Madabout Kitcars Forum

Madabout Kitcars Forum (
-   Marlin 5exi builds (
-   -   IVA fail: emissions (

MadDogMoggy 1st June 2013 23:52

IVA fail: emissions
Engine is a K20A2 out of a 2004 Civic Type R. Emissions were very high and lambda was low on IVA test.

I have the 'standard' honda engine exhaust, ie 4 into 1, into CAT, into box with two tailpipes. dbs were fine (average of 97).

I'm using a Hondata ECU with the standard K20A2 map as supplied by Hondata. Fuel line runs from tank (standard Rover pump) to rear of car. Fuel regulator has an input, regulated output (which runs to fuel rail) and a return (which runs back to the tank). Single line from regulator to fuel rail, which I believe is set at 42 PSI.

CAT is as supplied by MArlin and 'smelt' like it was working correctly. I have had low voltage errors with the lambda sensor, and am now getting a primary O2 heater error having played with a second sensor today.

MadDogMoggy 1st June 2013 23:55

Whilst monitoring the data on the K-Pro software today, initially I was getting no errors and a primary O2 voltage of between 0.01 to 0.2 volts, fluctuating quite a lot. The current was reading as -20 to -35 mA (yes minus!)

Later I was getting Heater errors, but the current was positive (I think) and the voltage about 1.4v

Anyone know what the correct readings should be? Initial thoughts are it could be the sensor, (or the wiring) and I was wondering what readings I should be getting!

alackofspeed 2nd June 2013 19:59

Tim, a few questions / comments:

- What short term fueling trim is the ecu showing?
- The ECU is in closed loop mode?
- Factory pressure is 43 PSI, but 42 should work.
- Lambda current is only for the stock wideband on non-euro cars.
- Stock zirconia narrowband sensors are 0.2v lean, 0.8v rich. If in closed loop dithering mode (which it is unlikely your car is achieving going on your lambda result) it'll switch between about 0.8 v and 0.2 v cyclically at a frequency proportional to engine speed, circa 1 Hz at idle.
- I found the stock map too lean in open loop. If you're maxing out short term fuel trim in closed loop to try and bring the fueling back to stoichiometric, then that would register as lean. Beware the Hondata map, it is only a basemap, despite it purporting to be a UK calibration replica.

MadDogMoggy 3rd June 2013 21:14

I'll check trim at the weekend. I've not made any changes to the map, so I assume this will be the default setting with the Hondata base map.

It sounds as though my voltages are too low, with a minimum of practically 0v and max of 2 to 2.2v. I believe I'm in closed loop, as I've not changed it to open loop.

This could therefore be the problem - the ECU senses a value of 0.2v or less, and therefore thinks that the engine is running lean. It then chucks in more fuel. The tester noted that after failing the test, the CO reading started to slowly rise. this again would relate to a low voltage: ECU thinks it's lean, adds more fuel CO goes up. Voltage is still low, it ads more fuel, CO goes up some more.

I'll play around with the connector some - I think the wires might not be making good contact in the 4 pin connector I've used to attach the lambda sensor. It's currently coming up with a heater error on the AF/primary O2 sensor.

If the weather's good at the weekend and you fancy a flying visit you'd be more than welcome!

As always, thanks for your input, John.

alackofspeed 3rd June 2013 21:48

I disabled the O2 heater and ran a switched 12v supply to the sensor. Try taking a constant 12v to the sensor heater and see if that helps - if for some reason the sensor isn't hot enough it'll not work.

My recollection is you took a copy of the map I run on my car. As mentioned when I came over in the Marlin, the stock Hondata map is pretty terrible - lots of holes in it and some curious steps in, well, everything. I'd suggest giving my map a try in closed loop. I know from my wideband that the fueling will go to perfect stoich when run in closed loop as the trim required is small. Some correction will be required by the closed loop, as I deliberately leaned off the mix slightly for very low loads to improve fuel consumption.

