Idle Problems in CRX after trans swap
 

I've been fighting this problem on and off for about a month now (mainly because I don't drive the car much during the winter). So far the guys at honda-tech and hondaswap haven't been able to help... I'm hoping somebody here has a few ideas:

The car: 88 Honda CRX DX. Originally AT car, recently swapped to stock DX trans (ECU also swapped).

The problem: After the swap, the car seems like it never comes off the warmup cycle. Idle is always high (1100-1700 depending on who knows what), but will drop to 750 if I blip the throttle like you would with a carbureted car to knock it off the fast idle. The next time I rev the engine it goes back to the high idle. It also pulses between 1100-1400 on rare occasions (only twice in the past month). The RPM also seems to drop more slowly than it should (or not at all) between shifts as if the ECU is keeping it artificially high.

What I've checked:
1) engine codes - none
2) coolant system - properly bled and filled
3) ICV - cleaned
4) linkage - inspected and lubricated. No binding or other problems. Vacuum actuators, appear to be working properly, but it's hard to tell for certain.
5) Sensors - need to test, but can't find anything to compare the values to.

I'm down to thinking that it's either a bad temp sensor, or maybe a bad O2 sensor that's spoofing the ECU. Aside from that I have no ideas except maybe that there was a wiring difference between the auto and manual ECU that needs to be changed. All I swapped was the Neutral cutoff and backup light wires. Everything else was just plug and play (or maybe I'm wrong on that?).

Any help or troubleshooting advice? Now that the weather is warming up again, I'd like to get this resolved so I can drive the car on a regular basis.

I'm guessing you swapped in a MT trans? Maybe the TPS is different between the AT and MT models(they sometimes are) and the MT ecu isn't getting the right readings. Often, vehicles will open the IACV(Idle Air Control Valve) when you aren't at idle to increase coasting distance. But, that also makes me now think that the speed sensor might be the culprit. If it thinks you are rolling it might keep the engine in a high-idle. But that's harder to explain then a bad TPS because something would have to be pulsing to the ecu at idle making it think there was movement, which could be cross talk from the ignition system but that kind of thinking can get you off on a whole different tangent.

Another thing to consider is a bad ecu or the pinout between the two ecu units weren't exactly the same causing any of the above issues.

Thanks DK,

Yes, I swapped to a MT (sorry, I didn't specify that in my first post). I'll I'll check with the dealership and see if the TPS is different.

Any idea where the speed sensor is on the car?

What about adjusting the TPS? Is that possible, or are the ranges completely different?

The speed sensor is plugged into the top of the transmission. Honda tps's for dpfi and mpfi have the two outer pins reversed, and when I used an mpfi tps sensor on my dpfi 91 hatch std, I had to reverse the outermost pin on each side. Your situation may be different. I would not think there would be any diff b/w 5spd and at in regards to the tps sensor.

Huh? The only sensor I remember in the top of any of the manual transmissions (I ended up w/ two bad ones before I got the one I installed) is the reverse light switch. Maybe the problem is a missing speed sensor? I dunno but I'm grasping at straws at this point.

Any chance you can pop your hood and snap a photo of the speed sensor for me? At least then I can really home in and take a close look to see if there's a spot for the switch.

typically auto trannied cars idle higher than manual trannies so there could be a sensor issue,

also loose/missing/cracked vacum lines can cause high/erratic idle too

You don't have a speed "sensor" on your transmission. Your car has a cable operated speedometer - it's purely mechanical, aside from the small electronic bits at the gauge cluster.

Yeah, a bit of googling revealed that there's a VSS on the back of the instrument cluster. I'm guessing that if it was good when I had the AT in, it's good now since it doesn't appear to be transmission specific.

Vetteowner, I've checked the vacuum lines carefully and they're all nice and pliable. None cracked, cut, or disconnected.

I have some testing to do, but it will have to wait since I'm getting ready to leave for a hunting trip and will be out of town till next weekend. But I found a website with the sensor values and a lot of other good information for this vintage Honda.

https://geocities.com/chipman_13/ecu.html#codes

Honestly, I'm beginning to think that I might have a bad ecu (it was tested before I got it, but it's also the only other piece of the puzzle that got changed out w/ the transmission. Maybe I'll pick a replacement up off ebay and see if it makes a difference.

I'd recommend getting the sensor values for the car the ECU was originally in and then spending some quality time with your multimeter under your hood before calling that ecu dead.

BTW, what did it come out of anyways?

I'm going from memory now (I bought the ecu like 7 months ago, and I didn't pull it myself), but I think it was from an '88-89 Civic DX. It was a DPFI car for sure, and it had supposedly been tested. Everything should be the same as my CRX, but If I can find the stock ecu number for an '88 CRX DX 5 speed I'll compare the two to confirm that they're the same. I guess it's possible that the ecu is compatible (more likely than that it's failed at least), but until I get a chance to do more on the car (probably weekend after next due to my schedule) I'm out of new ideas.

