is it better to run a little rich or a little lean

[kayakwesty]

I need to adjust my AFM and I'm showing just a little bit rich...and I mean just little bit rich, and was wondering if it was worth fooling with
_________________
1984 Westy 2.0 Jetta ABA Motor/1.8 head converted to auto tranny, with 180,000 on body, 55K on engine and transmission

B.5.5 Passat Wagen with 30v V6-67000 Miles

DAS AUTO

CHANGE YOUR FUEL LINES!

A post without pictures is useless

http://www.kayakwesty.com

http://www.waldensridgewhitewater.com/

[Dogpilot]

Are you speaking of the idle adjustment or the tension on the flapper? One of the misconceptions of adjusting the AFM spring tension is it will adjust the running mixture. This would work if there was no O2 sensor. The ECU will adjust the mixture, best it can back down to what the O2 tells it is right. Now if the ECU has hit its limit for pulse width on the injectors, then it would have an effect.

You would need to disconnect the O2 to let the ECU run from its memory and measure the output gas with a gas analyzer.

The culprit normally is the O2 sensor, as it ages, it responds slower and the signal is to the rich side. It can be killed off by a couple of things beside age, coolant and silicone.

an excerpt from this web site:


http://www.picoauto.com/applications/lambda-sensor.html
Drivability Symptoms

A lambda sensor’s normal life span is 30,000 to 50,000 miles. But the sensor may fail prematurely if it becomes clogged with carbon, or is contaminated by lead from leaded petrol or silicone from an antifreeze leak or from silicone sealer.

As the sensor ages, it becomes sluggish. Eventually it produces an unchanging signal or no signal at all. When this happens, the Check Engine Light may come on, and the engine may experience drivability problems caused by an overly rich fuel condition. Poor fuel economy, elevated CO and HC emissions, poor idle, and/or hesitation during acceleration are typical complaints.

If the average voltage from the lambda sensor is running high (more than 0.50V), it indicates a rich condition, possibly due to a bad MAP, MAF or Air Flow sensor or leaky injector. If the average voltage reading is running low (less than 0.45V), the mixture is running lean possibly due to a vacuum leak or because the sensor itself is bad.

If the lambda sensor continually reads high (rich), it will cause the engine computer to lean out the fuel mixture in an attempt to compensate for the rich reading. This can cause lean misfire, hesitation, stumbling, poor idle and high hydrocarbon emissions (from misfiring).

If the lambda sensor continually reads low (lean), it will cause the engine computer to enrich the fuel mixture. Injector pulse width will increase causing fuel consumption and carbon monoxide emissions to go up. Constant rich fuel mixture can also cause the catalytic converter to overheat and it may be damaged.

If the lambda sensor’s output is sluggish and does not change (low cross counts & long transition times), the engine computer will not be able to maintain a properly balanced fuel mixture. The engine may run too rich or too lean, depending on the operating conditions. This, in turn, may cause drivability problems such as misfiring, surging, poor idle, and high emissions.

If a heated sensor has a faulty heating circuit or element, the sensor can cool off at idle causing the system to go into open loop. This usually results in a fixed, rich fuel mixture that will increase emissions.

Sometimes an apparent lambda sensor problem is not really a faulty sensor. An air leak in the intake or exhaust manifold or even a fouled spark plug, for example, will cause the lambda sensor to give a false lean indication. The sensor reacts only to the presence or absence of oxygen in the exhaust. It has no way of knowing where the extra oxygen came from. So keep that in mind when diagnosing oxygen sensor problems.

