Slartibartfast

I suspect you're watching the clutch face (looks like the front of the pulley, isn't) instead of the pulley itself. The pulley spins freely, independent of the clutch face, until the magnet kicks on and draws it in. Then the clutch face locks to the pulley and transfers power from the pulley, through the clutch, to the shaft going into the compressor. When the magnet kicks off, the clutch releases, the face stops, and the pulley keeps spinning. If the belt was going to slip, I'd expect it to do so when the clutch was engaged, not when it wasn't (under load vs spinning an idler pulley). If the clutch is dragging a little, it'll tell you all about it when the compressor isn't engaged. Recently I helped a friend put a compressor on his S10 and the clutch dragged on the pulley right out of the box, which we noticed as soon as we fired it up (godawful noise). We found a spec for what the air gap between the clutch and the pulley was supposed to be, improvised a puller, and pulled the clutch assembly out on the shaft a little to get the gap correct. That shut it right up. There may be a similar process to set up the R50 aircon clutch.

I took mine off trying to track down an unhealthy-sounding noise, and left it off because I realized there was no gas in the system anyway. Clearly there's gas in yours if the clutch engages. Hopefully the belt is all it needs!

AFAIK it should go into mode 2 whether it's currently in closed loop or not, though obviously it won't tell you much until you get to closed loop. The one time I messed with it, I'm pretty sure I set it while parked and it flashed when I hit closed loop just like mode 1 does. Someone who was good with coding could probably work out what each pin did (using the service manual) and set up an Arduino or something to take that data and pass it along to whatever type of scanner they liked. There's also a product called Nistune that opens up older Nissan computers for tuning and diagnostics; they don't list the W/D21 ECUs, but they do list other vehicles with the same engine, and My1Path figured the system would probably work even without the M30 ECU swap he did on his. I'm not sure anyone's actually done it, though, and looking at their installation documentation (and price list!), it's not a rabbit hole I'll be crawling into any time soon. And yeah, no shock that the computers in newer models are completely different. I'd be surprised if anything cross-referenced from the newer rigs, other than maybe a few fasteners or trim clips or something.

If you find a legit adapter for the OBD1 port that doesn't cost an arm and a leg, I'd love to hear about it! It could take a lot of the guesswork out of this sort of thing. I was unable to find the right plug end when I did my flashing lights hack and ended up trimming down some blade terminals and hot gluing them into place. The fender has a bit of seam sealer on it from the factory. I didn't bother to reapply it when I put the fenders back on, figuring it didn't do much and would just complicate taking it apart again the next time. Yeah, no easy fuel system test point on these, unfortunately. EF&EC-139 of the '95 manual shows an in-line pressure gauge hooked up between the fuel filter and the rail. Some testers come with a T adapter for this sort of thing. I don't see those adapters sold separately, but I do see similar adapters meant to have a gauge screwed straight into them, so you could just buy that and a gauge if you don't want to buy a whole kit. Just make sure you get the right gauge, the ones for carb setups max out at 15 psi. Stored codes stick around until cleared, whether the fault remains or not. I'm guessing the CEL only illuminates while the ECM senses a fault. The CEL and the red light on the ECM illuminate together (watch the CEL flash with the red light next time you pull codes). My guess, given that it runs fine and cycles normally when driving around, is that it's not rich by much--just a little too much for the computer to compensate for when it's running with no load (again, more load means more air and fuel, which would dilute the effect of a small leak). So you're holding it at 2k, it enters closed loop, it reads rich, shoots for lean, and realizes pretty quick that it can't get there. It stores a code and turns on the CEL. Something about letting off the throttle causes a lean condition, the computer sees the lean condition it's been waiting for, it assumes the fault has resolved itself, and it turns off the CEL. Maybe the fuel cut triggers briefly and the computer mistakes that for the lean condition it's looking for, I don't know. I'd be interested to try mode 2 and see what that thinks of your mixture when you're driving around.

