Unstable IAT Signal and Connection Issue

[Sodokat]

Greetings, 

I have been running my G4X on my '92 MR2 turbo for about 6 months now.  When I first got the car running I didn't have any issues with connection drops while PC logging or at least they were very rare.  The rate that I would see a 1407 connection drop steadily increased over time until it is now at the point the connection only holds if the engine is not running or below 3k RPM.

Strangely I have also started to see one of my sensors misbehave at the same RPM threshold as my ECU disconnects from the laptop. Almost exactly 3100 RPM both the ECU will disconnect and the sensor will start to increase in voltage (decrease in registered IAT temp).  The sensor's data if plotted shows a large jump in voltage and a lot of noise, the readout for the sensor drops about 40 degrees and varies around +/-15 degrees at about 10HZ

The sensor is a standard GM IAT fast response sensor (Delphi IAT).  I know that its position is not optimal for airflow, it lives in the stock cold start injector location and is a little shrouded, but it is showing a 30-40 degree drop in IAT when I get above the "breakup" RPM.  When I get a little bit higher than that it seems to lose some of the signal noise.

Things I have tried so far: Cleaning all the grounds to the intake manifold, checking to make sure my sparkplugs were firmly connected and not leaning on any other wires, I am in process on trying another IAT sensor as a double check (the old one had correct continuity and increasing in voltage seems illogical for a sensor with a bad connection.)

Does anyone have any suggestions? I can use the ECU logging to grab whatever data but neither the 5V reference or the other sensors seem to show any erroneous behavior.

Thanks in advance,

Caleb

[Adamw]

Sounds fishy.  I suspect this is two separate issues, but some of your observations of behavior are a bit odd.  

The temp input voltage increasing/temp decreasing can only be an increase in resistance on the ground side of the sensor.  With ignition on but engine off, if you shake/yank/wiggle all of the bits of related loom you can access do you see any variations in reported voltage/temp then?  

Does it have resistor spark plugs?  

[Sodokat]

I will double check that tonight and report back, on the wiring jiggle. I will also look again at the noise to see if it looks cyclical with the rpm.

Yea I didn't mention that, but after this issue started I figured I would try new plugs so I switched from oem iridium back to NGK v powers. The car seems to like the copper plugs more.

I know there was an update on the firmware, but I am pretty sure I have the latest. I tried a ferrite cable for the USB, I might try another laptop as well. The connection drop is quite annoying. I read up on a couple other threads that had similar issues with 3sgtes and this ECU. Any suggestions there?

[koracing]

Running the cable outside of the car taped to the body for tuning works best for me typically, and some cars are just noisier than other.  Also using an active extension cable (like Belkin or Tripp lite) also seems to provide a more stable connection.  Plugs, plug wires, grounds for the engine (primary ground from trans to chassis particularly) all have an affect it seems.  The grounds should all have good bare metal contact to chassis.  I don't think OEM never ran iridium on the 3sgte - factory used a platinum Denso plug as far as I am aware.  That said the NGK copper BKR7E are my go-to for this and frankly most sport compact motors.  I've had some minimal success also playing with the slow connection toggle setting and changing from auto to manual connection, and messing with the data transmission rate.  Make sure to also set up ECU internal logging to capture data you need as you can always get that off of the ecu with key on and engine off.

Would be nice if this issue could be resolved in a robust way before the units receive more price increases.

[Sodokat]

I will try the a couple more things with the USB cable. I just don't understand why it would get worse over time, I was thinking degradation of a ground? Who knows but I will check the transmission one next.

My mistake on the spark plug wording, it is whatever the non copper one that the owners manual call out. I followed your/ ATS instructions on plugs so I will be sticking with copper, engine idles much better imo at least when cold.

I was fidgeting around with knock control which made it helpful to see realtime data on the laptop. I can do it with logging but it takes a bit more time. Also was doing some freeway eco tuning from the passenger seat where it is very useful to edit tables real time

[Sodokat]

Well here is the vig guys.  Pulled the sensor and when you would flick it, it would show a decreased temperature (this makes sense for an internal short on the sensor but only under vibration). So I put in a real Delphi sensor, and the issue went right away.  I still think that the sensor is not far enough into the flow to be getting a good reading so I might tinker some more with that.

Here is the interesting part.  My logging ability also came back to the computer.  I went on a longer drive, probably 10 minutes, and didn't once lose connection.

I do seem to have a little bit of an issue breaking up at higher RPM, but now that I have logging back, I can look at that more closely.