AF-XIED Beta for Husqvarna TR650

[Roger 04 RT]

I think the more efficient the combustion, the LESS O2 in the exhaust.

Who said anything about the Brisk plug injecting fuel?

However the characteristics of the Brisk may in fact improve the ignition & efficiency of combustion of that fuel, possibly leading to higher O2 in the exhaust

The AF-XIED "tricking" the ECU into "chasing" the 13.8 A/F

I think we also need to take into consideration valve over-lap

With a more complete combustion, the cylinder scavenging can be more efficient, inlet air being drawn into the exhaust, thus increasing O2 content

Not to mention reversal in the exhaust can

Hasenpfeffer is right, the more complete the combustion, the less O2 in the exhaust for constant conditions.

But the ECU won't allow an excess or shortfall of oxygen to exist for most states of engine operation. Closed Loop mixture adaptation will come into play. In the case that's hypothesized, incomplete combustion leading to an excess of O2, closed loop will add fuel and bring the O2 level back to that specified by the O2 sensor (plus AF-XIED). So assuming that hypothetical case, the bike would use more fuel but due to the richer mixture at the point of combustion might perform the same. Very tough to predict.

Now let's look at the other hypothetical case, a type of plug results in perfect combustion. It needs very slightly less fuel and perhaps combusts the fuel sooner. That might lead to peak cylinder pressure earlier than ideal (by maybe a degree or two). That might sound better but could result in a slightly less efficient work cycle due to the peak cylinder pressure being closer to TDC.

Yes, I'm speculating upon speculation. Do any of these hypothetical cases matter to us, only measurements will tell.

 

[engineerk9]

Wayne's data collection showed that the AF-XIED had failed, the Lambda voltage being "zero" the ecu reverting to the "limp-home" defaul settings to compensate

 

[drzcharlie]

How did it fail? How can the ECU negate it? What collection method is being used? 911? Can you point me to the data please? Thanks.

 

[engineerk9]

How did it fail? How can the ECU negate it? What collection method is being used? 911? Can you point me to the data please? Thanks.

Easy really, the Lambda voltage from the AF-XIED to the ecu was "zero" 0 volts, it would appear that the ecu reverted to the default "limp-home" mode

All of Wayne's data shows the O2 heater on, but 0 volts to the ecu

This also explains the incredibly high fuel consumption

The change from #8 to #9 ?

I've asked Wayne to visit again with his 911 to check if there's any damage to the ecu, or if there's any other "challenges"

 

[mag00]

Easy really, the Lambda voltage from the AF-XIED to the ecu was "zero" 0 volts, it would appear that the ecu reverted to the default "limp-home" mode

All of Wayne's data shows the O2 heater on, but 0 volts to the ecu

This also explains the incredibly high fuel consumption

The change from #8 to #9 ?

I've asked Wayne to visit again with his 911 to check if there's any damage to the ecu, or if there's any other "challenges"

I am so thankful for the work Wayne has been doing. I think with a good reset, your bike should be fine. I don't know that the GS911 has enough tools to get in deep enough for that. Hopefully it does, it probably should if you fix the o2 issue.

And E9, thanks for your participation.

Does the AFxied have a power lead? If it doesn't, there shouldn't be any major damage. Just a confused ecu, LOL.

 

[Roger 04 RT]

Easy really, the Lambda voltage from the AF-XIED to the ecu was "zero" 0 volts, it would appear that the ecu reverted to the default "limp-home" mode

All of Wayne's data shows the O2 heater on, but 0 volts to the ecu

This also explains the incredibly high fuel consumption

The change from #8 to #9 ?

I've asked Wayne to visit again with his 911 to check if there's any damage to the ecu, or if there's any other "challenges"

Ek9, Can you post a link to the data you're referring to?

There's no worry about damage to the ECU, its inputs and outputs are well protected against the worst that can happen, shorts to ground or battery. This protection is needed since the sensitive O2+/- leads are in the same shield with a switched-ground and power lead, to which they could easily be shorted if the lambda sensor was damaged.)

In this case, even if the O2+ input is held "lean" or to a low voltage, there's no worry. (BTW, you provide the AF-XIED's ground, and it get's power from the O2 Heater + lead of your O2 sensor, which as I mentioned above is in the O2 sensor cable. And power is needed for the AF-XIED to run its digital-signal-processing microprocessor.)

