Where do I start When learning how to tune?
06-28-2012, 06:50 PM
#51
Simple method:
Take some logs, the technique described in MMs thread is fine. You just need data from idle, cruising, 2nd gear full throttle, and 3rd or 4th full throttle.
Cleaning the MAF is a good idea.
Scaling the MAF table at the low end is the easiest way to eliminate fuel trims. Reducing fuel trims is important, because the long-term fuel trim (LTFT) that is developed under highway cruising conditions in closed-loop will affect full-throttle open-loop fuel delivery.
LTFTs are set over three ranges.
1-9 grams/second
9-20 g/sec
20-200 g/sec
Just select each range, hit m, and multiply the values by the error you are seeing in the log for that range.
For the last range, there is no need to scale the MAF beyond 100g/sec.
Iterate until the trims approach zero.
Adjust the fuel tables for gears 1-2 and 3-4 based on your full-throttle logs.
Iterate until you see the AFR you want. Best power is usually in the mid-13s, it is may be safer to aim for high 12s.
I find a visual log viewer like MegaLogViewer is a much easier way to view logs.
For the OMP, just increase the values in the load-based table.
I can't find it right now, but I have a Mazda research paper in which they measured oil delivery rate vs wall temperature. Temp decreases with increasing oil delivery, and then flat-lines.
In other words, there is an optimal point, but the only problem in going beyond the optimum is increased oil consumption. I have an OMP adapter/tank, and I like the smell of premixy exhaust, so have this table adjusted liberally.
Adjust the coil dwell times.
Done.
Not the way I do it, but I think this method is easy enough, and will probably give 95% of the improvements possible on an NA car.
Take some logs, the technique described in MMs thread is fine. You just need data from idle, cruising, 2nd gear full throttle, and 3rd or 4th full throttle.
Cleaning the MAF is a good idea.
Scaling the MAF table at the low end is the easiest way to eliminate fuel trims. Reducing fuel trims is important, because the long-term fuel trim (LTFT) that is developed under highway cruising conditions in closed-loop will affect full-throttle open-loop fuel delivery.
LTFTs are set over three ranges.
1-9 grams/second
9-20 g/sec
20-200 g/sec
Just select each range, hit m, and multiply the values by the error you are seeing in the log for that range.
For the last range, there is no need to scale the MAF beyond 100g/sec.
Iterate until the trims approach zero.
Adjust the fuel tables for gears 1-2 and 3-4 based on your full-throttle logs.
Iterate until you see the AFR you want. Best power is usually in the mid-13s, it is may be safer to aim for high 12s.
I find a visual log viewer like MegaLogViewer is a much easier way to view logs.
For the OMP, just increase the values in the load-based table.
I can't find it right now, but I have a Mazda research paper in which they measured oil delivery rate vs wall temperature. Temp decreases with increasing oil delivery, and then flat-lines.
In other words, there is an optimal point, but the only problem in going beyond the optimum is increased oil consumption. I have an OMP adapter/tank, and I like the smell of premixy exhaust, so have this table adjusted liberally.
