are talking about, and I don't wish to start an argument however, doesn't the crankshaft turn two revolutions for every cycle of a cylinder? 

 

This would make the math:

 

48 times 4 equals 192 teeth.

Then, 360 degrees divided by 192 equals 1.875 degrees per tooth.

So, 6 times 1.875 equals 11.25 degrees advanced.

 

Magnetism is something of an interest of mine, and, having had a look at the sprocket with some magnetic viewing film, I can attest that it is not permanently magnetised. However, for it to be inductively magnetised by the sensor, the sensor would have to be a solenoid. It is not, as the scope trace demonstrated to me, it is not constantly powered. Thus, it is not magnetised by the sensor, but rather, by the ground plane of the alternator. The large spike in the scope trace was due to there being two teeth in close proximity, effectively doubling the flux the sensor "saw", and thus increasing the signal. I don't quite understand why the levels were so different on the sensor though when the induction clip was on number 4 coil lead as compared to number 2, unless the clip itself was transferring some energy to the number 2 lead, when hooked around the number 4 lead, and it was feeding back into the sensor. I didn't try hanging the clip on the other side, and it was resting on the number 2 lead.

 

The field induced in the sprocket would not be a very strong one, but it is enough for the lobes on the edges to trip the hall effect sensor. These are very sensitive devices, able to pick up small changes in magnetic fields, and, when delineated by such edges as are found on the sprocket, they can act as enough of a switch to operate the transistor powering the coil packs. At least, that is my understanding, but I could be wrong.

[Daniel33]

Of course, I made a mistake....I was counting the crank revs for the cam, when the cam is half.....so, yes 22.5 degrees. Man, how does it run so good??

[GeneralDisorder]

I can tell you FOR SURE it isn't 6 teeth advanced on the valve timing if it runs. More than 2 teeth on a camshaft will bend intake valves. It also would run badly if it's 1 or 2 teeth off. Milling the heads has no significant impact on this. If the intake still bolted up without modification then they didn't get milled too much, and it's not possible for any of the sprockets to be off by 6 teeth and have the engine run. You must be using the wrong timing marks.

 

GD

Edited October 5, 2017 by GeneralDisorder

[Gloyale]

...............doesn't the crankshaft turn two revolutions for every cycle of a cylinder? 

 

This would make the math:

 

48 times 4 equals 192 teeth.

Then, 360 degrees divided by 192 equals 1.875 degrees per tooth.

 

you are counting all "4" strokes as rotations.  they are not.  1 rotation is 2 of the 4 strokes, so 2 rotations is 4 strokes.

 

 

But then even, the Cam rotates !/2 the speed of the crank.  So 48 teeth is the full rotation of 4 strokes.

 

so the math is 

 

360/48 = 7.5 degrees per tooth.  6 tooth advanced is 45 degrees of camshaft advancement.  22.5 of crank 

 

Which if IIRC, is about the distance between the arrows and the hash marks.

 

my guess is you are basing yer stats off of the arrows.

 

Because like GD said, it will not run well 6 teeth off.  The Cam and Crank sensors would be so far out of sync it would throw codes.

 

And "too aggressive with the sander"?  you'd have to take of massive amounts of material and then intake wouldn't bolt on. And why the hell are you using a sander on the heads?

 

Hope all goes well for ya cause something is off.

[Fairtax4me]

 

Magnetism is something of an interest of mine, and, having had a look at the sprocket with some magnetic viewing film, I can attest that it is not permanently magnetised. However, for it to be inductively magnetised by the sensor, the sensor would have to be a solenoid. It is not, as the scope trace demonstrated to me, it is not constantly powered. Thus, it is not magnetised by the sensor, but rather, by the ground plane of the alternator. The large spike in the scope trace was due to there being two teeth in close proximity, effectively doubling the flux the sensor "saw", and thus increasing the signal. I don't quite understand why the levels were so different on the sensor though when the induction clip was on number 4 coil lead as compared to number 2, unless the clip itself was transferring some energy to the number 2 lead, when hooked around the number 4 lead, and it was feeding back into the sensor. I didn't try hanging the clip on the other side, and it was resting on the number 2 lead.

 

The field induced in the sprocket would not be a very strong one, but it is enough for the lobes on the edges to trip the hall effect sensor. These are very sensitive devices, able to pick up small changes in magnetic fields, and, when delineated by such edges as are found on the sprocket, they can act as enough of a switch to operate the transistor powering the coil packs. At least, that is my understanding, but I could be wrong.

