Difference in CO readings between A and B bank

Dear all,

After playing around with different ECUs on my XJ12 HE 1983 (emissions B: europe – no cats, no lamda, no O2 sensors), I finally decided which to use in the future. Unluckily, this ECU (a 16CU) didn’t come with the car, so I thought it would be a good idea to calibrate the ECU using the built-in potentiometer and a CO measurment device (my choice: Gunson’s Gastester G4125). It turned out that the ECU was set a little too rich (4% CO), but turning the pot a few clicks counterclockwise did the job.

So far, so good. However, I noticed an interesting difference between the two banks regarding the CO readings. After the calibration (I aimed at 2% CO – get rich or die tryin’, eh?), the right bank provided a reading of 2,3% CO, with 1,9% CO on the left bank.

Is that an acceptable variance, or something to worry about? I was thinking: Could the location of the idle adjustment screw on the left intake manifold result in the left bank running slightly leaner than the right bank? I know that the left and the right intake manifolds are connected to each other, but even air may like to take short cuts??

I should add that there was a problem with an injector on the right bank, which lead to cylinder A4 running terribly rich. That is now solved since a few hundred kilometres. I also used some additive to clean the system (valves and injectors – at least, that’s what the ads claimed), but, of course, it could be that A4 is still a little more carbonized than the other cylinders.

Everythin else has been double and triple checked (and in many cases: repaired) during the last weeks for different reasons, all of them being more or less related to a careless PO: throttle plates, injectors, injector hoses, injector wiring, spark plugs, distributor, ignition timing … It really took some effort, but now the car runs perfectly smooth under all conditions, the same applies to the idling.

So, what do you think? Should I just leave it that way, or dig into the ignition system – again?


Good to see you are making progress!

First of all, are you measuring the CO² in closed loop or open loop ?
In Open loop (no 0² feedback) if one injector over-fuels you will get higher reading in that bank.
In Closed loop (O² feedback) if one injector over-fuels, the ECU will lean the entire bank until the O² sensor says it’s ok.
If you have different readings in Closed loop you it could be that you have a weak O² sensor.
There is an extra plug beside the ECU and you can switch between the two modes by unplugging the bridge cable.
Keep in mind that if you don’t have heated O² sensors they need to be warm to work. If they are not the ECU will switch to Open loop by itself.

By default Closed Loop is disabled at P and N.
Measuring CO² between the two modes could maybe help you identify what is really happening.

There is a detailed chapter about this on Kirbert’s book.
Also: http://www.jagweb.com/aj6eng/lucas_efi.php

Hi Aristides!

Thanks for your reply!

Your question is a bit difficult to answer. I have emission system B. As far as I know, there is no distinction between open and closed loop. There is absolutely no feedback mechanism installed, no O2 sensors, etc. pp. So, in other words: ECU operates continually in some sort of open loop mode, but I don’t know if it is 100 percent comparable to the open loop mode with different emission systems (especially the US cars). At least the vacuum advance mechanism is not the same.

Hence, the difference between the banks can only be due to differing ECU signals for the banks (not very probable), functional variance within the injectors (let’s call them best agers), ignition (distributor, cables), combustion (spark plugs, compression), air intake (e. g. misadjusted throttle plates), exhaust (small leaks). All of these mechanisms have been thorougly tested by me, albeit by means of the limited devices had at hand.

I am just asking myself: Is the difference in the readings acceptable, or does it tell me that I do have to go into all that stuff again? Is there some sort of definition as to how much variance is tolerable? What is your experience?

And, I am still wondering if the location of the idle adjustment screw at B bank could effect CO readings at idle speed. It seems that CO values basically represent unburnt fuel, so the less air the cylinders get, the higher the value should be.


Doesn’t B bank draw all air at idle and feed A bank thru the balance tube? Maybe that creates the imbalance at idle. Have you measured at higher rpms?

Yep, that is exactly what I was thinking: B bank draws the air and isn’t interested in providing A bank with an equal share. You are right, Chris, I should experiment with higher rpms. That should validate the hypothesis. (The neighbours in suburbia will hate me.)