When you go from setting 9 back to 8, resetting the ECU may speed the process of re-adaptation, but it isn't mandatory.

The real question is, what was going on? Let's get a further look at the data and decide.

 

[engineerk9]

I am so thankful for the work Wayne has been doing. I think with a good reset, your bike should be fine. I don't know that the GS911 has enough tools to get in deep enough for that. Hopefully it does, it probably should if you fix the o2 issue.

And E9, thanks for your participation.

Does the AFxied have a power lead? If it doesn't, there shouldn't be any major damage. Just a confused ecu, LOL.

Removal of the AF-XIED, fixed the problem, Wayne ran more tests, removed the Fault that the O2 had created, voltages measured at the O2 sensor

Bike now, Pod-mod, GPR exhaust & Brisk plugs (I feel that they made a huge difference)

Will keep riding/testing as is, let's see how things go - today was 50+ klms in heavy rain & Sydney night-time rush hours - bike did well

Ek9, Can you post a link to the data you're referring to?

There's no worry about damage to the ECU, its inputs and outputs are well protected against the worst that can happen, shorts to ground or battery. This protection is needed since the sensitive O2+/- leads are in the same shield with a switched-ground and power lead, to which they could easily be shorted if the lambda sensor was damaged.)

In this case, even if the O2+ input is held "lean" or to a low voltage, there's no worry. (BTW, you provide the AF-XIED's ground, and it get's power from the O2 Heater + lead of your O2 sensor, which as I mentioned above is in the O2 sensor cable. And power is needed for the AF-XIED to run its digital-signal-processing microprocessor.)

When you go from setting 9 back to 8, resetting the ECU may speed the process of re-adaptation, but it isn't mandatory.

The real question is, what was going on? Let's get a further look at the data and decide.

The data shows "0" (zero mV) from the Lambda O2 signal throughout the tests

I rode across to Wayne's this afternoon, for further tests - I was naturally concerned - tests proved positive, no apparent damage & 800-900mv from the Lambda at tick-over (maybe a little lean, but liveable with)

I had removed the AF-XIED unit yesterday afternoon

It would appear that the Brisk plugs really are of benefit

 

[mag00]

Kieth, is the pinging gone now or is that still an issue?

The Brisk Multispark made a significant drivability improvement for me.

It is great news to hear that the GS911 is working. Also, it confirms some thoughts about why each and every bike runs different with different spoofs. Seems that the o2 signals may vary enough from bike to bike to be of note. Not enough to throw a code though.

 

[Roger 04 RT]

Removal of the AF-XIED, fixed the problem, Wayne ran more tests, removed the Fault that the O2 had created, voltages measured at the O2 sensor

Bike now, Pod-mod, GPR exhaust & Brisk plugs (I feel that they made a huge difference)

Will keep riding/testing as is, let's see how things go - today was 50+ klms in heavy rain & Sydney night-time rush hours - bike did well


The data shows "0" (zero mV) from the Lambda O2 signal throughout the tests

I rode across to Wayne's this afternoon, for further tests - I was naturally concerned - tests proved positive, no apparent damage & 800-900mv from the Lambda at tick-over (maybe a little lean, but liveable with)

I had removed the AF-XIED unit yesterday afternoon

It would appear that the Brisk plugs really are of benefit

The new GS-911 Wifi has had some teething issues, on the R1200 I've seen lambda control factors, which should range from 0.80 to 1.20, read either a binary 1 or a binary 0; and seen other factors just sit at 0. A reading of exactly 0 mV doesn't really make sense and may well be an anomaly. Resetting the ECU could as easily have fixed the issue.

If you can email me the CSV file I'd be happy to compare it with the hundreds of others I've got in my database.

 

[mag00]

Here is a vid you may be interested in watching. From Haltech.

View: https://www.youtube.com/watch?v=Zzn3-ygH-v8

 

[engineerk9]

The new GS-911 Wifi has had some teething issues, on the R1200 I've seen lambda control factors, which should range from 0.80 to 1.20, read either a binary 1 or a binary 0; and seen other factors just sit at 0. A reading of exactly 0 mV doesn't really make sense and may well be an anomaly. Resetting the ECU could as easily have fixed the issue.