Code:
3.00 3.00 3.00 4.00 5.00 5.00 6.00 7.00 8.00 8.00 9.00 10.00 10.00 11.00 12.00 15.00 19.00 19.00 21.00 5.00 5.00 5.00 6.00 7.00 7.00 8.00 8.00 9.00 10.00 10.00 11.00 12.00 13.00 15.00 18.00 21.00 25.00 29.00 7.00 7.00 7.00 8.00 8.00 9.00 9.00 10.00 11.00 11.00 12.00 13.00 13.00 15.00 17.00 21.00 25.00 30.00 36.00 9.00 9.00 9.00 9.00 10.00 10.00 11.00 11.00 12.00 12.00 13.00 14.00 15.00 17.00 20.00 24.00 29.00 35.00 42.00 11.00 11.00 11.00 11.00 12.00 12.00 12.00 13.00 13.00 14.00 14.00 15.00 17.00 20.00 23.00 28.00 34.00 39.00 45.00 12.00 12.00 13.00 13.00 13.00 13.00 14.00 14.00 15.00 15.00 16.00 17.00 19.00 23.00 27.00 33.00 38.00 42.00 46.00 14.00 14.00 14.00 15.00 15.00 15.00 15.00 16.00 16.00 17.00 17.00 19.00 21.00 25.00 31.00 36.00 41.00 44.00 48.00 16.00 16.00 16.00 16.00 17.00 17.00 17.00 17.00 18.00 18.00 19.00 20.00 23.00 28.00 34.00 39.00 43.00 46.00 49.00 18.00 18.00 18.00 18.00 18.00 18.00 18.00 19.00 19.00 20.00 20.00 21.00 24.00 29.00 35.00 41.00 45.00 48.00 51.00 20.00 20.00 20.00 20.00 20.00 20.00 20.00 20.00 20.00 21.00 21.00 23.00 26.00 31.00 37.00 42.00 46.00 49.00 53.00 21.00 21.00 21.00 21.00 21.00 21.00 21.00 21.00 21.00 22.00 23.00 24.00 28.00 32.00 38.00 43.00 47.00 51.00 54.00 22.00 22.00 22.00 22.00 22.00 22.00 22.00 22.00 22.00 23.00 24.00 26.00 30.00 35.00 40.00 45.00 49.00 52.00 56.00 22.00 22.00 22.00 22.00 22.00 22.00 22.00 23.00 23.00 24.00 26.00 29.00 33.00 38.00 43.00 47.00 50.00 54.00 57.00 22.00 22.00 22.00 22.00 22.00 23.00 23.00 24.00 26.00 27.00 30.00 33.00 38.00 42.00 47.00 50.00 53.00 56.00 58.00 22.00 22.00 22.00 22.00 23.00 24.00 26.00 28.00 31.00 33.00 36.00 40.00 45.00 50.00 53.00 56.00 58.00 60.00 60.00 22.00 22.00 22.00 22.00 23.00 26.00 30.00 34.00 38.00 41.00 44.00 49.00 54.00 59.00 60.00 60.00 60.00 60.00 60.00 22.00 21.00 22.00 22.00 23.00 27.00 33.00 37.00 42.00 45.00 49.00 54.00 60.00 60.00 60.00 60.00 60.00 60.00 60.00
Done.
Not the way I do it, but I think this method is easy enough, and will probably give 95% of the improvements possible on an NA car.
06-28-2012, 08:27 PM
#53
If you're logging for test purpose, most likely what you want to do is to compare the running AFR value with its value for that RPM and CalcLoad found in the corresponding entry found in the Load-Based AFR table. To do so, the car must be running in open loop. (Closed loop tells you nothing about these tables). Probably from Mazda Maniac, the classic logging point is 4th gear, 4200 OBDII-measured rpm (about 4400 on the tach), lights and AC (IIRC) on. This generally results in steady open-loop operation on level ground. STFT will read -0.16 in open loop mode. The PCM takes the rpm and load values, uses that to get a number from the table, adds or subtracts the percent value represented by (LTFT+STFT), then commands the injectors to provide what it thinks the right amount of fuel is.
Ideally, one would test each of the hundred or so values in each of the tables. Not at all practical to test this way, of course, in steady cruising mode unless one has a true brake dyno handy.
06-29-2012, 06:27 AM
#56
I'm not sure exactly what Team means either ... but there is a hard Redline at 10,000 RPM
Perhaps that is what he is referencing.
You cannot (in ATR anyways) enter a Redline higher than 10k rpm
edit --
Just select each range, hit m, and multiply the values by the error you are seeing in the log for that range.
For the last range, there is no need to scale the MAF beyond 100g/sec.
Not the way I do it, but I think this method is easy enough, and will probably give 95% of the improvements possible on an NA car.
@Wreakloose
I kept Oltmann's last line in here because I think the point of his post was "Basically" or "Simply" it was NOT written for a new person learning to tune to follow verbatim.
