When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
I'll have to look at this again. I thought when I looked at it originally a couple years ago even accounting for the compressed voltage range at the end the taper still didn't make sense...
If you look at it in the AP (can see it better in M/E) ..... it doesn't look skewed because the chart software automatically ajusts the spacings to suit . Plotting it in XL doesn't do that however so leaves you with that skewed end .
Yeah but I *thought* I had done mine based on rate of increase from previous cell, not just visually looking at a graph. I'll need to dig back into it...
If you look at it in the AP (can see it better in M/E) ..... it doesn't look skewed because the chart software automatically ajusts the spacings to suit . Plotting it in XL doesn't do that however so leaves you with that skewed end .
I'm still looking for my original MAF scale spreadsheets I did a couple years ago. In the meantime I looked at percent rate of change of my current calibration vs the stock calibration for that region and to me it seems to follow. I have tweaked mine over time, so it's not quite the same as it once was.
Curious, how did you get yours to display correctly in Excel while overlaid with mine? I figured they would both be skewed?
I'm still looking for my original MAF scale spreadsheets I did a couple years ago. In the meantime I looked at percent rate of change of my current calibration vs the stock calibration for that region and to me it seems to follow. I have tweaked mine over time, so it's not quite the same as it once was.
Curious, how did you get yours to display correctly in Excel while overlaid with mine? I figured they would both be skewed?
When I extrapolated the voltage scale out to 5v the spacings became the same as the previous ones on the chart so the curve retains it's natural shape when plotted in xl.
I figured that, but also that it would have skewed or straightened both out if you overlaid with the spacing adjusted is all... But I guess not if my initial work was off... I just got off a plane so still need to work on this....
as far as how i got the values ............... i go for what looks the same as the original curve Mazda plotted . I'm sure there is an equation that works it out but doing it by eye in M/E (which has much better graphics than the Cobb) seems to work fine.
Here is the discrepency .................... your maf calibration in blue mine in orange
I just did a scatter plot of my MAF calibration, which automatically spaces out the values based on magnitude. I see a couple areas of error in what appears to be an otherwise smooth curve But it doesn't seem to appear to be of the magnitude that your graph showed. How are you getting your spacing? Can you scatter plot your MAF calibration? Can you share yours so I can overlay and view directly?
Some extra feedback I got from talking with the fuel injector guy when I picked up my Bosch injectors last week - he was quite sceptical about downscaling the MAF sensor using the resistor mod. He said something along the lines of, "Sure, it could work nicely, but while it brings your Voltage levels back below 5V, it won't account for the maximum airflow readings that the sensor element itself can take, so the figures at the top of this scale may be inaccurate." This doesn't seem to have been an issues for Brettus (as far as I know).
Injector guy recommended the following "industry standard" methods for upscaling the MAF sensor:
1. Increase the MAF tube diameter, or
2. Drop in a MAF sensor that can read higher airflow rates
Both of which obviously require recalibration of the MAF scaling graph.
So if you start seeing funky behaviour at high flows, and you're using the resistor mod, you could swap out the resistor mod for one of the above options.
I just did a scatter plot of my MAF calibration, which automatically spaces out the values based on magnitude. I see a couple areas of error in what appears to be an otherwise smooth curve But it doesn't seem to appear to be of the magnitude that your graph showed. How are you getting your spacing? Can you scatter plot your MAF calibration? Can you share yours so I can overlay and view directly?
I thought i sent you my values ? Not sure if the last 4-5 voltage settings are the same as yours ..... didn't check that but i know all the values up to there are the same . Mine are equidistant voltage spacings ,your's don't appear to be . Which could explain why your chart veers off when plotted in xl.
So if you start seeing funky behaviour at high flows, and you're using the resistor mod, you could swap out the resistor mod for one of the above options.
Nope , works well .However, I only dropped voltage by 10% because I was concerned about what errors that could arise .That gave me exactly the headroom i was after . There are actually products on the market that do exactly what I did so it's not like it's anything new or untried.
Interestingly , it appears that Slash's maf setup is reading much the same as mine even without the voltage change . I suspect his maf tube is a little larger than mine .
I thought i sent you my values ? Not sure if the last 4-5 voltage settings are the same as yours ..... didn't check that but i know all the values up to there are the same . Mine are equidistant voltage spacings ,your's don't appear to be . Which could explain why your chart veers off when plotted in xl.
