SR20DET MAF adventure

Paint, body preparation and modification, interior work and electrical
User avatar
jovial_cynic
Supporter
Posts: 628
Joined: 23 Jun 2003 15:49
Location: Central Valley, CA

SR20DET MAF adventure

Post by jovial_cynic »

I *think* I've narrowed down my probem to the MAF. When unplugged, the engine runs very well, and my wideband O2 gauge reads around 12:1 to 13:1... so a little rich, but ok. When I plug the MAF back in, the engine leans out entirely, causing the engine to stumble and die.

The first thing I did was try to find any potential leaks, because I'm aware that unmetered air will cause a lean condition. Test #1 was the soapy-water + air compressor test. I sealed everything up, opened up one of the ports on the intake for the air compressor, and sure enough, I found a few spots here and there that bubbled up from the soapy water. Knowing that I might not find all the leaks, I went out and built a smoke machine so I could find any additional leaks, and sure enough, I found a couple of them and promptly sealed them. But this has not resolved the issue. When I plug the MAF in, the engine still leans out and dies.

Figuring that the problem exists between the MAF and the ECU (either the sensor itself or the wiring leading to it), I needed to rule out the wiring. I read that the MAF puts out around 1.7v on the sensor wire at idle, so I built a voltage regulator circuit and directly injected 1.7v down the pigtail on the sensor wire and fired the car up, and sure enough, it worked. Just to make sure the car wasn't in limp-mode, I adjusted the voltage to see if my O2 gauge would register the difference, and it did. So the wiring is good.

So that leaves me with the MAF. I spent some time cleaning it (CRC, rubbing alcohol, gentle q-tip massage, etc.) and then ran power to it out of the car to see if I get a sense of the readings. With some alligator clips, I ran power to the MAF and grounded the appropriate pin and then checked the voltage on the sensor pin. With the car not running, the MAF sends out 1.2v. With the car running (but running with the 1.7v directly to the ECU, not with the MAF), and the MAF bolted back in place, the MAF voltage increased from 1.2v to 1.8v... which puts it within spec. So everything seems like it should work.

HOWEVER.

After disconnecting the voltage injection circuit and plugging the MAF in and starting it up, it runs fine for about 5 seconds and then leans out again, causing the engine to stumble and die.

So... back to the drawing board. I didn't measure the signal from the MAF while it was plugged it in, but that's my next step. I want to watch to see if there is a voltage drop after the first 5 seconds, causing the engine to lean out.

The only thing I could think of is that the regular O2 sensor (the one connected to the ECU, not the wideband I'm reading) is throwing something out that's creating some kind of feedback loop, causing the MAF to drop voltage... and maybe the only reason I'm able to maintain a constant idle with my 1.7v injection is because it's not going to change with any O2 readings.


Hm. So maybe the first thing I need to do is to unplug the O2 sensor and see if that resolves things. It's a pity I'm at the office right now and can't provide an instant status update on this.
71 510 2-door SR20DET (S13)
------------------
http://www.myagentjosh.com
http://www.newprotest.org/510
User avatar
S15DET
Supporter
Posts: 2561
Joined: 03 Dec 2004 17:02
Location: Waxhaw, NC

Re: SR20DET MAF adventure

Post by S15DET »

Nice troubleshooting! Looking forward to your results.
I wish the SR had an air temp reference, because it seems like I have seasonal headaches with the mixture and I end up cleaning the MAFS.
User avatar
jovial_cynic
Supporter
Posts: 628
Joined: 23 Jun 2003 15:49
Location: Central Valley, CA

Re: SR20DET MAF adventure

Post by jovial_cynic »

Thanks. I LOVE troubleshooting. And I look forward to getting this tired SR20DET (which needs a rebuild anyway, sigh) back on the road. I've had it running (terribly) a few times before I had any gauges installed, so I'm pretty sure I was beating the motor up even more. Now that I'm fine tuning everything (ie., micro-managing), I'm a lot less tolerant of less-than-ideal conditions for the engine.

I built this a while ago:

https://www.youtube.com/watch?v=Mw6Cf_TSfcA

I can make it display the readings on whatever I want and can extend it to an arbitrary number of sensors so I can view everything. Instead of bogging down the arduino code with each individual sensor, I going to reserve one input that'll pull from a multi-switched source. That way, I can have tons of sensors at the push of a button and display whichever one I want, but keep the "main" ones as a part of the multi-gauge screen.

As far as an air-temp reference... obviously you can't make the ECU do extra math, but I wonder if you can throw in a small voltage-tweaking circuit onto the MAF sensor wire so you can just slightly elevate or lower the voltage with a dial on the dash. Being able to adjust the MAF signal up and down seems to be a fairly simple solution as long as you keep an eye on the O2 readings.