PS: There is, I think, always a little bit of air passing the throttle plates on both sides, even with the throttle fully closed, but the idle screw obviously is a relevant factor.

The theory of the AAV giving more air to the B bank is quite possible.
We need some Flow Dynamics experts here.
I also don’t know if this amount of difference is an issue or not.

Yes, there is always some air passing from the throttle plates. Maybe worth re-checking the plates gap ?
A test on higher RPMs is a good idea. It eliminates the AAV and slight differences in the throttle plates. If the difference is still there I suspect one, or more, fuel injectors are giving a tad more fuel thar the others. Or one, or more, have a somewhat restricted flow ?
Did you ever tested their flow rates ?
Did you ever check their filter screens ?

I don’t know if the electronics inside the ECU can get “tired” and thus result in slightly different pulse signals.
@RogerBywater / AJ6 Engineering will know better.

The only problem I can see is that as you can only measure the average and not every individual cylinder, one, or more, cylinders might be running lean, not a very good thing. But then again, it might be minimal, don’t know.

I got a slightly better idle (1988 US version XJS) by setting the throttle stop screws about 1/3rd turn more open than .002". I then had to turn aav idle screw in a few turns and reset throttle linkages.

My thinking is, the A bank (and B bank?) now gets a bit better airflow from center of intake manifold as opposed to back of manifold. Of course, you cant do this if your AAV idle screw is almost all the way in. But i had plenty left to turn.

Its also important to have a functioning pcv valve. I measured about 200rpm at idle that it contributes. The intake manifold draws that air from the crossover pipe, nicely located at center of intake manifolds.

I doubt you will ever get it balanced with no O2 sensors. That ECU has no reference to lean or richen a side of the engine with long and short term fuel trims as in later cars. Coupled with no mass flow sensors and only a single vacuum reference, the ECU has only it’s built in mapping to go by. It has no way to tell one side of the engine from the other. The best you can do is ensure there are absolutely no vacuum leaks to upset the mixture. A cheap smoke test machine is a good investment.

And even with O2 sensors like my xjs has, when setting ecu fuel mixture i can see both banks going back and forth quite a bit.
These engines just aren’t that well tuned with 80s technology.

Sometimes I’m just happy that it turns over, and will get me to the shops. Fastest pint of milk I ever buy :wink:

1 Like

Hi guys!

Last weekend I did some more testing with Gunsun’s gastester. As some of you expected, some variance seems to be a principle of this non-feedback ignition system – and the gastester I used is no precision device for sure as well.

What did I do?

  1. Warm idle, again about 1.9% CO left bank, 2.3% CO right bank.
  2. 2.500 rpms: 5.0% CO left bank, 3.9% right bank. (The gastester is not intended for use with higher rpms, so one shouldn’t take these readings too seriously. Generally, CO-levels are supposed to be lower at higher rpms compared to idle readings, but, who knows.)
  3. Warm idle, directly after the 2.500 rmps session (which took some minutes – the gastester prefers slow measurements): 2.0% CO left bank, 1.7% CO right bank – which is to say: The banks switched positions regarding their CO level!

After that series of tests I let the car sit for about 30 minutes and looked at some of the spark plugs – all in perfect condition, no indications of running lean or rich. Then started the Jag again (engine of course still warm) and did a final idle test, which showed the same results as test 1) and the test I did the other day.

I can’t say what causes the differences and the shifts between the banks – my first guess would be a slight variance in air flow --, but I suspect that they don’t point to an issue that can be solved given the ignition system the car uses. If you think I’m wrong, please let me know!


PS: Vacuum system has been overhauled a few weeks ago, all elements tested, all vacuum lines renewed, everything double checked afterwards and tested for leaks – negative.

PS: I gave the fuel injectors a service recently, all filters etc. are new, fuel rates have been checked by means of a primitive diy device. Fuel rates seemed to fall within a very small bandwith, if I measured correctly.

