12 replies to this topic

[lambs]

I've put this in the CRX section, because that's what the engine is from, but it's in a 3G Civic

Anyway I have the EW5 from JDT's spares car in my daughters 3G Civic. Of course the engine slotted in, but the loom took quite a bit of work. Basically I've cut down the CRX loom to the EFI basics and spliced it into the civic loom where needed. Finding a wiring diagram was a big problem, but it turns out that the EW5 uses almost identical wiring to a Honda City turbo and with help from a CityTurbo forum member I was able to get it sorted.....mostly).

In any event I got everything installed and fuel supply sorted last Friday. It fired up straight away and everything seemed fine, but after maybe 20 secs running it started to get quite rough.....and that's the problem I still have even after checking everything possible. All sensors check out within spec and ignition + cam timing are spot on. The ECU is showing no error codes. I found that the factory set throttle stop had been backed off completely, but resetting that plus the TPS only helped a small amount. One significant indicator is that there is a lot of black material coming out the exhaust, suggesting extremely rich running and I suspect this is the cause of the problem but why it's happening is a bit of a mystery. Being a CVCC makes this a bit more complicated as I have no experience with this type of engine.

I've narrowed the potential causes to 3 possible areas:

- spark plugs. Most of the ign parts are new or known to be in good condition. I've noticed that the spark plugs are Denso type W20EP-U, which is the equivalent of NGK BP6ES. According to the NGK Oz website, the recommended plug is BUR7EB-11, which appears to be quite different in terms of specs. Are these engines sensitive to plugs ?? I've ordered a set of the recommended plugs.

- injectors. As I understand it, this engine has been sitting around for some time unused, so it's possible the injectors are gummed up, etc. I pulled of the fuel rail and checked the injector tips. I couldn't see any obvious leaks when fuel rail pressure was applied, but realise that isn't any sort of reliable indicator. The CVCC injector also seemed fine, although there is one o-ring that needs replacing. I may have to bite the bullet and have these cleaned, but it's an expense I'd rather do without.

- wiring. With the wiring diagram help I had it turned out to be fairly straight-forward, however there is one area that may or may not be a contributing factor. The CRX has a type of speed sensor fitted to the ECU. Basically it signals when the car has exceeded 15kmh and then again at 45kmh (it also activates a buzzer at 105kmh, but that's not relevant here). The amplifier module for the speed sensor in the speedo also connects to the ECU. There are 2 wires connected to the ECU at pins A19 and A20 and according to the wiring info I have one is associated with the 15kmh trip an the other with 45kmh. Exactly what effect they have on the ECU is a mystery, but I am wondering whether there is an effect to change injector timing for some reason. This is where owners of EW5 powered cars might be able to help.....eg back probing the ECU terminals mentioned above to determine voltage changes, or even disconnecting the amplifier module to see the effect on engine operation....

Of course I'm open to other suggestions. Mechanically I think the engine is fine as a simple power balance check suggests all cylinders are operating similarly in terms of compression. I also replaced the cam belt and tensioner, plus all seals and the water pump. Valve clearances are spot on.

Despite not having an air cleaner (it's coming....) I took the car for a test run tonight.....and it ran fine. No sign of missing, etc at all when under load. Indeed all seemed normal, but for the miss and general rough running at idle.

Any help appreciated.

Cheers (sorry for the long post, but unavoidable).

[cbstdscott]

Have you looked at the plugs (normal color brown, black=rich, white =lean) to confirm that you are running rich?

Have you replaced the O2 sensor? If that sensor is not working right, you will get a rich condition.

Scott

[zakats]

didn't this car start out as a carbureted version? you might consider double checking the injector wiring to see if its out of order

[3GCVC]

i didnt know about the speed thing in the ECU on the CRX's where did you find out about that. im not sure if the civics have that or not.

does the car run fine everywhere else except for at idle?

possibly check the vac advance mech in the dizzy and check both the vac pipes to it one is vac and the other electronically controlled from the black box isnt it? maybe try applying power to the pump in the box see if that will smooth it out.

other than that it is a tough one to try to figure out what could be going on. AFAIK CVCC and NON CVCC injectors cant be swapped so it shouldnt have those in it by mistake, but the later CVCC manifold might swap. what CVCC motor do you have a 8485 or 86-87 they are a little different plug wise

if its a wiring fault i do still have the EW5 loom from my sedan if that would make things easier.

[lambs]

Thanks for the responses -

Scott: the plugs are showing uniformly sooty black, which is rich, of course. As I understand it, with CVCC engines the plugs fire a rich mixture so even with the engine running normally it might still be possible to have black plugs. The material coming out the exhaust definately indicates rich running. The oxy sensor is new, so that shouldn't be a problem. I have a Motec lambda meter and wide band sensor so may use that to check the actual mixture, although any misfire can skew the readings.

