AFR confusion
Aug 11, 2005, 06:59 AM
#46
Originally Posted by shiv@vishnu
Hi guys,
Please don't deviate from the "happy zone" when dialing in your Fuel table. 17-19% corresponds to around 10.5:1, not around 9:1 as suggested by the innovate calibration data.
Please don't deviate from the "happy zone" when dialing in your Fuel table. 17-19% corresponds to around 10.5:1, not around 9:1 as suggested by the innovate calibration data.
This is because the XEDE does not display a 0-5v input as a perfect 0-100% signal. There appears to be a small offset, especially at lower voltages which cause the %-to-voltage relationship to deviate a bit. This doesn't matter when replicating a signal (as the XEDE is designed to do). But it can lead to confusion when trying to interpret voltage as a percentage as many have been trying to do.
Perhaps I should post up the calibration notes I acquired from my session on the dyno (when I correlated an1 input percentages (from the LC-1) to AFR (from the wideband on my Dyno). I'll dig this up tomorrow and post them. Sorry for posting the Innovate calibration data which lead to all this confusion.
To be honest, I structured the whole SMART fuel system upon AN1 %, and NOT around nominal AFR since that will differ from wideband to wideband brand. It was my attempt to avoid this confusion that resulted in all this whole other confusion.
So, once again, please don't think you are running super rich. Those AN1 targets perfectly replicate the AFR trace we tune for in all our baseline maps, as we have been for the last 3 years. Yes, they are a bit conservative. But in no case are they stupidly rich.
So, once again, please don't think you are running super rich. Those AN1 targets perfectly replicate the AFR trace we tune for in all our baseline maps, as we have been for the last 3 years. Yes, they are a bit conservative. But in no case are they stupidly rich.
d
Aug 11, 2005, 07:25 AM
#47
Newbie
Join Date: Apr 2004
Location: Detroit, MI
Posts: 79
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by donour
I have to step in here and point out that people claiming 17% -> 9:1 are simply wrong. 17% = 9.91:1, just like it should be. 20% is 10.35:1. After my runs last night (the software is so COOL by the way), I'm happy that fuel targets are pretty much what they should be.
Originally Posted by x99percent
17-19 ~= 9.9-10.2:1 AFR
Aug 11, 2005, 09:13 AM
#48
Evolved Member
iTrader: (20)
Join Date: Mar 2003
Location: Danville/Blackhawk, California
Posts: 4,941
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by donour
To reiterate my previous post. With my correctly calibrated (true free air) LC-1, using the correct Innovate conversion factor, I got about 10.35:1 at my richest point. Seems ok to me.
That's what we tuned Dustin's car for. It should hit around 12:1 by 4000rpm, then drop into the mid 11s and then drop into the mid 10s in the last ~1500rpm of the rev range. The SMART Fuel tables in the race gas maps should be approximately 1 point leaner at all RPM points. We do not design these maps to hold a constant AFR from 4000rpm to redline. Although the AFR trace from Dynojets would suggest otherwise. I've never figured out why they tend to read leaner than most widebands I've compared them to. Oh well. I guess this drives home the point that not all AFRs are created equal and comparing AFRs measured on different widebands with different sensor placements (tailpipe probe vs. installed in downpipe bung) is going to be problematic. This is also the reason (attn: Zues) that we packaged the SMART system with the LC-1 and the LC-1 alone. As you can see, there is enough confusion on these matters without having to account for different types of widebands.As for the SMART templates provided, it they be made a bit more agressive one is careful and knows what to look when doing so. I'll be updating the tuning tips in the SMART xede tutorial in the following days with this info.
Regards,
Shiv
Aug 11, 2005, 10:31 AM
#51
Originally Posted by shiv@vishnu
It is okay. That's what we tuned Dustin's car for. It should hit around 12:1 by 4000rpm, then drop into the mid 11s and then drop into the mid 10s in the last ~1500rpm of the rev range. The SMART Fuel tables in the race gas maps should be approximately 1 point leaner at all RPM points. We do not design these maps to hold a constant AFR from 4000rpm to redline. Although the AFR trace from Dynojets would suggest otherwise. I've never figured out why they tend to read leaner than most widebands I've compared them to. Oh well. I guess this drives home the point that not all AFRs are created equal and comparing AFRs measured on different widebands with different sensor placements (tailpipe probe vs. installed in downpipe bung) is going to be problematic. This is also the reason (attn: Zues) that we packaged the SMART system with the LC-1 and the LC-1 alone. As you can see, there is enough confusion on these matters without having to account for different types of widebands.