Well, I'm back from hunting (got my pig btw), and decided to check something out this evening.

I'm not sure, but I think I may have the wrong ECU. I have a 37820-PM5-A080, and I think the stock ECU for a manual 88-91 CRX DX is a 37820-PM5-A060.

I can't find a breakdown of the ecu's by year/model, but I'd be interested in finding out if anybody knows if the 'A080 is in fact the correct ecu. That could explain things and would be easy to fix if it is the problem....

Is the throttle cable adjusted too tight? The ~4 foot long DPFI throttle cable actually shrinks and grows depending on engine bay temperature. If you adjust the cable to zero play when the engine is cold, it will hold the throttle open once things are warmed up.

It was suggested earlier that some of the sensors may be different. The are not.

The things that first came to mind for me was the possibility of a lose ground connection on the thermostat housing (a common occurrence when reassembling things after major engine work), or maybe you bumped and cracked a sensor when pulling the old or installing the new transmission.

Throttle cable is slack when at idle so that's not it. Besides the problem is definitely worse when its cold out.

I did some checking and the ECU is correct for a manual trans CRX. So unless it's a MPFI ecu and not a DPFI ecu, that's not the problem (still haven't figured out how to tell that).

I'll be checking the sensors for damage this weekend, but I'm pretty sure that isn't the case. I'm pretty careful about that kind of thing.

All PM5 (the middle group of 3 characters) ECUs are designed to operate the DPFI D15B2 engine.

The MPFI D16A6 engine (found in the CRX Si, Civic EX, Si and AWD wagon) is controlled by either a manual transmission version of the PM6, an automatic transmission version of the PM6 (as far as I can tell, found only in the automatic AWD Civic wagon) or a PS9 ECU (found in automatic Civic EX sedans).

The MPFI DOHC (dual over head cam) ZC engine is driven by a manual or automatic transmission version of the PM7 ECU. The ZC was never available in the US market... The closest thing we got was the D16A1 found in the '88-89 Acura Integra, which has significantly lower power output and a different bell housing/engine mount bolt pattern.

The MPFI D15B6 (found in the CRX HF) is controlled by a PM8 ECU... No automatic variants.

The DPFI D15B1 (found in the base model Civic hatchback) is controlled by a PM9 ECU... I'm pretty sure there are automatic variants, but not 100%.

Thanks Bobski. That jibes with the replies I got on the honda forums.. So I can eliminate the possibility of the wrong ecu.

I've been out of town all week till last night, and will be leaving again tomorrow. So no time to work on the car. Probably until next week I'll just be playing "armchair mechanic"... So if anybody has any other ideas, please let fly. Maybe I'll have a laundry list of things to check when I get the chance. For now all I have is to check for bad connections/broken wires, and to check the sensor values for a bad sensor. I think I've accounted for everything else.

Does your TB have a FITV? (Fast Idle Thermal Valve)

https://forums.trinituner.com/tech/fitv.jpg

Plug the bottom port with your finger inside of the throttlebody with the engine warmed up and running. If it idles normal after doing that than your FITV needs to be replaced or repaired.

That's an MPFI throttle body... DPFI engines don't have an FITV.

Correct. I wish it was that easy. In my experience, these first generation fuel injection systems are all a pain in the butt. My old '85 mercedes (Bosch Jetronic fuel injection) has a similar symptom and with that car, it can be caused by any number of problems. I'm converting it to EFI eventually, so it's not worth messing with.

.... Maybe I'll just bite the bullet and do the MPFI conversion on my CRX.....

Clean/replace the PCV valve? It could be getting stuck open, acting as a vacuum leak.

I wouldn't mind doing that but I've actually been unable to locate the PCV valve to date. I have one that I purchased when I got the car, but there's literally no matching part anywhere on the engine.

Other than that, I have a little update. I thought I had found the problem a few days ago when I noticed that the linkage seemed to be binding in the last few hundredths of an inch of travel before it closed. When the idle is high, I can manually close the throttle linkage and it will drop to base idle. But lubing the linkage didn't get rid of the problem, and I noticed that if I get the idle down, I can shut off the engine and then restart it immediately without touching the throttle, and it'll be back up again. I checked the two vacuum motors on the pass. side of the TB and they don't seem to be bumping the throttle, so unless there's another solenoid that I haven't found, or unless the TPS itself is binding up I don't think the sticky linkage is the problem.

Also, the idle is consistently hunting by about 50-75 rpms. With the airbox open, I can literally hear the injectors firing for a second or so, then stopping, then starting again, then stopping..... I'm starting to wonder about the O2 sensor maybe getting a bad value for the mixture and throwing things off. The wires are intact, but I haven't removed it to inspect it yet...

