TH3_5T1G

iTrader: (15)

I'll try to keep this as short and sweet as possible without missing any major details

I bought a built longblock that was unrun, come with nearly all sensors, all timed up etc.

I put all the ancillaries and put the new engine in the car. Fitted all that needed to be fitted and had Simon (JGM) come and put a base map on it. All was good bar a fault code - P0021...

Now, the more knowledged ones of you might already know that this code is stating that the AVCS is too advanced (more than the subaru tolerances between both banks) on bank 2. At the time of mapping it really wasnt too bad and we'd both agreed it was most probably the AVCS solenoid sticking.

So i fitted one of the sensors from my old engine (same part number) from which i knew to be in perfect working order. Still the problem persisted. I carried out the banjo bolt screen filter removal. In fairness, the filter on the drivers side had fallen out of the bolt and into the solenoid, crucially, this was one the drivers side which was not having any problems so the issue still remained.

Yesterday i thought i'd drive the car to my place of work so i could work on it in a suitable environment. Only now the issue has worsened. It now no longer idles unless eased off very slowly, so every set of lights saw it cutting out, every restart needed some persuading. After less than a mile i voted in favour of turning back. Here's where it gets interesting. I pulled my diagnostic machine out from the back of the cupboard and plugged her in once more (used to obtain the P0021).

This time running comparison graphs on the AVCS advance at 1000-2000 revs:


As you can see the left cam is advancing circa 70* all the time.

Anyone experienced this before? My worst fear is it's a cam sprocket seized for whatever reason

TimH

iTrader: (11)

My understanding is that the cams are advanced at low rpm - I thought by about 40 degrees - and the advance is wound out as revs increase.

Looks to me like the right is not advancing at all, and left is advanced perhaps too much but is working?

Possibly an electrical fault, so worth downloading the manual and looking for a diagnostic procedure perhaps?

TH3_5T1G

iTrader: (15)

I also assumed it had some advance at low rpm too but over on NASIOC they had comparisons and some were on 0 degrees too. Apparently its down to load x rpm but i "think" the right is ok as is. I can't rev it it too hard as its a fresh engine but imagine if i could the right would start to advance.

On the electrical side of things, i had one of the solenoid go open circuit on me about 2 years ago and it came up with a completely different code

Jolly Green Monster

iTrader: (4)

What happens if you unplug solenoids?

EngineMapper @ Group B Motorsport

On Idle the Cams should be reading no advance at all as they do exactly that, make the engine run rough and hard starting if they are advancing at low load. The cams are advanced usually aggressively on low rpm with increased load and then retarded back to wherever the map numbers tell it to go at high rpm with increased load.

It sounds as though the polarity could be wrong on a solenoid or as Simon says, see what happens if it's unplugged. If it idles and runs nicely after being unplugged then the solenoid is likely to be wired incorrectly, if it continues to run badly then the solenoid is stuck, but they are not the only two answers that may come from the effect of unplugging the solenoid!

Graham

TH3_5T1G

iTrader: (15)

Hi Simon, tried disconnecting one, running, checking codes and then doing the other side. The readings stayed the same, except for the Amp draw. Both do move in synchronization. But when one is disconnected and then reconnected it takes a little while to settle down to the same level. But as they're both drawing the same amount of current, i don't think either are sticking.

Cheers for input Graham. Solenoids have been disconnected, swapped over etc. I'm contemplating loosening the oil feed pipe to the passenger side solenoid, disconnecting the crank sensor and then cranking to see if there is oil supply to the solenoid.



Anyone know if the cams are advance through oil pressure or lack of oil pressure? In other words, is the sprocket in the advanced position and the oil pressure retards it back to 0 degrees or is it vice versa?

TH3_5T1G

iTrader: (15)

Simon, may be a daft question but did you disable AVCS for the run in map? Only reason i ask is i watched the cam advance again while just driving it 1/2 a mile to put a little bit of load on the engine and the right cam still didn't move from 0 degrees. Or is this the ECU preventing any advance as it know AVCS is suffering with some issues?

As for the theory of no oil to the left solenoid, i cracked the nut off, disconnected the crank sensor and turned the key. The pressure was so intense the spray even reached the bulkhead
Nice new engine a bit dirtier now but atleast i know there's pressure to the solenoid.

Are there anyways of carrying out actuator tests on the solenoids? Are they switched via a 12v feed or a 5v feed?

If it really isn't the solenoid and the pressures all good, i can only imagine it's either the cam is out of timing on the belt or something has seized the cam sprocket.

Jolly Green Monster

iTrader: (4)

Solenoid energises regardless of polarity so cant be wired wrong way around.

When did running in map I assumed it was solenoid but as you have swapped them from old engine it has to be physically stuck etc

Jolly Green Monster

iTrader: (4)

No not disabled

TH3_5T1G

iTrader: (15)

I did think, although it was a bit of a longshot, that maybe the one i had in there was stuck and so was the one i put in after. So i swapped them over and it's still the same side advancing like mad.

Do you think it's most probably the cam sprocket?

Jolly Green Monster

iTrader: (4)

You changed speed sensors too?

Jolly Green Monster

iTrader: (4)

Do both cam wheels bolte have the oil feed holes in?

merlin24

iTrader: (1)

Un-likely to be an oil feed problem to the solenoid as the oil feed also supplys the oil to the front cam journal on the inlet cam - would have siezed the cam/journal by now if you had a lack of oil.
The AVCS pulleys can stick if the engine has been stripped and the pulleys have been sat around for some time - be worth giving the AVCS pulley a 'gentle' tap with a rubber/rawhide mallet to see if you can release the internal locking plunger on the pulley.

Mick

TH3_5T1G

iTrader: (15)

No i haven't touched speed sensors. Could this cause issues with one cam and not the other or are you thinking about the right hand one not advancing?

