Poll about "surging" on K-Jet with Lambda models

You are probably right.

 

Yes, in most cases.

In areas that require emissions testing, 'sometimes' you can squeak by without them, depending on the levels required to pass and the condition of the Cat. Plugging in the sensors on a car where everything is working correctly and the base settings are Properly set... it near always results in higher HC readings with the sensor on line.

By the very nature of its design, the system cycles from Too rich to Too lean. Given these engines are aggressively cammed to be able to run a power range approaching 8K RPM.... it gets to splitting hairs at idle when it comes to emissions. Throw in just one cylinder running out of spec, be it lean or rich, and it will throw the entire system out of whack by changing the sampled 'average'. On Futureman's car..... (1) one year old injector spraying a less than ideal pattern, after a host of other small issues were resolved, made the difference between a sensor reading .03 volts and it reading .92 volts (while unplugged), the full range of a normal early sensor, while averaging all 8 cylinders.... that on a car that ran quite well before any work.

We are faced with a rather unique challenge here being as we have no oxygen to ignite and burn this non flammable swill fuel..... that while we are faced with ever decreasing limits on what is considered passing an emissions test. Have the O2 sensor take the system to the lean side of the bell curve with just one cylinder already running on the lean side of normal and it ends up a dead lean miss, with the HC skyrocketing..... often with Cats glowing.....the result is the surge felt when the system heads the other direction towards too rich and the lean miss goes away. The base cause can be found without the individual sample tubes but it damn sure takes a lot of time to do so! Make everything run as designed or live with a heavy surge, it becomes a monetary decision IF you can pass the emissions test. IF you can feel it seat of the pants....... the car has more to offer by fine tuning it.

 

How long would that take? (assuming you have no cat) it only takes a couple of minutes to sniff it at the tail pipe and make the one "average" adjustment.

But what about doing it at each tube? One of my banks still has the tubes attached. How much of a waste of time would it be to try those and make any necessary adjustments without the other bank?

 

How long is a piece of string?

Once the tubes are uncapped, it takes no longer to get a reading from a sample pipe than the tail pipe.

Learning about how one bank is functioning is a start, it might even show a trend. You wont be able to identify cam timing out of spec and you will only have 50% of the data required but at least it is a start towards complete knowledge.

 

I am going to do this and post my numbers.

BTW I misunderstood your post. I thought you were saying it took a long time to test and adjust correctly from the pipes which is why I was puzzled at first

 

My mistake Tommy, sorry about that. It takes as much time to remove the sample tube caps as it does to take readings. Finding and identifying the problems (or even an accurate fuel setting) when individual cylinder readings are not available.... THAT is a labor intensive effort that involves a lot of guess work.

Raise the bar and make it run right.... the rewards make it worth the effort in 95% of the cases.

 

Just trying to understand.

So the O2 sensors are there in earlier cars to protect the cats and pass emissions by bring down the CO?

I might add, I had I had some surging at idle on my F40. After sending the injectors out and, through David Helms assistance sourcing the replacing the O2 sensors, the surge is very minimal. My limited understanding of this is that the oxygen sensors are trying to bring the CO reading down to about half a percent at idle vs. where it would like to run at about 1.5% and that is what causes part of the inherent surge issue as the O2 sensor cycles as David pointed out.

Happy to learn more.



Drew

 

I would love to but I have that problem with the front (I think front) pipe removal.

Depending on the point it was closed off, and let's assume it is the best case scenario, is bat**** crazy to attempt to weld a new set back on down there?

 

Drew, I realize I am not answering your question - but just to elaborate a bit more about this surge thing; You mention your's was at idle. What I (and I assume most here) am referring to is not at idle. my issue was at speed. I could be driving steady and it literally felt like I was tapping the brake about every 5 seconds or so. It is the most annoying thing. I don't remember paying any attention to it hunting around at idle. Maybe it did but I just don't recall

 

Yup, Bull Goose Lonney. Once they are broken off and welded/pinched off, the header needs to come out to replace them. They are NLA and need to be fabricated, so that too is another investment.... been there already.

That process waits until you have a need to be working back there on something else. Start with the one functional bank and see if you can learn something that can be equated to the other bank.... "Better" is just fine, rarely do we find anyone that wants to "perfect" the system, getting it acceptable is still a marked improvement for most cars.

