Honda / Acura K20a K24a Engine Forum banner

Running Rich at Idle, Stumble on Tip-in

41K views 101 replies 10 participants last post by  ekracer98 
#1 ·
I am trying to solve a few problems with a K-swapped Fiat.

CAR:
1977 X1/9 (not that that matters)
Z1 engine-trans-harness
A2 KPRO
K-Tuned Adapter harness

PROBLEMS:
Very rich idle (11.5:1)
Stumble on tip-in
Timing stuck at -1 degree at idle (timing advances with RPM)

Let me start by asking that PLEASE nobody say "call Hondata." Anybody who has dealt with them knows that they are a bunch of unhelpful pricks.

The car idles around 11.5:1 in spite of having a target of 14.59:1 and fluctuates as one would expect it to in closed loop. Just as an experiment, I turned the fuel pressure down to 35 PSI and the computer adjusted to keep mixture around 11.5:1. It also has a wicked stumble on tip-in (just off idle is by far the worst, but it will hesitate a bit revving from say 4K to 8K. During the stumble, it seems for a split second to go terribly lean, but the value changes so fast that I can't say how lean. I could swear that I see a 2x.x:1 reading. Otherwise, when holding any RPM, it will run just as rich as at idle.

I loaded KManager onto my laptop and started up the engine. The car runs well and idles great. ECT reaches 195 degrees and holds within a couple of degrees. O2S is new NGK and I quadruple checked my O2S wiring. Stock 310 injectors. Fuel pressure is static at 50 PSI.

I am very confident that all of my wiring is correct, but I am not too proud to answer questions about my wiring work.

I have tried twice to load a "New Calibration" in KManager. I used the "K20A2-rsx-stock-tuned.kal" map. The problem has persisted. Actually, the reload ADDED the retarded timing problem. Idle timing HAD been about what I would expect (and fluctuating at idle, which I also would expect) before loading the "K20A2-rsx-stock-tuned.kal" map.

What is going on here?!

Please help, and I am happy to answer questions about my setup and situation.
 
See less See more
#4 · (Edited)
You said when you load kmanager it runs great. I am confused by this.

I'd really like to see a datalog of the conditions and a copy of the map you are currently running.

Was the kpro purchased new or used? Xenocron could test the ecu but I wouldn't do that just yet.

You might want to try another set of injectors - hopefully someone is understanding and let's you borrow them (not buy if you don't need) you could have an injector stuck open but your real rich. Are all the plugs saturated with fuel? Your getting spark on all 4?

Not proud of wiring? Are you running and but connectors or is everything soldered? How many grounds and where? K ecus are very finicky. Where are you drawing constant power from for you o2 relay? Personally I only use oem denso rsx-s o2 sensors with k-pro (worse case run external aem, plx, innovative wide band but k-pro has to be setup properly)

Edit: most importantly are you getting any error codes. You sure map connector is going to map sensor and tps connector going to tps because they are identical

I'll check up on you later tomorrow way past this old mans bed time :p
 
#6 · (Edited)
I had something similar and it was caused by accidently swapping the MAP and throttle harness connectors. Worse, because the connectors have reversed wiring, if you power up the car that way it almost always ruins the MAP sensor, so even if you catch the error and swap them, it still won't work.

What does MAP read without the engine running, with just ignition hot? It should be near zero. Also, move the throttle and confirm that TPS moves in KManager.
 
#7 ·
Thanks Kurt
I asked the OP to check the TPS value in KPro and I was told it is working properly. The engine also idles well considering the A/F.
Reversing the MAP and TPS plugs causes the engine to run pretty rough.
At least the ones I've seen.

Hopefully we will get an update soon...
 
#8 ·
curious, stock PRB IM? TB? Header?

after reading this again - it may seem like a/f is just not reading correctly. curious if the car is actually running rich or the sensor is reading wrong. You'd smell raw fuel coming out the pipe? or smell it on the plugs?

any change from closed loop to open?
 
#9 ·
curious, stock PRB IM? TB? Header?

after reading this again - it may seem like a/f is just not reading correctly. curious if the car is actually running rich or the sensor is reading wrong. You'd smell raw fuel coming out the pipe? or smell it on the plugs?

any change from closed loop to open?
That is the best/quickest test for checking the sensors :up:
 
#10 · (Edited)
I said I wasn't too proud to answer questions about wiring, not that I wasn't proud of my wiring. I take my craftsmanship very seriously and go very careful with wiring so as not to release the magic smoke.

The only spade connectors in the system are at the relays. Everything else is soldered. I hate butt connectors. I raced RC cars when I was a kid where every milliamp counted, so I solder everything.

The ECU is wired exactly according to the Hondata diagram except for the speed sensor, which I already know will not work in my application as Hondata indicates.

I seriously doubt that I have a stuck injector. The idle is very smooth AND the ECU is TRYING to keep it around 11.5:1, indicated by the searching mixture around 11.5:1. If I had a stuck (open) injector, I would not have a smooth idle, it would not be content to circle around 11.5:1 and it would not go desparately lean the second I opened the throttle. I do have a set of spare injectors I could test, but I would need to see a strong argument to replace them first.

11.5:1 is terribly rich, but not rich enough to blow unburned fuel out the tailpipe or foul the plugs.