The lambda sensor is also grounded through the exhaust manifold. If rust and corrosion of the manifold gaskets and bolts is creating resistance, it may affect the sensor's output. To rule out a bad ground, use a digital volt meter to check for a voltage drop between the sensor shell and the engine block. More than 0.1v can cause a problem.
_________________
Geology with a Syncro rocks!
86 Syncro Westy AKA "The Bughunter"
98 Disco I
08 Range Rover SC
08 VW Rabbit S
1951 O-1G

[kayakwesty]

Actually I bumped the hex screw setting on my AFM,(the factory plug is missing ) when cleaning my K&N cone air filter (I have to remove the instake boot and AFM to clean it) and my Ken Lewis O2 meter is now showing a little rich, and I was wondering if I was better off not touching it, or slowly tweaking it, I don't want to burn a valve,but I also want better performance and mpg...and not to pollute also
_________________
1984 Westy 2.0 Jetta ABA Motor/1.8 head converted to auto tranny, with 180,000 on body, 55K on engine and transmission

B.5.5 Passat Wagen with 30v V6-67000 Miles

DAS AUTO

CHANGE YOUR FUEL LINES!

A post without pictures is useless

http://www.kayakwesty.com

http://www.waldensridgewhitewater.com/

[tencentlife]

Is the Ken Lewis thing a wideband controller? CO sniffer? or just a narrowband voltage display?

It takes 2 to 3 full turns, or more, of the CO screw to alter idle AFR by one full part air (e.g. 13.7:1 vs. 14.7:1), so a small inadvertent change in the screw position isn't likely to produce any substantial change in mixture, and anyway you really have to turn the screw a lot to put the system out of lambda response range if you have a good working o2 sensor. You are almost always better off erring on the rich side, within reason, and the published CO adjustment tables indicate that that is the preferred tune. If lambda is working properly, though, the only times you'll notice are when the engine is during warmup, especially warmup idle, and when you're at WOT under load. Being off by one full part air could mean the difference between having a smooth warmup idle and a stumbly one until it goes closed-loop; the difference will be much less obvious under WOT load but there is a noticeable drop in full-load power when you go from 12.5:1, which is what you want, to something like 14:1, with an otherwise well-tuned engine.

But in any case, something like a quarter-turn of the CO screw is probaby not worth being concerned about; it's not a substantial change in mixture.
_________________
Shop for unique Vanagon accessories at the Vanistan shop:
https://intrepidoverland.com/vanistan/

Please don't PM here, I will not reply.

Experience is kryptonite to doctrine.

[kayakwesty]

Dogpilot and 10centlife...once again you rock

My O2 sensor is brand new and the gauge shows it's healthy

and

[tencentlife]

That looks like it's just a narrowband LED voltage readout. These are of limited usefulness in analysing actual mixture, but they can tell you when it is at lambda, and whether it has fallen off lambda toward rich or lean, but not by how much. The V output curve of the narrowband sensors (the type most vehicles including our vans use; in the last decade more and more vehicles incorporate wideband, a completely different animal), is extremely non-linear. When mixture varies more than about 0.2 full part air (meaning 14.5 on the rich side, 14.9 lean) the signal V runs right to the end of the range. So a narrowband sensor acts like an off-off switch, where the ECU can only tell that mixture is wrong, and in which direction, but the range of V where mixture is correct is so narrow that it's like walking a knife-edge, it can't actually hold mixture at the correct level. So if you represented mixture graphically, it would be drawn as a sine-wave of varying amplitude but always crossing the ideal line. How often it crosses the line is actually the best expression of the sensor's responsiveness.

That type of display merely reads the sensor's output V and represents it as an LED scale, with a bit of signal damping so it doesn't look too wild. If the LEDS in the center are lit, it's hovering close to lambda, doing what it's supposed to do, but when they go full rich or full lean, you do know that it's richer or leaner than lambda, but it can't tell the difference between 1/3th of a part and three full parts. So it's useful for knowing if the lambda system is working, but you can't really tune by it. If your display went to rich, it could be that your mixture is anything from 14.5:1 to the outer limits of richness, there would be no difference in the display because outside of that narrow range the sensor's signal V goes full-low (0 volts) if it's too lean, to full-high (1 volt) if it's too rich.

