Fuel gauge issue (was...Anyone know of a circuit diagram for the MIU?)

[sgould]

A few days ago SWMBO filled the Carlsson with fuel because the gauge was showing a quarter full. After topping up the tank and turning the ignition on. The fuel gauge stayed on zero and the fuel light was on. The gauge stayed on zero for a couple of days and then recovered.

Today it was filled with fuel again and when the car was started I noticed that the gauge was on zero again. When it got dark, I noticed that the panel lights were off as well. I would say a circuit fault. My electronic skills are low. But I could prod around with a circuit diagram.

Any ideas?

[aerojon]

you might be better hooking up on tech2,as the MIU is part of the I bus along with 14+ other modules..Lights and gauge are 2 different circuits,gauge side i wouldn't rule out sender unit..

[sgould]

We’re away for the weekend.  During the next short trip, the gauge was still dead, but the instruments were lit up.

I hate intermittent faults

[phoenix]

Both the fuel fuel gauge and the illumination rely on information from the I Bus but my instinct is that they're not connected as a fault on the I bus would affect far more than just those two.

I'd go with a fuel sender problem to begin with- it is an input to the Motronic ECM which then feeds I bus info the the MIU for the fuel gauge.

[sgould]

I’m beginning to suspect the sender. We are on our way home from Devon. The gauge was OK until we filled up. But seems to recover once the fuel level drops to about ⅔ full.

I will risk a refill soon while there are two of us in the car, and do a refill with the ignition on. Then I can see if the gauge stops reading as it gets high.

[sgould]

I tried filling the car with the ignition on, but the gauge didn’t rise with the filling.  Maybe we should have tried without turning the engine off.  We stopped the engine and turned the ignition back on without starting.  So back to square one with a full tank and the gauge on empty.   The gauge does move up a tiny bit with the ignition on.

The fuel consumption reading works all the time, but the DTE only shows a reading when the gauge wakes up.  It’s looking more like a sender issue.  It’s too wet to go and investigate at the moment.  Not that I can, as she’s taken the car to church…

[carrera]

A non bus suggestion (not got there yet):

Check the resistance from the sender (46)

According to WIS there's a connector on a bracket by the tank strap on the front edge of the fuel tank (H10-3)

You can measure betweens pins 2 and 3 from the tank, while there should be 5v from the ECU feed

Sounds like an opportunity to get the car lift out again ?

[phoenix]

There's a quite a long time constant applied to the gauge reading to prevent variations with fuel moving around in the tank so you'll barely see it move when filling if you leave the ignition on. It resets as soon as you turn the ignition off then on again.

[sgould]

Thanks for all the info.  I’ll try and look into it soon, but I’ve woken with a streaming cold and may spend time in bed instead.  It’s a pity because today is going to be the warmest day this week.

[sgould]

I've had a look at the circuit diagram, sitting in the warm, and I see that the XWD has two level sensors.  I'll check out the connector once I'm well enough to get under the car.

[carrera]

I didn't realise your car was AWD, lack of 9-3 product knowledge on my part

I assume that the tank is a different shape to allow for the drivechain intrusion, so two connections to check, rather than one. Intrigued as to how the levelling works & why the need for two sensors exists if the levelling between compartments works

[sgould]

It's an odd arrangement.  The tank sits over the propshaft like a saddle.  The pump is in only in one half of the tank.  There are pick-ups on both sides.  The fuel in the side with no pump is sucked through into the pumped fuel stream at a venturi in the pump delivery.  Very Saab...

[sgould]

It’s been a while, and I’m going to investigate this soon.  I may have suffered the local “100 day cough” that’s been going around here. 

The current regular, repeatable situation is that the fuel gauge will stay on empty after the tank has been filled.  That is, it twitches off the bottom stop and rises a tiny bit to empty, and the low fuel light comes on.  It stays like that until the car has been driven 60 miles or so, when it all wakes up and appears normal.

So it looks as if the gauge is OK and it’s getting false info.

[carrera]

Well, now I have an AWD there's a decent chance there is a similarity in approach / parts so if you need any comparable measurements perhaps I can help

Sadly, having looked, it seems that the NG 9-5 has the same defect of requiring the fuel tank to be dropped to replace the fuel pump 

The petrol 9-5 also has a "for life" fuel filter. This does not count as progress from my perspective

[sgould]

No. I've had a further look at all the bits on WIS and EPC.  I now think it must be a fault with the level sensing. But there are two level sensors, one in each half.  They are separately connected to the ECU, but share an earth return, also via the ECU. They are claimed to be variable resistor types.  I was expecting them to be PWM like a lot of Saabs, but then remembered that it's a Bosch system in the V6, so maybe there's a difference.

A sthe ECU is involved, I may be able to find some info via Tech2.  I was wondering why the gauge dropped to zero when there were two  level sensors, but maybe it's a failsafe system controlled in the ECU, which also bodes well for a Tech2 check.