Road tuning VE's different route - different results

[Karl H.]

Ip is current. The widebands are actually measuring current. Pressure is in the exhaust.

Steve is only quoting the extremes. The 12.8 number is at 2 bar at 10000 feet. The 13.5 is at the other extreme, .5 bar at 0 feet, which is exhaust backpressure. Every wideband controller I've seen explicitly states it will be inaccurate if you have excessive exhaust back pressure. Here is the relevant section from the TwinScan docs:

Excessive exhaust back pressure. Wide-band sensors are affected by back pressure. Excessive back pressure causes exaggerated AFR indications under rich and lean conditions, but has little effect at 14.7 AFR (stoichiometric). Motorcycle exhaust systems are relatively free flowing and problems with exhaust back pressure are not likely.

Thanks alot for clarification! 

[whittlebeast]

I generally only worry about exhaust pressure  on the o2s when I find out it is a turbo install with the o2s between the engine and the exhaust.  I have the owner move the o2s.

Andy

[Steve Cole]

...Attached is a spreadsheet with the parts of the data you left out but are clearly part of the Bosch documents.

What is Ip? Is "Pressure" the manifold pressure? If so, how do you get 2 bar without turbocharging?

I see Frank continues to leave out the important parts of this issue. First his claim of 0.1 AFR is BS even at 14.83 AFR. That is due to the temperature correction missing from the data still.

As everyone can see I only plotted the data at two points so you could see what is happening. The further you move from Stoichometric the errors get larger and larger and do not stop. Try plotting a fuel mixture of 11.0 AFR like the Supercharged or turbocharged engines run and see what happens. One needs to remember this is all based on pure gasoline too, think of what happens when you move to E10 or E85!

Ip is the current from the sensor, as this sensor works on current flow.

Pressure is the Absolute pressure not gauge pressure. There is a very large difference between the two measurements.

Absolute pressure is the atmospheric pressure plus/minus measured pressure. So in the case of a engine at or near seal level you have to start with atmospheric pressure then add/subtract what is measured in the exhaust and this must be done each and every time the sensor is read, with a sensor capable of taking accurate readings at a minimum of once per engine cycle per cylinder. So if the O2 sensor is read more than once per firing cycle you must also read the pressure at that same rate and time for it to be accurate.

Pressure swings in the exhaust very greatly as the exhaust valves open and close. We have measured swings of +/- 5 psi gauge pressure in HD Drag Pipes and +/- 10 psi in pipes with mufflers and converters. It will vary as engine speed increases and decreases along with engine build to engine build. So while you may think the range on our provided charts is large, a HD engine uses most all of it.

So when you have no idea of how things are working in the real world of the running engine, what does the measured data mean? You have to have all the surrounding information to make it useful at all. Since the aftermarket Broad Band systems do not measure any of this data to know what is going on, there is no way possible for them to be accurate and repeatable.

[whittlebeast]

I wish someone had a MS3 Squirt up and running on one of these bikes.  We could do a tooth log of the pressure in the exhaust.  A MAP sample every 10 degrees of crank rotation.  Just for fun we could do a O2 read at that same sample speed.

Andy

[Tsani]

Motorcycle exhaust systems are relatively free flowing and problems with exhaust back pressure are not likely.

Really?! Could have fooled me. So I take it that torque dips are all in our minds? Pipe manufacturers are going to love hearing this.

[whittlebeast]

For people looking a MegaLogViewer HD, there is a tab called <ignition Log Viewer>  That is used in the Megasquirt world for just this sort of thing. You have really fast data logging over there.  We use that for things like watching a a motor crank that we have never seen the cranking characteristics.  We watch the timing of the crank sensor teeth compared to the timing of the cam sensor (if there is one) and the pulsations at the MAP sensor.  With that info, we can determine the logic for the wheel decoder section of the code.  Way cool stuff.

Andy

[FLTRI]

I'm going to sign off this discussion now. I'll let Steve continue attack mode. Make your own conclusions.