The nice thing about working on cars is that you don't generally plummet twenty thousand feet if something goes wrong. The plug by the hood release is OBD1. It sounds like Nissan had an adapter for their Consult system (or some version of it, anyway) to make it work with that plug. They also had a more primitive scanner that could plug into that port, which just had two lights and a button, and works just like the lights on the ECM. (The advantage is that it's not under the seat.) Some of the guys on Infamous worked out how to rig one up and I used their notes to build my own. It only used a few of the available connections. Presumably the Consult adapter (or aftermarket equivalent) could access more information from those other pins. Whether that info would be useful in this case, I don't know. But yeah, I'd be shocked if an ECM designed in the '80s to communicate via blinking lights had some kind of hidden wireless capability, or if someone put the effort into making a bluetooth dongle for an increasingly obscure OBD1 system. I took the fender off to access the resonator on mine. The screen I used was a bowl-shaped stamped metal piece meant to hold a foam air filter element for a carburetor (similar to this). I used it mostly because I had one handy. Not a bad idea to clean out in there, but I'd be surprised if something upstream of the MAF (short of a rat's nest blocking airflow) was causing mixture issues. At this point I'd be moving on to the leakdown test. One other long shot. Have you checked the ignition timing since you put it back together?

No worries man, good luck with it! Hopefully the dealership guy can at least get into live data or something. Funny story, mine had a rat's nest in the air filter box when I got it. PO had no idea why it was down on power, and apparently his mechanic hadn't checked the computer, because it was coding for the MAF. All the nest material from the filter box was piled up against the screen in front of the MAF. After finding the remains of the rat in the resonator box in the fender (where's that smell coming from?), I screened off the intake (also behind the fender) to prevent a repeat performance.

Sounds like you nailed it with your theory that a mixture issue was preventing it from cycling in closed loop. Makes sense that a small leak (or whatever's causing the rich condition) would cause more trouble when free-revving (no load) than when you're actually driving around (when there's enough air/fuel going through the engine to dilute the effect of the unwanted extra fuel). Nissan put out a service bulletin (NTB95-021, pretty sure I found it in the '93 Pathfinder section on Nicoclub--I'd link it but I can't get their site to load right now) saying that the ECM throws an injector leak code any time there's a rich condition that it can't do anything about. They suggest checking the MAF, checking the fuel pressure (34 engine running, 43 engine off), watching the fuel pressure with the key off to see how quickly pressure drops (should take around four hours to hit zero if all is well, or ten minutes if it's leaking), then check the temp and oxygen sensors if everything else checks out. (Sounds like you've run through most of this already!) Once you've got it figured out, change the oil. The bulletin also suggests tightening up the pins in the MAF harness, which was enough to resolve an intermittent stalling issue with mine (though I did the MAF, TPS, and one of the idle plugs at the same time, because they all felt loose... and I didn't do it quite as professionally as what the bulletin shows, because I did it on the side of the road with a pocketknife). The additional ground shouldn't be necessary, but shouldn't hurt anything, either. The bulletin doesn't specifically mention the WD21, though this is probably because it has you verify the repair using CONSULT, Nissan's super-special OBD1 checker box that these aren't set up for. It does list fuel pressures for the D21, though, so I assume the rest of it still applies. It's been a while since I had my intake off but IIRC the only thing you'd get from the injectors not seating right would be a vacuum leak (at the manifold) or an external fuel leak (at the rail). An injector leak would come from the injector itself. If it fails the leakdown test, I'd pull the spark plugs and check them to see if one of them is darker or wetter than the rest, or smells like gas, to track down which injector is incontinent. IIRC the plug for the TPS open/closed switch in mine has a third pin in it, too, though I don't know that it goes to anything. Strangely enough, there's a wide open throttle switch on the gas pedal, inside the cab, which tells the auto trans to kick down. I'm not sure why Nissan added another switch for that rather than tying into the existing one. Easier to adjust, maybe? In any case, mine drops out of closed loop long before it reaches wide open throttle. IIRC the turbo cars used the wide open switch for something, but don't quote me on that.

^Good point on the different O2 sensors. I don't remember if the bung size is the same but IIRC the wrench size isn't (the oxygen sensor socket I got for my '95 doesn't fit my '93). Hopefully the scanner helps narrow things down. I'd be interested in whether the temp sensor and the oxygen sensor readings match reality. Sometimes when an oxygen sensor fails, it reads lean all the time, causing the engine to run pig rich as the computer tries to fix a lean condition that doesn't exist, but I'd expect it to run like poo whenever it tried to go into closed loop if that was the case. There is a code the computer can throw if it thinks there's an injector leak, but the '90 manual says only California-spec models check for that, so yours might not notice or report if that's the case. New injectors suggest it's not that. If you haven't, check that the vac line from the fuel pressure regulator to the intake doesn't have fuel in it. And +1 on cleaning and checking ground points! I've had weak grounds show up as good with a meter but fail to power a load. Mode 1 is the oxygen sensor monitor (green light flashes as the oxygen sensor cycles in closed loop). Mode 2 does the same thing, but the red light flashes with the green one if the mixture is between 5% rich and 5% lean, and stays on or off if the mixture is outside of that range. (Worth noting, modes 1 and 2 only tell you anything when the ECU is in closed loop--either light staying on or off when the green light is not flashing does not indicate a rich or lean condition.) Mode 3 spits up trouble codes. Mode 4 lets you test the speed sensor, starter switch, and soft closed throttle position switch. Mode 5 watches the MAF, ignition signal, and camshaft position sensor (for tracking down intermittent issues). Edit: one other idea. Old gas can get pretty nasty. Have you filled it up with fresh fuel since you un-mothballed it? It's a long shot (especially if it wasn't parked for that long,) but my dad's old Ford ran (and smelled) a whole lot better on fresh gas than it did on the seven-year-old varnished garbage it had in it when we dragged it out of a field.

