The sort of thing that gets my attention

[Steve Cole]

Guess they do not know how to handle it properly, or they feel the market segment is too small.

[wurk_truk]

The code is the same for BTs and Sportsters.  It is NOT busted.  That has been made very plain.

But, in the same vein, a sportster is NOT a BT.  They generally are ridden harder than a BT.  The cam events influence things along with head design.  The biggest factor, I think, is pipe design.  BT pipes are light years ahead of Sporty pipes.  These items make MOST tuners wish to open loop these bikes, simply because it is easier to make em run decent by doing so.  I imagine it is fairly involved to make one, with upgrades, perform well in CL without some work being involved.  Like I have said, to you, in the past, it would REALLY behoove you to check out those Bobcats.

Some BT owners will leave their pipes alone, but I cannot think of one sporty with stock exhaust.  I had a 59 sporty, and loved that bike with drag pipes.  Of course, the shifter was on the 'wrong' side from BTs and it had a magneto one spun around for kick starting, but it was a hoot to ride...  back in the days when they REALLY had leaded 100 octane!

I don't know about anybody else, but even though I think you post off the wall crap, I also think you are truly trying to learn.  I WANT you to succeed in this tuning crap and see you with a strong running sporty and use closed loop.

I have a friend, FBRR, a member here, who told me about the enleanment deal in open loop.  GM thought it would be stupid to allow the AFVs to make an engine leaner when in OL.  AFVs effect the WHOLE tune, and if things aren't dead right with pipes, the tune degrades over time.  With the pipes available...  MY (and only mine) thoughts are it's the pipes that make folks shy away.

Also, understand that Bean tunes with 5 gas and NOT AFR and he does NOT tune CL bikes AFAIK.  It's not that he wouldn't tune ur sporty CL, it's that he simply doesn't tune hardly ANY bike CL.  See?

I, again myself, feel that a smaller CL section is called for when using iffy pipes.  Since the AFVs WILL effect all open loop areas, why not let them do that with a smaller section of CL to start with.  I would NOT do idle areas, and I would not do anything over 50-60 MAP either.  I would find an area where the pipes DO work good and keep THAT as the cl area.  If the CL is solid, it will NOT make a tune turn to crap after a few hundred miles.  So, I like picking an area that IS solid, see what I mean?

Here is a thought.  Everybody gets in Steve's ass all the time, and Steve fights back, etc.  Ever wonder about that?  PV,EMS, et. al. leaves OTHERS to defend those products.  Steve does not.  Steve messes up at times, etc, but he IS passionate about his products, as opposed to 'bean counters' from Dynojet, etc.  You MAY wish to consider here who actually has quite a bit of experience, Steve, or others.

Me, I am like, and do like, Andy.  I also fiddle around, etc and the folks on here tell me how dumb I am with some of my ideas.  Yeah, I'm dumb.  And I tend to listen to Steve for guidance and when I don't believe him, I search it out with others.... even Andy.  A LOT of times 99%...  I discover I AM wrong and Steve was right, OK?

I think it is time for you to start playing around with v-tune on your bike.  Start with an unmolested base tune and see what happens.  Last time...  everything was a 'mystery' to you about tuning, but now you are gaining a handle on it.  BIG difference...  Even if in the end, you cannot get one to tune to your liking... at least you 'get' the process and just maybe Steve, Mayor, etc will help you along the way?  You will always have your present tune t fall back on if things go awry!

[Sporty 48]

Steve,

I like reading your responses and really appreciate your time. Steve does also represent a good product/company and has a responsibility from that point of view. So I respect the responsibilities.


Wurk Truk,

Funny, others suggested how strong the open loop is but have not made concrete suggestions. You have, thanks.
You also remembered the D&D Bobcats and the concerns they bring to the tune, good thinking. I wonder about how to check the effect on performance by the pipe? Am I looking for an instability in the data log? How would I spot it?
"BT pipes are light years ahead of Sporty pipes," thank you Wurk Truk for making me smile about off the wall comments I might make.


