EJ22 random idle, pig to start - what else?

[silverbullet]

Used to play around with old Rovers and efi's - I did wonder if I could perform the old flap meter fix of bending the contact finger to increase pressure and improve the s/n ratio, but with the cam for the switch on top you can only just about clean it anyway.
I will lift the inlet manifold too, just in case I'm getting a leak there. It's definitely much worse when hot and I know from my race engine work on Subaru manifolds that they do warp a bit with age. Should have thought of it before!
Will report back - all good for the symptom spreadsheet.

[RJES]

If you don't mind it being a bit risky you can easily check for induction leaks by spraying carb cleaner around anywhere that there is likely to be a leak with the engine idling, withouttaking anything to bits. The high vacuum at idle will draw it in through any leaks, and you'll hear the rpm increase.

[silverbullet]

Found it!
Was all fired up to take off the idle valve and as I took off the breather connections...the short 20mm bore hose right on the crankcase split right round, level with the bottom clamp! Must have been going for a while...
Explains the "worse when hot" symptoms as the rubber softened up and allowed it to breathe more.
Dropping the engine a bit (is my scooby engine too high?) must have taken some pressure off the intake pipe that was managing to keep the split (mostly) closed.
The two breather pipes that feed the PCV on the plenum were scaled up with carbonised cack too. Some of this was flaking off and finding its way up to the PCV - maybe causing it to get stuck occasionally? Can't have helped.
Live and learn. Again. What a muppet.
Now to fit that new throttle cable...

[RJES]

Glad you found the problem. These sort of problems are usually caused by simple stuff like split or missing hoses, loose connectors, etc. The kind of thing that gets caused by disturbing the engine from it's original home in the Subaru.

Thinking about it, I took one of those breather hoses off the black plastic T piece in the top of the block, and it cracked all around. The rubber had gone hard and brittle. I thought nothing of it at the time, but maybe it's something to look out for in the future?
One for the Troubleshooting section on my site I think. I seem to be spending more and more time trying to help people sort this kind of problem out. I'm wondering whether I can come up with some sort of troubleshooting flow chart.

[silverbullet]

A flow chart for trouble shooting would certainly help everyone out. The oem subaru manual's 1 to 15 priority list doesn't help all that much.
BTW I haven't checked to see if the code 49 has cleared - it could have been caused by the unmetered air from the split hose falsing the ecu. Will let you know.

[RJES]

Yes, I would imaging that a big induction leak is likely to cause an MAF error code.

[silverbullet]

Latest in the saga. Idle much improved but still has a tendancy to rise up to 1300 rpm when hot, but will suddenly drop back down to 850/900 after 5-10 seconds, but not always. No more air leaks, have checked everything!
Reading through my downloaded manual carefully, I found just one line that says that VSS were fitted to early air conditioned legacys as an exception. Then it dawns on me that my motor used to have the compressor fitted and working but was removed by the previous owner bar one.
Must be that I need a VSS after all.

[RJES]

All Subaru's from 1990 --> have a VSS. What is the part number of your ECU (the long number which starts with 226)? Some 1990 - 94 models run without a VSS quite happily, but not all. After that they pretty much all need one, although the severity of the symptoms you get without one varies massively between different ECU's. I keep a list of which symptoms you get with each ECU without a VSS and how severe they are.

[silverbullet]

[quote="silverbullet"]I'm working with this:

ecu 22611 AA670 A11-000R97 (very early, integral fault code lamp)

quote]

Thanks Richard.

[RJES]

I've not had much feedback about that ECU. Under exactly what conditions does the idle rpm increase?

[silverbullet]

OK. Starts first time and idles at about 900/950 (bit quick I know) Doesn't seem too rich from experience.
Blip the throttle and the revs are slow to return to idle, drops to 1300/1400 ish immediately but then takes about 6-8 seconds to get right down to 900 again.
As the engine warms up while driving, the idle speed slowly creeps up each time you have to stop, until the thermostat is open and then if you get caught in traffic and the fan is cutting in, the load from the fan pulls the revs down by about 200. ISCV still at fault?
Sometimes (but not always) the revs will stay up for about 10-15 seconds, then drop down to 800 but the idle is then lumpy. Drive on for a bit, and all's well for a bit but not long.
(I have fitted one of your throttle cable kits and spent about 6 hours extracting the lash-up and routing the new one correctly. Car also has new HT leads, Fuel filter and fuel lines, but the fault pre-dates all of that.)

[RJES]

I If the idle drops from 900 - 950 rpm when cold to more like 600 once warmed up (with the engine just idling, and not moving), then that suggests the ISCV and the rest of the idle speec control system is working OK. So the idle only becomes instable after driving? If so it's probably speed sensor related.It doesn't sound like any of the three common symptoms of not having a VSS with an ECU whic isn't happy without one, but it wouldn't surprise me. No error codes?

[silverbullet]

Will check for codes maybe tomorrow as it's out turn for the monsoon right now.

The other odd thing is that the throttle pot (4 wire) has a value of only 2k ohm (give or take) when wide open. The book says 3-5k ohm, but I can't adjust it more than a few hundred ohms before the idle switch goes out of range. It's an A22000R25.

It won't stand full throttle under load as it starts to hesitate. Back off just slightly and it's ok. Don't know why.

[RJES]

All the manuals I've seen set the throttle position sensor by measuring the voltage output rather than the resistance. As far as I know, they are all wired as a potential devider with 0V at one end and 5V at the other, so that actual resistance of the track is cancelled out. There's a thread on vwkd called something like 'MOT fail - rich mixture' at the moment which I posted the Subaru procedures for setting the throttle position sensors (3 and 4 wire versions) a couple of days ago.

If it doesn't work properly at full throttle, then I'm pretty sure there is something more wrong than whatever is causing the idle problems.

[silverbullet]

Have joined in the "failed mot emissons" discussion on vwkd as someone also has the same full throttle misfire / hesitation, wondering about the fuel pressure regulator's health now.