Oh... actually... if you wanted to get really clever, I'm sure somebody has done the math already for it... you can build a circuit that modifies the MAF signal on the fly for you based on air temp.
71 510 2-door SR20DET (S13)
------------------
http://www.myagentjosh.com
http://www.newprotest.org/510
User avatar
PoorMtnKid
Supporter
Posts: 1219
Joined: 28 Aug 2010 14:43
Location: Las Vegas NV

Re: SR20DET MAF adventure

Post by PoorMtnKid »

I upgraded to a z32 maf just a few months back so if you are looking for a known working sr maf its all yours.
looking for pass. side arm rest
User avatar
jovial_cynic
Supporter
Posts: 628
Joined: 23 Jun 2003 15:49
Location: Central Valley, CA

Re: SR20DET MAF adventure

Post by jovial_cynic »

Sweet. Once I yank the O2 sensor and see if that resolves the issue, I'll let you know. From my readings, it seems to be behaving correctly (unless there's a big voltage drop after a few seconds of running... but again, that SEEMS like an O2 sensor issue, not a MAF issue), but we'll see. Thanks!
71 510 2-door SR20DET (S13)
------------------
http://www.myagentjosh.com
http://www.newprotest.org/510
User avatar
okayfine
Supporter
Posts: 14154
Joined: 12 Nov 2007 23:02
Location: Newbury Park, CA

Re: SR20DET MAF adventure

Post by okayfine »

I don't see an O2 sensor interfering with the MAF signal. ECU can run the engine without an O2 (or with a bad one) in open-loop - and in fact does so on those early emissions cars for most of the operating range. IIRC it only looks at the O2 during cruise, and maybe idle. Which I know is the problem you're trying to diagnose...

If you can pull voltages from the MAF, can you also do the same for the O2? You should see the sensor values cycling pretty quickly. A dead MAF will go into limp mode, but a dead O2 shouldn't kill an engine, in my experience.

Also, the FSM has a MAF troubleshooting section. Might cross your checks already done with the checks they specify.
Because when you spend a silly amount of money on a silly, trivial thing that will help you not one jot, you are demonstrating that you have a soul and a heart and that you are the sort of person who has no time for Which? magazine. – Jeremy Clarkson
User avatar
jovial_cynic
Supporter
Posts: 628
Joined: 23 Jun 2003 15:49
Location: Central Valley, CA

Re: SR20DET MAF adventure

Post by jovial_cynic »

Yeah - I can pull voltages from anywhere.

So, I know the MAF isn't dead - it's providing the right voltage when I'm reading it directly while it's bolted up but unplugged from the ECU. It's reading 1.7, 1.8v which is precisely where I want it to be at idle. And while it was sitting there running, the voltage never suddenly dropped (which, if it was connected to the ECU, would cause a lean condition).

What else would cause it to go lean after about 5 to 10 seconds of starting the car while connected to the ECU? The only thing that makes sense to me is that it took the O2 sensor about that long to heat up and somehow advise the ECU that it was running WAY too rich... which would cause the engine to lean out and die. Taking out the O2 sensor, in theory, should, should prevent it from sending a rich condition to the ECU if that's what is happening.

Sometimes a bad sensor is worse than a dead sensor, because the ECU has a map for dead sensors. It doesn't always compensate well for unexpected behavior.
71 510 2-door SR20DET (S13)
------------------
http://www.myagentjosh.com
http://www.newprotest.org/510
User avatar
PoorMtnKid
Supporter
Posts: 1219
Joined: 28 Aug 2010 14:43
Location: Las Vegas NV

Re: SR20DET MAF adventure

Post by PoorMtnKid »

Shoot a PM over to jeff ( Icehouse ) . He is pretty amazing when it comes to these SR engines and their electronics. He helped me find out that my ECU was the issue after sever hours of troubleshooting the TPS. Turned out my ECU had been in a flood and was covered in mud inside and all over the board.
looking for pass. side arm rest
User avatar
jovial_cynic
Supporter
Posts: 628
Joined: 23 Jun 2003 15:49
Location: Central Valley, CA

Re: SR20DET MAF adventure

Post by jovial_cynic »

Welp, unplugging the O2 sensor did nothing. So that's not likely the problem. ha.

I tested it a ton, and it looks like it is pushing a constant 1.7v at idle through the volt meter, but once I touch the MAF signal terminal to the signal wire, it leans out. Interestingly, after the car has warmed up quite a bit, it doesn't lean out so much that it dies... just enough to make the O2 sensor freak out as it goes 15:1... 16:1... 17:1... and then pegs off the proverbial chart. No bueno.

So... I think what I'm going to do first is provide some constant voltage that underlies the MAF signal. Maybe my volt meter isn't picking up some kind of voltage loss down the line that I'm missing. If I provide that underlying voltage, and the MAF only adds voltage to the ECU, maybe that'll keep it above the lean condition I keep experiencing.

And also... maybe the fact that I only recently hooked up my alternator (which means I've been running just slightly less than 12v until today) is a variable I didn't consider. Sigh.
71 510 2-door SR20DET (S13)
------------------
http://www.myagentjosh.com
http://www.newprotest.org/510
User avatar
S15DET
Supporter
Posts: 2561
Joined: 03 Dec 2004 17:02
Location: Waxhaw, NC

Re: SR20DET MAF adventure

Post by S15DET »

This is awesome!! I love troubleshooting too, but haven't been able to do it much lately with other demands. The Arduino project is GREAT! I'm an EE, and have been wanting to learn more about these easy microprocessors, raspberry pi, arduino, etc.