In this cool weather the Sportster is so smooth and fast, I love it.
A State Trooper gave me an early Christmas present last night, so remember boys, be polite, sit up straight and they have rear looking radar.

[glens]

Sporty, I thought you were asking me about settings recommendations for use while "tuning", which is why I mentioned the tuning guide for reference.  Personally, I like to drop hot idle areas just slightly out of closed-loop.  I leave the upper bounds where they were in the calibration.   You could certainly play around with it, but there are some possible ramifications, even if only in my head.

I'm going to throw a couple notions out here for consideration.  I don't have the definitive answers to the questions this will raise, but perhaps we can come to learn them by discussing it.

There are at least two ways the AFVs could affect open-loop operation (in the good way).

For purposes of this discussion, speculate that you have per cylinder an AFV table just like your VE tables, only much more sparsely populated (fewer cells at greater intervals).  Just like in the VE tables, an area of operation that falls outside the center of any given cell uses not only the information from that cell, but proportionately also the information from the adjoining cell(s) in the direction(s) away from center of the cell in which you're operating.  So if you were operating at the intersection of four cells, each of those four would be used with equal weight, averaged together for the value to use at the moment.

While the VE tables obviously go all the way to 100% throttle in a TPS/RPM table, the AFV table would not reach that high since there'd be no time the engine would be running closed-loop there, thus no way to update an AFV there.  Okay?   So the right side and the bottom of the AFV table would fall short of both 100% throttle and maximum engine speed.  But when you venture into those areas in operation, the last-known AFV from the direction you came would get used where you are now.  That would be one way (speculatively, but certainly the way I would do it) in which an AFV would affect open-loop operation.

Another way would be while in an area which is normally closed-loop, you've whacked the throttle open, calling for alternate operation which is itself not closed-loop.  The AFV(s) which would be influencing that spot while closed-loop will influence it still just the same.  I'm fairly confident this notion is in fact fact.

Now, back to that first, speculative, "AFV table".  Let's assume the cell locations are hard-coded.  Without knowing where the cells are, it would be possible, by alteration of parts of the calibration over which you have control, to decrease the scope of closed-loop operation.  Doing so could conceivably "strand" the last column or bottom row (or parts thereof) of AFV "cells".  If this were to happen, the always-open-loop areas which rely on them could never benefit from any proper updates to the AFV table.  If you'd switched from straight gas to corn gas, you could wind up leaner than you'd figured in those outlying areas until such time as you ran that fuel through and replaced it with what you'd used while tuning.  (Naturally, that problem slaps you right in the face in the event you're running open-loop everywhere.  If you'd so tuned to the corn gas, then you'd be richer everywhere while running straight gas.  Or maybe it'd be something unguessable when running whatever it was you could get for fuel some times.  And what would happen if your fuel pressure regulator got a little sick with no AFVs to compensate?)

You might also be able to "AFV-strand" areas on the left and top of the table if it's a fixed table.  As well certain areas within the table.  Now you might want to do some of that on purpose for good cause, but we still don't know the particulars of the table, so any of this could happen in a way you'd like, or in one you wouldn't.

Maybe the AFV "tables" are dynamic, based upon the closed-loop area called out in the Fuel table?  That could be a sweet deal.

We (at least I) don't know just how this is coded.  It would be nice to know, if possible, to be able to master the situation as/when needed.  And to avoid trouble.

Perhaps some of this is why some tuners opt to completely forego closed-loop when they encounter something not entirely closed-loop-friendly?  Just throw the baby out with the bath water?

Closed-loop operation has some very real benefits, for sure...

[Sporty 48]

Glens,
Thank you for the thoughtful explanation. It is far too complex and nuanced for me to understand all at once, so I will print it out and think about it.
In defense of the open loop guys, since I am still one foremost but secondly in 10 hours of dyno tuning and teaching, Bean gave me an open loop tune with room for improvement that put the faith back into me that the Harley was worth keeping and not a fat old pig.
Now it is a fat screaming boar. (Note: Wurk Truk, for a Sportster that is)

[wurk_truk]

Sporty will kick ass on a BT.