I looked through my emails and didn't find your calibration, but maybe it got lost on the way across the Pacific
My thought was that if you overlay our graphs in a standard line graph in Excel using the same X-axis values then they will not align because the data sources in our respective calibrations for the X-axis are different. Yours being equidistant and mine bunched up similar to the stock calibration.
At any rate I only got excited about this as an academic exercise, I didn't mean to stir up trouble, ha! For all I know I did my math wrong, I wouldn't be surprised in the slightest I just like to talk it through and figure it out. I was just hoping we might find some clues to help us get the science down further, but maybe it's already been figured out and I just need to catch up. Ultimately, the smaller margins of error aren't likely worth chasing our tails over. Like you just alluded to, it could be that my MAF tube ID is slightly larger, which even by a millimeter would impact the readings. Along with differences in MAF tube dimensions is MAF sensor accuracy tolerance. I have gone through 5 MAF sensors since I have had the car and none of them have read exactly the same...
I looked through my emails and didn't find your calibration, but maybe it got lost on the way across the Pacific
It is on the xl spreadsheet that you sent to me and I sent back to you.
Originally Posted by slash128
My thought was that if you overlay our graphs in a standard line graph in Excel using the same X-axis values then they will not align because the data sources in our respective calibrations for the X-axis are different. Yours being equidistant and mine bunched up similar to the stock calibration.
At any rate I only got excited about this as an academic exercise, I didn't mean to stir up trouble, ha! For all I know I did my math wrong, I wouldn't be surprised in the slightest I just like to talk it through and figure it out. I was just hoping we might find some clues to help us get the science down further, but maybe it's already been figured out and I just need to catch up. Ultimately, the smaller margins of error aren't likely worth chasing our tails over. Like you just alluded to, it could be that my MAF tube ID is slightly larger, which even by a millimeter would impact the readings. Along with differences in MAF tube dimensions is MAF sensor accuracy tolerance. I have gone through 5 MAF sensors since I have had the car and none of them have read exactly the same...
There are a few things I like to use as indicators to how accurate my maf scaling/tune is .
*The maf curve itself should be uniform just like Mazdas' curve.
*NA loads are around 100% (at 0psi) so it follows that at 14.7psi loads should be approximately 200%. Maybe a little over that after 6500rpm as APV opens and improves VE.
*At a constant boost pressure it follows that loads should be relatively constant also as boost and load are roughly proportional . Except when the system gets very inefficient for some reason.
*Loads should also be proportional to torque , so if you are seeing a big spike in load but not a corresponding increase in torque ..... the load spike isn't real.
*With all this present and correct .......... afrs commanded vs actual should be reasonably close. If they aren't , I look at injector scaling . I generally don't play with VE too much believing that should reflect reality and hover somewhere close to 100%
*All this assumes the big enemy of maf tuning (boost leaks) is taken care of.
It is on the xl spreadsheet that you sent to me and I sent back to you.
There are a few things I like to use as indicators to how accurate my maf scaling/tune is .
*The maf curve itself should be uniform just like Mazdas' curve.
*NA loads are around 100% (at 0psi) so it follows that at 14.7psi loads should be approximately 200%. Maybe a little over that after 6500rpm as APV opens and improves VE.
*At a constant boost pressure it follows that loads should be relatively constant also as boost and load are roughly proportional . Except when the system gets very inefficient for some reason.
*Loads should also be proportional to torque , so if you are seeing a big spike in load but not a corresponding increase in torque ..... the load spike isn't real.
*With all this present and correct .......... afrs commanded vs actual should be reasonably close. If they aren't , I look at injector scaling . I generally don't play with VE too much believing that should reflect reality and hover somewhere close to 100%
*All this assumes the big enemy of maf tuning (boost leaks) is taken care of.
Missed that attachment, got it now. So it appears we have some mismatch in our MAF voltage row as well. I haven't figured it out yet, the numbers seem to be the same. However, when I use my MAF voltage numbers my curve looks much better than when I use your MAF voltage numbers. Conversely, when I plot your MAF calibration against my MAF voltage numbers your calibration has a dip where mine has a spike when plotted against your MAF voltage numbers.