Zakats: the original engine in the Civic was carb'd, but I've transplanted the entire EW5 engine from JDT's CRX and the necessary wiring. I'm confident the injector wiring is correct and the fact that the car ran more-or-less normally on the brief test drive would seem to confirm.

3GCVC: I've pulled apart the dash cluster from JDT's JDM CRX and it definately has a sensor in the speedo and an associated amplifier. It's a simple stepped chopper arrangement that passes between hall effect sensors (or something along those lines). A bit of wire tracing showed that 2 wires from the amplifier go to the ECU. The City turbo wiring diagram confirms that those connections exist and even the wire colours match - it's what signals the send to the ECU and what the ECU does with them is the mystery (a city turbo forum member - bigelboe - confirmed that he is running his car without those wires connected and it runs fine). I suspect they might have an effect on ignition timing. The car runs more-or-less normally under load, no missing, etc, it revved quite cleanly. In fact I was quite pleased with the performance (OK, the auto kills performance to a large degree - let's say it ran quite well).

Vac and mech advance are working normally. Yes, one of the pipes is connected to the manifold so sees vac, but the other goes to the black box on the firewall. I'll do some m/meter work and see what is powered there. I think the car is an '87, but JDT may be able to confirm. Hang onto the loom.....I may yet need it :-)

[GENONE]

are all the vacuum lines connected correctly between the sensor box and the relevant engine points?

I've got a copy of the diagram you can follow to make sure thats all correct?

[lambs]

are all the vacuum lines connected correctly between the sensor box and the relevant engine points?

I've got a copy of the diagram you can follow to make sure thats all correct?

I left the black box and all its hoses connected to the engine when I did the changeover, just so as I didn't run into any problems when re-connecting things after installation. That said, I am assuming that everything was connected correctly in the first place, so anything you have that might help in that direction would be welcome.

[3GCVC]

it may come right once you have an EW5 sedan cluster in it which has the correct controls for the ECU possibly.

there is an idle mixture adjustment screw in the black box behind a rubber bung, not sure which way is lean or rich but i just remembered it and it may help you out.

[JDT]

The car is an 87 model

I don't know if it applies to this engine but I had a near identical problem with my ZC. Bad idle began shortly after car was started but ran fine when throttle was applied.

I pulled the TPS sensor out of the car and couldn't find anything wrong with it. When I plugged it back in it ran fine, I assume it was dirty contacts.

James

[anjin]

lambs

My EW5 does exactly the same thing - runs rough at idle like its missing, but as soon as you accelerate it's fine. I also have rough idle at low revs, and the car runs a bit jerkily. I have not managed to definitively work out why.

I did have a loose wire into the ecu for injector 3, but that's fixed and the rough idle continues.
I do get a tps error on the ecu (but no pgmfi light), but after checking out the settings its about right - possible a liitle under voltage at idle. So it might be a tps problem at low revs. Playing with the settings hasn't done much however.
I was going to replace the O2 sensor, which could be causing a rich mixture. But I didn't get that far as the engine is due to be replaced (later this week unless something else happens) and I haven't bothered to solve the problem.

In terms of what to try, cleaning the grounds is a recommended general fix. With a maximum of 12 volts and often a lot less good grounds and contacts are important.
Try disconnecting the o2 sensor and see what happens?

One of the benefits of this condition is that the engine is pretty peppy under acceleration, so its not all bad.

And your air box will be in the mail this week.

[lambs]

3GCVC - The IMA sensor is wound fully counter-clockwise. Playing with it just made things worse, which fits with the rich running situation. I must say I'm really hoping to get away without having to do more wiring, but so be it if that's what's needed.

JDT (James) - I seemed to remember you saying it was a 1987. I think the TPS certainly plays a part. After finding the factory throttle stop had been played with, resetting it plus the TPS helped to smooth things out, but not completely. I think that is where part of the solution lies.

anjin - interesting about your car, I was actually quite happy with the engine ran under load. I'm not getting any errors at all, but the fact that the throttle stop has been played with, plus I did get some improvement after re-setting, suggests that the solution lies in that area. The CVCC o-ring I mentioned in the first post is also playing a part I suspect because it seals the base of the injector holder and any air leaks in that area would stuff up the CVCC mixture potentially leading to poor overall fuel ignition (missing, etc.....well, it's just a theory). I have tried disconnecting the EGO sensor, but with no noticeable effect. I've also made sure there are multiple good earths, even having the main engine earth connected back to the battery and an extension of that cable earths the main ecu terminals.

Thanks for all the inputs, very useful.....Steve

[lambs]

Just an update. Today I replaced the o-ring at the base of the CVCC injector, plus put in a set of the recommended spark plugs and, lo and behold, I have a smooth idle. I strongly suspect the o-ring was the main problem. I've also reset the factory throttle stop more accurately using a vac gauge as well as the TPS. All in all it's running much better and I'm much happier with it.

[anjin]

Again I learn something. Well found.