[Sporty 48]

Wurk Truk,

After the tune,
So much depends upon the clutch and throttle operator, the power to weight ratio and when above 100mph the Sporty is a little twitchy.
And I just got the message from God to slow down a bit.

[wurk_truk]

A lot of us have received this message in one form or another.  I wrecked in Tennessee and was lucky lucky lucky I didn't really get hurt.

Twitchy is no fun anyways.

[whittlebeast]

Here is an interesting high resolution screen shot.  These traces where taken off a totally stock Sportster running the stock tune.  The tune was a fresh install and the fuel trims were cleared.  The bike was ridden for about 30 min and the killed.  About 30 seconds after killing it, I started the PowerVision data logger and motor was started and logged for about 15 min.

On the left, I plotted MAP, RPM and the trace identified Rear AFF.  On the right is MAP, RPM and a function that is VE New/VE  Both traces are in the same scale.

Notice that both traces give very similar (but not identical) results.  Also notice how the VE are fat almost everywhere.

http://www.ncs-stl.com/PVTune/AFF_VENew_HighRes.bmp

The original data can be found at

http://www.ncs-stl.com/PVTune/StockInstall.zip

You can see the o2 driven injector swings if you elect to download and trace it.

Have fun tuning

Beast

[Sporty 48]

MegaLogViewer colors are opposite from the TTS.
IS that to be consistent with other Power Vision data plots?
The vagaries of the closed loop system are giving me fits......

[glens]

Here is an interesting high resolution screen shot.  These traces where taken off a totally stock Sportster running the stock tune.  The tune was a fresh install and the fuel trims were cleared.

What do you mean by "the tune was a fresh install"?  If everything was totally stock, including the ECM calibration, all you'd need to do would be to reset the AFVs.  By mentioning a "tune" installation you immediately call into question the basis for your current statements / implications.

The bike was ridden for about 30 min and the killed.  About 30 seconds after killing it, I started the PowerVision data logger and motor was started and logged for about 15 min.

Much more interesting would be a log of those first several minutes, not so much after you'd run it half an hour.  But I'll assume that "log0048_FreshStockInstall.csv" was in fact from the initial run and "log0049_StockInstall2edStart.csv" is the only log you mention making just there.

On the left, I plotted MAP, RPM and the trace identified Rear AFF.  On the right is MAP, RPM and a function that is VE New/VE  Both traces are in the same scale.

Much more visually informative is the "strip-chart" traces overlayed when both scaled from 75% to 125%.

Notice that both traces give very similar (but not identical) results.  Also notice how the VE are fat almost everywhere.

More on the "AFF" trace later, but probably the main reason they're similar but not identical is that the "VE New" isn't good data.  I mean, (again, assuming the first .csv was from the "fresh" run) "VE New" shows changes, albeit minuscule, before the O2 sensors even come on-line!

And how do you gather the VEs are "fat"?

You can see the o2 driven injector swings if you elect to download and trace it.

You mean injector-driven O2 swings, right?  'Cause that's what's happening, not the way you worded it.  It's textbook, Andy.  Pure textbook.

So PowerVision shows AFF (why not AFV as it's properly called?).  I suppose you consider that to be an unveiling of one of the "many" "Harley Hidden Tables"?  How about the "Gear", "Econ.Inst" [Economy (Instant)], and "TripA.Econ" [Trip A (Avg Economy)] "channels" in those log files?  Are they also HHTs?  Or are the whole lot of them "calculated channels" provided as "value added" stuff by PV?  (see the WinPV Help file):

"Power Vision provides insightful, valuable information on how your bike is running.

        •  Display all J1850 H-D vehicle data as well as wide band AFR and various calculated
           channels (such as MPG instant and trip MPG)
..."

In any event, THERE'S NOT ENOUGH DATA!