If you end up not needing MtnKid's SR MAFS, I would sure like to have it for troubleshooting and experimenting.
User avatar
okayfine
Supporter
Posts: 14154
Joined: 12 Nov 2007 23:02
Location: Newbury Park, CA

Re: SR20DET MAF adventure

Post by okayfine »

jovial_cynic wrote:I tested it a ton, and it looks like it is pushing a constant 1.7v at idle through the volt meter, but once I touch the MAF signal terminal to the signal wire, it leans out.
What happens to the voltage when you do this? Seems that the ECU is interpreting the signal wrong (leaning out) or the MAF is not doing what it should when connected to the harness.

How are your grounds?
Because when you spend a silly amount of money on a silly, trivial thing that will help you not one jot, you are demonstrating that you have a soul and a heart and that you are the sort of person who has no time for Which? magazine. – Jeremy Clarkson
User avatar
jovial_cynic
Supporter
Posts: 628
Joined: 23 Jun 2003 15:49
Location: Central Valley, CA

Re: SR20DET MAF adventure

Post by jovial_cynic »

PoorMtnKid - I'll for sure take that MAF. I want to at least swap it out to see if it makes any difference. If it doesn't, I'll send it your way, S15DET.

okayfine - the MAF's output appears to be 1.7v as expected when air is sweeping across the sensor. I could be wrong, but I imagine that the grounds are fine because my voltage regulator circuit works perfectly; the grounds there have to be good, and the ECU grounds have to be good in order for that to work properly. And just to make sure that there isn't some tiny bad connection between the plug and the sensor itself, I ran alligator clips between the three terminals to the plug and simply connected/disconnected the sensor wire itself to simulate plugging it in and unplugging it. Same problem. I believe that rules out any wiring issues.

But... I do think something is up w/ the MAF. When I directly injected the 1.7v PLUS alligator clipped the MAF, it idled just fine, but when I opened the throttle a little bit, it bogged and died on account of no additional voltage going to the ECU (ie., no fuel being injected). I'll fiddle around with it a bit more this morning to see if I can discover more, but in the meanwhile, it doesn't make a ton of sense to me.

Just for fun, I'm going to throw in an LED between the MAF and the plug so I can see right away if there's a significant voltage drop. With my 1.7v voltage regular circuit, the LED shouldn't change at all, but with the MAF, it should light up initially w/ the 1.2v with the car not running, brighten up when the MAF is supposed to put out 1.7v, and in theory, brighten up even more when I open the throttle a bit. I'll see visually what happens.
71 510 2-door SR20DET (S13)
------------------
http://www.myagentjosh.com
http://www.newprotest.org/510
User avatar
okayfine
Supporter
Posts: 14154
Joined: 12 Nov 2007 23:02
Location: Newbury Park, CA

Re: SR20DET MAF adventure

Post by okayfine »

jovial_cynic wrote:and in theory, brighten up even more when I open the throttle a bit. I'll see visually what happens.
This is what I was suggesting. Perhaps when connected the MAF voltage changes to indicate to the ECU that less air is passing through, hence it needs less fuel. Because the engine's still actually turning idle rpm, air/fuel goes lean. Or in some other way presents the ECU with something it's not expecting. Since everything bad happens when the MAF is connected... 8)

LOTS of grounds in the system. MAF even has a ground. Parts of the system can work fine, but one circuit will short across another seeking ground.

Brad, if the OP keeps PMK's MAF, I have an extra SR MAF - it has a dent in the top shield (top plastic cracked). Doesn't look like the board itself is damaged, but I never tested it since I went with the 240sx harness/B13 ECU and MAF route.
Because when you spend a silly amount of money on a silly, trivial thing that will help you not one jot, you are demonstrating that you have a soul and a heart and that you are the sort of person who has no time for Which? magazine. – Jeremy Clarkson
dislexicdime
Posts: 652
Joined: 09 Aug 2004 18:26
Location: tigard oregon

Re: SR20DET MAF adventure

Post by dislexicdime »

Do you have a consult cable? Best thing i ever bought for trouble shooting my sr
L series only have one header!

i need another garage mine is full of part's
User avatar
okayfine
Supporter
Posts: 14154
Joined: 12 Nov 2007 23:02
Location: Newbury Park, CA

Re: SR20DET MAF adventure

Post by okayfine »

From the FSM
Pages from SR_ECU_Page_1.png
Pages from SR_ECU_Page_1.png (467.69 KiB) Viewed 7520 times
Pages from SR_ECU_Page_2.png
Pages from SR_ECU_Page_2.png (332.56 KiB) Viewed 7520 times
Because when you spend a silly amount of money on a silly, trivial thing that will help you not one jot, you are demonstrating that you have a soul and a heart and that you are the sort of person who has no time for Which? magazine. – Jeremy Clarkson
Post Reply