The sensors and computers are shooting to get everything at the ideal air/fuel ratio... "Of the Day". The major problem is the ideal AF ratio has changed as the fuel changed.... but the computer software doesn't know that. The days of 14.1/1 are long past gone but the computers are still striving for that number. It is for that reason that the last 8 years of my research has been towards dealing with exactly this problem. Why did I design the GCK? I needed everything operating Exactly as designed so I had a steady and reliable baseline to build up from. Sure it also dealt with a host of other problems but that was never my intended focus, it has always been a means to an end.... that doesn't stop with the O2 sensor issue either. Once I get everything working right with the current fuels...... Then, and only then, can I play with Performance. Attempting anything towards that end before that is done amounts to wasted money, while creating a host of new problems. There is a method to my madness that few have even considered.... the ongoing testing is showing extremely promising.

I have posted it here dozens of times before and I will state it again, the times have changed! How the cars are maintained, how they are tuned/timed, hell.... how the engines are built.... Ignore any of those and its turning a blind eye to a major problem. Build or maintain based on Yesterdays knowledge with no eye to the future and the problems have only begun. It's a whole new game now, close no longer cuts it, and Yes, you can make extremely good power with this fuel if everything is "right".

 

David - appreciate your comments as always.

I had not thought at all about the fact that most fuels today contain 10% ethanol and that alone makes the tuning different.

Live and learn.

Tommy - did not read thread carefully enough - thought we were talking about at idle - did not know this was at speed.



Drew

 

Sorry, no, this is completely incorrect and misleading: there's no software involved, the lambda control loop contains no assumptions as to fuel type, and the control loop has no notion of the proper AF ratio!

The O2 sensor and "ECU" on these cars don't give a rats patootie about 14.1 or the AF-ratio to any extent within a rather wide range. Instead, the O2 sensor senses a rich-lean condition solely on the basis of residual oxygen in the exhaust; the O2 sensor doesn't care how / why the O2 got into the exhaust, whether from combustion air, or oxygenated fuel (ethanol, MTBE, etc), or even NOS, or an air leak: it's all the same to the O2 sensor. If the O2 sensor says there's too much oxygen in the exhaust, the lambda control loop richens the fuel delivery; conversely, if too little oxygen in the exhaust, the lambda control loop leans out the mix. It's that simple, and no notion of "striving for 14.1/1" is involved.

The "computer" (the system on these cars is actually a very simple analog control loop that hardly qualifies as a "computer" -- there's no software involved) adjusts the fuel feed - richness - to try to keep the O2 sensor output in the correct target range; these systems don't know - and don't need to know - anything about fuel composition or 14.1 or any of that nonsense. The control loop makes no assumptions about oxygen content of the fuel, nor does it need to, and thus can work with any fuel.

There is a limit to how far the control loop can adjust the richness, so these systems need to be properly tuned for a baseline; however, once tuned, the O2 lambda control will work equally effectively whether you are burning plain pump gasoline, some esoteric hand blended gasoline, or gasoline with any level of oxygenates added: it's all the same to the lambda control loop.

(Hunting on these systems is a problem of feedback oscillation; understanding control systems and the causes of oscillation will lead to understanding of how to correct these problems. *****ing and moaning about fuel composition and making up nonsense about "striving for 14.1/1" won't help in this regard.)

-frank

 

Which is why I mentioned earlier how this setup is subkindergarten on how to manage it correctly. It was the most affordable way to try and fit a round block into a square hole back beginning in about 1980/81 for many car manufacturers, not just Ferrari. It was an add-on far after the original development of the car. Frankly, it or the entire CIS really doesn't belong there to begin with.

This is why you can unplug it on a 308 and get away with it but you cannot do an equivalent move on a 458 - whatever that O2 sensor equivalency may be in one of those things

 

Tommy, understood and generally agreed.

I do think it's a little unfair to call these systems 'subkindergarten', tho; CIS with and without lambda control is actually a very sophisticated and well-engineered system, particularly when you consider the state of automotive electronics and sensors back in the day.

