CIS Basic -> Invent Jetronic

SteinOnkel

Got it working.

The recommended coils are junk. I don't know if all of them or just the two I have here. I plucked a coil off and ignitor off my wife's car and it sparked immediately.

Happy Sunday everybody!

P.S: New issue. It is injecting so much fuel, that it drips out of the exhaust. What should I look at first?

JHM

SteinOnkel
I just checked the coil I have in the office, it's a cheap ERA 880001 replacement coil.
It draws 25mA when pulled low.

Anyway, we will redesign the PCB to address this inconsistency of coil parameters.

About dripping exhaust - probably it has a lot of fuel from previous start attempts. Warm up the engine to evaporate the excessive fuel, and then check the lambda corrections.

SteinOnkel

Good tip about the dripping exhaust. Could be, although we were spinning the distributor by hand with the fuel pump off while testing. The car ran for about 10-15 minutes yesterday. We decided to shut it off when we saw the dripping fuel, that seemed a little risky 😉

Is there anything in my settings (attached) that do not look right for a 2.0l 16v that could cause this?

way-too-rich.imf

10kB

Also, please find attached a screenshot of the non-inverted transistor circuit we were testing. Works really well, pulls 12.7v to 0 no problem. Not sure what the amps were, I think we calculated 100mA.

Larttis

SteinOnkel Hello, what is the purpose of this circuit you were testing and did it work?

JVR

Are you sure you have the 107 injectors installed? Also is the coolant sensor reading correctly?

SteinOnkel

Yes, quite sure. The fuel injector part numbers are: 0280158107. My pressure regulator is the one invent gave me (4 bar, no vacuum).

The temperature sensors are both reading correctly.

It will not stop dripping. I've disconnected and aired out the exhaust overnight. The strange thing is that the car runs beautifully. Starts great, revs up easily. But the fumes and smoke make it impossible to drive. Total runtime so far is about 20 minutes. That should be enough to burn it off, no?

Hmm. Very puzzling!

SteinOnkel

New update:

Found an injector stuck open.

Question to the developers: the blue wire coming from the invent on the injector harness is constantly grounded. Confirmed with multimeter and test light. This is why injector #2 was stuck open.

I suspect that a coil driver may be bad (wires ohm out okay). I tried using the orange wire to trigger the injector, but that is never grounded unless I tell the software I am running 5 cylinders. Then the ignition timing is wrong and the engine won't start.

Can I use one of the working injector trigger wires to drive two injectors? Or will that burn up the ecu? I am running batch, so the timing of injection is irrelevant. This would be two injectors in parallel and ohm's law says the resistance is then half. That should be okay or...?

Progress at last 🙂

SteinOnkel

Hello!

We are back with another update.

We have checked and double-checked the wiring. Everything is good.

If we plug an LED into injector connector #3 and turn the ignition on, the LED lights up immediately and stays on. If we repeat the same test with the injector connector #4, the LED stays off until the starter is activated. Then it flashes. There are other symptoms that I won't go into which tell us that injector #3 was always on.

We have now opened up the ECU. On the board, we can see 7 mosfets, all in a row. 6 are identical with one being different.

Our theory is that Q3 is the ignition driver and the others are the injector drivers.

After we look at some datasheets, we will measure and see if we can locate a fault with the driver for injector #3. If we do, we will replace it with one of the working ones we do not need (i.e. injector driver 5 or 6).

If they measure out good, that indicates the fault lies with the input side. It is beyond our ability to desolder and resolder microprocessors. In that case, we would like you to send us a replacement ecu.

Thanks 🙂
Manuel (landscape architect) and his dad (electrical engineer)

SteinOnkel

One more update for tonight.

Q1, 2,3,5,6,7 have the following measured values:

Source to drain: 0.548V
Drain to source: OL

Q4 has the following measured values:

Source to drain: 0,309V
Drain to source: 0,309V

Q4 goes to A5 header pin -> Injector #3, our problem child.

