Miata Turbo Forum - Boost cars, acquire cats. - Keep braking LSU 4.9 sensors, help.

Hi,

I tried getting useful info on Fakebook, but I haven't found the solution and the Fakebook topics are quickly deluted by irrelevant info, often. I hope someone here can help me out. I run Speeduino on a 2004 1.8 VVT engine I fitted into my 1998 NB, ditching the old worn 1.6. I have a Spartan 2 controller and the Bosch LSU 4.9 sensor.

Sensor 1 gave up, only giving full lean signal. I blame my bad tuning, this was when I first got the engine going and my tune was all over the place. Also, I started the car 3498572398584208 (approximately) times, before getting it running properly and driveable, sometimes coughing up black clouds. There is still a black mark on the garage wall behind the car, I keep it as a memory.
Sensor 2 gave up, again only giving full lean signal, a few months later. I was told it was from the cold starts. I should wire and program the system so it would only fire up the Spartan 2 20 seconds after the engine started, to avoid condensation in the exhaust hitting the hot sensor element. OK, I programmed an extra output (pin 23) and wired a MOSFET to that pin, which in turn switches a solenoid powering the controller. It is programmed to switch 20 seconds after the engine goes over 500 rpm (and stays there). Meanwhile, I have a pretty good running tune except for cold start.
Sensor 3 gave up in the same way. I was told the blame lay with running fuel cut on decelleration, which gives a cold shock to the sensor. I doubt that, since factory cars do the same and run this sensor without issue.
Sensor 4 gave up again in the same way. Because, well, I don't know.

I was told three more common failure causes are all related to cold fuel dropplets hitting the hot ceramic sensor element. Startup, warm-up enrichment and accelleration enrichment would be the causes. With the startup delay, I avoided the first. My warm-up enrichment setting are slightly richt, but nothing terrible. May accelleration enrichment was way off, but to the lean side. This is next to fix on my list, although someone on the Speeduino Tuning Fakebook group already advised me a change in my setting which greatly improved driveability, so I think I'm on the right track here even before measuring.

Those were all Bosch sensors from a reputable source. I tried two AliExpress sensors, since they cost € 20,- instead of € 90,- for the Bosch ones and a fellow Dutch Speeduino user had success with them. Both gave bad readings right out of the box and I found similar problems in another topic online. So, we will not count those.

I am currently thinking in three directions:
1) Ditch the current controller and sensor, and go for an LSU 4.2 or LSU ADV setup. Those supposedly are more robust. But it would be a sizealbe investment again, and what if that doesn't solve it, the cause remains, I keep braking those sensors, too?
2) I put the Ali sensors to good use as temperature gauges. I have checked yesterday in the kitchen, using a multimeter, that the resistance of the heater element changes significantly with temperature. Of course with the sensor completely disconnected from any controller or power source. The readings I got were:
4 degrees C (in the fridge); 3,5 ohm
100 degrees C (in the water cooker to make tea); 4,9 ohm
200 degrees C (in the oven, very untrusworthy value, but anyway); 5,7 ohm
250 degrees C (same untrustworthy oven); 6,3 ohm
So next I will today measure the other Ali sensor, which is still in the car to plug the bung, and see how quickly the sensor is heated by exhaust gasses alone. I have the 20 second startup delay, maybe that is too short. If the sensor reads something above 5,5 ohm, I think I can be sure all moisture in front of and at the sensor location is gone.
3) Sensor position. I have the 4-2 stock manifold from the 1.8 VVT, without a sensor bung. I have the 2-1 downpipe from a normal 1.8 (a 10th aniversary, to be exact, bought second hand), which also came without a bung. I welded in a bung on top of the tube, after the merge. I think it is not only pointing the sensor down (which is good) but also maybe slightly forward. Meaning whatever is flung at the sensor is blown into it. If I cut and reweld the bung still on top of the tube, but at an angle so the sensor points backwards (wires forward), maybe less dropplets of anything are flung into the sensor. I know the gasses will swirl around and still enter the sensor, I hope any droplets will be too heavy to swirl and continue on their way.

So, that is where I am now. I will to the warm up measurement this morning, and maybe remove and cut/weld the downpipe somewhere this holiday week as well. Still, I fear I haven't found a culprit, no smoking gun. I doubt all my sensor were duds, Bosch makes milions of those. I also have some claims of people using the Spartan 2 and the sensors succesfully for many years, still on their first sensor, even without the warm-up delay. One even on a Speeduino controlled MX-5 1.6, quite local to me.

Please let me know if you know the issue, have a suggestion, etc. Or join the search if you have a similar issue, maybe a common denominator can be found more easily with more faults collected in this topic.

Cheers,

Hugo