I've pulled the passenger side head (drivers ones proving to be annoying!) but i'm curious about some of the findings.

The engine was pressurising the cooling system, slightly, and using coolant, but drove fine and you could do 100's of miles in it a day without issue. Lots of short trips seemed to guzzle coolant, but a long trip would only use a little, suggesting it was leaking more when cold perhaps?

Spark plugs on 1 and 3 would turn orange over time. Its been like this for several years and i've finally decided to fix it as it seems to (finally!) be getting worse with some signs of water on the other bank.

I'm particularly curious having removing the head though, of how exactly its ended up the way it has.

Examining the head and pistons, you can see that 1, 3 and 7 all appear to have signs of steam cleaning, with the clean areas of the piston crowns and clean areas of the head. 5 looks pretty normal.

Moving over to the gasket, cylinder 1 has obvious rust on the firing ring, so we can imagine thats certainly a path between the water jacket and cyl 1. Oddly enough for the cylinder with the most obvious water route, its the dirtiest of the three cylinders. Cyl3 has a small amount of rust on the fire ring too. I guess i'm pondering how water got into these cylinders in the first place to rust the fire ring though... Especially 3, it doesnt appear to be near any water... Cyl 7 fire ring looks fine, infact theres nothing obvious at all around cyl 7, so not sure why its showing signs of steam cleaning. It is ofcourse adjacent to the water jacket, so it certainly could be leaking across from there.

Is there anything else i can check while its apart? Cylinder liners being the obvious question mark.

Ive put a stream of photos on imgur: https://imgur.com/a/PyyyvxJ

Its strange that they seem to act differently...

With mine, with the bad MAF connected up, it would start then immediately die. If you tried to keep it running with throttle it would produce clouds of black smoke and then typically still die a few seconds later.

However if you unplugged the MAF, then some portion of the time it would also start and immediately die, however if you gave it just the right amount of throttle it would stumble then run fine. A lot of the time it would just fire up and run normally.

I drove mine around all winter with the MAF unplugged, sorta forgot about it TBH, and you wouldnt have known for the most part. Then one day, wife drove it to work and it was utter crap. Stalling, stumbling, jolting, basically undriveable. I dont know why it changed. I plugged the faulty MAF back in and it started perfectly and drove home fine. That was the kick i needed to finish the converter though, and with the converter and Audi MAF installed its been spot on ever since.

I've stripped mine just now for head gaskets, was pondering re-routing the coolant for the vaporisor to use the 1/2" feed that used to supply the throttle body... Anyone looked at that route? The mess of T'ed pipes really bothers me, i bought new hoses and am loathe to hack them up, but series mode also gives concern as the vaporisor itself only has 10mm inlets, which feels like it would be very restrictive for the matrix.
Using the throttle heater loop seems like it could work?

I drove my GEMS for months with the MAF unplugged. Ran fine, and fuelling was more or less normal. A bit flat perhaps but certainly nothing untoward.

So yours not running properly with it unplugged suggests something else might well be going on...

With my bad MAF connected, it would often massively overfuel at startup and either chuck clouds of black smoke or not even start at all.

I ended up building a converter to use a MAF from an Audi TT, which has been in for a few months and seems to be working well enough.

I got it off, was definitely a tinting film. Horrible to remove though, it broke up into tiny pieces and wouldnt pull cleanly. Also left loads of horrible residue. Been over it twice with "Elbow Grease" cleaner and a going over with brake clean which has improved things, but i need another few passes over it to get it all off.

Okay. I suspect it's tinted then! Looking out it has a very distinct blue shade.
Must be a professional job as even with the plastic trims off I couldn't see any edge. The only giveaway (other than the peeling damaged but!) Is some halos around the "dots" at the edge of the glass.

I actually quite like the dark tinted look now it's on the car, but given it's damaged and only one window I guess it's better to remove it for now

I've just replaced by tailgate as the old one was rotten.

The glass on the new one is much darker, and has "opticool" on the info plate.

However there is a damaged area about the size of a thumbnail right in the middle where what looks like window tint has peeled off.

Anyone know if the opticool glass has a coating applied that could peel off? Or is it just normal glass and someone has applied standard window tint?

I don't want to start trying to peel it if it's supposed to be part of the glass! But if it's tint ill probably just peel it.

the O2 fault is specifically for the sensor heater. Suggesting either a wiring fault or a dead sensor.

The more you look into it, the less it makes sense.

For some rough figures, charging a battery and driving and electric motor from your electricity is around 90% efficient.

Using that electricity to create hydrogen, compressing it, shipping it to somewhere near your house and then transferring to your car, and running it thru fuel cell to power an electric motor is about 30% efficient.

So even if all those intervening steps cost nothing, it's 3 times the energy (which in the real world means cost).

Burning the hydrogen is even worse as instead of the fuel cell your using a combustion process which wastes 60-70% of the energy. So your down to what, 10% total efficiency compared to the simple battery EV.

Additionally, storing it onboard the car isn't easy, it requires special pressure vessels and the result is poor capacity to volume. A fuel tank can be moulded to fit under the seat etc, can't do that with a hydrogen tank. So even if you magically made loads of the stuff for nothing, fitting it all in a car to give even a couple hundred miles of range isn't possible, especially if your burning it rather than using a fuel cell.

The main use I can see is doing something with actually free energy. For example wind turbines and solar farms offen get pushed into "curtailment" which is essentially a mechanism to balance the grid by turning off generators. If you could instead divert that energy into a hydrogen electrolyser and store it up, it's literally free hydrogen. Nowhere near enough for the countries transport needs but still potentially a useful storage method for medium term energy storage.