If you can email me the CSV file I'd be happy to compare it with the hundreds of others I've got in my database.

Resetting the ecu between AF-ZIED every time #7 - #8 - #9 then - #8 then to #6 & resetting after Wayne's initial test didn't make any difference, the CONs reading on the dash went from 0.1 to over 16 within a few klms

 

[engineerk9]

Kieth, is the pinging gone now or is that still an issue?

The Brisk Multispark made a significant drivability improvement for me.

It is great news to hear that the GS911 is working. Also, it confirms some thoughts about why each and every bike runs different with different spoofs. Seems that the o2 signals may vary enough from bike to bike to be of note. Not enough to throw a code though.

Certainly the AF-XIED did throw a code, the GS911 tests showed that a fault code had been present, however had been removed, but not cleared, when I removed the AF-XIED

Wayne's very first GS911 test had showed error codes on both AIT & O2, which he cleared prior to test runs, it was those 2 early tests which showed , zero Lambda O2 volts

Thr Brisks certainly made a huge improvement in rideability, the pinging at WOT 3,500-4,500 rpm appears to have gone - the engine I feel seems very lean, however I'll put a few tankfuls under the wheels before doing any other mods - "interesting"

 

[Roger 04 RT]

Ek9,
I went to the GS-911 forum to check the status of my GS-911 WiFi. It appears that your bike was scanned with a beta release of the next revision which may or may not be ready to scan your bike. (See below) I can't tell from WayneC's posts if Hexcode is ready to support or not on TR650, only that WayneC tried it and got something.

This helps me to understand why you might have been told the O2 sensor values were zero.

Until we at least see release notes from hexcode telling us what does and doesn't function I'm going to hold off asking for the data from your bike.

Have raised it with Hexcode and they will put up the release announcement and release notes, they ar a little busy coding away at the moment

I just tested it with the Husky Terra and Nuda, no Series Menu but autoscan finds them and ECU info/fault codes/real time recording are all there

 

[mag00]

Ek9,
I went to the GS-911 forum to check the status of my GS-911 WiFi. It appears that your bike was scanned with a beta release of the next revision which may or may not be ready to scan your bike. (See below) I can't tell from WayneC's posts if Hexcode is ready to support or not on TR650, only that WayneC tried it and got something.

This helps me to understand why you might have been told the O2 sensor values were zero.

Until we at least see release notes from hexcode telling us what does and doesn't function I'm going to hold off asking for the data from your bike.

I'm curious to your thoughts on the pinging.

Seems a pretty straight forward test. Pings when installed, doesn't ping when removed. We now have two documented cases. One, the first guy to run the AFxied on a TR650, had a blown engine, from pinging. And now a second bike. Both bikes are from down under, so that brings up the next issue of stock ecu programming. If bikes down under are not liking the AFxied, why is that? Hmmm, maybe a different tune from the initial Moss? That being said, it is evident that the AFxied is screwing with the tune, much more than just fooling the o2 input.

But I am still curious for the empirical data about the AFxied. You know, a dual scope reading, one reading the o2 signal before the afxied, and one scope reading the output after the afxied has done it's thing.

And what I would be looking for in the over lay is the interval difference, not amplitude shift.

 

[Roger 04 RT]

I'm curious to your thoughts on the pinging.

Seems a pretty straight forward test. Pings when installed, doesn't ping when removed. We now have two documented cases. One, the first guy to run the AFxied on a TR650, had a blown engine, from pinging. And now a second bike. Both bikes are from down under, so that brings up the next issue of stock ecu programming. If bikes down under are not liking the AFxied, why is that? Hmmm, maybe a different tune from the initial Moss? That being said, it is evident that the AFxied is screwing with the tune, much more than just fooling the o2 input.

But I am still curious for the empirical data about the AFxied. You know, a dual scope reading, one reading the o2 signal before the afxied, and one scope reading the output after the afxied has done it's thing.

And what I would be looking for in the over lay is the interval difference, not amplitude shift.

magoo, I don't know what the emissions programming differences are between the countries but it would be interesting to know.

I see no reason to believe that the AF-XIED has anything to do with pinging on either bike you reference. And it is certainly unrelated to the blown engine, even that rider agrees that it didn't.