No the sprocket is not magentized. The crankshaft sensor IS magnetized.

The crankshaft position sensor is not a Hall effect sensor. It is a magnetic pickup sensor. It has a permanent magnet core, with a copper coil. It is not powered. The sensor produces an AC waveform as the lobes of the sprocket pass through its magnetic field. That waveform gets sent directly to the ECU. The ECU uses that waveform to determine crankshaft angle and speed of rotation.

The crankshaft sensor does not trigger the ignition coil. The ECU determines crankshaft angle and commands the ignition control module (IIRC is built into the ECU on that year car) to trigger spark from the coil.

If you probe directly at the pigtail connector for the coil (I don't recall exactly which 2 wire2 off hand) you'll see a DC square wave as the ignition module pulls the coil circuit to ground. The break in the ground causes the coil to spark.

 

Probing the crankshaft sensor wires directly at the sensor connector for an AC signal you'll see the AC waveform from the sensor. With the proper settings on your scope you'll be able to easily count all of the teeth on the sprocket by looking at the waveform.

 

 

With an inductive clamp on the spark plug wire, you're seeing spark voltage going to the plugs, not ignition trigger from the ECU.

Spark voltage will vary depending on many conditions, but primarily by plug gap and cylinder compression. At idle you'll often see the spark voltage increase and decrease with slight changes in idle speed. If you see a constant difference in spark voltage between one plug wire and another, that's usually because the plug showing the higher voltage has a wider plug gap, and the energy in the coil has to build to a higher voltage before it can jump that gap.

[Daniel33]

Hi Guys.

 

Thanks for all the great info.

 

Firstly, met me say, I am not rich, and it is pretty common practice to use a steel bottomed belt sander on cast iron heads and blocks to remove old gasket material around these parts.....it may not be too cosher, but it works. I thought I would try a softly softly approach on the aluminium as well. It seems to have worked OK. I have done it to two cars now. One has traveled about 250 ks with no signs of trouble, and the other has done a few thousand by now.

 

The crank angle sensor is not something I am not familiar with, so I am listening to other's advice and wisdom on that. I am only a hack with a scope...I couldn't get it to show me an AC waveform from the sensor at all, just a DC spike. Perhaps (most likely) I had the settings wrong.

 

But, the timing of the cams, I can do.

 

I had the 1 cylinder at the top, crank mark on the sprocket (the small arrow, not the large line, as determined by piston position) lined up at the top, with the vertical line of the block seam, (no idea what the deep line mark is for, as it comes after the cylinder is at TDC) and had the 1 and 3 bank cam sprocket lined up with the mark at the top and the seam at the head/camgear transition. All OK so far.

 

Then, I used a socket on the other bank to attempt to align the timing mark, and the valves were hitting. Now, "that's weird" I thought. So, I moved the crank back, (counter-clockwise) about 30 degrees (being mindful of the open valves on number 3), and set the cam to where is "should" be and brought the piston up slowly. Still hitting......damn. "What is going on???" I thought. So, after checking and re-checking everything, I went to previous knowledge.

 

Remembering where the Forester engine was set when I pulled it down to do the heads, This engine had someone spend money on good plugs, and extractors, I tried rotating the cam gears on both banks 6 teeth counter-clockwise from the standard position, thinking I had perhaps removed too much material from the heads with the sander. (I don't know what the tolerances are for the head to valve clearance).

 

I then turned it over gently by hand a few times, feeling for valves crashing into the heads, and loss of compression too early or building up compression too late, (as best as I could). For this, I used a T-bar and socket, so I could control the throw over TDC on each cylinder. I couldn't feel any issues, so, confirming the timing was the same on both sides, I put the timing tensioner and bottom pulley on, and the intake manifold, plugged it all in, and gave it a kick in the guts.

 

It fired straight up after priming the fuel rail with a few short hits on the starter, and ran, so I continued on my way, until the miss-fire issue emerged once everything was properly back together and it had warmed up, prompting this post. (The 5 minute epoxy seems to be holding fine BTW).

 

The engine runs superbly. It has a definite torque increase at approx 3,500 rpm in second gear. (haven't booted it in 1'st, as it would just leap too fast). It pulls well in each gear, with the curve shifting it's harmonic upwards in each successive gear by about 1100 revs. (Hard to gauge while driving, but I can't afford nor justifiy a dyno test). It revs to 6500 rpm easily. I can hear no sign of, and it is not throwing any codes for, a miss-fire nor backfire. 