Busy this weekend I'm afraid: XC Friday night, DH-ing at UK bikepark on Saturday, XC on Sunday - gotta get the bike out whilst the weather is good!

My Marlin is looking sorry for itself at the moment - sat the garage with missing and broken front suspension waiting for an injection of time to make new wishbones.

If you want to borrow my wideband to see for certain what your fueling is doing, that's an option - you'd just need to put a second lambda boss on the exhaust to mount it to.

MadDogMoggy 4th June 2013 13:57

Lots to ponder.

I may well try wiring the O2 heater directly to the battery (via an ignition live relay) to see if that helps.

I do have a copy of your map, so I'll have a play with that as well - although our setups are a bit different I would imagine it's a much better platform than the hondata base map (especially considering the time you've put into testing it).

My exhaust does have a second boss for an O2 sensor (post CAT) so I may take you up on your offer of borrowing the wideband - I'll see how I get on at the weekend.

Sorry to hear about your Marlin - hope it's up and running again soon (especially with this nice weather). I'm out biking tonight (only some basic cross country/singletrack). Hope you have a good weekend.


MadDogMoggy 10th June 2013 12:38

Just a quick update:

Had a play with the connector at the weekend - stripped the contacts out of the plastic housing and connected it all up.

Still had an error on the O2 heater, so I'm going to miss out the ecu on this altogether (and turn off the error warning) and wire it straight to the battery via a relay (for ignition on).

The actual O2 sensor seems to be behaving, with a voltage of between 0.3v and 0.5v, so that could have been down to a bad connection, and may further improve with the re-wiring of the heater.

Once the heater is re-wired I'm going to take it to a garage to get the emissions checked, both with the original Hondata map and then with John's.


alackofspeed 10th June 2013 16:34

I'd hope you can achieve acceptable closed loop control with either map. I think for the test I'd go with whichever has the least fuel trim to get it to stoich.

Ric H 13th June 2013 00:01

Tim, I also found the lambda heater needed feeding directly with a switched live. Initially I wired it up with a relay switched by pin 8 on the E connector like in the Hondata engine swap diagram on the website but it doesn't work for the euro type R. I got hold of the Honda workshop manual and the schematic in there confirms that the primary lambda is fed 12v direct on the orange/black or yellow/black from a switched live rail.

I'd already fitted the relay by this stage so it got re-purposed as part of an immobiliser!


MadDogMoggy 13th June 2013 20:01

I had another quick play yesterday evening.

The switched live that feeds the heater had blown its fuse, so the heater wasn't working. Replaced it with a larger one (went from 10A to 15A). This stopped the O2 heater error!

Voltage of the sensor (bizarrely) is back to fluctuating between 0.01v and 0.23v. Not sure why it was reading 0.2 to 0.5 at the weekend.

Riachard - any reason why the relay didn't work? I was thinking of using the original switched live to operate the relay (earthed directly, not through the ECU) and then have a direct feed from the battery through the switched part of the relay to the heater, again earthed to a suitable earth point.


alackofspeed 13th June 2013 20:56


Does your post mean that last night the heater was being fed a constant 12v with fuse replaced?

If the lambda signal voltage was still dubious with the heater working, I'd suggest the wisest route would be to put a wideband in and check to see if the narrowband is reporting correctly.

Incidentally, if the narrowband sensor is functional, if whilst logging the output using the hondata you were to blip the throttle, you should see the lambda voltage change to rich with throttle tip-in enrichment.

What brand of lambda are you running by the way? I first fitted a universal sensor I had lying around when I put my car together. Although it appeared to be working, for reasons I never bothered to fully investigate the ecu / engine didn't like it, the evidence being a minor miss-fire on idle. It suggested the lambda reported stochiometric incorrectly, causing the fuel trim to be lean. I put in a sensor from my donor car (damaged wires, but bodged to work) and all was well. My car for this reason has a NTK sensor fitted currently.