Thank you

[Steve Cole]

I've gotten a few calls asking for me to better explain the pressures being used, so I will try again

Absolute pressure is the Atmospheric pressure plus or minus any addition measured pressure in the exhaust. So at Sea Level that is ~ 1 bar or 100 kPa. So if you were to look at the chart for 13.23 AFR and find the graph line for your altitude you would be at the 1 bar position on the "X" axis. Now once you start the engine the pressure in the exhaust pipe rises above and falls below Atmospheric as the engine runs. At low RPM the pressure rise is less than it is as the engine is at say 2500 RPM. This causes your position on the graph to swing back and forth on the "X" axis. The range that it will swing is going to be different based on your engine combination and exhaust system. So if the exhaust pressure drops below Atmospheric when the O2 sensor is read you will move to the left of the 1 bar reading and if the exhaust pressure rises above Atmospheric when the O2 sensor is read you will move to the right of the 1 bar reading.

Since we have measured HD drag style pipes with pressure swings of +/- 5 psi or +/- 0.344737864 bar this is going to cause an error in the O2 reading from exhaust pressure ONLY  from 13.447 AFR to 13.142 at sea level only. Since these systems do not account for pressure (Exhaust or Atmospheric) you have no idea where in that range you really are at any given moment in time. If your exhaust system has greater pressure swings as most street driven HD's do, that range increases. Now, add to that, if you are riding the bike, the pressure changes due to altitude and you can begin to see how the mixture could be not moving at all (in this one case 13.23) yet the data shows it moving all over the place.

So now let's think about Whittlebeast comments of how the data showed the mixture changing in a Pikes Peak data log, did the mixture really change or not? As I recall the Pikes Peak race starts at near 5500 ft and climbs to 14,000 ft at the end of the race. The Atmospheric and Exhaust pressures are certainly changing as the engine runs through the race! With no means of measuring any of it, the data from the O2 becomes useless for accurate measurements!

Now, if your Dyno tuning, Atmospheric pressure stays constant, so you only have to deal with Exhaust pressure swings but that still leaves an area greater than 0.5 AFR on most HD exhaust systems that you have no way to know what is really happening. If thats good enough for you, that's fine, but let's not mix good enough for you, with what it really is, or the errors that are caused by people not following what Bosch says needs to be done to get accurate readings. I have yet to find a Aftermarket Broad Band system that measures any pressure or temperature data or applies the necessary corrections that Bosch supply's with the sensor data. When you step into lab quality equipment I have found one that allows for it but then your looking at $10K in equipment.

[Admiral Akbar]

So Steve, could you post the data sheet?

[Karl H.]

For data sheet see reply #38

BTW: The discussion starts gaining quality 

[Admiral Akbar]

I all get out of this is that the afr accuracy of the can vary depending on conditions.  Knowing the issues, the device measuring the AFR can't be stupid.

Add.. Who cares if the variance is +- 0.5 points?? Still better than a carb..

[FLTRI]

I all get out of this is that the afr accuracy of the can vary depending on conditions.  Knowing the issues, the device measuring the AFR can't be stupid.

Add.. Who cares if the variance is +- 0.5 points?? Still better than a carb..

👍 the only criticism I have is reaction speed of the LSU4 not accuracy. Close enough on accuracy but there is a bit of difference in quality of rideablility between alpha-n and speed density. No back to back comparison quantified in a lab environment but based on my own personal experience riding bikes tuned with both systems over the years. Should be a slight fuel mileage advantage to speed density as well.
Bob

[Steve Cole]

I all get out of this is that the afr accuracy of the can vary depending on conditions.  Knowing the issues, the device measuring the AFR can't be stupid.

Add.. Who cares if the variance is +- 0.5 points?? Still better than a carb..

So how does one expect the system to work then? If you tell it to control to a moving target then it will always be moving and never get to a final adjustment. Then experts say it must be xxx number or the tune is no good. You must plot the data and it look like a straight line, and on and on and on. The truth is they have no idea what it really is and have never spent the time to learn what really happens and why it is.  Data is a great thing when you understand what it is and how to use it, properly. Used improperly all you get is pretty pictures that do not mean much of anything.