Nice of it to pick a convenient time to let go! The side of the stud looks like it was loose for a while. Speaking of funny noises, my driver's front wheel bearing has been grumbling at 60 since the Gambler (and howling if I go much faster), so there's what broke on mine recently. I tore it down to see what I was up against and the grease for the inner bearing is black and the grease seal's all torn up. Not sure when/how that happened (not like I bombed it around the sand dunes and clonked the front end on a whoop or anything ), but I know what it needs next.

I would be surprised if they still had the equipment kicking around, much less a tech trained to use it. I've been told there's an adapter for the Snap-On Modis reader that will work with these, but of course it's ludicrously expensive. The smog test is an interesting idea. They probably won't know what to do with the computer either, but their smog sniffer on a stick could give you some idea of the mixture. I'm not sure what that would tell you at this point, though. There are some California vs federal differences in ECUs, but AFAIK mode 1 is the same across the board. Are you sure it's in mode 1? Long shot, but if it somehow ended up in mode 5, the green light will flash in twos if the MAF is acting up or in fours if the ignition signal is malfunctioning (EF&EC-43 of the '95 manual). I'm puzzled, too. Other than the light not flashing, does the truck show any obvious symptoms? Hard starts, black smoke, rough idle, anything like that?

Does yours have the three-speed or the four-speed? 88+ got a four-speed, which is easily replaced with the later four-speeds (including the one from the first-gen Xterra). If you've got the three-speed, you may have a bit of a project on your hands converting the wiring and computer to run the four-speed, and sourcing parts to rebuild the three-speed might be a whole other challenge. If you want to go manual, the '96+ transmissions are slightly revised to hold up better, so you might as well take advantage of that. (Earlier boxes tended to have issues with the countershaft bearings; later boxes moved the fill hole up slightly to increase the oil level.) Any Pathfinder, Frontier, or Xterra with the 3.3 should have the updated trans. Takes a little screwing around to do an auto/manual swap (though there's at least one good thread on here from someone who's done it), but the good news is the engine computer won't notice or care.

The part number you put up appears to be for a Mitsubishi, so I wouldn't count on that working. I had a look at Rockauto and the booster they show for '01 is out of stock, but they have different ones for '96 and '04 (the only other years I checked, there may be other options yet). Both cheaper than what you were looking at, and much more likely to bolt up. When I thought the booster on my '93 was malfunctioning, I got a replacement from a Hardbody pickup at the wreckers. The number on the sticker was a little different, but it bolted right up and works fine.

I was impressed by how well Rat Trap did in the sand. And yeah, that Festy was something! I think we could've gotten a few more people over there and carried it out if plan A hadn't worked.

EF&EC 17 and 18 of the '95 manual covers open and closed loop, but yeah, you've pretty much got it. Checking sensors, connectors and vac lines is a good place to start. Clean the MAF if you haven't (you can get special MAF cleaner, but I used alcohol and a rag very carefully and got away with it--just don't damage the filaments or use a cleaner that leaves a residue). Might as well just replace the vac lines if they're original. It probably won't solve the problem, but it's cheap and easy to do and lets you rule them out. ECM failures in these aren't common, apart from people drowning them because they didn't think the puddle was that deep.

Mine looks like a Power Wheels next to yours!

LOL is that the overlook I couldn't find?

I'm no stranger to making wiring harder than it needs to be. I unwrapped a section of my harness once chasing what I thought was a weird high-resistance short to ground on one of my high beam wires that turned out to be the indicator bulb in the cluster. Just checked mine, warm, and it flashes if held steady at 2k rpm. Is yours throwing any codes? Could be an issue unrelated to the oxygen sensor preventing the ECU from going into closed loop, though I don't think the TPS is essential for that.

