Good WideBand Info

[Jack B]

A couple of us have been logging a/f for quite a while, at the same time we have had many comparative dyno pulls. One thing that was obvious was the fact that the Innovate wideband (our cars) and the dynojet wideband were almost 1 to 1.5 points apart. I have posted that same comment several times, but, the lack of response seemed to indicate that as a group the viper community wanted to believe in the dynojet values.

The proof is out, a real nice article was just published in the "Ford Muscle" magazine. There is a link on the Innovate web site. They took eight of the most popular widebands and used laboratory calibrated gas samples. They also welded all eight (in a circle) around an exhaust pipe along with a $5,000 laboratory wideband. They used the lab samples to verify accuracy and the 8 welded units to display response rate (latency), this was on an engine dyno.

It wasn't pretty for the dynojet, it was inaccurate and it was the slowest to respond, in affect, it was so slow you would be displaying the a/f in the wrong rpm column or cell.
By far the Innovate showed the best results.http://www.fordmuscle.com/

Here is the link:
http://www.fordmuscle.com/

 

[FE 065]

Thanks for reporting the comparo Jack.

Most of the Viper community probably isn't wired for data logging, so we pretty much have to go with what the dyno tells us.

So are you saying the dynojet is consistently leaner or richer in its' error?

I'll save this thread in my Favorites 'Viper' folder for future reference...!

 

[Jack B]

Thanks for reporting the comparo Jack.

Most of the Viper community probably isn't wired for data logging, so we pretty much have to go with what the dyno tells us.

So are you saying the dynojet is consistently leaner or richer in its' error?

I'll save this thread in my Favorites 'Viper' folder for future reference...!

It runs leaner (dynojet), it is off the scale at the beginning and it gets closer, but, still leaner at high rpms. The problem is the latency, it is reading history not real time.

 

[Joseph Dell]

direct link: http://www.fordmuscle.com/archives/2007/06/WidebandShootout/index.php

 

[Purdue_Boiler_Viper]

Interesting. Thanks for the post.

 

[ViperBing]

I currently run two dynojet units. I bought the second one slighlty used and noticed that the second one had a much quicker response time than the first. It turns out that the smoothing on the second one was set to a different setting. The default smooothing on first one seemed to slow down response times. Once I adjusted the first one to the same smoothing as the second, they both respnded identically and seemed to offer a much quicker response time. So, I wonder if this was the case in their tests?

 

[GR8_ASP]

My Dyno jet unit responds quickly as well. Ican only state the accuracy at stoich and it is dead on there. I know of a study done a few years back and the essence was the wideband O2 sensors did not provide the accuracy needed and that gas evaluation was the only accurate guide.

I have not read the link info yet but wonder if the comparisons were made to a known O2 sensor or to a gas analyser.

 

[Jack B]

My Dyno jet unit responds quickly as well. Ican only state the accuracy at stoich and it is dead on there. I know of a study done a few years back and the essence was the wideband O2 sensors did not provide the accuracy needed and that gas evaluation was the only accurate guide.

I have not read the link info yet but wonder if the comparisons were made to a known O2 sensor or to a gas analyser.

They had a $5,000 sensor that was the base sensor and they used laboratory gas for the actual a/f accuracy. The engine dyno was only for secondary tests to measure response.

 

[GR8_ASP]

Just read the article. Couple things that struck me odd. One they tested with gases and did not indicate that they were preheated. That means the sensors heater circuit was of greater significance in an accuracy measurement than normal. Second, the engine test was done with leaded fuel. Not sure how that impacts the results, but given that leaded fuel is rare today you have to wonder about that selection.

Their statement about accuracy did not indicate accuracy at what point. Obviously I care little about the accuracy at 17:1, but am very critical of it near 12:1.

All said I am not quite certain of the overall analysis accuracy. Kind of like reading consumers reports about sports cars where they are most concerned with the a/c performance and the like.

 

[99 R/T 10]

It was unclear to me how it was inaccurate( I haven't read the article), but on Viperalley, Jack said when you get a reading of say 12.0, it's probably 11.0 or so. I feel better that if it's going to be off, at least the correct number is richer(safer) than what it shows.