As for the SMART templates provided, it they be made a bit more agressive one is careful and knows what to look when doing so. I'll be updating the tuning tips in the SMART xede tutorial in the following days with this info.
Regards,
Shiv
That's what we tuned Dustin's car for. It should hit around 12:1 by 4000rpm, then drop into the mid 11s and then drop into the mid 10s in the last ~1500rpm of the rev range. The SMART Fuel tables in the race gas maps should be approximately 1 point leaner at all RPM points. We do not design these maps to hold a constant AFR from 4000rpm to redline. Although the AFR trace from Dynojets would suggest otherwise. I've never figured out why they tend to read leaner than most widebands I've compared them to. Oh well. I guess this drives home the point that not all AFRs are created equal and comparing AFRs measured on different widebands with different sensor placements (tailpipe probe vs. installed in downpipe bung) is going to be problematic. This is also the reason (attn: Zues) that we packaged the SMART system with the LC-1 and the LC-1 alone. As you can see, there is enough confusion on these matters without having to account for different types of widebands.As for the SMART templates provided, it they be made a bit more agressive one is careful and knows what to look when doing so. I'll be updating the tuning tips in the SMART xede tutorial in the following days with this info.
Regards,
Shiv
It is very apparent now with all of the confusion why you are sticking to one wide band for now. I initially thought the system was going to be more user (read dummy) friendly.
Aug 12, 2005, 06:09 AM
#52
I wasn't going to say anything but you comments below drove me to it 
The installation directions has one connect the LC-1's analog reference to the ECU's mounting bracket, not the Xede's ground. This can cause an offset at the digitizer (Xede) that can account for small offsets in the digitized number as seen by the Xede's processor. These small offsets obviously only count when the LC-1's output voltage is small.
I connected the LC-1's analog reference to the Xede's ground on the grey connector. While this could causes differences between my setup and how the maps were created, in runs yesterday logging the LC-1's view of the world (via Logworks) the resulting AFR tracks your goals very will. Thus in practice I'm not sure this makes much of a difference.
The installation directions has one connect the LC-1's analog reference to the ECU's mounting bracket, not the Xede's ground. This can cause an offset at the digitizer (Xede) that can account for small offsets in the digitized number as seen by the Xede's processor. These small offsets obviously only count when the LC-1's output voltage is small.
I connected the LC-1's analog reference to the Xede's ground on the grey connector. While this could causes differences between my setup and how the maps were created, in runs yesterday logging the LC-1's view of the world (via Logworks) the resulting AFR tracks your goals very will. Thus in practice I'm not sure this makes much of a difference.
Originally Posted by shiv@vishnu
Hi guys,
This is because the XEDE does not display a 0-5v input as a perfect 0-100% signal. There appears to be a small offset, especially at lower voltages which cause the %-to-voltage relationship to deviate a bit. This doesn't matter when replicating a signal (as the XEDE is designed to do). But it can lead to confusion when trying to interpret voltage as a percentage as many have been trying to do.
Shiv
This is because the XEDE does not display a 0-5v input as a perfect 0-100% signal. There appears to be a small offset, especially at lower voltages which cause the %-to-voltage relationship to deviate a bit. This doesn't matter when replicating a signal (as the XEDE is designed to do). But it can lead to confusion when trying to interpret voltage as a percentage as many have been trying to do.
Shiv
Aug 12, 2005, 09:26 AM
#53
Originally Posted by freedom
I wasn't going to say anything but you comments below drove me to it
The installation directions has one connect the LC-1's analog reference to the ECU's mounting bracket, not the Xede's ground. This can cause an offset at the digitizer (Xede) that can account for small offsets in the digitized number as seen by the Xede's processor. These small offsets obviously only count when the LC-1's output voltage is small.
I connected the LC-1's analog reference to the Xede's ground on the grey connector. While this could causes differences between my setup and how the maps were created, in runs yesterday logging the LC-1's view of the world (via Logworks) the resulting AFR tracks your goals very will. Thus in practice I'm not sure this makes much of a difference.
