Motorsports Village

How a Wideband gets tricked to read wrong AFR
Page 1 of 1

Author:  shrinker [ Sun Jan 27, 2013 12:42 am ]
Post subject:  How a Wideband gets tricked to read wrong AFR

I had an experience the other day I thought I'd pass on. I and my engine builder were on the engine dyno with a SBC 400 tunnel rammed race car engine. It has a pair of quickfuel 750's on it I had setup for it. We started the first day by running the engine in on avgas, well it was supposed to be run in by the time the dyno guy said it was, but i never believe that. So anyway we started some pulls and used Sonoco 112 race gas. The dyno is equipped with a meter box type Lambda gauge, I dont know the exact brand etc, it doesnt matter, it reads the same as my innovate stuff when Ive hooked that up there before so i just used his being on leaded fuel etc. I took my picoscope (a laptop data logging oscilloscope) to the job just to watch the combustion from that point of view. It was interesting only analyzing the tuning from the point of ignition traces alone. I tell you everything you need to know is shown on a scope. If your not getting the first part of the process right, like the firing KV and the duration and flame kernel development your not going to make for it on the rest of the stroke.
Anyway the first few pulls were showing up really rich at lambda .83. Dyno guy was grizzling at me its too rich, so I pulled the plugs out and they were white, well I'm not going to lean this sucker off am I. I told the dyno guy its not that lambda; that's an incorrect reading, the engine is making too much CO (I could see it on the plug) from inefficient burn and its inferring your reading, there's noting wrong with the carby setup. Ive told you guys on forum many times about plugs reads that are white with grey tinge when the AFR is supposedly rich and how that possible to happen. So I say "nah not leaning it off yet, go change the timing". So we give it 2 degrees more and its still reads .82 and the plugs are still white and the power went up a tiny bit. So I look at the scope and think for a bit and say this fuel is not right for this engine as it stands now, we send the owner off to go get more Av-gas. The scope was showing the crane Hi-6 was only achieving 2/3rd the duration time of spec, the systems working great but there is way too much ringing left over in the coil so there is plenty of energy but its wasted. The firing KV's were 12kv and that ridiculous for a supposed .83 lambda or even the lambda I knew it was.
Anyway we get more av-gas and give it a pull and it makes 10hp more than the sonoco 112, the dyno guys still rabbiting on about his lambda gauge and saying its too rich. Still reading 0.83. Pull some plugs out and its getting some tan now so we keep fiddling. I blocked the bottom e-bleed because that was part of my plan, I deliberately left the bottom bleed in in-case it went above a power level I wanted to make sure it went rich. So no problems put a blank in the hole and swapped the e-bleed into the IFR so it leaned out the bottom end. That all worked good made it run better down low but it still read 0.83 lambda on a pull. The firing KV's were down to around 8kv now with the av-gas, so that was getting better but it was erratic not consistent firing KV. I was a little bothered by the scope readings, its telling me the engine is insufficiently compressing the mixture, but we charged on. Did some timing runs and plug gap tests etc and got the power up another 10hp, but then it started to ever so slightly pepper the plugs. Then we did a compression check on the motor and low and behold the compression had come up 25psi, So much for saying the engine was run in before huh.

I could tell the engine was now making too much HP per cube for av-gas, so we swapped back to the sonoco 112 and low and behold it now loved it, the plugs reads were good, and the scope was full duration arc with consistent 7kv firing on the fuel. Everything was now falling into line.
And guess what the lambda was reading 0.87
The dyno guy shut up because It was fixed by NEVER changing a main jet, just like I told him at the start.

Author:  shrinker [ Sun Jan 27, 2013 5:02 am ]
Post subject:  Re: How a Wideband gets tricked to read wrong AFR