For example, you should never just pick an entire range of the maf and just scale it up with out doing some horizontal interpolation to smooth out the maf curve. I would also strongly recommend to NEVER just multiply the maf scale by the LTFT value, do it in small increments and re-log. You'll often find you don't need the entire LTFT amount to eliminate the trim.
I know Oltmann that you know this ... my point is WreakLoose don't got making big value changes to anything if you must tinker with your tune.
Small changes mean small failures
Perhaps that is what he is referencing.
You cannot (in ATR anyways) enter a Redline higher than 10k rpm
edit --
Just select each range, hit m, and multiply the values by the error you are seeing in the log for that range.
For the last range, there is no need to scale the MAF beyond 100g/sec.
Not the way I do it, but I think this method is easy enough, and will probably give 95% of the improvements possible on an NA car.
I kept Oltmann's last line in here because I think the point of his post was "Basically" or "Simply" it was NOT written for a new person learning to tune to follow verbatim.
For example, you should never just pick an entire range of the maf and just scale it up with out doing some horizontal interpolation to smooth out the maf curve. I would also strongly recommend to NEVER just multiply the maf scale by the LTFT value, do it in small increments and re-log. You'll often find you don't need the entire LTFT amount to eliminate the trim.
I know Oltmann that you know this ... my point is WreakLoose don't got making big value changes to anything if you must tinker with your tune.
Small changes mean small failures
Last edited by wcs; 06-29-2012 at 06:39 AM.
06-29-2012, 07:16 AM
#57
I kept Oltmann's last line in here because I think the point of his post was "Basically" or "Simply" it was NOT written for a new person learning to tune to follow verbatim.
For example, you should never just pick an entire range of the maf and just scale it up with out doing some horizontal interpolation to smooth out the maf curve. I would also strongly recommend to NEVER just multiply the maf scale by the LTFT value, do it in small increments and re-log. You'll often find you don't need the entire LTFT amount to eliminate the trim.
I know Oltmann that you know this ... my point is WreakLoose don't got making big value changes to anything if you must tinker with your tune.
Small changes mean small failures
For example, you should never just pick an entire range of the maf and just scale it up with out doing some horizontal interpolation to smooth out the maf curve. I would also strongly recommend to NEVER just multiply the maf scale by the LTFT value, do it in small increments and re-log. You'll often find you don't need the entire LTFT amount to eliminate the trim.
I know Oltmann that you know this ... my point is WreakLoose don't got making big value changes to anything if you must tinker with your tune.
Small changes mean small failures
Logs had the LTFTs at 4.5 / 5.3 / 6.9 for the 1-9, 9-20, 20-300 g/sec ranges respectively. I left the first two entries in the table alone (because some claim changing them causes problems) and multiplied the rest by 1.06. LTFTs look pretty much as expected afterward. There's some evidence of spikes at the lower rpm as it comes out of closed loop. What I'll do is concentrate on logging pulls done very smoothly and set my closed loop exit rpm very low (1500) to keep it from complicating the log plots.
Even the free version of the MegaLogViewer is worth its weight in premix, btw, use it!
Last edited by HiFlite999; 06-29-2012 at 07:28 AM.
06-29-2012, 09:31 AM
#58
Thanks, but its what any of the experienced guys around here will tell you. Jeff's enforced this, Kane and Brett as well.
This is what we are lead to believe however I've not actually tried it.
If your going to do it that way ... also adjust the Closed Loop Exit Load table A, B and C (depending on year)
Or you can unplug your Front O2 sensor and use you wideband ... if its setup to log and you trust it.
The other alternative (Kane showed me this) is to do logging at 500 rpm intervals starting at 1500rpm
Take the average LTFT and the average STFT ... add the two together and adjust the Maf area accordingly with a little bit of horizontal interp.
Repeat for each 500 rpm log.
Or you can unplug your Front O2 sensor and use you wideband ... if its setup to log and you trust it.
The other alternative (Kane showed me this) is to do logging at 500 rpm intervals starting at 1500rpm
Take the average LTFT and the average STFT ... add the two together and adjust the Maf area accordingly with a little bit of horizontal interp.
