Hello, relative EFI n00b here....

Free and Open Source Engine Management discussions excluding more specific things on this board.
Post Reply
day7a1
TO220 - Visibile
Posts: 5
Joined: Fri Dec 26, 2014 7:01 am
Location: San Diego, CA

Hello, relative EFI n00b here....

Post by day7a1 »

Been reading through some posts, and have a few questions and requests for comment regarding how my particular ECU works.

My motorcycle is a Triumph, and we have the benefit of a 32x20 cell, alpha-N and SD switched ecu that has already been hacked and is accessible with TuneECU software via an OBD connection. I also have a SLOW datalogger hooked up to that connection. I have not performed any tuning based on the info, but have recently used WinLog view and previously OpenOffice calc to investigate some tunes that Triumph and various tuners have made. You can see the bulk of the results here: http://www.tiger1050.com/index.php/topic,29471.0.html

I feel I know way more than the guys over there....but absolutely nothing compared to what I've read over here.

I've decided to get a wideband so I can actually get some reliable feedback on my datalogging. Exactly which system I'm not sure yet. Strongly leaning towards the 14point7 products, just trying to decide between them.

First, there's no way I can hook up a 0-5V analog output to the narrowband sensor input to datalog the wideband the same way I currently datalog the narrowband, right? 5V....1V....pretty sure I can't but maybe I'm wrong. I can switch off the MIL.
Second. Is the simulator output sufficiently accurate on a wideband that I can get reliable data from the O2 V from the ECU? Obviously I can't for the narrowband, but does a particular simulated voltage give accurate results from say, .88-1.15 lambda?

Third, and probably more difficult......
My ECU has SD tables labeled as L and alpha-N tables labelled as F. There is a 1x8 row of cells, at seemingly random but increasing RPMs (up to about 60% of max rpm) that is labelled in this software as F-L switch. The available entries range from 0 to 50....not 100, not 255.....50 is max.

There is no debate as to what this actually does, because no one has any freaking idea. Some have shown it's not a hard switch between tables at that particular TPS as changing either table seems to affect the results at any [RPM,TPS] or [RPM,MAP]. Though I have not personally tested that theory, for the most part that info is reliable.

So, you EFI gurus and EEs, what could this 0-50 actually refer to, any ideas? If else, any recommendations on how I proceed to determine what it refers to?

Thanks guys....like I said, total n00b.
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Hello, relative EFI n00b here....

Post by Fred »

Gidday and welcome!

Good call on the 14point7 decision. May as well go spartan 2 with newer 4.9 sensor and plug for swap and changeability.

Re the "switch" or blend table, I'd love to know how they can prove that 0-50 is the correct scaling... because the raw values are almost certainly a standard data type. Do you know what the raw type is? And are you telling me that maxing or zeroing the table doesn't eliminate one algorithm type? If not, is it the tuning Def limiting you from achieving that? Something doesn't add up there :-) Can you post the factory raw data from that table? IE without any scaling or logic applied?

Regardless of the above, it could be anything at all that the firmware is doing with it. This is why I hate OEM tuning, there's always an element of uncertainty involved.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
day7a1
TO220 - Visibile
Posts: 5
Joined: Fri Dec 26, 2014 7:01 am
Location: San Diego, CA

Re: Hello, relative EFI n00b here....

Post by day7a1 »

Using WinLogView I was able to compare the fuel trims and average voltage of the NB against the [RPM,TPS] table of the same tune with different "switch" points. There didn't seem to be any appreciable difference in the parts between the curves that would be affected, but that's hardly conclusive. I could see a difference between two maps that has the same alphaN but different SD tables...it's a bit hard to think in so many dimensions. Maybe I need to stare at it longer.

If 0-50 is the wrong scaling it would make more sense. The tables are almost certainly blended, I would think a hard switch between the two tables would show up pretty clearly when looking at logs from maps where the switch is the only difference. Maybe not. A few people with dynos and WBO's have stated that they're blended, but haven't explored the nature of the blend.

One of those were maxed out above 3500 rpm, but I'd have to develop a scheme to positively prove that its eliminated. I'm not real sure what that proof would look like. +12% on SD table then move switch all the way towards alpha-N, go from there?

Not sure about posting factory raw data....I think that's a bit above my paygrade.

Regarding the 14point7, I think I want the SLC DIY2.1, which doesn't exist! It's a bit overpowered for my needs, but I really like the programmable simulator switchpoint and the display type. I've emailed Alan but he hasn't gotten a chance to get back to me yet. Not sure if the 4.9 is worth waiting for?
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Hello, relative EFI n00b here....

Post by Fred »

Ideally both tunes will be complete and usable such that in the event of a sensor failure, either can be used without the blend. Hence changing the switch point could easily make no difference to the actual fueling and EGO at all.