The O2 value and air/fuel ratio oscillates like that under normal operation. The O2 sensor actually detects a very narrow range of values (something like 14.5:1 to 15:1, so when the ECU makes a small adjustment to the mixture, the O2 reading will swing from one extreme to the other. The ECU isn't able to peg the center ideal reading on the O2 sensor, so it's constantly adjusting and re-adjusting.

Is the spring still in place on the idle adjust screw?

The PCV valve is a metal fitting screwed into a port on the intake manifold between the 2nd and 3rd cylinder runners, next to the intake temperature sensor. It should have a U-bend tube attached to it that runs down between the 2nd and 3rd runners to the breather box on the back of the engine block.

Bobski, the spring on the idle adjust screw is still in place, but what difference would that make?

Also, we confirmed this morning that the extra air is definitely coming through the ICV, so that pretty much eliminates the PCV as the major cause as far as I can tell...

Anyway, I thought I had the problem solved, but alas! No luck....

This morning I tested the temp sensors, they're both good and in range. When I checked the TPS on an oscilliscope though, it was pretty wacked out near idle. After finding out that nobody in town stocks TPS sensors for this car (???), I tracked down a good used one, tested it to confirm it works, and installed/calibrated it with high hopes that it would solve the problem.... NOPE! No change at all. The idle is still high and erratic.

I think I'm back to either a bad ecu, or a wiring problem (remember, I did an auto to manual swap and ecu swap)...

Can anybody post links or pinout diagrams for an automatic 88 crx dx and a manual 88/89 CRX dx manual? I have looked and haven't been able to find anything that differentiates between the two.... The only differences I know about are the backup light and neutral lockout, and that's all I changed when I did the swap......

As for the spring, I was just throwing ideas out there... If the spring was missing, I could see vibration and/or the throttle cam tapping on the end of it possibly causing inconsistent idle speeds.


You didn't accidently switch around the IACV and tandem control solenoid valve connectors, did you?


Wires that differ at the ECU:
- Lockup control solenoid
- Interlock control signal

Elsewhere:
- Shift indicator wiring
- Interlock control unit wiring and associated shifter and ignition switch solenoids
- Neutral / clutch safety switch wiring

The main wiring harness should be identical between manual and automatic DXs. It's the engine, dashboard and rear wiring harnesses that vary.

I have a similar problem on my 89 DX (no swap). Doesn't always idle high, but has the hunting problem, and if I give it gas it's like I have a dead spot - then it will rev back to life wherever the accelerator position is. Checked and the TPS is fine.

I will be replacing my fuel pump and fuel relay. Timing belt is brand new, btw. I am wondering if it's a faulty O2 sensor as well. Bobski or anyone - know of a cheap place to get a new O2 sensor for an 88/89 DX auto? I'll be reporting my results.

BTW I have heard it can be the fuel relay - that's why I'm changing both. The fuel pump was about $40 online, too. Chances are you still have the 20+ year old fuel pump in yours - and the relays are known for going bad.

Thanks guys.

Bobski, by your description, it sounds like I don't need to worry about any wiring differences, right? Lockup Control solenoid wire is obviously disconnected at the trans since it's not there anymore, and by interlock control signal, I'm assuming you mean the wires at the shifter that I had to jumper together to disable the neutral lockout so I could start the car....

I'm going to throw my old Auto ECU back in when I get home and see if that changes things. (Fingers crossed!)

101mpg, I'll have to check my records, but I don't recall ever seeing the fuel pump/relay being replaced in the car's history. It's worth a try, but still, since the problem cropped up on the first drive after a trans/ecu swap, I'm of the opinion that all the things I'm checking at this point are long shots anyway.

Where's the relay at? If it's easily accessible, I can bench test it today.

Oh, and by the way, If you suspect your O2 sensor, you can usually remove it and burn it clean with a propane torch. Get the metal shroud around the sensor cherry red so it burns out any accumulated carbon, let it cool, then replace. If that is your problem, this should at least improve things. It's a trick that I learned from while having similar problems with the old mercedes I own (also first generation fuel injection).

If you use any anti sieze on the threads, be extra careful that you don't get it on the sensor. It can foul things up and give you bad readings.

Crap...

It's not the ECU, and it's not the O2 sensor.

Could it be a bad tach sensor or lead maybe?

I'm out of ideas at this point. I hate to give in and take it to a shop (even my cousin's shop). I've never had to do that before!