So what really happened with this engine? We never changed a main jet and yet the AFR went from lambda 0.83 to 0.87 (12.2 to 12.8 on a standard gas scale) Thats a 4% change in fueling with actually changing the fueling. When I changed the IFR size it didnt change the lambda reading so it wasn't that change that did it. So what did change? People are taught by every Wideband manufacturer in the world that they are accurate tools for reading the AFR. And yet here is an example of not actually doing anything that actually has bearing on the AFR and yet it changed. What happened was the engine became RUN IN. And the timing was adjusted for best power.
So why does that result in a different reading?
Its all to do with the species of gasses output from the chamber. You have to look at the percentage of each species to understand why. As a burn progresses the chemicals produced go through stages until the final outcome is for a perfect burn just water and CO2. Our engines don't get to that stage but they get to the stage of producing water and CO and CO2 and a little bit of NOx and having some O2 left over. The point to consider is that it takes time to go through the stages to get to the final result.
The most important stage is the preparation prior to ignition. When that is not right, as in this engine, the flame kernel doesn't grow fast enough and the result is a lower cylinder pressure at an equivalent crank position to a good kernel. The flame kernel is a growth of laminar conditions to around 20mm diameter, then the conditions change to turbulent flame and the speed of burn is much faster. What you have to know is its droplet size and the vaporization that influences the speed of both laminar and turbulent flames. So when an engine has large droplets its needs a lot of heat to vaporize them in the time available.
When I say to you guys that a particular engine hasn't got enough vaporization that is what I'm commenting on. How it looks inside the cylinder is like this, the AFR at ignition time is leaner than the average of the cylinder, lets say the average is 12.8 but at ignition time the AFR around the plug is AFR 14, the flame is going to grow slow and some of the flames energy is going to used in vaporizing the fuel thats not gas yet. So the flame kernel doesn't generate enough heat and you dont get a TAN plug, you also run the risk of extinguishing the flame if turbulence is poorly directed. The kernel uses up about 10% of the mixture then the conditions transition to turbulent. Once the flame is turbulent it cant be extinguished. However the conditions toward the remainder of the chamber ahead of the flame kernel front are still under vaporized, therefor requiring energy from the advancing flame front to vporize mixture ahead of the front, not an efficient way to go. It's the burn angle duration between 10 and 90% that matters. Good burns are in the range of 15 to 25 degrees. Burns of engine with poor vaporization can be as long as 45 degrees. THAT'S WHY some engines make best power at timings like 38 to 45 degrees compared to another that best at 28 degrees.

The burn time determines the peak cylinder pressure. Its the peak pressure that is one of the principle determining factors for the production of CO into CO2. If its not high enough you wont get the conversion underway so the exhaust contains high values of CO. If the CO is produced with a high oxygen consumption then the exhaust stream will be low in O2. Now a WB theorist will tell you that the WB will use the unused O2 to convert the CO into CO2 etc and complete the burn to perfection and report a result. Well that's correct but the reality of this type of burn is that the O2 is used to make excess CO and forgo the conversion to CO2 because it just didn't get hot enough for long enough. This type of burn is far from correct as the theory books only want to consider. There isn't the normal correct balance of O2 to CO of a rich burn. The CO2 levels are low and the WB catalyst needs the CO2 to break it apart and get some O2 to reform the CO to CO2, the reaction of which is O2 neutral. So the WB has to get the O2 from the outside air and that causes it to report a high O2 requirement thus richer than truth.
I fail to understand why the industry doesn't educate people to correct this anomaly. The only manufacturer who is game to come close but they dont realize the reason why is Innovate, their system goes into error 8. But they fail to list this cause in their error reasoning help chart.

Author:  shrinker [ Sun Jan 27, 2013 8:54 am ]
Post subject:  Re: How a Wideband gets tricked to read wrong AFR

I've found that Widebands report erroneous readings when the gas in the exhaust is high in CO and also high in O2. That is an unnatural situation. HC levels are not a problem to a WB, its the CO that is the trouble maker. The problem is without a gas bench you have no way of knowing if the reading is correct or not. A gas stream of high CO and high O2 is a dangerous one, it means the engine is not making its power potential and it means there are going to be excessively lean areas in the chamber. So when you look at the plugs and they are not a dark enough tan for the AFR your running you can pretty well assume the engine isnt preparing the fuel enough prior to ignition. So whatever you do don't tune off the WB, tune off the plug. The engine of this thread is not a bad one, its pretty good actually because I've seen WB's up to 1 or even slightly over ratios out of truth.
You cant put nitrous on an engine with that sort of issues, well people do but they rely on the fuel to get them through. I've run engines in perfect combustion scenarios with correct gas balances and even with perfect combustion as in zero CO and O2 and HC and only CO2 and Ive seen how plugs look. I have experience at ranging the AFR around the plug from the plug color. I can tell you this issue is the most common thing I see.

Author:  Slowpoke [ Sun Jan 27, 2013 9:31 am ]
Post subject:  Re: How a Wideband gets tricked to read wrong AFR

Very good read, I've got to get a gas bench!


Author:  treyrags [ Tue Jan 29, 2013 7:28 pm ]
Post subject:  Re: How a Wideband gets tricked to read wrong AFR

Slowpoke wrote:
Very good read, I've got to get a gas bench!


I agree, thanks for taking the time Bruce.

Page 1 of 1 All times are UTC - 5 hours
Powered by phpBB® Forum Software © phpBB Group