The AF-XIED does not affect the tune, period. It just lambda-shifts the input to the MM ECU with a completely normal looking signal. BTW, the LC-2 implementation used by several riders in this forum functions in the same way.

If you did attach a dual channel, scope with persistence so that you could watch the 2 second period of the O2 sensor waveform in closed loop you would see a waveform that looked identical but would be shifted in lambda. I have done this with a digitizing recorder to satisfy myself for the AF-XIED, LC-1 and LC-2.

Here's what you would see, working with the >700mV to <250 mV to >700mV normal stock narrowband O2 signal sequencing. For example, starting with the O2 sensor in a >700mV state, as the MM ECU reduces fueling to lean the mixture, the AF-XIED would get its signal to full lean sooner, signaling the MM ECU to start richening the mixture sooner. When the MM ECU tries to richen the mixture, it will take longer to get the AF-XIED to it's full rich >700mV state (so the O2 sensor get's to a rich state ahead of the AF-XIED. The waveform would look the same in both the stock and AF-XIED conditions, but the AF-XIED would be shifted forward in time. It does this using patented and patent-pending proprietary Digital Signal Processing algorithms.

So a simple description is the AF-XIED get to lean sooner (forcing more fuel) and then it takes longer to get back to a rich signal (also forcing more fuel). And remember, the amount of extra fuel is controlled and precise, unlike the EJK which forces Open Loop, and the PC which allows any amount of added fuel and which definitely messes with the tune.

 

[engineerk9]

I'm curious to your thoughts on the pinging.

Seems a pretty straight forward test. Pings when installed, doesn't ping when removed. We now have two documented cases. One, the first guy to run the AFxied on a TR650, had a blown engine, from pinging. And now a second bike. Both bikes are from down under, so that brings up the next issue of stock ecu programming. If bikes down under are not liking the AFxied, why is that? Hmmm, maybe a different tune from the initial Moss? That being said, it is evident that the AFxied is screwing with the tune, much more than just fooling the o2 input.

But I am still curious for the empirical data about the AFxied. You know, a dual scope reading, one reading the o2 signal before the afxied, and one scope reading the output after the afxied has done it's thing.

And what I would be looking for in the over lay is the interval difference, not amplitude shift.

I posted a number of dyno charts, with AF-XIED on #7 & then #8, both showed a shift in A/F from rich 13 to lean 15 at between 3,500-4,500 rpm with pinging at that point

There is a distinct possibility that neither Trooper Lu (the dealer who sold me the bike) or the Paul Feeney Group, Husky importers at that time (following conversations with their "technical" manager) had any idea about programing the ecu with the latest firmware prior to delivery -

In theory, wouldn't the ignition be slightly retarded by the ecu at WOT from 2,000rpm until higher rpm, to prevent engine damage due to over torque?

All very "interesting" - contact Wayne for copies of his GS911 results

I'm expecting to be "offshore" again till end of May, will undertake more rides "unspoofed" either AIT or O2 on my return

 

[DeLewis]

I have to chime in here and say that my experience with AF-XIED has been very positive. I experienced pinging occasionally prior to installing the AF-XIED but have never since. What I do notice is a surprising degree of variability in engine performance on different rides. Air temp may have some influence but the bike also seems to adjust to riding style to a degree. I don't know whether it is a deliberate feature of the ECU (I doubt it) but when I ride aggressively (lots of throttle) it seems to enrichen the mix across the rev range but also at all throttle settings. The result that the bike becomes more responsive and there is some exhaust popping as well.

 

[PaulC]

I have to chime in here and say that my experience with AF-XIED has been very positive. I experienced pinging occasionally prior to installing the AF-XIED but have never since. What I do notice is a surprising degree of variability in engine performance on different rides. Air temp may have some influence but the bike also seems to adjust to riding style to a degree. I don't know whether it is a deliberate feature of the ECU (I doubt it) but when I ride aggressively (lots of throttle) it seems to enrichen the mix across the rev range but also at all throttle settings. The result that the bike becomes more responsive and there is some exhaust popping as well.

+1

 

[mag00]

magoo, I don't know what the emissions programming differences are between the countries but it would be interesting to know.

I see no reason to believe that the AF-XIED has anything to do with pinging on either bike you reference. And it is certainly unrelated to the blown engine, even that rider agrees that it didn't.