You can use it to dial it in to very near lambda (14.7:1), but depending on your altitude, the base (open-loop) mixture is probably supposed to be a little rich to very rich. So using your LED display to adjust that isn't going to be very accurate since it would display 12:1 the same way as it would 14.4:1. You need an exhaust gas analyser, which measures percent CO, or a wideband o2 controller, which returns actual AFR, to accurately tune the base mixture.
_________________
Shop for unique Vanagon accessories at the Vanistan shop:
https://intrepidoverland.com/vanistan/

Please don't PM here, I will not reply.

Experience is kryptonite to doctrine.

[kayakwesty]

[tschroeder0]

Yeah Tencent, what Kayak said and...

I know I'm asking a loaded(no pun intended) question but is there any reliable way to self adjust without a meter and get things "right" just curious what you would say. Thanks, Todd.

[tencentlife]

The instrument that gives mixture as AFR is called a wideband controller. There are many that you can buy, but they're a bit steep in price for just casual use. Some are a handheld shop tuning aid, like the Innovate LM1 I have. There are ones designed more for hard installation in a vehicle, like Innovate's LC1, an NGK unit, and lots of others.

The LM1 is consideredone of the best overall by many tuners:
http://www.innovatemotorsports.com/products/lm1.php

Any of these control a Bosch or NGK wideband o2 sensor which returns a signal of 0-5V that is accurately proportional to mixture. These sensors are OEM on many later cars, but don't think you can just install the sensor and read the voltage like you can with a narrowband; the wideband needs a special circuit to heat and control the sensor. Without the controller circuit, the sensor by itself is useless.

Here's one you can build yourself from a kit, very cheap, or buy prebuilt for a little more:

http://www.14point7.com/Widebands/JAW_1_03/JAW_1_03.htm

I'm going to build one of the JAW units to hardwire in. It can feed a voltmeter, LED bar, or numbers display to show actual AFR in realtime. It also has some datalogging ability, and a couple other sensor inputs.

Most all the controllers integrate some datalogging capability, and the ability to extend the controller to be the heart of a more extensive data collection and logging system. Usually they can be connected to a notebook PC so you can see more engine data displayed in realtime, and recorded for later analysis. If the datalogger has enough channels, you can read any other type of engine data there is a sensor for: temperatures, pressures, timing events, switching control events, acceleration force, just about anything.

To identify the two types of sensors, a wideband will have 5 or more wires; a narrowband has 1 to 4 wires. That's the easy way to tell which type you're looking at, because otherwise they look the same. 5 or more: wideband. 4 or less: narrowband.

Auto repair workshops typically had an exhaust gas analyser, a big roll-cab machine with a tailpipe probe. It showed exhaust CO as a percentage. Not practical for home use, and I think a lot of shops now are going to using wideband for tuning because of its cost, accuracy and portability.

Those are the only ways I'm aware of to accurately tune mixture. Of course you have to accurately set ignition timing as well, because it will also affect mixture. That's easy with a strobe, though.

Other than the old tuning by ear, and looking at the burn on spark plugs, and a real dynamometer or the seat-of-the-pants dyno, there's no other ways to tune. Those old methods work if you have a feel for them, but they will never be "accurate."

Luckily, as far as mixture, lambda-based systems allow more leeway, because you only have to be close and lambda will take over under idle and cruise once it warms up and goes closed-loop. But where CO is set has a big effect on open-loop warmup and WOT.
_________________
Shop for unique Vanagon accessories at the Vanistan shop:
https://intrepidoverland.com/vanistan/

Please don't PM here, I will not reply.

Experience is kryptonite to doctrine.