Carbs can do a very good job at mixture management when properly tuned, but have little ability to maintain proper mixture in the face of ambient air pressure / density changes (altitude and weather conditions), or with significant aging / deterioration of the engine (carbon acumulation and minor air / exhaust leaks). Like a carb, CIS is also a form of analog computer, but by changing the control paradigm (air mass as opposed to air velocity), CIS achieves significantly better mixture control in a broader range of operating conditions. As we all know, this was essential to meeting the strict emissions regulations of the day. The integration of lambda control to CIS is actually well done, and not at all any sort of bastardization of the CIS paradigm; in fact, the integration is very simple and well engineered to fit with the operation of CIS.

CIS is also a vast improvement over earlier forms of mechanical fuel injection (plunger pump systems); those systems are relatively poor at mixture control even within relatively narrow operating condition bounds. Plunger pump systems were greatly favored on racing engines (e.g.: Cosworth DFV) as they were very reliable at delivering very large volumes of fuel in harsh operating conditions and these systems are simpler to tune and maintain than an equivallent carb system. Additionally, the ability to use slide throttles, along with no need for aux venturis, provided improved airflow, especially at WOT.

At the time CIS was developed and commercialized (mid to late '70s), the sensors required to implement a fully electronically controlled fuel injection system were simply too costly and thus unavailable; it took decades to refine automotive electronics to the point where these systems, with their myriad sensors and connectors, became reliable and inexpensive. Interestingly, the problem was not with the control unit: analog electronics were sufficiently advanced in the '70s to implement fuel injection control inexpensively! (And many of us implemented and experimented with analog EFI over the years.) In fact, I have often mused that a near-ideal EFI system for a performance vehicle could be built by replacing the metering plate of CIS with appropriate electronic sensors controlling the distributor plunger via an electronic servo driven by an analog control loop (along with eliminating some other components).

In short, CIS is often unfairly maligned: it's a sophisticated and well-implemented system given the technology of the day, IMHO.

-frank

 

I don't mean to bash CIS. Not at all. K-Jetronic was on the very first car I ever owned (1980 VW Scirocco) and I STILL have the car and it STILL has the same injection system on it. It is such a beautiful and dependable system that it has powered 3 completely different longblocks in that car during the past 27 years that I have owned it. Not one single time has that CIS given me a moment's problem. Never. Same with my other CA version of that car, or with the 16V model sitting next to it this very moment. Or with my 308 - all with it disconnected.

That O2 sensor was added to the K-Jetronic years AFTER it was put into successful production on millions of cars. As an example VW added this CIS to the Scirocco in July 1977. They ran fine on plain K-Jetronic until 1981 (with the exception of CA that had it in 1980). My point was (and still is) is that it simply is just not a necessary component to that particular system if you do not want it to be because it didn't really start out that way. We all here know why is was added and again is was, more or less, an add-on after the fact. Because it was, it is a "dumb" system, if you will. As you correctly pointed out, it cannot process this O2 presence intellegently like we can now. It makes very rudimentary adjustments (as compared to todays systems) that the engine can live without if you want it to just like all the K-Jet systems did before the O2 was stuck in.

The ONLY reason it is off my 308 is because my exhaust tubes are missing on one bank. I can't adjust it correctly even if I want to other than of just ahead of my hyperflow for that average we have already discussed here. It just will not run smooth no matter what we do to it - until I just unplugged it. Then it was all back to normal and ran perfectly smooth again.

Some of us don't have much choice so unplugging it is the only real option. And it's okay if you do.

 

As a European car mechanic I dealt with this problem frequently back in the day. It wasn't unique to Ferrari though it was usually noticeable. It was particularly bad with Audi, not so bad with Porsche. Sometimes the cars came in for other service, and I noticed the surge when the car owners had not. In most cases it was a matter of getting things in balance. The O2 sensor has to be in good condition. when in doubt, replace it/them. If tuning the mixture in the shop doesn't do the trick, a graphing voltmeter on the O2 sensor where it can be seen from the driver's seat can enable you to do it in the "kamakazi dyno" mode. But when it's in balance, the surge shouldn't be noticeable.

 

Replaced O2 sensor during fuel system work...no surges, nice regular buzzing.

The old one was old (*hint*).