I never saw in this thread that you did the correct procedure to set the idle. If you did, I apologize. That is to disconnect the EACV, set the idle with the air screw to ~650rpm unnder no load (no lights, A/C, etc). Shut the car off, connect the EACV, and reset the ECU by unplugging he hazard fuse from the main fuse box for 10 seconds, and then start the car and let it idle for a few minutes to let the ECU adjust the EACV on the idle. If you just adjust the air screw without resetting the ECU, it will get confused as it has a "memory" as to where the EACV valve is supposed to be from when the last time the ECU was reset.

The idle hunting like your description is what the ECU does when the idle is set too high. The ECU is programmed to shut the injectors off when the rpms are over 1100 rpm and the throttle is closed (like you were coasting in gear). If the idle is set too high, the rpms will go above 1100 rpm, and then drop back below it when the ECU cuts the injectors off. The idle basically bounces.

Good point. I don't know why I didn't think of that... I've explained it to others in the past. Another characteristic of the idle-up/coast-down cycle is a relatively fast idle-down portion of the cycle. The idle-up speed depends on how high the idle has been set or how big the vacuum leak is.

I was actually about to post on another forum and ask what the procedure for setting the base idle is just to make sure that I had covered that base as well.

FWIW, I did slow the base idle down by a few rpm when I first started fighting this problem, but didn't follow the procedure you outlined.

So is the EACV the same as the Idle control valve? I'm assuming it is. Also, is this procedure for the DPFI system, MPFI system, or both (I've got bad advice from a lot of people who don't realize that depending on the model, '88-91 civic/crx's had two completely different fuel injection systems).

Would an incorrectly set idle also account for the exceptionally high idle when cold and not dropping to base idle nearly as fast as it should?

The EACV is the idle control valve. The CRX manual I downloaded does not specify which DPFI or MPFI in the section to set the idle.

I'm not sure about your last question, but my understanding of how the idle is set is that the ECU is trying to determine what the proper voltage to send to the EACV valve to get the idle speed correct.

The pdf manual I have is huge, if you want a jpg of the procedure PM me your email address and I'll send you a screenshot

Maybe and maybe. Take it one step at a time.

Probably, I'll have to recalibrate the TPS as well. I'm guessing that if I have to slow things down, The tps voltage at idle will drop to <.45V.....

You should be calibrating the TPS to between 0.5 and 0.45 volts with the throttle completely closed. As in, back the idle set screw off completely and then adjust the TPS.

So not 0.45V when the car is at idle or when the butterfly is at the "normal" idle setting? Well I got that wrong, but it turns out that it really doesn't matter much anyway. My idle speed screw was only about one turn from being backed all the way out anyway.

With the ICV disconnected, I can only get the idle down to about 750, and plugging the idle back in causes the same surging to start again immediately.

Also, I think the symptoms may be changing a little bit over time. Yesterday I noticed that when I the car was sitting at the high idle (about 1700 at the time) it was also pulsing a little bit on occasion. It would drop by about 200 then recover. It only did it a few times though, then stopped.

The voltages for my temp sensors were all good when I checked them, but when I disconnected the air temp sensor it didn't cause the idle to change at all. I'm still a little suspicious of it. Can somebody else with a DPFI car maybe disconnect that sensor momentarily and see if it affects their idle drastically?

I had a similar high idle problem on my 87 Acura integra (but it was MPFI).

I ended up disconnecting the ICV, but then my "check engine" light came on because the computer didn't see the ICV. So I looked up the resistance specs in a repair manual and bought some resistors of the correct ohms at radioshack and plugged those into the wiring harness to fool the computer thinking the ICV was still in place. It worked OK- but when the AC and lights were on, the engine did idle a little too low.

I tried that out of desperation last week. But with the ICV disconnected, the car bucks badly when the gas is applied.

Take the cover plate off the top of the throttle body. Start the engine, disconnect the IACV and back off the idle screw completely. Block off the port on the driver's side of the chamber you just uncovered... The port that feeds air to the IACV (use your thumb, it wont hurt you).
If the engine stalls, you've got a bad IACV. If not, you've got a pretty big vacuum leak.

It sounds like a combination of a bad or dirty IACV and the TPS not properly calibrated. I had the same bucking issue in my Civic a while ago.

Start off by cleaning everything. After you've cleaned everything, then calibrate the TPS.

Bobski,
The IACV appears to draw air directly from the airbox, through a flexible tube to the intake manifold and then to the IACV, not from the throttle body like an MPFI car, but I'll double check that this weekend just to be sure.


I've checked for vacuum leaks with a can of starting fluid (ended up using nearly the whole can... I checked pretty thoroughly) as well.

Optix,
I've cleaned/inspected the IACV twice now, and I recalibrated the TPS again yesterday night. I couldn't tell if it made a difference since the idle wasn't acting up (for once) before I checked/adjusted the TPS. I still can't get the Idle down below 750 using the idle screw and the IACV disconnected though.