 

Fuel use around town shows high 7's, but the long range average is at 14.2 k's per hundred. Not too sure how long that will take to reflect a true average, as it has only burned through around 20 liters of fuel so far.

 

The engine had done about 240 000 k's before going into this car. It had been abused by the previous owner, who ran it with cheap crap oil, for a long time, with leaking head gaskets. I changed the sump and oil pickup tube over.

 

All I did to it was re-seat the valves and clean off the carbon from the top of the pistons.....oh, and took off the back of the valve seats with a dye grinder on the intake ports where they were protruding into the port.(best to leave any on the exhaust ports, as the ridge actually assists with creating a vacuum)...that is all the porting I did.....ohh, and I shaped the back of the valves to remove any ridges as well....all by hand with a dremell and grinding stone. I had to swap over a couple of valves too, but I reground all the valves by hand and eye, and set them into the seats with grinding compound and a drill......took me a few hours, but I enjoy that kind of thing, and the results are worth the effort.

 

I don't have a workshop manual, and so have to think my way out of everything. It usually works, but can get me in trouble too, and cost me time. Luckily, I don't have to pay for that. The one I can't do too well, (by that I mean, at all), is electronics.

 

Daniel.

[Fairtax4me]

The notches are the timing marks.

Arrows are used only for TDC for setting valve lash. Not used for timing. http://www.subaruoutback.org/forums/#/topics/39436?page=2&gid=1&pid=1

[Daniel33]

Hi Fairtax4me.

 

So, I used the correct mark for the valves then. There is only a single divot on the cam sprockets. I used the arrow on the crank sprocket, as that is where the piston was at TDC. So why does the notch on the crank come so much after TDC if it is for ignition timing? Do these engines use controlled advance that has that large a curve? Interesting if that is the case. So, the timing curve begins at full retard and is computed forward depending on fuel load and engine load. This would allow for deceleration ignition timing so far before TDC, there would be no chance of any ignition, and no back-fire, wouldn't it?.

 

However, for it to do this, the computer would have to be one cycle behind real engine time, using the previous cycle's data to fire the engine.

 

Perhaps that is what the flat spots are for on the cam sprocket? To allow the computer to register crank position. I saw no sign of these blank segments on the scope trace, but like I said, I probably had the scope set incorrectly.

 

All a bit of a mystery to me but thanks for the input.

[Fairtax4me]

It's not for ignition timing. It's just a marker for setting the cam timing when installing the timing belt.

With the crank at 90° off TDC the pistons are all at half stroke, so no chance of bending valves/interference if you need to turn the cams.

Also with the cams set at 45° off TDC they can be left in position without needing a special tool to hold them. Makes belt installation much easier at the very least.

 

Every other manufacturer makes you set cam/belt timing with the engine at TDC, which for many engines makes sense, but in many cases also requires special tools to hold the cams in place while the belt is off. SOHC heads are usually not much problem, but DOHC heads are always an issue with other manufacturers. On Subaru there's nothing to them, you just line up the marks and put the belt on.

[Mike104]

Crankshaft Alignment (Note:  This shows a Crankshaft gear that may be different than yours, but the alignment points are the same)

 

 

 

Camshaft Alignment

 

 

Edited October 10, 2017 by Mike104

[Daniel33]

Ok. Thanks for that. Seems like today I learned something new...

 

So, the TDC mark is followed by the cam timing mark, which is 90 degrees behind it.

 

I incorrectly used the TDC mark, and set the cams 6 teeth advanced, as that is where the other engine I did was. (I didn't pull this engine down, I left that to the kids). This worked out to be 22.5 degrees advanced on the cams. BUT, now it is 90 degrees retarded on the crank.....so, the question becomes, where is the timing set to??

 

Seems like it might be correct after all??

[Daniel33]

Well, we have removed the radiator again, and checked the timing. Turns out that the arrow mark on the crank sprocket works out to be 6 cam teeth away from the correct valve timing mark, so it is all good.

 

First time I have ever come across something like an engineer actually thinking about ease of fitment of this level. Outstanding!!

 

I have fitting the cam belt with two people down to a fine art now. The next one will be a breeze!!

 

Thanks for all the help and advice everyone. Those pictures were the clincher.