RomRaider

This problem was an absolute headache to deal with, so I hope this helps someone avoid all the hairpulling

04 usdm wrx, hybrid block, 02 usdm ecu with carberry 4.2

- maf voltage while plugged in (via RR) 4.98v
- maf voltage while unplugged (via RR) 0v
- p0103 and p0104 (maf voltage high and low respectively) would not set
- p0113 (iat high) would set when unplugged
- would start with maf plugged in but bog and die once it finished cranking fueling IDC would shoot up, rich AFR
- would start with maf unplugged, but would simply fuel cut after cranking fueling. IDC would just go to 0

So, answers: the sensor reading 4.98v was due to a bad sensor ground. That much was known. The no start was caused by carberry. As far as I can tell, the maf codes p0103 and p0104 have either had their triggers set outside of possibility, or the logic is broken/disabled. As a result of not having a code, no failsafe fueling is enabled. I believe that carberry's ability to switch fueling modes and support for blending the two fueling modes ultimately is why failsafe fueling is nonfunctional. To get around this you can flash back to a factory logic tune OR switch to SD temporarily. I used port temp estimation and it was... ok. It works fine in a pinch, but getting consistent fueling with out a port temp sensor is a pipe dream.

Wiring resources I used:

- https://www.xcceleration.com/pinouts.htm
- https://www.wrxinfo.com/service_manuals/

My wiring nightmare. I was replacing the shielded signal wire for the maf signal. The section I was replacing went from the maf sensor to about where the loom turns and runs along the frame rail. I depinned the sensor so I could re-sleeve the harness and generally make it look nice once I'm done. The repair went well, the shielding was continuous and nothing was shorted. No start happened. This was caused by the bad maf ground. I would like to note here that backprobes for a multimeter are very helpful. In the end I tested all the maf sensor wires for continuity between the sensor and ECU - all passed, tested for shorts against all the other ecu pins and ground - all passed. At the maf plug, there was no continuity from the ground pin to ground. Plugged in and IG on, backprobing the maf ground and checking voltage to ground showed 10v. Bad ground for sure. Ultimately the break in continuity was not caused by the wiring, but by an ecu failure. (This is the time to click that image if it didn't load)

What caused that? I don't know for sure. What I think happened, is that when I took a break in the middle of the job to charge my dead phone i turned the ignition on, instead of just going to accessory. Because I had depinned the connector, I think that allowed the 12v maf supply wire to short to the maf ground wire. Its not supposed to see that kind of voltage, especially with 150ohms to ground and I think that broken pin is the result. Fortunately, I went to the computer repair/random 3d printing stuff shop near me and they were able to move the bottom leg up to meet the horizontal leg and solder them together. I don't know if this will have any lasting impact or if there's any other damage I don't know about. The maf ground now reads 100ohms to ground, vs 150ohm with my old ecu. I have needed to slightly rescale the maf, I don't know if that's due to the ecu oopsie or due to repairing the bad signal wire shielding, because I think I was right about that shielding causing my maf scaling issues. Since the repair I haven't been having any more weird scaling/fueling inaccuracies.

with 7volts on the MAF 5v rail it's possible you may have damaged the ECU internally and that's why its no longer reading any voltage (assuming the wiring is good).

I don't have a 16bit car handy but next one i see i'll try and remember to unplug the MAF and see what codes are generated. It's possible although the codes are on the high/low input codes may be have their values set to unachievable numbers so the code will never flag.


tuning speed density isn't that tough, it will just take you a little longer and of course without the air temp sensor it won't be 100% right but you should be able to get it running and add the air temp sensor later. I'd be more worried about damage to the ECU though.


as for the stock rom for the o2 ecu - have a look and see if its listed:
https://www.romraider.com/forum/viewtopic.php?f=35&t=7591

Ecu damage is something I'll keep in mind. Pretty sure the ecu is fine but its something I've considered so I did find my 2 spare ecu for the car. So I'll be able to work around that, but I'll lose carberry for a bit.

The maf is reading 0v right now because I very deliberately cut it behind where I thought the issue was. The real concern is the lack of the code, but I suspect your right about the codes being altered so they wont flag.

Going to try to cobble together a VE map that'll play well with my hybrid block and give that a shot. Frankly if that works I might end up fully converting to SD. Otherwise, stock rom or old ecu.

Hopefully I'll be able to post an update tonight.

OK, so confirmed that the car can run in SD mode

Replacing a large section of the maf signal wire, and confirming the shielding was intact for the full length did not cause the maf voltage to drop. Upon testing, I found 10v between the ground pin on the maf and a nearby chassis ground. I'm going to check the maf to ecu ground and see if I find high resistance.

I'm pretty sure the maf sensor works with a voltage divider, so if the ground on the divider went down voltage would get dumped to the maf signal line. I could be completely wrong, I have the MOST rudimentary understanding of electrical engineering.