If I were trying to prove it wasn't used, I'd zero the SD table and blend fully to AN and verify that it still revved out.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
day7a1
TO220 - Visibile
Posts: 5
Joined: Fri Dec 26, 2014 7:01 am
Location: San Diego, CA

Re: Hello, relative EFI n00b here....

Post by day7a1 »

See.....that's why I came here. No one ever mentioned the (now) obvious likelihood that if it is a hard switch and you lost a sensor then it would only work either above or below the switchpoint.

Your method of proving it isn't used is probably more foolproof, but I already started a test. I've blended fully to SD and moved SD +12. My fuel trims are -8 or son throughout the rev range. I already loaded a map where I've blended fully to SD but moved AN to +12. Once I look at those results, I'll probably do yours, since it will likely look like the map where I had both tables normal but blended fully to SD, and I'll be wanting further proof of the result.

You know, while typing this, I realized that the Throttle is a 5V sensor. I wonder if the 0-50 is really a 0-5.0V? I've always had a theory that it switches to AN at a certain MAP value, and back to SD at a certain throttle value.....

clearly your method is the only way to test that.
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Hello, relative EFI n00b here....

Post by Fred »

Keep us (me?) posted on your results! :-)
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
day7a1
TO220 - Visibile
Posts: 5
Joined: Fri Dec 26, 2014 7:01 am
Location: San Diego, CA

Re: Hello, relative EFI n00b here....

Post by day7a1 »

Oh, I guess I should have posted this yesterday but here goes...

AN all 0:Switch to 0
AN normal:Switch to 0
SD all 0 :Switch to "50"
SD normal:Switch to "50"

Bike wouldn't run under any of those conditions. I'd get a few fires, then immediately go dead once the starter was not longer being pushed. I had previously warmed up the engine. I'm guessing the blend mechanisms aren't accessible via the program, which wouldn't be surprising. Doesn't answer what that setting does, though. I wonder if it's response times to inputs? The program is a hack of a hack of a tool. If I called a person that it would be a grave insult, but I can't think of a better way to describe it. It's not meant to actually control anything in depth.

I've got a few more tests planned, the idle trim values are pretty high, -8%, so there is some tuning that can be done there. I bumped AN down at idle 9% and moved switch to 50...I'll see what effect that has on that one particular set of values, rinse, repeat with more scenarios.

Still waiting to hear back from 14point7 on some questions I had. I'm guessing he's taking a holiday. Just weird that he answered back immediately on christmas! Otherwise, I wouldn't even be expecting a response until next week. :lol:
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Hello, relative EFI n00b here....

Post by Fred »

Alan has always been a bit slow and sporadic to respond to things. No one gets special treatment there :-) Sometimes you get one right back, other times I guess he's snowed under. Given he's in Canada, that could very well be literally :-D

As for your tests, I can think of these scenarios:

AN all 0, SD normal, Switch 0
AN all 0, SD normal, Switch 50
AN all normal, SD all 0, Switch 0
AN all normal, SD all 0, Switch 50

Is this the same as what you tried? It's not clear what was in the other table for each test, from what you wrote.

Additionally, you might be triggering some special logic or bug by having dead zeros in there, rather than some small value that won't run the engine. Might be also worth trying 10 or 20% of what the happy/normal values are instead of zero. This won't be enough to run/run anything close to right, but *should* (not guaranteed) dodge any funny logic or bugs.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
day7a1
TO220 - Visibile
Posts: 5
Joined: Fri Dec 26, 2014 7:01 am
Location: San Diego, CA

Re: Hello, relative EFI n00b here....

Post by day7a1 »

That's what I already tried, but today I tried moving the values to 1% of normal. I also realized that I can view the actual pulswidth. There was no change with the AN tables set to 1% on 0 or 50, but with the SD tables set to 1% it would not start. I had previously stepped down the AN values and showed no change at idle. Pretty sure now that the "Switch" is not a blend algorithm. Someone with resources about 7 years ago mentioned that it is more of a delta control that controls the rate of change from the SD to AN. This is preliminarily confirmed by my testing today. With the switch set to 0 a small change killed the engine. With the switch set to 50 I was able to rev it up quite a bit if I took it slow, but a quick change killed the engine. So I'm working with the theory that the blend % is TPS based and unable to be changed, but the rate of transition from 100% SD to, say, 50% SD/AN is what the switch controls.

Does that make any sense whatsoever?
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Hello, relative EFI n00b here....

Post by Fred »

Yes, some, however I'm a little hung over this morning :-D Keep at it. This is all just reverse engineering, be a good sherlock and put in the time and you'll get the results. It's just a matter of time and dedication. :-)
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
Post Reply