I don't think I reversed the connectors on MAP and TPS because I labeled them before I removed the harness from the engine. It is possible that our source for the engine reversed the connectors before he crated the engine to send to us, but that doesn't seem likely.

No codes. Wouldn't crossed MAP and TPS connectors throw some vicious codes?

Stock everything from the throttle body through the head. Header with 1-11/16" OD primaries. 2.5" exhaust with Magnaflow muffler. Exhaust is approximately 6' long.

EDIT: Checked wiring colors for intake cam, exhaust cam (also identical connectors,) MAP and TPS against a complete K20A2 sitting here at the shop. While the car I am working on has a Z1, all colors match my labeling and installation of the wires on the current car.
 
#24 · (Edited)
... The ECU is wired exactly according to the Hondata diagram except for the speed sensor, which I already know will not work in my application as Hondata indicates...
I'm almost positive that there is an error in that Hondata drawing... I ran across it myself but don't remember which pin it was.

Wouldn't crossed MAP and TPS connectors throw some vicious codes?
No it does not since both readings are still within range... what does MAP read with ignition-hot but not started?
 
#14 ·
My logic regarding stuck injectors:

If the car has a stuck injector, one cylinder will go terribly rich. Because all the ECU knows is average AFR across four cylinders, it will lean out to compensate and try to get to 14.59:1. Thus, one cylinder would be very rich and three would be very lean. The resulting idle would be rough, not smooth, and, in all likelihood, the ECU would achieve an average of 14.59:1 by starving the three cylinders it can actually control.
 
#17 · (Edited)
in k-manager:
-under closed loop is 2ndary o2 checked? if so uncheck, save, upload
-multiplexer is disabled

its not easy for me to go thru - without being connected & in front of me running but.

under graph tab - click on 1 of the 4 (you may only see 2 for the time being) blank rows. right click one of them & select "define template". you can now define 4 sensors for each row. select different colors for each set of 4

RPM, CLV, O2, AF, INJ, DUTY, INJ, ECT, MAP, IAT, IGN, TPS, BAT, ELD, S.Trim, Knock Count. then "close" open back up & make sure its saved.

with the ecu connected to K-pro you should be able to datalog. click on the lighning bolt to go live & click on red round icon next to it. you should see the recording progess under the graph tab - on the template you defined.

click the red round icon to stop (but don't hit it twice) under Datalog (up top) select save file as. save the file on your desktop. it should be a kdl file

try to record idle, some blips of the throttle for the time being. maybe start cold & watch it warm up.
 
#19 ·
in k-manager under closed loop is 2ndary o2 checked? if so uncheck, save, upload
Yes, unchecked. That was one of the changes Mike walked me through yesterday. I am getting to be more confortable with this system. Yesterday, I knew nothing. I still haven't solved the problem, but I am getting more confortable navigating and making changes.

BTW, Mike, I uploaded a stock-stock-stock OE A2 calibration last night. No changes.
 
#20 ·
Notes: it positively will not come down to its target idle speed of 900. Coolant at 197, it's idling at 1150, give or take.

Mixture is now between 12:1 and 13:1 at idle. No change in stumble.

This has always been there as well: sometimes I get a niftly little blue flame out the tail pipe on break-up.
 
#22 · (Edited)
so its a/f now flucuates but its higher so thats good right? steady but high idle?

under parameters -> idle try moving the slider about 1/3 or 1/2 each way upload. does the idle change. what is your MAP sensor reading at idle. BTW as practice I always save cal then upload.

please go back - & read my instructions I typed for datalogging. I gotta run, may or may not be on later tonight. but will be back the latest tomorrow

ps glad to hear you have that denso & not ngk :nod: thats the right one
 
#21 ·
Just for the nuts of it, I tried adjusting the target idle speed and mixture. Went as low as 600 RPM on target idle speed and as lean as 20:1 on target mixture. This thing doesn't give a flying flip what target speed and mixture are. It's 1100RPM and 11:1, take it or leave it.
 
#30 · (Edited)
Mike,

When I built my first K20 converted X1/9, I encountered the stuck IAC problem after the car was assembled. Since then, it is my standard practice to disassemble the IAC and thoroughly clean and lubricate it before I assemble the car. I always make sure that the IAC barrel spins as freely as a bicycle wheel.

Yes, factory parts throughout.

The throttle stop screw still has the factory "don't mess with this screw" gunk on it. I have not touched it.

When I cover the IAC port with my finger, it idles very low. It will stall coming down from higher revs with the port covered.

Strangely enough, disabling the IAC has little to no effect on idle speed and consistency.
 
#33 ·
Strangely enough, disabling the IAC has little to no effect on idle speed and consistency.
I just want to note that in order to disable the IACV properly, the sensor must be removed completely and covered with a block-off plate.
Disabling it in KPro alone still leaves and air leak in the system when the valve is still bolted in place.
In your case, you can keep the IACV functional like you have in the past.

I just adjusted the IAC duty cycle to the left a bit and now it holds a nice 900 RPM idle (target idle is set at 900 RPM.)

Mixture still rich and it still has a wicked stumble on tip-in, but it is idling at the correct speed now.
Glad to hear the idle is stable now.
As for the rich A/F, I am still unsure and waiting to hear what will fix it.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top