Good evening everyone, I’m a new member that just joined yesterday with an 86 XJS. I’m also an licensed mechanic for 40 years. Maybe I can shed some light on CO, HC, NOX, and O2 sensors. All single wire O2 sensors produce their own voltage by the gas that passes by it, between .1 volt to almost 1.0 volt. .5 volt is a happy well burning stoichiometric balanced. You can simply pull the wire off the O2 sensor connection place your VERY GOOD DIGITAL METER start the car, let it warm up and watch it zig-zag up and down quite rapidly. If it’s slow then the O2 is no good. If it reads always above .5 volt then you have a rich bank, if under .5 volt it’s lean and you have a vacuum leak or bad injector possibly exhaust leaks or even valves. The meter leads should be,red on the wire of O2 and the black on a good ground. Also please do not play with the throttle valves, set it at .002!! . always check fuel and ignition before you try to hide the problem with throttle adjustment. HC is unburnt fuel or oil. CO is not completely burnt fuel or oil… remove the linkages and verify RPM with a basic Tach or Timing gun. Then once warmed use the volt meter and start removing an injector plug or lean it out with a vacuum hose which will determine that bank may be burning more than fuel or oil.
I hope this helps
Pierre (Pete)

Hello Pierre - Welcome to Jag-Lovers - when you get a chance, add your country to your profile so that a country flag will display by your name and so that others close to you might be able to receive, or give, some help (since this is an international site) - this sounds like a good explanation of troubleshooting the O2 signal and good reminder to have the throttle butterfly setting correct - Tex Terry, II - 1991 XJS V12 Classic Coupe, 1986 XJS V12 Coupe - sent 11/16/2020 2036hrs. EST USA

Good evening Tex Terry, thank you. I don’t know how or where to add my profile. I live in Canada, I’ve had my 86 XJS for nearly 20 years and now we gave it to my 18 son as a gift. We began restoring it, all custom chrome pieces, all aluminum polished , new interior, shaved door handles , etc. It’s the greatest feeling working with your son on a car. Happy to help.

Hello Pete - go to the top of the front forum page, on the right hand corner, were your “C” in the circle is located - click on the “C” and then use the drop down menu to update your profile - when you put in your country then the flag should show up on your next post, after your name - there are many members on here from Canada, not just in the XJS section, but also in some other sections - navigate through the sections for more information on usage of the site and enjoy - if this is not correct, then maybe someone else will chime in to guide you - glad you are keeping alive the interest by passing along the learnings to your son - Tex Terry, II - 1991 XJS V12 Classic Coupe, 1986 XJS V12 Coupe - sent 11/17/2020 0007 hrs. EST USA.

Hello Pete - I have a lot of work to do on my 1986 XJS, such as transmission change, engine bay area clean-up, headliner replacement, interior cleanup, floor pan repair, jacking points replacement, IRS service, so I may well be asking questions - I have been reading many posts here while I am healing from knee surgery in hopes that I will be ready to go in a very planned manner for my repairs - Tex terry, II - 1991 XJS V12 Classic Coupe, 1986 XJS V12 Coupe - sent 11/17/2020 0015hrs. EST USA.

I think your spark plugs are the best indicator that your engine works fine.
If you want to be 100% sure remove them all and look if there is one that’s darker or lighter than all the rest.
But the fact that also you don’t get one bank being consistently richer than the other leads me to believe that considering the 80’s technology you are dealing with, it is as good as it gets.

After all the work and testing you did I think so too.
Don’t forget that it’s far from a “perfect” engine (like all automotive engines).
It’s all a compromise, packaging, cost etc. etc.
Different intake runners lengths with the center cylinders being straight behind the throttle, slightly different compressions between cylinders, the AAV favouring one bank, different valve gaps, and dynamic flows that are constantly changing in different rpm’s.

Time to drive and enjoy!!

Yep, seems so: Time to drive! Thanks for the encouraging feedback, Aristides! Just to close this subject: Today I installed another working ECU I have, and guess what? Slightly different readings again: Warm idle gives nearly 1% difference in CO readings between the banks, with the left bank running richer this time. I take that as a final proof that the ECU interaction with the engine is not a matter of precision. But, hey, that’s a leading principle of these cars, isn’t it? Being, roughly, perfect.

1 Like