Hydrogen for passenger cars is a white elephant.

The only reason its being pushed is the fossil fuel lobby wanting to keep their monopoly on transport. The vast majority of Hydrogen comes from steam-reforming natural gas.

Burning hydrogen is even more nuts than using it in a fuel cell. You put a tonne of energy into splitting the gas into hydrogen in the first place, then just burn it? You'd be far better off just burning the natural gas you started with, and its easier to move around and store.

The only place i see it being actually feasible is in things like marine or aircraft. Maybe some heavy vehicles in certain locations.

The fob is broken on my '94. It works fine with the key.

The only time problems start is if you lock it with the fob.

If you lock with the key, then it will unlock with the key and start just fine.

If you lock with the fob, and then unlock with the key, you end up immobilised and need the EKA.

At one point i was locking with the key and unlocking with the fob, that also worked fine, however at some point the BECM went in a huff and stopped listening to the fob and i gave up trying to get it working again.

Dragging this up from the depths...

In short, its working!

I'll need to upload the code to github or something.

In essance, the arduino reads the output from the Bosch MAF on one of its ADC inputs, uses that to lookup the correct value in a lookup table, and then sets the DAC output to the required voltage.

The Arduino's 5v regulator sends a 5v reference to the bosch MAF and to the DAC board. I've also incorporated a 9v DC DC converter on the input to the Arduino, just so its not having to deal with full alternator voltage (the included linear regulator has a max input of 15v).

Connected it up yesterday, and it works, drives great, if anything it feels a little more snappy on the throttle, lets hope it keeps working :)

sensor all swapped today

Cluster is installed and working as expected

GEMS: 260ohm at 90c (engine fully warmed up after a drive)
Thor: 280ohm sitting in a cup of water out the kettle (will be close to 90c but perhaps a little less)

Seems close enough for me! Same reading at 5c and same reading at 90c.

Now i just need to find time to fit it!

bad fuse box perhaps?

Seems like a lot of unrelated circuits losing power at the same time?

on a "traditional" vacuum referenced system the pressure will absolutely fluctuate, thats the whole point of the vacuum reference. On boosted cars the swings are even larger due to the boost reference. If there is 1bar of boost in your manifold, the fuel pressure rises by 1 bar to match it, to ensure injector flow rate remains the same.

My A4 runs a base pressure of 3 bar. however at idle its pulling 20 inhg of vacuum (about 0.6bar) and the rail pressure drops to around 2.5. At max boost it runs 1.5bar and the rail pressure rises to 4.5bar.

Early returnless systems, ran a fixed pressure, and the ECU compensates for the flow variations using a MAP sensor and a calibrated model of how the injector responds to the varying manifold conditions. Newer returnless systems have full PWM control over the pump and can adjust pressure at will as defined within their engine mapping.

Stick a pressure guage on your GEMS engine and you'll see it for yourself. Fuel pump running and engine off (no vacuum) will see about 2.5bar of fuel pressure. Once the engine starts and the vacuum acts on the regulator, it'll lower the pressure to around 2bar. Open the throttle and the pressure will rise back towards 2.5bar. Pull the vacuum line off the regulator and it will rise back towards 2.5bar. At wide open you'll see 2.5bar, and part load will vary between 2 and 2.5 depending on the exact levels of manifold vacuum.

I suspect my engine is a bit knackered 🤣

Were up to three rusty coolant stained spark plugs now, and one oil-caked plug

Head gaskets getting done as soon as my Audi gets an MOT and see where it goes from there?

Gilbertd wrote:

You'll need to add an extra ground too. The early GEMS had a single wire to the gauge sender with the ground through the body (later GEMS have a different sender with two pins so has a ground wire to it), the Thor one has a 4 pin connection, two signal and two ground wires.

Yeah I noticed the extra wire, I guess I'll fashion up a wire to the manifold to provide that. Still haven't managed to fully test the Thor sensor yet. Maybe tonight!

Pete12345 wrote:

Surely the easier method is to make sure the correct engine type is selected in the BECM. The BECM converts the analog reading from temp sender & fuel sensor to a value from 0-255, which then drives the gauges. It's more likely the A-D conversion changes than the gauge settings ?

But then there are many different instrument pack part numbers

My suspicion is the becm simply measures the voltage and sends the signal across to the cluster as a 0-255 digital value. It doesn't do anything to the reading other than measure and pass it on.

The cluster takes that digital value and applies a lookup table for the guage deflection.

The reason I suspect this is the becm diagnostic doesnt actually show the temperature, it only displays the digital value.

I tried resetting the gems engine type and writing settings again just incase doing that would somehow pass it on to the cluster, but it didn't make any difference.

is the fuel pressure regulator manifold referenced? If so its normal that it will fluctuate around.

The aim is to maintain say 3bar across the injector. If the manifold is under 0.5bar of vacuum, the pressure in the rail will drop to 2.5bar. If there is 1 bar of boost in the manifold, the pressure at the rail will increase to 4 bar.

Returnless systems are not referenced like that, but the ECM can alter the fuel pump speed via PWM, so you may also get similar variations in pressure as the pump adjusts its speed to meet demand.

I'd be looking at the three odd cylinders and figuring out whats going on. Might be worth getting the injectors properly ultrasonically cleaned and tested? Its usually fairly cheap and can be a good thing on an older engine anyway. A free test would be to move the injector to another cylinder and see if the issue follows.