So is the new rule of thumb going to be that fuel mixture only needs to be +/- 0.5 AFR? So a tuneup can be at anywhere between 13.0 and 14.0 at WOT and it's good enough? The truth is you have no way to know any better! If that's good enough then there is a ton of dyno sheets that got post here on this site that people said we no good and now they are just fine!

[Admiral Akbar]

I all get out of this is that the afr accuracy of the can vary depending on conditions.  Knowing the issues, the device measuring the AFR can't be stupid.

Add.. Who cares if the variance is +- 0.5 points?? Still better than a carb..

👍 the only criticism I have is reaction speed of the LSU4 not accuracy. Close enough on accuracy but there is a bit of difference in quality of rideablility between alpha-n and speed density. No back to back comparison quantified in a lab environment but based on my own personal experience riding bikes tuned with both systems over the years. Should be a slight fuel mileage advantage to speed density as well.
Bob

Not sure what the reaction speed is for the LSU4 but there are ways to declare the samples good or bad in ECU firmware..  I've not seem a whole lot of difference in ride-ability except for partial throttle stuff where Tmax needed to bring out injector timing.. Once they did that certain builds got better.. Many builds ran fine as is..

One thing that Tmax lags on is using throttle based ignition timing instead of map.  I would think that this is one of the main reasons for mileage being less and a partial reason for possible low speed manner issues.  Closed loop systems need to to have a good AFR sample to to make a decision on whether to add or remove fuel.. The problem is that the combustion process at low speeds can be irregular cycle to cycle.. The ignition timing adjustments at light loads become way more critical for the closed loop system.. Also the data needs to be filtered / closed loop response adjusted under these conditions.. 

[whittlebeast]

My wiz-bang, $10000 Motec 150 ECU has throttle based timing.  There is a MAP based correction on top of that.

Throttle based is more "predictive".

Andy

[Admiral Akbar]

I all get out of this is that the afr accuracy of the can vary depending on conditions.  Knowing the issues, the device measuring the AFR can't be stupid.

Add.. Who cares if the variance is +- 0.5 points?? Still better than a carb..

So how does one expect the system to work then? If you tell it to control to a moving target then it will always be moving and never get to a final adjustment. Then experts say it must be xxx number or the tune is no good. You must plot the data and it look like a straight line, and on and on and on. The truth is they have no idea what it really is and have never spent the time to learn what really happens and why it is.  Data is a great thing when you understand what it is and how to use it, properly. Used improperly all you get is pretty pictures that do not mean much of anything.

So is the new rule of thumb going to be that fuel mixture only needs to be +/- 0.5 AFR? So a tuneup can be at anywhere between 13.0 and 14.0 at WOT and it's good enough? The truth is you have no way to know any better! If that's good enough then there is a ton of dyno sheets that got post here on this site that people said we no good and now they are just fine!

As usual you sound technical but really don't say anything that I understand.. Sorry I'm slow so you'll have to esplane' it to me..

What system? 

Who are the experts? 

What is  the XXX number? 

What target are you talking about?   AFR? 

Plot what data?  Do data plots have to be a straight line.. If you know what the curve is, you're half way there. 

You talking about the data collection piloting programs?  I've not looked at them much but it seems to me they be some value. At least there can be a collection of data that cam be tied helping with diagnostic issues.. Definitely not a controlled environment, with better equipment but better reading plugs.. 

So at WOT maybe you want the default to be 13.2  since you are WOT need a little more accuracy.. Figure out ways to make it more accurate. Filtering, taking other measurements and using the data in the data sheet to give you a better number.  If you have fixed bias in the system take it out..   What about partial throttle?   Maybe a swing from 15 to 14 is OK?   

if the transducer is well characterized and consitant does that mean because the output is not a ploted straight line that it's inaccurate? 

[glens]

Not sure what the reaction speed is for the LSU4 ...