Our calibrations plotted against your MAF voltage row. Blue line is you, orange is me. I have a spike when plotted against your MAF V numbers. Also notice that the x-axis represents the MAF voltage multiplied by not quite a factor of 10...
Our calibrations plotted against my MAF voltage row. Here the blue line is me, orange is you. You now have a dip when plotted against my MAF V numbers, and my spike is gone. Also notice that the x-axis represents the expected voltage values...
I emailed the file back so you can play with it direct if you want. I don't have time to dig further at the moment, but possible there is something silly with the number formatting when copied from our sources and pasted into the spreadsheet.
Missed that attachment, got it now. So it appears we have some mismatch in our MAF voltage row as well. I haven't figured it out yet, the numbers seem to be the same. However, when I use my MAF voltage numbers my curve looks much better than when I use your MAF voltage numbers. Conversely, when I plot your MAF calibration against my MAF voltage numbers your calibration has a dip where mine has a spike when plotted against your MAF voltage numbers.
Our calibrations plotted against your MAF voltage row. Blue line is you, orange is me. I have a spike when plotted against your MAF V numbers. Also notice that the x-axis represents the MAF voltage multiplied by not quite a factor of 10...
Our calibrations plotted against my MAF voltage row. Here the blue line is me, orange is you. You now have a dip when plotted against my MAF V numbers, and my spike is gone. Also notice that the x-axis represents the expected voltage values...
I emailed the file back so you can play with it direct if you want. I don't have time to dig further at the moment, but possible there is something silly with the number formatting when copied from our sources and pasted into the spreadsheet.
I didn't give you my voltages ............... Up to the last 5 they are the same as yours though.
The last five are : 4.53,4.65,4.76,4.87,4.99
So at this point, all I can fathom is something got goofed up in the file shuffling. If I scatter plot my MAF calibration from the original csv I posted in the link above it looks pretty smooth:
There are a couple bumps that need to smooth out but nothing near as drastic as the spike that shows up when I scatter plot based on the file you sent back with your MAF g/s values added:
The numbers in the MAF Voltage row appear to be exactly the same, yet something changes both the graph and the x-axis displayed values...
Ok Brettus... here is a combined scatter plot of stock, yours and mine, now that I have your proper MAF voltage values. That spike is gone from mine so I feel better now
Both our calibrations start out very similar, slightly above stock. Around 4V yours slopes down back towards the stock line while mine continues on its original trend above stock. At the top end at 4.77V mine slopes down in error. I plan to tweak my top end.
As you mentioned they are both working for us, so this is just informational from my perspective. I need to pull my MAF tube out and get a precise measurement compared to stock. A single millimeter larger diameter would increase the volume of air being measured.
Yes , in hindisght .... I assumed your voltages would be equispaced like mine ... doh !
Yeah, that's what I was getting at when I said I was trying to follow the stock spacing. Mazda bunched up the values at the end and I wasn't sure if there was something magical about the numbers they chose... It appears not, based on seeing your numbers
All this troubleshooting got me motivated to see if I could figure out a more scientific approach to extending the MAF. I found this software called Curve Expert which will find the formula for a given data set. From there you can have it calculate out further data points mathematically, no guesswork Attached is the resulting Excel file and data. Of course if anyone wants to use this you will need to pull some columns out to make room for the extended values in ATR. Also, since each MAF and setup has some margin of deviation the values will likely still need to be massaged. But at least this is a mathematically accurate curve.
All this troubleshooting got me motivated to see if I could figure out a more scientific approach to extending the MAF. I found this software called Curve Expert which will find the formula for a given data set. From there you can have it calculate out further data points mathematically, no guesswork Attached is the resulting Excel file and data. Of course if anyone wants to use this you will need to pull some columns out to make room for the extended values in ATR. Also, since each MAF and setup has some margin of deviation the values will likely still need to be massaged. But at least this is a mathematically accurate curve.
02-21-2016, 10:21 PM
I just like to talk it through and figure it out. I was just hoping we might find some clues to help us get the science down further, but maybe it's already been figured out and I just need to catch up. Ultimately, the smaller margins of error aren't likely worth chasing our tails over. Like you just alluded to, it could be that my MAF tube ID is slightly larger, which even by a millimeter would impact the readings. Along with differences in MAF tube dimensions is MAF sensor accuracy tolerance. I have gone through 5 MAF sensors since I have had the car and none of them have read exactly the same...
