3 SU-carbs: oscillating AFR-value

[danxke]

Hello,

I've fitted 2 wideband oxygen- ("lambda")-sensors into my 4.2L. The sensors are located between engine and main silceners. The sensor-controllers are from 14point7.
My engine has been rebuilt to standard specifications (from VSE in Wales) and comes with a 123jag-distributor. Exhaust and carbs have no (air-)leaks.

I find it pretty convenient to inspect the AFR-values on my mobile while adjusting the carbs.
I've got UM-needles, the carb-springs are untouched, i.e. they might be wrong, or faulty.
I managed to get a decent idle @750RPM (although with a rather rich mixture) and the mixture becomes leaner when revving up but not too lean.

I noticed that the AFR value oscillates slightly during idle, also the RPM is not 100% constant.

I've uploaded a screen capture from my mobile App which I recorded during my attempts to get rid of the oscillations:

What did I try?
- adjusted the mixture of the center carb
- fixated the piston of the center carb to avoid oscillations of the piston
- checked the damper oil levels

Apparently my carbs suffer a bit of sticking pistons, despite the "click"-test...

What do you guys think about the oscillations?

Greetings
Dan

[PeterCrespin]

You may find they coincide with the orbit of Pluto, in which case you can only drive 15 nights a year. Either that or the gravitational pull of Ready-Mix concrete trucks driving past.

More seriously, I am looking forward to seeing how this thread develops as I am considering setting up carbs with O2 sensors and an app, as my Souriau analyser needs a station wagon to move it.

[MarekH]

Sadly, I can't see your video, but...

The AFR and rpm will move if either the air or the fuel quantities move about, and/or also if the amount of advance is moving. A flattish advance curve with a dip at idle will tend to stabilise idle.

You will probably find that a stable idle will be slightly rich, e.g. about 13.5AFR, whilst it'll tend to want to take off and race if it is lean. Similarly, retarding advance will tend dampen the engine.

Different cam profiles like different advance amounts, e.g. a fast road cam may happily tolerate a huge amount of advance but be very lumpy under a low advance regime at idle.

Your observations are complicated by having three carburettors monitored by two sensors. Using something like a Colourtune spark plug on each cylinder in turn might help you assess whether any of the cylinders are leaner than each other.

kind regards
Marek

[danxke]

Hi Marek,
thank you for your input!

The AFR and rpm will move if either the air or the fuel quantities move about, and/or also if the amount of advance is moving. A flattish advance curve with a dip at idle will tend to stabilise idle.

I've got the 123jag distributor (with curve F) which has no advance below 1500 RPM or so.

You will probably find that a stable idle will be slightly rich, e.g. about 13.5AFR, whilst it'll tend to want to take off and race if it is lean. Similarly, retarding advance will tend dampen the engine.

RPM is also oscillating a bit, but not really noticable. So you are saying that probably the engine takes off and stalls?

thanks.

[christopher storey]

The oscillation might well be caused by the float levels being slightly too high. Thus when the needle valves are open, the fuel level rises and the mixture richens ( slightly) , and vice versa when the needle valves close. However, in general I would not worry at all about your figures. The XK engine likes a richer than usual mixture at idle and an average of 12.5 stoichiometric ratio at idle is about right, equating to a CO level of about 5% to 5.5% , leaning away once the engine is held at steady fast idle. This appears to be what happens, if I interpret your video correctly

[danxke]

Hello Christopher,

thank you for your feedback!

The lambda-value oscillation range is around λ0.09 (between λ0.84 and 0.93), i.e. AFR 1.3 (AFR 12.35 - 13.67).

Interesting that you come up with the float chamber levels as probable suspect.

I am using a modern (non su-type) fuelpump, which might build up too much pressure without really flooding the jets...

What do you think?
best
Dan

[christopher storey]

Dan : the XK engine ( indeed all SU carburetters ) is definitely fuel pressure sensitive, and the pressure should not exceed about 0.2 bar = 2.9 lbs/sq.in . many people who have fitted e.g. Facet pumps fit a variable pressure limiting valve for this reason, as many of the aftermarket pumps give out 4 to 5 lbs pressure

[Heuer]

Even with a stock SU pump I use a Filter King set to 2 psi in place of the standard filter. Although the SU pump is rated at 3.8psi it pulses at 6psi! Full Filter King details here: http://forum.etypeuk.com/viewtopic.php?p=711#p711

[danxke]

I noticed an immediate improvement at idle with a far less 'splashy' exhaust note meaning a less rich mixture. The car revs perfectly through the range, even with my modified engine, with no misses suggesting 2 psi is ample pressure for the SU's.

WOW that sounds promising!

Thank you, Mr Heuer :-)

[abowie]

Are you sure that the "oscillations" you are seeing aren't normal?

With the caveat that I don't know anything about how combustion and valve timing affects measured AFR in exhaust gases, I doubt that it is constant throughout the exhaust cycle.

It may simply be that you software setup has no damping/dwell/averaging built into it to smooth out variations into a nice neat average number.

[danxke]

Hi Andrew,

you are right: my data acquisition does not calculate the rolling average over the sensor values. I could do that and would achieve a smooth line of AFR/lambda values.

The problem is that I don't know if a) these oscillations are a data acquisition artefact, b) a symptom of a problem in my car or c) totally normal.

I know only of one person who is doing AFR-logging currently; he has posted a screenshot in the German forum:


You can see similar oscillations (with smaller amplitudes than mine) if you look at the values labeled "lambda123", but his other bank (lambda456) appears to be smooth like the river Rhine in a boiling summer ;-)

Anybody else here doing AFR-logging?

Greetings from Germany :-)
Dan

[mgcjag]

Hi.....you said in a post above....I've got the 123jag distributor (with curve F) which has no advance below 1500 RPM or so.
As far as im aware quite a few on the forum with a 123jag use setting 0-1-2 ...which will give a max advance of 34deg @3600 ...setting F will give 42deg.....also any of the settings give 12deg advanve between 500-1200rpm
Have you tried the lower settings.....Steve

[danxke]

Hi Steve,

I've got the 123jag distributor (with curve F) which has no advance below 1500 RPM or so.

I gave wrong info from the top of my head, sorry. I am using curve "0".

.....also any of the settings give 12deg advanve between 500-1200rpm

What I meant was: the 123 doesn't add avance degree to the static advance below 1200 RPM or so. The point was that the advance curve is flat @idle and that is probably not the reason for the AFR-oscillations.

:-)

greetings
Dan

[danxke]

Hello there,

upon Mr Heuers recommendation I checked my fuel pressure, which is ca. 5 PSI (ca. 0.34 bar):


I've replaced my Pierburg pump with a new SU pump, and installed a fuel pressure regulator between filter and carburetors. I set the fuel pressure to ca. 3 PSI.

The process of adjusting the mixture and hunting for the most even idle can be seen in my video:

(Audio alarm!).

Thank you Mr Heuer, with your help I have been able to solve a mystery which is 20 years old (really!).
I am owing you a beer!
Dan