The AF-XIED does not affect the tune, period. It just lambda-shifts the input to the MM ECU with a completely normal looking signal. BTW, the LC-2 implementation used by several riders in this forum functions in the same way.

If you did attach a dual channel, scope with persistence so that you could watch the 2 second period of the O2 sensor waveform in closed loop you would see a waveform that looked identical but would be shifted in lambda. I have done this with a digitizing recorder to satisfy myself for the AF-XIED, LC-1 and LC-2.

Here's what you would see, working with the >700mV to <250 mV to >700mV normal stock narrowband O2 signal sequencing. For example, starting with the O2 sensor in a >700mV state, as the MM ECU reduces fueling to lean the mixture, the AF-XIED would get its signal to full lean sooner, signaling the MM ECU to start richening the mixture sooner. When the MM ECU tries to richen the mixture, it will take longer to get the AF-XIED to it's full rich >700mV state (so the O2 sensor get's to a rich state ahead of the AF-XIED. The waveform would look the same in both the stock and AF-XIED conditions, but the AF-XIED would be shifted forward in time. It does this using patented and patent-pending proprietary Digital Signal Processing algorithms.

So a simple description is the AF-XIED get to lean sooner (forcing more fuel) and then it takes longer to get back to a rich signal (also forcing more fuel). And remember, the amount of extra fuel is controlled and precise, unlike the EJK which forces Open Loop, and the PC which allows any amount of added fuel and which definitely messes with the tune.

I don't share your sentiments.

Below is a real crude sketch/graph of different signal the o2 may produce under the three conditions. It by no means is accurate, merely a representation of concept.

Blue being the stock, non altered o2, red being the amlituded shifted o2 of the afxied, and green the interval shifted signal. Both the red and green shifting produces an altered voltage, but the signals are completely different.

It is and has been my contention all along, that the narrow band o2 system reads the signal as an alternating current voltage, not a dc current voltage. The longer pulses of the green line never spike above the trigger voltages the ecu was programmed with at the factory. If those voltages spike, the ecu is going to detect a malfunction.

Imagine the red lines moving up and down as you adjust the afxied from the lowest setting, (closest to stock blue) and highest, moving up.

When the engine runs through the different rpm, the ecu will draw from the mapping programmed. If the mapping calls for more fuel, it will adjust based on the pulses given by the o2. The pulses, not voltage, is what the ecu reacts to. A longer duration, as in the desired (green), would be the proper signal for the ecu to receive, not a higher voltage of same duration.

As mentioned, both would produce a higher voltage average, but it is the pulse that is needed, not the voltage as a wideband o2 would need.

This picture is for visual only. It does not represent an actual scope reading in any way.


But it would be nice to think that the only thing affected by this sort of spoofing is the fuel. I just don't buy that. The MM ecu is advanced enough to also adjust timing for any number of conditions. If the signal is telling the ecu that the engine is performing in such a way that it needs more spark advance, it is going to advance the timing.

If that timing change is improper for the rest of the engine conditions, it could ping. Prolonged pinging can and will blow an engine. Carbon build in a high compression engine is also deadly.

TWO ozzie bikes...coincidence? These are not bmw, but TR650. These (TR650) are higher compression engines and different engine management than the BMW. And Oz tuning appears to be different as well, we know this because of de restriction etc. We don't do that stateside.

 

[Roger 04 RT]

This is not a matter of sharing sentiments, this is a matter of correctly understanding how a system works.

The simplest thing to say at this point is if your blue graph is actual, your red graph is not what the AF-XIED does (i.e. It does not shift voltage) the green graph is not desired (it would create leaner running).

Read my description more carefully and you will see that what is desired is a green line (on your graph) that leaves the high state earlier and returns to the high state sooner.

The ecu does not read an AC waveform, it samples a waveform that varies. And the variations are changes that the ecu creates by modulating the the final steps of the fueling calculation.

I am glad that you're making an effort to understand how closed loop control works. A site with very good description, and the source code that runs the program, is the megasquirt site and the megamanual. If you had a stronger understanding rather that trying to imagine it, it would be more obvious to you how lambda-shifting and the AFXIED and LC-2 work together with the ecu. I am willing to help you learn if you want.

Let me say again, your chart is wrong, the AFXIED does not work by shifting voltage. It works by shifting lambda.