Even if they're as fast as plain old switching sensors there's going to be a minimum of another engine cycle delay in the report at higher RPM.  Got one closed-loop system (the sensor) feeding information to the host closed-loop system.  This could be mitigated by asking for the report sooner so the answer comes at the right time, but then the sample wasn't taken at that right time.  It would've been take a while ago when it maybe wasn't quite pertinent.

The minimum change detection time of the sensor itself ought to be easily found documented somewhere.  The speed of its controlling system will obviously vary from manufacturer to manufacturer, even if they're all merely (independent?) implementations of the same (forget the term - something akin to "proof of concept") circuit design.

On a dyno this would matter quite a bit less than on a road-only-tuning system for most every user.  Now we got Bob, a dyno user, saying he doesn't like their speed.  But he didn't quality whether that pertained to his use of them on the dyno or to their use in a general-purpose (so to speak) end-user on-the-road auto-tune system.  Or both.  Or maybe it's just the racer in him disliking anything that's slower than something else

[hrdtail78]

A man needs to know his limitations, or understand the limitations of what he is using.  How does it really effect us at the end of the day.  The standard in the industry is either the stock sensors or the LSU4.2's.  This is what we have.  I target 13 for 80% and 100% tps.  I have it dialed in.  The sensor is off by .5.  I am actually at 13.5, but I noticed on this build that target 12 gives me more power.  Does it matter what I am targeting if the end result is that I tuned it for best power?

Knowing and understanding helps explain this.  Also helps weeding through the propaganda from companies that sell to this industry.  The manufacture of the sensor states 4 times a second.  2000-6000 sweep that takes 7 seconds is going to be sampled 28 times.  At 4000 rpm's the bike is seeing about 33 firing events.  Easy math shows we will be missing some.  That is why you make that pull 5-6 times.  Get an average, and go from there.

I recently did a SE tune on a low compression build.  Spark activity when I enabled wanted to pull 8 degrees in several places in higher kpa area.  I adjusted the timing tables to make the ION sensing system happier.  Couldn't get all the spark activity gone even after pulling 15 degrees.  I lost 10 and 10 and the trace looked like it wanted a bunch of timing.  I never heard it detonate once.  Understanding the data goes a long way.

[whittlebeast]

God bless a man with solid logic.....

Andy

[Steve Cole]

I all get out of this is that the afr accuracy of the can vary depending on conditions.  Knowing the issues, the device measuring the AFR can't be stupid.

Add.. Who cares if the variance is +- 0.5 points?? Still better than a carb..

So how does one expect the system to work then? If you tell it to control to a moving target then it will always be moving and never get to a final adjustment. Then experts say it must be xxx number or the tune is no good. You must plot the data and it look like a straight line, and on and on and on. The truth is they have no idea what it really is and have never spent the time to learn what really happens and why it is.  Data is a great thing when you understand what it is and how to use it, properly. Used improperly all you get is pretty pictures that do not mean much of anything.

So is the new rule of thumb going to be that fuel mixture only needs to be +/- 0.5 AFR? So a tuneup can be at anywhere between 13.0 and 14.0 at WOT and it's good enough? The truth is you have no way to know any better! If that's good enough then there is a ton of dyno sheets that got post here on this site that people said we no good and now they are just fine!

As usual you sound technical but really don't say anything that I understand.. Sorry I'm slow so you'll have to esplane' it to me..

What system? 

Who are the experts? 

What is  the XXX number? 

What target are you talking about?   AFR? 

Plot what data?  Do data plots have to be a straight line.. If you know what the curve is, you're half way there. 

You talking about the data collection piloting programs?  I've not looked at them much but it seems to me they be some value. At least there can be a collection of data that cam be tied helping with diagnostic issues.. Definitely not a controlled environment, with better equipment but better reading plugs.. 

So at WOT maybe you want the default to be 13.2  since you are WOT need a little more accuracy.. Figure out ways to make it more accurate. Filtering, taking other measurements and using the data in the data sheet to give you a better number.  If you have fixed bias in the system take it out..   What about partial throttle?   Maybe a swing from 15 to 14 is OK?   

if the transducer is well characterized and consitant does that mean because the output is not a ploted straight line that it's inaccurate?