MadDogMoggy 13th June 2013 22:06


I've used a couple of Lambda sensors: the first was from the donor but it had a longer lead, so I think it was actually the secondary O2 sensor from the car.

The second (and currently fitted) I bought off ebay and came from a civic type R. I don't know if there's any real difference between the two - they look identical apart from the stamped on part numbers and a shorter lead.

At idle the injector duty cycle is about 1 or 2%.

When the voltage reading was between 0.2v and 0.5v it was dependant on throttle position.

Currently the voltage (at idle) is about 0.2v to 0.23v and is pretty stable, although it occasionally drops down to almost 0v. From memory a blip didn't have much effect but revving at about 3k made a change.

I think I need to be a bit more scientific in noting down what's happening now I've got the heater running! The plan is to rewire the heater on Sunday and then check the wiring for the sensor (in case there's a dodgy joint or something) and then see what the situation is. Using a wideband to actually see what the emissions are is a sensible idea - at the moment I'm just speculating over voltages!

Just as an after thought, what's the part number of your sensor, and what are the wire colours (I have 2 white (heater) a grey and a black)?

Ric H 13th June 2013 23:16

The relay was fine, it was the earthing through the ECU that didn't turn it on - I guess the type R ECU just doesn't include for heater switching. I didn't bother using a relay as there didn't seem any point - I could easily run an adequate direct feed. The installation in the civic doesn't use a relay I don't think (not according to the schematic I have anyway).

alackofspeed 14th June 2013 22:16


Originally Posted by MadDogMoggy (Post 44477)
Just as an after thought, what's the part number of your sensor, and what are the wire colours (I have 2 white (heater) a grey and a black)?

Not a clue about the part number, I just bought an OEM spec' NTK narrowband sensor from the somewhere on the internet about 5 years ago. It has the standard zirconia sensor colours: 2No. white, 1No. grey, 1No. black.

MadDogMoggy 16th June 2013 21:45

Had another play with the O2 sensor wiring today.

Rewired the heater using a switched 12v to power a relay, which in turn feeds the O2 sensor's heater. Turned off the error reporting on the ecu (since the heater circuit no longer connects to it) and checked the heater was getting a supply - all working fine.

I managed to find the part number for the correct primary O2 sensor for the type R. Turns out I have 2. I removed the one I currently have fitted using a 4 way connector I bought from CBS and replaced it with a spare I had which still has the original connector on. I found the female part and wired that into the loom.

Warmed the car up and checked the voltage of the O2 sensor. At idle it fluctuates from about 0.16v to 0.20v. Revving or blipping the throttle causes it to drop as low as 0v and then it recovers back to about 0.2v.

This is the 3rd sensor that I have used which gives this low range, instead of an expected 0.2v to 0.8v.

The car's booked in to a local garage on Saturday to get the emissions checked. With the heater now working they may be all right, but I still think the ECU is measuring a low voltage. If this is the case, the only thing I can think or is that there must be a large amount of resistance in the sensor circuit somewhere which is reducing the detected voltage: perhaps I have a poor connection in a plug or a wire not fitted right.


alackofspeed 17th June 2013 16:52

Sounds like ecu is seeing a lean signal, correcting by maxing out the +ve fuel trim, hence the lambda value and emissions result seen at the IVA.

I think sadly that Saturday's test will prove nothing other than no change since IVA.

I think the best test from here is to rig up a AA battery with a potentiometer, set the voltage using a multimeter, then put that known voltage to the engine bay sensor plug (with sensor disconnected). If the ECU reports the correct voltage, then at least you know the ECU isn't telling porkies / the wiring is okay.

Shouldn't have thought there'd be a voltage drop in the wiring from the sensor to ecu, as theoretically there's no current flow in the signal wire.

To confirm, you have the following?