Q7 appears to go to A2 header pin -> Injector #6, not needed on this engine.

We will now attempt to desolder and change Q4 for Q7.

Thanks
Manuel

P.S: In my post #26 I accidentally said Injector #2 is stuck open. This is an error, it is injector #3.

JHM

Wow, so many updates for a single day 🙂 You're really skilled.

I'm sorry that you encountered a faulty MOSFET. We test each ECU at the bench with a burn-in MOSFET test. It seems that Q4 survived the test and failed just afterwards.

It's right decision to swap MOSFETS.
I hope it will finally run as it should

SteinOnkel

Hello,

you say I'm skilled - my smd soldering tells a different story 😃 The dead mosfet fits nicely with the dead ignition coil, doesn't it? So hard to find good components these days.

Tried to desolder Q7, a pin broke off. Tried to desolder Q4, the trace came off the circuit board. We thought that was game over and we had killed it. As a last effort, we soldered a jumper wire from the output of R13 to the input of Q6.

And it actually works! The car runs on all four cylinders now.

I will report back if we find any other issues, but for now everything is work as it should.

JHM

That was a close call 🙂

I'd recommend to remove R21 to avoid interference, and to secure the wire with thermal glue

SteinOnkel

Hello!

A few more questions:

The big check valve by the fuel pump, is it still needed with invent? Or can I remove it?

And the lambda adaptation, can that be done with logs? It’s very difficult to drive this car with a laptop.

Just completed its first hour on the race track 🙂 It’s so fast now!

Habe a great weekend.

Thanks
Manuel

JHM

I don't know which exactly valve do you mean, but I guess it can be removed, as long as fuel is delivered.

You can adjust VE table by logs, but it is not safe to drive a car on a track with a lean mixture.
Adaptation won't cover full-throttle area.

SteinOnkel

JHM

We ran the day with Adaptation disabled. It performed really well. I am tempted to do nothing at this point - never change a winning team 😉

The valve is actually called the "fuel accumulator", as seen here: https://www.stoddard.com/47720908300.html

The fuel accumulator has two functions:
– After the engine is switched off, this accumulator keeps the fuel system under pressure to promote a warm start.
– The accumulator dampens out the fuel pulses generated by the fuel pump. A large spring-loaded diaphragm to keep fuel pressure up with the engine off also dampens pulses from the pump.

The problem is that here in California, because of ethanol, these things like to leak. And they are very expensive to replace.

SteinOnkel

Larttis

The purpose was to pull a small pulse (5-10ma) and pull it to ground with a higher current (tested and works to 100ma).

It did work, but it was not needed. Two ignition coils were bad.

JHM

The fuel rail is kept pressurized by the pressure regulator in the filter. So you should be fine without accumulator.

SteinOnkel

That sounds good! If it leaks again, I will get rid of it.

So, we still have a mystery with the ignition coil. I ordered two more cube coils (intermotor part # 12916). No spark with either of them 🙁

The only coil/amplifier my ecu likes is VAG 357905115 that one always works and the car runs beautifully.

Very strange! Any ideas?

As for the adaptation: the ecu simply always reads lambda fail and the value is stuck at 0.45V. I've used multiple sensors and triple checked the wires. I will look with my wideband 02 sensor and gauge what the value is at full throttle. If it's around 12.5:1 AFR, we will not worry about adaptation or lambda. For a race car perfect values really are not needed, just as long as it never runs too lean.

JHM

It looks like the only coil which is made according to specs is VAG. To say more, I need a scope data from your coils.

Why won't you use WBO signal in the ECU? Narrowband isn't useful for a race car at all .

SteinOnkel

The WBO is in a little engine tester I made. We use it to diagnose our cars and I don't want to spend $250 on a wbo for the race car right now.

Which oscilloscope data would you like us to measure for you?

JHM

SteinOnkel
I need EMUL output signal, connected to the coil that don't work.

« Previous Page Next Page »