Oil pressure fluctuation and AVCS fluctuation

[Chartreux]

Hello,

I'm back again with questions

I'm facing to strang oil pressure fluctuation. It don't happen every day, but sometimes, one morning, oil pressure decide to fluctuate a lot. One or 2 day after it's perfect.... I don't understand what happen

Firstly I though about a faulty oil pressure sensor, but I noticed that AVCS sensor RH (VVT1) read the same fluctuation but not VVT2.
So I assume oil pressure sensor is not faulty.
I've changed AVCS camshaft pulley by OEM Subaru ones and I changed both AVCS solenoid too by original new ones.

Oil pressure sensor is under the alternator, directly on the block.
VVT1 mean right side (driver side for RHD cars) for AVCS sensor
VVT2 mean left side (passenger side on RHD cars) for AVCS sensor
Oil pump is 12mm OEM

For the moment I have no "VVT fail" and ECU don't go in safe mode when running, even hard.

Any idea about the cause of such erratic readings ? maybe the relief valve of the oil pump is faulty ?

Here my readings :

[stevieturbo]

What oil pressure sensor ?

Is that or the VVT wiring routed near any HT wiring ?

If you unplug that VVT sensor, and you log the raw AN input....does it flatline, or are you getting any readings that may indicate interference ?

[Chartreux]

Oil pressure sensor is STRI :
https://www.stri-racing.com/replacement ... ure-sensor
Oil/Fuel pressure sensor , suitable to most universal application/gauges and other measuring devices* * 5VDC 0 ~ (+1,000) kpa 1,000 kpa (approx 10Bar) 0.5 ~ 4.5 VDC

I have coil on plug from STI version. The 4 coil wires are on a separate loom, dedicated to coil.

When I made the engine wiring, I routed the AVCS sensor together to the original GC8 grey plug. Then, from the car grey plug to the ECU, it's routed through the OEM GC8 harness

Today I made a new log, with raw Input Log :
AN03V = VVT1 = LH side
AN04V = VVT2 = RH side
I don't find any significant differences when I'm looking to raw values voltages, but still some fluctuation in the VVT signal



Sometime, signal is clean


And sometime, not

[TimH]

My hunch is this is similar to issues we see with coil-pack conversion looms supplied for 99/00 cars. Once fitted. there are some cars simply won't run, or idle badly, due to interference with the cam/crank signals - but no reason it similarly effect other sensors. It's only the 99/2000 cars that suffer this and rarely on those running Syvecs or OEM (I won't mention on this forum which ecu suffers the issue most!) - but there's definitely something "iffy" with this MY's engine loom earthing and ECU pinout.

The root cause (in our case) is down to the OEM earthing of cable screens being a bit suspect and/or it failing. The fix (in our case) is to locate the earthing joint for the cam/crank/knock sensor cable screens and add a new wire down to chassis earth.

In your case I would be suspecting a grounding issue of some sort but without knowing the method you've used to wire in the AVCS sensors it's difficult to be sure - and, to be fair, it's odd that it's an occasional issue, so it could be just an intermittent earth problem perhaps?

[Chartreux]

Ok, good idea. I think it could be the moment to make a grounding kit and see if it solve this issue.

TimH, based on your experience, does your customer solved their problems or, at least, improved their situation with grounding ?

Thanks guys

[TimH]

An engine "grounding kit" isn't going to fix this as sensors have their own analogue ground that goes back to the ECU and is - and should remain - independent of the main chassis ground.

Screened cables (cam, crank, AVCS and knock) need their cable screens connected to chassis ground back at the ECU and *not* to analogue ground, so any interference picked up on the screen is directed to chassis not into the ECU.

Yes, the fix has been used many times and always works when we see issues with our conversion looms (1999/2000 cars only, and rarely on Syvecs). But it does suggest that the OEM wiring is a bit susceptible to interference especially if you modify things!

[Chartreux]

I'm back again with my problem and now I have regulary VVT Fails and I'm very often in Limp mode.

I already changed both AVCS Solenoid and pulley by new OEM ones. I double checked the wiring and the connexion. Everything seems OK.
What I don't understand is that suddenly the VVT1 signal always stick to 55-56, even during very light cruise. I don't really understand why it suddenly stick to this value... and sometime all goes well....
I'm disapointed with this issue... any idea ???

[stevieturbo]

The two are not matching anywhere there, so what are the control duties for each during this time ?

And have all the correct trigger/latching points been setup so they may be matched ?

[Chartreux]

The mapping was not made by me, but by an official mapper.

I compared the values in the tables and for both VVT1 and VVT2 they are absolutely the same. (brake points, target, duties etc etc)

I zoom out the same curve and see that there is clearly a problem with VVT1. VVT2 is respect what is in the map, but I don't understand why VVT1 decide to stick to 54-55° value from on moment to another.


Earlier in the log, sometimes the VVT1 signal is OK and then very erratic and after come back OK..... after erratic.... and after few cycles decide to stick to 54-55° value which don't match the value in the map. For the breakpoint (map and rpm), VVT2 is ok, but VVT1 is not.


One thing interesting on this morning log :
When engine is cold, both signals of VVT1 and VVT2 are perfeclty in phase.
When EOT reach the temperature to exit limp mode, then VVT1 signal is very different.


Later on, same issue : VVT1 is suddenly stick to 54-55°.... I loose my mind

[Chartreux]

If it can help; this is the same log with different VVT parameters logged. Same horizontal axis scale