I pulled up the ECM pinouts for '90 and '95 and looked at them side by side, and the location of pin 19 is the same between them. Everything is the same with the exception of pin 106 being present in the '95 and crossed out in the '90. Weird that it's lighting up a different pin. The wiring diagram shows a straight shot with no other connections, so if the oxygen sensor and computer are unplugged and you have no continuity between pin 19 at the ECM connector and B at the oxygen sensor connector, but continuity from B to somewhere else, there's something wrong with your harness. Which pin is lighting up? The diagram shows the wire in question having a grounded shield, which is wired to pin 48 of the ECM (and to joint connector A, which ties it to an engine ground). If the wire is crushed somewhere, I wouldn't be surprised if it shorted to its shield. The ECM has multiple ground pins, so you'll find multiple pins with continuity to pin 19 if this is the case.

I don't have a '92 manual, but I've got a '90. The oxygen sensor circuit shown is different than the one in the '95 manual. The '95 grounds the oxygen sensor through a black/yellow wire to pin 106 of the ECM. The '90 grounds the oxygen sensor to the engine via a black wire (ground point 35M, via joint connector A). The black/white wire should still have +12v with the ignition switch on, and the white wire should still go to pin 19 of the ECM. I checked the oxygen sensor plug in my '93 and found white/black, white, and black, same as '90. Assuming the wire color changed when the wiring did, this suggests that mine and yours are both wired like the '90, not the '95. The test procedure says to check for power between the black/white wire and ground (ignition switch on), check continuity between the white wire and pin 19, and check continuity between the black wire and ground.

Gotta say I like either of the stock R50 tail lights or the QX4 tail lights better than the 2fast2furious Altezza aftermarket stuff.

Hmm. Yeah, that's more pulling (some of) the alarm's teeth than removing it entirely. IIRC the R50 alarm is built into other things to the point where it may be difficult to nuke in its entirety.

42 is the throttle sensor circuit (TPS) according to my paper copy of the '87 service manual. The full list: 11, crank angle sensor circuit (sensor in the dizzy) 12, air flow meter circuit (MAF) 13, Cylinder head/water temp sensor circuit 21, ignition signal missing in primary coil 33, exhaust gas sensor circuit (oxygen sensor) 42, throttle sensor circuit (TPS) 43, injector circuit 44, no malfunctioning in the above circuits Later models added a few more codes, and, just to be annoying, moved the last three around. Later manuals list 43 as the TPS circuit, 51 as injector circuit, 55 as no malfunction. 42 doesn't appear on the later code charts, so I'm not sure why you're getting conflicting info there. I've found no reference to a fuel temp sensor in any W/D21 service manual I've gotten my grubby hands on. A loose plug sounds likely. If you look into the plug (assuming your '87 uses the same electrical plugs as my '93), the female pin is formed like a rectangular C with the ends touching. I tightened up a few loose plugs on mine by sticking the end of a knife between the pin and the plug body and pushing those ends inward just a little to make them grip the male pin more tightly. This probably isn't the right way to do it, but the plugs felt much tighter afterwards, and the idle issue I was chasing went away, so I'm calling it a win. There is a fail-safe that will drop the rev limiter to 2800 rpm. EF&EC-42 says it'll kick in if it's not getting a signal from the MAF sensor or if a fuel injector fails to chooch more than four times in a row under certain conditions, but the only mention of the TPS on that page is that the ECU fudges MAF data with TPS data if the MAF isn't working. So that's a little odd. Not sure what to tell you, but I'd check the MAF plug while you're in there, and pull codes right away if it goes into limp mode again. Good luck!

Well now that is just bizarre. If you lock the rear gate again, does that lock out the starter again? And how did you go about disabling the alarm?

I'm not aware of stronger aftermarket CVs for these, but there are two different types of CVs out there. One has a tripod joint inner (for four-cylinder trucks), the other has six balls (for V6 trucks). I'm told the six-ball type is stronger. IIRC the tripod joint style CVs have a smooth shaft between the boots, and the six-ball type have a little narrowed section in between. Either will bolt up. In fact I've got one of each on my truck right now (had one set of each, went to re-boot and found I had one usable one in either set), so I guess I'll know which is stronger if I manage to snap one. Also, I'm buying CVs next time instead of rebooting. What a PITA that was.

The lines are just a shortcut and a visual aid for getting the timing correct when you're first assembling it. If you've got 40 teeth between the dimples on the cam sprockets, and 43 teeth from the driver's cam sprocket to the crank sprocket, then it's timed properly.