[kayakwesty]

[Dogpilot]

I got one of these loggers. It runs on its own or can stream through the USB. It already has a bunch of A to D converters and several digital I/O lines. It programs in C ++ and the compiler is freeware. Additionally, it can run under MAC OSX or Windoze. It only runs about $40 for the unit, and you can get plug in daughterboards to prototype on. I picked up a bunch of opto-isolators and some power conditioners for a few bucks on eBay. I was hoping to wire it to my version of Digitool through the pass through connector. You can also drive an LCD display for real time viewing or use your laptop. My summer was gutted by work, so I haven't got to it yet.
Resources:

http://www.arduino.cc/en/Main/ArduinoBoardDiecimila

http://www.adafruit.com/index.php?main_page=product_info&cPath=17_21&products_id=51

http://www.sparkfun.com/commerce/categories.php

http://www.arduino.cc/playground/Main/Resources

_________________
Geology with a Syncro rocks!
86 Syncro Westy AKA "The Bughunter"
98 Disco I
08 Range Rover SC
08 VW Rabbit S
1951 O-1G

[tencentlife]

[kayakwesty]

Okay I hate combining thoughts and posts here...esp from Dogpilot and 10cent...btw you guys rock ...and I'm loving what you have going here and making me smarter...and my mind is racing a million miles an hour here and it's past my bedtime here on the east coast. (run on sentences coming right up) ..but here is what I'm thinking...direct me if I'm heading in the wrong direction. I have a Darrell Bohler Digitool with spare settings not being used and I like the idea of a all in one box.Also the digitool needs upgrading to keep me from having Bostig envy.

For our newer Sambanites here, the Digitool Darrell once provided the community is similar to the kit that van-cafe now sells

http://www.van-cafe.com/vanagonparts.jsp?pa=ip&ip=345100050

here is Darrell's link

http://www.loam.org/vw/

And in my 2 cents is a must have for the DIY Vanagon owner

so back at the ranch

[Dogpilot]

When I made the pass through connector, I drew out all the harness wires in a 25 pin cable. I did this so I could look at any line and so I could pass the signal through something in the future and fool the signal to the ECU. THe original Digitool is cool, but I didn't want to permanently alter the ECU. The other item was the Voltmeter idea was good, but of limited utility. Putting an oscilloscope was more useful than the random voltages. That way you can measure pulses/second and such to get real data, like fuel flow for milage, timing advance numbers, the list goes on. If you want to build the pass through, get the PDF file write up I did on my public folder:

Digifant Test Tool.pdf
http://homepage.mac.com/WebObjects/FileSharing.woa...mp;lang=en
_________________
Geology with a Syncro rocks!
86 Syncro Westy AKA "The Bughunter"
98 Disco I
08 Range Rover SC
08 VW Rabbit S
1951 O-1G

[tencentlife]

Which is brilliant, by the way.

Integrating something like the Jaw into the Digitool would be pretty easy, I would think. It'd just be another module, and would have to have its dedicated cable to the sensor. But it could patch it's two output signals to the Digitool display, at a minimum, since they're configurable V outputs and the display is V-driven, no? It could feed an output for controlling the lambda point back to the ECU from there. There wouldn't be the need for the stock NB sensor.
_________________
Shop for unique Vanagon accessories at the Vanistan shop:
https://intrepidoverland.com/vanistan/

Please don't PM here, I will not reply.

Experience is kryptonite to doctrine.

[Dogpilot]

Yes the possibility of modifying the signals does open up a lot of possibilities. The lambda signal is not actually a straight voltage, but more of a sawtooth signal. It can be synthesized, just a bit more programming. I was going to go for simple and work out to complex. You can measure the pulse width and frequency of injector pulses. This gives fuel flow. Couple that with speed from a GPS, or a calibrated VSS, you get milage. You also can delay the pulses from the ECU to the coil to change the timing. Biasing the inlet TEMP I and you can fool the ECU to charge density and so on. Lots of fun, but a dyno would be nice to check it empirically.