[cts1950]

Glen you can't take things at face value.
You know this sportster is several years old and has been tuned extensively. Andy did a restore of the original saved factory tune as saved by mastertune that = fresh install. Again clearing the fuel trims is nothing out of order when flashing the ecm he is simply satiating that all data registers are at a fresh starting point. I am not certain why you are always trying to pick a fight? What is the big deal what PV call one of it's data records as long as it is documented in its instructions. It is obvious you have a hard on  for any thing Andy wants to share. I would suggest if you don't like it, ignore this thread.

[glens]

Sporty, it's relative simple.  Really.  And the colors are immaterial.

In closed-loop operation, the ECM drives the fuel high, then low, then back again, then back again...  It monitors itself by reading the O2 sensors tracking the swings.  It zeros in, not on the specific AFR, but on the narrow range centered upon that AFR.  Over time as necessary corrections are noted, they're kept track of and utilized so as to keep the fueling swings nicely centered on the target output of the O2 sensors.  And these corrections can successfully carry over to the areas of operation which cannot be self-monitored via the simpler, faster, and cheaper type of O2 sensor we have in use: both wide-open throttle and sudden throttle transitions throughout the operating range, along with cold startup, to name a few.

This is an excellent method of maintaining the tune in the face of changing circumstances.  Circumstances like differing fuel make-up/quality; differing injector response over time; differing fuel pressure over time; etc.

A carburetor has better ability to adapt to running variations than does open-loop-only EFI.  And a carburetor can't hold a candle to EFI which is running with a self-contained system of checks and balances (closed-loop operation).

The only real drawback is that all the closed-loop running parts must work and work together in harmony.  Sometimes that doesn't happen, though, and the conventional wisdom says to just bypass closed-loop operation altogether when problems arise.  Some folks never even bother to look for any problems, they just step back several years in technological time as a matter of course.

[glens]

You know this sportster is several years old and has been tuned extensively.

Now I do, since you say so.  I'm certainly no mind reader.

I am not certain why you are always trying to pick a fight? What is the big deal what PV call one of it's data records as long as it is documented in its instructions. It is obvious you have a hard on  for any thing Andy wants to share. I would suggest if you don't like it, ignore this thread.

I can't ignore it now since I'm a part of it.

I'm not trying to pick any fights.  I'm trying to get Andy to put his intellect to good use and for good purpose in this matter.

He knows how closed-loop works yet he represents it otherwise and I can't figure out why he's doing that, if not just to stir things up needlessly.  I'll have to ask your forgiveness for my trying to be a voice of reason.

And as far as Andy sharing things, he should learn how to do it better.  Most of his stuff comes right out of left field.  He obviously has something in mind but often fails to actually state just what it is.  Like he's spent hours poring over a graph, pops on here to share it with us, and assumes that either we'll stare at it for hours too, or that whatever it is he thinks he's come to see just automatically jumps right out at us and tells the whole story; one which he has fleshed out in his mind but has failed to (re)produce for us here.

Then, when confronted with questions of most forms about it, says precious little.

I'm not trying to pick on him or pick a fight with him.  But some of his stuff is just plain wild and it can be a little difficult to be civil in the face of it and the way he does it time after time.

[whittlebeast]

Both the first and second ride on the stock tune are posted for download.  In the first ride, you will see the AFFs come on line about 10 min into the ride.  The O2s come on line at about 80 sec. 

On the second start, the o2s come on line in 20 sec.  The AFFs are there as soon as the bike is started as expected.

Beast

[Sporty 48]

Glens,

The colors are immaterial if they are consistent, otherwise confusing. But of course the color chart on the side tells the story.

A very good explanation of the ideal closed loop operation, one of the best I have read. 

"The only real drawback is that all the closed-loop running parts must work and work together in harmony," Glens.
I would point out that a lot of us Harley guys are not operating at top mental speed. We do our best with what we have to work with. When complicated systems cause problems the simplest working answer is the best solution. One of the reasons Harley lasted through the depression is the simplicity of solutions to many problems, a good hundred year old design like the 1911 Colt.

You should notice that Andy has been scolded from on high and is a chastened man. Enough said.