2No white - 12v from battery via a relay.
1No black - connected to ECU lambda pin (D14?).
1No grey - connected to the SAME earth as the ECU?

MadDogMoggy 17th June 2013 23:00

Ok this has got me thinking!

My assumption was that the ECU fed the O2 sensor with a current/voltage and measurd the voltage drop (ie the variance in the resistance of the sensor, which change with the amount of oxygen in the exhaust gas)

Bearing in mind the following is from my build notes - I've not actually looked at the car this evening.

I wired up the ECU following RichH's wiring diagram, which was invaluable.

The connections that I have are:
E1 Fuel relay coil
E4 O2 Sensor -ve (grey wire)
E7 Main relay coil
E9 Relay for O2 Heater
E26 RPM feed
E31 MIL feed

A1 O2 Heater (now disconnected)
A2 12v feed from main relay
A3 12v feed from main relay
A6 O2 Sensor +ve (black wire)
A18 Speedo
B8 Fan relay

The ECU doesn't appear to have an earth (not via a pin anyway), which is a bit confusing!

I've looked through a PDF of a manual which has some pin outs for the ECU, but I'm not sure if it cover the Type R. Interestingly:
A4 and A5 are listed as grounds for the ECM circuit
A16 is AFS sensor (-ve)

E4 is just listed as 'sensor ground'

Could this mean then that a) my ECU isn't earthed properly, and b) my O2 sensor is wired wrong?

There also seem to be a lot of other connections missing (like those to VTEC solenoid)!

I'm wary of adding/changing wires in case I damage the ECU!

alackofspeed 18th June 2013 15:48

I may have got the sensor wires totally wrong - don't rely on the pins I mentioned - I haven't touched the wiring on my car since 2006.

The comment about voltage remains though - with the narrowband, the Ecu is simply measuring the voltage.

If your ECU isn't correctly earthed that could explain the voltage offset and the problems.

I have a honda manual somewhere with the official wiring diagram within. I'll dig it out tomorrow if I have a chance.

MadDogMoggy 19th June 2013 19:10

Had a quick look at the actual wiring earlier on this evening.

The wires on my E connector are all fitted individually (after Marlin lost the one I had when they had my ECU and did nothing with it for over 6 months).

The A and B connectors have pugs fitted to them from the engine loom.

I have a diagram that seems to tie in with the Type R ECU, and it shows the AF sensor negative going to Pin A16. However on my A connector (not the ECU pins themselves) there is no wire attached to Pin A16 (and no 'female' contact inside the plug.

I'm therefore wondering if I have a slightly different model of loom, or that if the Type R doesn't use that pin. After all, both the Hondata wiring schematic and Rich H's wiring diagram both have the AF -ve going to SG3 (sensor Ground 3) which is pin E4.

Whilst disconnected from the ECU I check continuity (which is good) and resistance (which is low).

John - let me know if you find your wiring diagram - I'd be interested to compare! Meanwhile I'm going to try your suggestion of a battery and resistor!

MadDogMoggy 22nd June 2013 15:08

So rewired the O2 sensor heater with a relay, then checked through the sensor's wiring. Connection to ECU E4 pin was earthed, and A6 supplied about 0.4v.

Also contacted company I bought Hondata system from. Guy said they wire it the same as the diagram on the Hondata site. Also said he thought I'd be lucky to get it through emissions without getting it mapped properly.

Took it to dad's mate's garage this morning. Put it on their machine and it was still reading too high (but better). Ran the engine at 2500rpm for about 5 minutes then rechecked it, and low and behold it passed. By quite a lot (CO was registering less than 0.075 and it needed to be under 0.2). So, fingers crossed for IVA retest.

Managed to sort the headlight positioning at the same time. It required a bit of headlight fettling to reposition the whole light slightly.

Thanks everyone for your help, input and support.

All times are GMT +0. The time now is 14:25.

Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2021, vBulletin Solutions, Inc.
Copyright Madabout Kitcars 2014