So what part are you having trouble with? 1 + 1=2 or maybe it should be 1.5 or 2.5 

Who said anything about a transducer or for that matter a straight line?

"If you know what the curve is" what curve you talking about?

How would you know what the WOT is, if the sensor doing the measuring is incapable of telling you what was measured accurately, now your back to 1 + 1=2 or maybe it should be 1.5 or 2.5 

What it means is simple and even you know it already. Just because you see a graph on a piece of paper if the data that built that graph is wrong, the graph is wrong too! So those that want a magic mixture on a piece of paper to look a certain way, is completely meaningless, if you do not know and understand how the data came to be.

You can use bad data just about anyway you like and it's still bad data. Trying to tell someone that the answer needs to be any one thing from Bad data is like pissing into the wind and hoping you do not get wet.

If you make changes and the engine runs better, performs better and gets better fuel economy does it matter what the data from the control system says? How about it gets better fuel economy but runs hotter? Hotter than what? What's too hot for one maybe fine for another.

The truth to it all is in the eye's of the owner and if the outcome is what he/she wants does the rest of it matter?

[Admiral Akbar]

I all get out of this is that the afr accuracy of the can vary depending on conditions.  Knowing the issues, the device measuring the AFR can't be stupid.

Add.. Who cares if the variance is +- 0.5 points?? Still better than a carb..

So how does one expect the system to work then? If you tell it to control to a moving target then it will always be moving and never get to a final adjustment. Then experts say it must be xxx number or the tune is no good. You must plot the data and it look like a straight line, and on and on and on. The truth is they have no idea what it really is and have never spent the time to learn what really happens and why it is.  Data is a great thing when you understand what it is and how to use it, properly. Used improperly all you get is pretty pictures that do not mean much of anything.

So is the new rule of thumb going to be that fuel mixture only needs to be +/- 0.5 AFR? So a tuneup can be at anywhere between 13.0 and 14.0 at WOT and it's good enough? The truth is you have no way to know any better! If that's good enough then there is a ton of dyno sheets that got post here on this site that people said we no good and now they are just fine!

As usual you sound technical but really don't say anything that I understand.. Sorry I'm slow so you'll have to esplane' it to me..

What system? 

Who are the experts? 

What is  the XXX number? 

What target are you talking about?   AFR? 

Plot what data?  Do data plots have to be a straight line.. If you know what the curve is, you're half way there. 

You talking about the data collection piloting programs?  I've not looked at them much but it seems to me they be some value. At least there can be a collection of data that cam be tied helping with diagnostic issues.. Definitely not a controlled environment, with better equipment but better reading plugs.. 

So at WOT maybe you want the default to be 13.2  since you are WOT need a little more accuracy.. Figure out ways to make it more accurate. Filtering, taking other measurements and using the data in the data sheet to give you a better number.  If you have fixed bias in the system take it out..   What about partial throttle?   Maybe a swing from 15 to 14 is OK?   

if the transducer is well characterized and consitant does that mean because the output is not a ploted straight line that it's inaccurate?

So what part are you having trouble with? 1 + 1=2 or maybe it should be 1.5 or 2.5 

Who said anything about a transducer or for that matter a straight line?

"If you know what the curve is" what curve you talking about?

How would you know what the WOT is, if the sensor doing the measuring is incapable of telling you what was measured accurately, now your back to 1 + 1=2 or maybe it should be 1.5 or 2.5 

What it means is simple and even you know it already. Just because you see a graph on a piece of paper if the data that built that graph is wrong, the graph is wrong too! So those that want a magic mixture on a piece of paper to look a certain way, is completely meaningless, if you do not know and understand how the data came to be.

You can use bad data just about anyway you like and it's still bad data. Trying to tell someone that the answer needs to be any one thing from Bad data is like pissing into the wind and hoping you do not get wet.

If you make changes and the engine runs better, performs better and gets better fuel economy does it matter what the data from the control system says? How about it gets better fuel economy but runs hotter? Hotter than what? What's too hot for one maybe fine for another.