Andy made those posts in response to offline questions, he is trying to illustrate a counter point, to demonstrate in a clear manner what he has been working so hard to explain. And he has tried many times to demonstrate it so I see it and understand it. We both agree that we felt it, the shitty stock closed loop tune, now Missouri, show me.
Once again, although I have not yet opened the zip file, the way you respond to him clues me that there is something to what he says. Sorry, just the cynical retired investigator in me.
I really appreciate your efforts to bring me to the "closed loop" though. I can see advantages if everything is working correctly.

Sporty, it's relative simple.  Really.  And the colors are immaterial.

In closed-loop operation, the ECM drives the fuel high, then low, then back again, then back again...  It monitors itself by reading the O2 sensors tracking the swings.  It zeros in, not on the specific AFR, but on the narrow range centered upon that AFR.  Over time as necessary corrections are noted, they're kept track of and utilized so as to keep the fueling swings nicely centered on the target output of the O2 sensors.  And these corrections can successfully carry over to the areas of operation which cannot be self-monitored via the simpler, faster, and cheaper type of O2 sensor we have in use: both wide-open throttle and sudden throttle transitions throughout the operating range, along with cold startup, to name a few.

This is an excellent method of maintaining the tune in the face of changing circumstances.  Circumstances like differing fuel make-up/quality; differing injector response over time; differing fuel pressure over time; etc.

A carburetor has better ability to adapt to running variations than does open-loop-only EFI.  And a carburetor can't hold a candle to EFI which is running with a self-contained system of checks and balances (closed-loop operation).

The only real drawback is that all the closed-loop running parts must work and work together in harmony.  Sometimes that doesn't happen, though, and the conventional wisdom says to just bypass closed-loop operation altogether when problems arise.  Some folks never even bother to look for any problems, they just step back several years in technological time as a matter of course.

[glens]

Both the first and second ride on the stock tune are posted for download.  In the first ride, you will see the AFFs come on line about 10 min into the ride.  The O2s come on line at about 80 sec.

As you may have deduced, I had in fact fetched the files and had a look at them, and that I'd noticed "right off" both those points you just mentioned here.  Do you have any opinion on whether this "AFF" is a "channel" provided by the ECM or is one mathematically derived after the fact?  What's your opinion on the activity of "VE New" prior to any possible closed-loop operation? 

On the second start, the o2s come on line in 20 sec.  The AFFs are there as soon as the bike is started as expected.

Duly noted as well.

Even with knowledge of the source/derivation of the "AFF"s themselves, it would be hard to use them for anything meaningful without knowledge of any sort of "table" layout for their source in the calibration.  Unless you were able to identify where their cells (as such) were centered, and hold operation exactly there(s) for enough time to gather fair samples, all you can ever expect at any point in a log file is to see what a proportional average of up to four of those "cells" has become.

Are the values shown helpful at all?  Certainly.  They can indicate the relative accuracy of the calibration at the point in time the log was taken.  But it would be pretty futile to try to use them to accurize the calibration beyond below several percentage points in most every case.  That is to say, it'd be fair enough to consider them when checking on changes you'd made, but not so much so to use them as a source of changes.

[whittlebeast]

I now tune using the AFFs first to rough adjust the VEs.  If the code is not giving the AFFs then I use the VeNewFrontPctPV trick to essentially derive the Front and rear AFF.  Then I go to moving the target AFRs to 14.3 or so and chase out the lean holes. based on watching 30000 or so data point logs.  Last I play with the AFRs to see where the motor feels the best.  I have had very consistent results tuning this way, at least in the midrange where Harley riders spend the vast majority of their time.

I have no reason to think the AFF is anything but a raw channel from the ECU.  I could ask the PV programmers if we really need to know.

My wideband will be here Monday and provided the midwest weather holds out, we will get a few logs in soon to see how I did.

Over the years, I have found the narrow bands to be easier to "read" and the wide bands to be a little flaky.

Beast

[glens]

I have no reason to think the AFF is anything but a raw channel from the ECU.  I could ask the PV programmers if we really need to know.

Is there anything we don't really need to know?    By all means, ask and report.

[Steve Cole]

Let's look at the recorded data and tell me you cannot see anything that is wrong with it.