 

[Steve 00RT/10]

As I have 2 NGK AFX units.....and they finished last in the test, I wanted to get their take on the matter. They find the test to be lacking in several areas. NGK is the largest maker of O2 sensors in the world. They package the Bosch WB sensor with their unit for cost containment in bringing an affordable unit to market. Replacement O2s are NGK- NTKs---considerably more money. The same engineer that designed the test gear used to evaluate these sensors (ECM LambdaPro) also had a big role in designing the NGK AFX meter. NGK uses the Horiba system for their in house testing of their products and work contracted from the Big 3. They do use the Innovate system for comparison testing as it is a good one. NGK has found that the results in their testing between the two systems are not really discernable.

As for the analog wire not being in the kit, they said that's not really possible. The first 100 units produced last year did not have an analog out wire, but every one since has. When they realized what demand there was for data logging, they sent them out free for those who asked on those first 100 units. I bought my first AFX unit in 3/06 and mine had the wires (2 needed -- separate ground). These pigtails for analog out are only about 8" long and come stuffed inside the loom harness. Evidently the Ford Muscle guys didn't read the manual about where these wires were. How did they get the output correct from the NGK unit?

The article also states the Bosch sensors are factory calibrated. One of mine was not. .... Another feature only the NGK has: You can calibrate the sensor in open air. Checking my sensor this spring after 11,000 miles last summer showed it still in calibration. BTW, this is the sensor that I had to tune in last year.

Another point of contention with me is that the AFX unit is panned for ease of use, software, and display. It couldn't be any easier to use and I'm not a computer guru by any stretch of the imagination. As for the display, I have never had trouble seeing it, even in bright sunlight with the top off the car. The numbers are very large -- no glasses needed. I'm not sure what prompted this negative critique of the unit.

Here's a little more of NGKs comments in their own words:

If you navigate through this Ford Muscle website, you will see advertising from Innovate and F.A.S.T. I know it is difficult to have any kind of truly unbiased evaluation of anything, but it doesn’t look good when your top finisher is one of your advertisers. Also, if you look to the sidebar of the main page, they list related articles on this topic of AF tuning. Two of the three articles are about Innovate, and the remaining one is about F.A.S.T.

After reviewing the article, it seems there are many variables not taken into consideration here, nor are they explained. The so-called “accuracy test” is a test with bottles of simulated exhaust gases. Innovate designs their meter to match these simulated gases and for this article, rigged the accuracy test to be based on these gases.

The gases are just that, “simulated exhaust gases” and are used to calibrate gas emissions benches, not wide-band ceramic sensors. Using these gases to calibrate wide-band ceramic sensors is wrong. Wide-band ceramic sensors operate on a different principle than gas benches and use not just the CO and O2 in the exhaust but also the H2, water, and hydrocarbons that are not present in these synthetic gases to determine the AFR. Using these gases to determine the accuracy of a ceramic sensor is like using a ruler to size pistons for an engine.

I guess the bottom line for me is that I am, to date, completely satisfied with my NGK units. If my logged AFR readings were potentially off as much as the article claims, I likely would have broke my non-cooled, cast piston car long ago. Instead, I have 15,000 miles on the blower using the NGK meter for tuning. The Innovate system is definitely a good one, but it appears this test may lack some scientific credibility.

Steve

 

[Parisianviper]

The article also states the Bosch sensors are factory calibrated. One of mine was not. .... Another feature only the NGK has: You can calibrate the sensor in open air. Checking my sensor this spring after 11,000 miles last summer showed it still in calibration. BTW, this is the sensor that I had to tune in last year.

Going through Innovate documentation, Bosch sensors are to be calibrated in open air also....

 

[Steve 00RT/10]

Yup.....missed that....self-calibrated. I find it interesting that the Ford Muscle guys consider this calibration feature quite important.....Only 2 devices in the test can do it, yet one of them took last place.

Steve

 

[Jack B]

Just read the article. Couple things that struck me odd. One they tested with gases and did not indicate that they were preheated. That means the sensors heater circuit was of greater significance in an accuracy measurement than normal. Second, the engine test was done with leaded fuel. Not sure how that impacts the results, but given that leaded fuel is rare today you have to wonder about that selection.