The best world would be to put a knock sensor on the engine. Let the little processor monitor that and adjust the timing based upon Lambda and knock to run closer to lean than possible with the stock Digifant. The bulk of engine control and timing map would remain stock, it just gets trimmed by the aux processor.
_________________
Geology with a Syncro rocks!
86 Syncro Westy AKA "The Bughunter"
98 Disco I
08 Range Rover SC
08 VW Rabbit S
1951 O-1G

[RCB]

So whats the answer to the original question?? Is it better to run a little rich or run a little lean.

[kayakwesty]

as mr 10centlife stated it is better to run a little rich
_________________
1984 Westy 2.0 Jetta ABA Motor/1.8 head converted to auto tranny, with 180,000 on body, 55K on engine and transmission

B.5.5 Passat Wagen with 30v V6-67000 Miles

DAS AUTO

CHANGE YOUR FUEL LINES!

A post without pictures is useless

http://www.kayakwesty.com

http://www.waldensridgewhitewater.com/

[GWTWTLW]

If you want a dash mounted Lambda meter, check out this one. http://splitsec.com/products/airfuelmeter.htm . It looks pretty sweet.
_________________
89 Syncro Westy - GW 2.5, now with a double knob job
@gwtwtlw

[tencentlife]

Yeah their model 1 is just a narrowband voltage reader like we talked about above. It'll tell you when you're at lambda, or above or below it but not accurately by how much. Of limited utility for actual tuning but it does serve to confirm that your lambda system is working and how fast the response rate is (which tells you when your sensor is aged). Plus it's a nifty gee-whiz gimcrack thang to have on your dash and I'm as weak for that kinda stuff as the next guy.

Their model 5 is supposed to work with a 0-5V sensor. I think I've heard such a thing exists but never seen one, it doesn't seem to be the same as a WBO2 because WBO2 requires a dedicated controller to drive the sensor and resolve the output. maybe theirs is intended to work in cars that have the WBO2 stock, as many late cars do.

Since this thread has been reawakened I might as well add that I did do the mod I was describing earlier. I got an Innovate LC1, which is the same as the LM1 but without the built-in display and datalogger; it's just the controller circuit housed in a cable. I experimented with the LM1 for awhile first to see that the idea works satisfactorily, before buying an LC1 to hardwire into the van. The LC1 operates a wideband sensor, and its two analog output channels I fed to a tiny SPDT relay. The relay is operated by a vacuum switch to feed one channel into the ECU at higher manifold vac, and to switch to the other channel when vac falls as it does under load. The vac switch is tunable to adjust the switchover point.

It all works great! The LC1 has a serial cable that a laptop PC can be connected to, with which I can program the two analog channels to force the ECU to run at my two chosen mixtures, which I have at 16:1 for cruise and 12.6:1 for load. I can also watch the exact mixture on the PC in Logworks while I drive if I want to. The load response isn't instant, it requires the ECU to go thru enough correction cycles to reach the target mixture, so it's pretty slow at idle but only takes a couple seconds to reach 12.6 under load at highway rpms. At any rate, it's a far faster and more accurate load response than the stock system offers when the WOT switch closes; that takes over 30 seconds before it lets go of lambda and responds at all.

There is a minor problem of the LC1 rebooting as you go down the road, and I haven't been able to figure out why. There is an output wire that can be optionally used to power a single LED which indicates system status and shows error codes by flashing, just like an OBD1 CEL does (there's my gee-whiz dashboard thingy, it would be no fun without that!). I can see the LED flash like it's in warmup mode again, and it happens every so often while driving. Depending on the load I might feel a slight loss of power for a second or two while it boots back up. It could be a grounding reference problem, these controllers are very sensitive to ground offsets, or maybe I can alleviate it by trying some different program settings, but it's a minor annoyance. Meanwhile the system works very well, I have better cruising mileage and the load response is awesome compared to what the unmodified system offers, when I put my foot into it I actually get more power!.
_________________
Shop for unique Vanagon accessories at the Vanistan shop:
https://intrepidoverland.com/vanistan/

Please don't PM here, I will not reply.

Experience is kryptonite to doctrine.