The truth to it all is in the eye's of the owner and if the outcome is what he/she wants does the rest of it matter?

I don't have any problems with addition, thank you..

Whats wrong with calling an O2 sensor a transducer?  You did mentioned the straight line.. "You must plot the data and it look like a straight line,"  up above..   

I have no clue as to what curve/line anyone is talking about,, That is why I asking.

I used WOT as an example where a tighter tolerance is needed.. Doesn't the ECU kinda sorta know where the throttle is based on TPS and MAP?


I'm not really sure what you are saying at the end.. What are you talking about?  Are you saying all measurements of AFR are useless?   Are you saying one sensor sucks pond water?  If this is the 1+1 = 2 then tell me what each 1 is..

Are you talking about the data collection scheme?  Whittlebeast's stuff?    Are you saying his stuff don't work?

Tuning in general?   

Is this whole episode, the result of some other episode that I missed?  Remember I typically don't bother reading subjects that involve flash tuners.. This one may have reminded me why I don't use them..

[whittlebeast]

This is more or less the graph you are looking for.

http://www.nightrider.com/biketech/images/nbo2_output.jpg

Yes, for a very narrow band, they are fairly liner.  At 450mv or so, they are very repeatable.  That graph is implying 600 mv is the sweet spot.  I tend to find where the motor is switching the o2 the fastest and just use that voltage as 14.6 and forget it.

Once the VE table is set, Generally I then turn off closed loop and put the target AFRs where I want to make the motor happy....  within reason...

I shoot for about 14 on the level highway and fade to about 13 AFR at WOT.

Andy

[whittlebeast]

If you look at a population density map of the USA and an elevation map of the USA, it is quite apparent that the majority of the miles driven by Harley riders is under 1500 feet.  Most of this high altitude tuning stuff is simply academic as most of us may only do mountain climbing on a motorcycle once a lifetime.  If the bike makes it to the top and starts when you are done looking at views, we are good to go.

Very few of us actually turn on our data loggers at the base of the mountain just to watch the magic for the fun of it.

I have been involved a couple of times with a guy that want to melt tires on the short shoots between 200 degree turns.  Those motors are a blast to tune.  The snow mobile guys deal with this stuff all the time.

Andy

[HD/Wrench]

WOW still going..


Guys tune the phucking bike get the MPG and the power call it good,  IF you are not a tuner well then ride the bike.. I do not give a rats azz what the sensor can or cannot do or what it come's down to as for a spec point.. +/- whatever

Can I do anything about it?? nope so again who cares ,  Its the tool I have to work with and that's that..

And I disagree I moved from Prescott to flat  Texas. AZ is full of mountains and some of the best riding I have ever done. I miss that , but I can tell you that I have in fact logged countless hours of data.

But again if I log with that data collection system its wrong, or at least one person thinks so. and the other does not

PHX  up the mountain for a little day time short trip 1500 to 5300 feet well that's a 45 minute ride. oh over here up to 7000 then down to 1500 back up to 4500 back to 1250 all in a tiny little day trip inside of 100 miles  , well you will be shocked that in fact that massive loss in pressure happens all over the US and you have AZ, NV NM CA UT CO all in that area that are pretty much the same type of riding in many areas.


I think all of you are pissing in the wind at high altitude..with this entire topic ( you will gain a few FPS on your stream , data log that) .. Have a great day I have to tune a bike with a PV and and I am using win pep 8 . I bet I can get the job done with out all you engineering type's telling me how to use my TOOL and what its capable of.. may a little less measuring it and just knock it home .. Myself I have that area covered very well.. Have a great day guys.... ohhhh look something shiney.

[whittlebeast]

Keep in mind you owe all of that PV tuning ability to engineers that were willing to look at the data and figure out the patterns.  They then give the programming logic to the programmers to type up the code.  Then those same engineers verify that the code is doing as designed based at looking at all that data graph stuff.

It all comes down to PV=nRT 

http://en.wikipedia.org/wiki/Ideal_gas_law

Andy