Their statement about accuracy did not indicate accuracy at what point. Obviously I care little about the accuracy at 17:1, but am very critical of it near 12:1.

All said I am not quite certain of the overall analysis accuracy. Kind of like reading consumers reports about sports cars where they are most concerned with the a/c performance and the like.

Read the article again:

The leaded fuel was a used to age the sensor so that they were not working with a new sensor. One of the issues was that only two of the tested units were self calibrating. The others would have the base curve shifted over time.

The a/f accuracy was taken with the two mixtures of laboratory gas, therefore, at two different values. They were described in the article. Your value was bracketed

 

[Simms]

Great info. Thanks Jack.

 

[GR8_ASP]

I do not care why they used leaded fuel. It would place the sensor in an environment that it will never see in my car. Does the lead impact the sensor in other ways? Do they know or just assume not. If they wanted to age them then they should have created a method that did not introduce new unknown variables.

As to the use of the gas I am also very unclear as to the impact. Whether it be the temperature or the lack of full gas spectrum the method just seems wrong. I want the test with real gases and real temperatures. It does indeed seem if there was some bias in setting up the test.

As to the results where the innacuracy occurs is very important. Having a graph showing actual and measured (steady state values) would help a lot. If the innaccuracy occured below 11.5 or above 14.7 I would not care. But a .75 error at 12.0 would be alarming. These companies know that is the important range. However we have no clue where the error is. They might as well of had Consumers Reports do the testing.

Btw how do they know if a system has automatic recalibration. By the box or do they really delve into the code to verify it? At stoich mine has not varied an iota over 2+ years and 15k miles. Very good if it has no automatic cal like they suggested.

 

[Mr Hemi Head]

The Dynojet manual packaged with my system does not list a spec for accuracy.

Does this mean that gauge accuracy is primarily a function of the O2 sensor?

 

[Jack B]

I do not care why they used leaded fuel. It would place the sensor in an environment that it will never see in my car. Does the lead impact the sensor in other ways? Do they know or just assume not. If they wanted to age them then they should have created a method that did not introduce new unknown variables.

As to the use of the gas I am also very unclear as to the impact. Whether it be the temperature or the lack of full gas spectrum the method just seems wrong. I want the test with real gases and real temperatures. It does indeed seem if there was some bias in setting up the test.

As to the results where the innacuracy occurs is very important. Having a graph showing actual and measured (steady state values) would help a lot. If the innaccuracy occured below 11.5 or above 14.7 I would not care. But a .75 error at 12.0 would be alarming. These companies know that is the important range. However we have no clue where the error is. They might as well of had Consumers Reports do the testing.

Btw how do they know if a system has automatic recalibration. By the box or do they really delve into the code to verify it? At stoich mine has not varied an iota over 2+ years and 15k miles. Very good if it has no automatic cal like they suggested.

No flame intended, but, you have your mind made up and don't seem to want to read or try to understand what was done. Most of your objections are explained away in the article - no matter how you cut it the only way to accurately test any O2 sensor is laboratory gas - in ending the expensive laboratory sensor was there to backup the test's accuracy. I was not trying to make any point, merely pointing out a nice article. By the way, it is not usually the sensor that induces the error it is the way the data is interpreted - the NTK by itself is a good sensor.

 

[Jack B]

The Dynojet manual packaged with my system does not list a spec for accuracy.

Does this mean that gauge accuracy is primarily a function of the O2 sensor?

It is the sum of the parts, in affect, the system components.

 

[GR8_ASP]

Sorry if came on too hard Jack. It is that as an engineer I have seen many studies that provide incorrect results or conclusions. Setting up the test parameters is key, as is the interpretation of the results.

In my experience 9 times out of 10 a study like this is misleading. If it were conducted by an instrumentation evauluation group you would find many changes in the test protocol and results reporting. As it is I take it as garbage in and garbage out.

I understand you are trying to help. But it is like that oil filter internet post. The fact that it is garbage does not change the fact that people believe it to be the truth.