The installation directions has one connect the LC-1's analog reference to the ECU's mounting bracket, not the Xede's ground. This can cause an offset at the digitizer (Xede) that can account for small offsets in the digitized number as seen by the Xede's processor. These small offsets obviously only count when the LC-1's output voltage is small.
I connected the LC-1's analog reference to the Xede's ground on the grey connector. While this could causes differences between my setup and how the maps were created, in runs yesterday logging the LC-1's view of the world (via Logworks) the resulting AFR tracks your goals very will. Thus in practice I'm not sure this makes much of a difference.
d
Aug 12, 2005, 09:59 AM
#55
Evolved Member
iTrader: (20)
Join Date: Mar 2003
Location: Danville/Blackhawk, California
Posts: 4,941
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by Jorge T
Can I assume that your pump happy zone AFR is for 91 oct?
Shiv
Aug 12, 2005, 11:49 AM
#56
Not at all true. Analog / RF engineers world wide are cringing at this thought.
But as cited any offset appears to be small enought with respect to the signal present as not to make any difference.
But as cited any offset appears to be small enought with respect to the signal present as not to make any difference.
Originally Posted by donour
In practice though, ground should be ground.
d
d
Aug 12, 2005, 12:09 PM
#57
Originally Posted by freedom
Not at all true. Analog / RF engineers world wide are cringing at this thought.
d
Aug 12, 2005, 12:51 PM
#58
It's more than that given that the high resolution section of the table translates to 0.1v between cells.
I was going to say that 0.1v around the ECU isn't going to occur. I just checked the schematic for the ECU and its ground point is at the fuse/relay box in the engine compartment. (Xede's ground is the ECU's ground.) In the past I have measured 0.5v between the battery ground and the ground point on the left front fender wall.
Thus to be honest I'll bet that there is at least 0.1v of offset between a ground point in the dashboard and the engine compartment. I would run out and measure the ground offset but I'm sick of looking at that part of the car, at least for today....
I was going to say that 0.1v around the ECU isn't going to occur. I just checked the schematic for the ECU and its ground point is at the fuse/relay box in the engine compartment. (Xede's ground is the ECU's ground.) In the past I have measured 0.5v between the battery ground and the ground point on the left front fender wall.
Thus to be honest I'll bet that there is at least 0.1v of offset between a ground point in the dashboard and the engine compartment. I would run out and measure the ground offset but I'm sick of looking at that part of the car, at least for today....
Aug 12, 2005, 03:59 PM
#59
Evolving Member
Join Date: May 2005
Location: NorCal
Posts: 354
Likes: 0
Received 0 Likes
on
0 Posts
Then how about grounding the ECU bracket to the engine compartment ground? I would spend the time to run a wire to fix what could be a .1 to .5v offset. would this be feasable?
Originally Posted by freedom
It's more than that given that the high resolution section of the table translates to 0.1v between cells.
I was going to say that 0.1v around the ECU isn't going to occur. I just checked the schematic for the ECU and its ground point is at the fuse/relay box in the engine compartment. (Xede's ground is the ECU's ground.) In the past I have measured 0.5v between the battery ground and the ground point on the left front fender wall.
Thus to be honest I'll bet that there is at least 0.1v of offset between a ground point in the dashboard and the engine compartment. I would run out and measure the ground offset but I'm sick of looking at that part of the car, at least for today....
I was going to say that 0.1v around the ECU isn't going to occur. I just checked the schematic for the ECU and its ground point is at the fuse/relay box in the engine compartment. (Xede's ground is the ECU's ground.) In the past I have measured 0.5v between the battery ground and the ground point on the left front fender wall.
Thus to be honest I'll bet that there is at least 0.1v of offset between a ground point in the dashboard and the engine compartment. I would run out and measure the ground offset but I'm sick of looking at that part of the car, at least for today....
Aug 12, 2005, 05:15 PM
#60
In my opinion it is less trouble to strip back the insulation on the Xede's ground pin and move the LC-1's analog ground to that point rather than pulling a wire.
In any case let me make a measurement of the difference between the two points before you invest the time. But it's too hot outside right now to play. I'll do it in the morning and post it up here.
In any case let me make a measurement of the difference between the two points before you invest the time. But it's too hot outside right now to play. I'll do it in the morning and post it up here.
Originally Posted by Sackett
Then how about grounding the ECU bracket to the engine compartment ground? I would spend the time to run a wire to fix what could be a .1 to .5v offset. would this be feasable?