Let's use Andy's log0049_StockInstall2edStart.

Open the CSV in Excel and look at lines #180 - #190 Those 11 lines of data along with a bunch more are wrong and I am not saying this to start an argument but just so Andy can see what we have been trying to tell him for ages about the RAW data. In all those lines the Engine RPM is reported  as 1184 for 8 lines and 1516 for 3 lines yet the Injector PW is reported as 6.88 front and 7.52 rear the whole time! VE is reported as ZERO for the first 6 lines and 73.5 front, 72 rear for the next 5 lines. Then the Rich/Lean indicator is reported as 0 for the first 5 line and 1 for the last 6. These type patterns of BAD data follow all through the entire CSV file. As a matter of fact the pattern that comes to light if you spend some time with it seems to be about every 10 - 11 lines of data in bunches. A HD engine be it XL or BT will NOT sit at and hold 1 RPM for 10 - 11 data frames! Let alone jump from one RPM to another and hold it perfectly still. In this log it shows that it does not vary by 1 RPM for ~10 frames at a time.

Something is wrong and it is very obvious if you just spend a little time with the raw data looking at it in this CSV file. Look to see what makes sense with an understanding of how an engine runs and it's easy. I can tell you for sure this is not coming from the ECM this way as I've looked at raw data from HD bikes for too many years. The ECM will not send data as fast as these logs are showing so it would be my guess that the PV has an internal clock read that is happening this fast and that is what is being reported, but I have no way to know for sure.

[glens]

Looks like some fake "granularity" has been imposed on those charts.  How does that come about?

It'd be a simple-enough task from the command-line using standard unix text-manipulating tools to pull every 11th line from the file and reconstruct it with just them, but then some of the data changes within the 10 that'd get culled, so that's a no-go, too.

I'd sure like to hear about that "AFF channel".

[Steve Cole]

Take just the time stamp line and subtract the old from the new so you get to see the delta time between data lines and you will see a range of greater than 11 times! While the time is not huge with line by line varying 11 time the amount of time, god knows what goes with what. Now take that and put it against the fact that the ECM does not update each channel of data at the same rate and you can begin to see how impossible it is to work from the raw data without know something more about what is going on. Take Andy's made up formula's he posted earlier in this thread and you can begin to see why they do not work and you see it in the display.

As for your question about AFF I cannot answer that but I can tell you that the AFV's are not reported out in the data by HD if that's what your after.

When you see what's really going on its not so "way cool" any longer.

[glens]

Andy said he could ask the PV programmers about the "AFF channel" and I'd sure like to hear the response...

While you're here, Steve, what would the O2 voltage offset be on a 2012 touring?

[Sporty 48]

Anyone reading this should be impressed by the high sounding tag teaming techno talk.
I happen to have analyzed it as babble, very sure others see through it too.
Andy was trying to illustrate the not "broken code" but "jumbled code" that the Sportster people live with.
You two gentleman have just proven his point for him with your very elegant language skills.
Thanks.... 
Now, can we get down to fixing the problem or at least understanding it to make a work around?
Open Loop.
Bean, your tune is an example of hard work made to look like genius.

[glens]

Sporty, I'm not affiliated with Steve or his products in any way, shape, or form, and my only communication with him over the past two years (at least) has been openly right here in this site's forum.  Hell, he'd probably be offended by you thinking and/or suggesting that there's any affiliation between us!

As for myself, I don't care one way or the other what you might think in that regard.  The reason I'm here doing what I'm doing is because of a sense of civic duty.  If you want to disregard it that's fine; my conscience is clear.

If you're going to run open-loop throughout, removing all fueling adaptability ("protection") from the system which came with it, you'd best take the greatest care to get it right.  Using a method that involves only unstructured data logs just ain't going to get you there.  I don't know, maybe sportsters are cheap enough to experiment on.

By the way, the sportster code is neither broken nor jumbled.  Have a look again at the image I posted a few pages back.  The big twins exhibit exactly the same phenomenon.  It's textbook correct in both cases.