Now this is where it gets tricky. I'm really wanting to start stripping bits down but worried about tampering too much incase i'd be voiding any warranty if it did, on the off-chance, be a build issue. Or do you not think this would result in such action?

Certainly sounds a possibility, will have to give that a try and buy a rubber mallet tomorrow. Main worry is if there is a tiny bit of debris in the cam sprocket

merlin24

iTrader: (1)

If the in-line screen filters are fitted in the banjo bolts then its highly unlikely debris will be inside the AVCS pulley.
You may have to give the pulley a tap with the engine ticking over as it will have oil pressure applied to the advance/retard port of the AVCS pulley with the engine running - it may free it off if thats the problem.

Mick

EngineMapper @ Group B Motorsport

PWM output polarity has to be correct to drive in the right direction. Is this on a standard ecu? It definately must be stuck though if they are still doing it disconnected. They're really simple pieces of equipment, just pop it off and check it over?

Graham

TH3_5T1G

iTrader: (15)

Cheers Mick, will give this a try tomorrow. I was more concerned about a bit of debris getting in there when it was lying around during the build

I've cleaned them out, swapped them over from side to side and changed them over for the ones from my old engine. I'm more than confident they're not sticking now. It is on standard ECU for the time being but plan to move to a Syvecs soon

EngineMapper @ Group B Motorsport

Ok this is a little more confusing then. They drive in the advance direction when driven. So with them disconnected or not receiving a signal from the output on the ecu they will be fully retarded. If the engine is definitely visibly rough you think as an effect of this problem then your solenoid is definitely stuck, otherwise it is a reading issue, cam sensor wise.

Graham

TH3_5T1G

iTrader: (15)

This is why i'd been thinking of cam sprocket seizure. Either seized in advanced position prior to install or small piece of debris in it, then when it first moved to the advanced position, the debris jammed the sprocket in that position

EngineMapper @ Group B Motorsport

Or this! lol

TH3_5T1G

iTrader: (15)

Lol I don't want it to sound like i'm not grateful for the advice, because trust me, I am. I just think i've managed to confidently remove a sticking solenoid from my diagnosis

Jolly Green Monster

iTrader: (4)

It doesnt really idle like cam is advanced however I set idle high due to cams to aid running them in so not as easy to tell

TH3_5T1G

iTrader: (15)

When you mapped it, it was nowhere near as bad Simon so it's nothing you've done, that i know

It was only yesterday morning it worsened after driving it a very short distance. I'm sure the advance on the cam when you mapped it was far less. I can only really think that the cam will advance but not retard, hence it having now reached 70 degrees advanced and not returning, whereas as the time of mapping, advance may only have been, say 30 degrees or so. Sound feasible? Or would it only advance to 70 degrees when giving it the beans? (i've hit just about 3500-3600 once and prosport gauge peak reading is about 0.05 bar of boost so i've still been gentle)

Jolly Green Monster

iTrader: (4)

That makes sense as to why it was fine. I dont recall it being that far out but we knew there was an issue hence the cel was on etc

TH3_5T1G

iTrader: (15)

Being a man of such broad knowledge, i don't suppose you know the maximum advance value for a version 8 sti AVCS system? Been looking for the answer on the net but only value i've found someone come up with was 30 degrees!!!

70 does sound like a massive advance when considering it's an interference engine, but 30? I do hope 30 isn't the max advance or i'm in trouble

Jolly Green Monster

iTrader: (4)

the max you notmally target really is about 45degrees.. the max it will physically go to... not sure..

the max reading.. not sure.

on the one hand if seems unlikely it is really at 70... so you could consider the speed sensor is reading wrong, however you said it is now lump... so therefore the cam must also be advanced..

did you try unplugging the solenoids yet?

TH3_5T1G

iTrader: (15)

I did unplug the solenoids while actively watching the cam angle. Disconnecting either sensor made no difference to the cam angle at all.

The speed sensors being the ones towards the rear of the heads? I could try to swap them over tomorrow but cam must be advanced a bit for such poor running. The other thing i'll have to do tomorrow is see if the timing belt is still aligned correctly on all cams. Not sure whether the tensioner was new or old going on to the longblock. If it is slightly worn, it could have the passenger side cam running a tooth behind as it would be the furthest driven cam from the crank given the direction of rotation. If the tensioner had worn and allowed a tooth to skip, surely it would be on this cam? Atleast i think that makes sense

I'm gonna get to the bottom of this even if it kills me

Thanks for the help so far

EngineMapper @ Group B Motorsport

Physical max whilst driving it at 100% duty I've not seen anything over 55degrees. So 70 is a bit of an odd figure to have. Ok so my logical problem solving would go like this, unplug solenoids, (done), leave them unplugged and unplug the problem side cam sensor and make sure the anomaly 70degrees goes away at least on your ODB reader. That then means the cam sensor should be reading, just to check swap them over and make sure you read the same figures you had before on each side and the problem isn't transferred. If all of these tests check out as they should then it must fundamentally be a timing problem or the physical cam advance system on the sprocket has jammed itself fully advanced.

Graham

WMS

iTrader: (2)

Ive had a similar problem, same fault code anyway, after many hours I eventually found the black plastic electrical connector on the back of the solenoid had been damaged, although it looked fine when i gave it a 'wiggle' it came away in my hand. Changed it and all fine, worth a look?

rickya

iTrader: (18)

I thought I had AVCS isues myself a few months ago & so did some research & went through some of the checks you have too. My issue was different & turned out to be a dodgy turbo (my AVCS logs were low but low across all banks due to turbo).

-Have you checked Cambelt to see if it has'nt jumped or partly broken idler?
-A long shot but have you swopped Coil Packs around?