Repeat for each 500 rpm log.
06-29-2012, 12:30 PM
#59
The other alternative (Kane showed me this) is to do logging at 500 rpm intervals starting at 1500rpm. Take the average LTFT and the average STFT ... add the two together and adjust the Maf area accordingly with a little bit of horizontal interp. Repeat for each 500 rpm log.
Last edited by HiFlite999; 07-02-2012 at 05:35 PM.
06-30-2012, 08:45 AM
#61
Oh, not arguing at all here, but I'd already tried the other way. From a lot of other experience with ~feedback loops, as long as I'm not risking damage, I've come to prefer taking relatively large steps first to get some clear idea as to the response - often the effects of small steps gets lost in the data noise.
Sorry I didn't mean to make it sound like you were. I was just trying to give credit to those that had made me take small steps. However ... the lost in data noise is kind of curious.
I typically change one parameter only and re-log. Rarely is the change lost in noise.
Seriously ... can you give an example?
Sadly, I have no wideband - basing everything on the front sensor. I'd already tried some smaller changes to those tables, but the datalogs are showing that the settings which should kick it out of closed loop, aren't. It's possible that being calculated values, their response is not quick enough to keep up with fast-rising rpm during WOT. Do you know which table corresponds to which MAF range?
If the following changes do not drop you in to OL sooner ... I'll be a scurvy dog.
Thanks for that tip. I'm a bit leery of doing all corrections with Maf alone though. Resonances in the intake will cause more or less air to arrive during port opening even though the time-averaged airflow past the intake remains the same. Same with shutter valve opening/closing. Complicating things is that changes to the MAF calibration also change the calcLoad (I think). This paragraph probably falls into the "fine tuning" category, however, and I'm not really there yet.
But I typed it all out and don't want to delete my effort
lol
(oh a butter fly)
(All the corrections?
Resonances? Mmmmmm You mean the VDI dynamic supercharger effect? It's still air measure past the maf.
I'm not sure where you're going with all that ... Too be honest I'm a bit confused by it. This effect is still seen by the maf as more air moving past it.
The maf calibration shouldn't change because of the SSV opening. The ssv opens and this creates a void that needs to be filled, it's all still air past the Maf.
This is what we see in the dyno's. Basically Air is rushing down the intake in to fill the vacuum created by the intake cycle ... but then suddenly ahead of this the intake changes it's size by opening up space and creating a hole (vacuum) that needs to be filled. This creates a tiny delay in delivery of air to the intake cycle.
Once this space is filled more air is available to the intake cycle for that split second and can cause a lean spike. No the intake tract has changed and more air is available.
What this creates or what we use to compensate for this is the VE or Fuel Map. Depends on your style, I guess. Kane and Brett I believe use the VE table while Jeff is more of a Fuel table guy (in his webinar anyways).)
Last edited by wcs; 06-30-2012 at 08:59 AM.
07-02-2012, 06:53 AM
#63
Cool, TeamRX8. Why not a true expert, presumably you, post up a proper a,b,c,d ... step-by-step procedure to tune? 
Or is it too fundamentally too complicated? Or is it more fun for you to snipe at those trying to sort through the couple thousand posts here with a couple of books thown in, to distill out something useful for the lowly masses?
Or is it too fundamentally too complicated? Or is it more fun for you to snipe at those trying to sort through the couple thousand posts here with a couple of books thown in, to distill out something useful for the lowly masses?
07-02-2012, 09:22 AM
#66
I think the next confusing thing is the phantom rear O2 trim - will re-read Harlen's and Oltmann's posts on the subject on how to disable that function during tuning. I may also back off of scaling the entire MAF range by 1.06 (which also changes calcLoad) (though it pretty much fixed the LTFTs) and play with the injector size entries (which doesn't). Reshaping the MAF curve in a region that isn't right didn't change the problem ...
07-03-2012, 08:47 PM
#69
If I set my dwells to this table I should be good to go, right? (BHR Coils)
Been meaning to bump this. I made a spreadsheet to convert the dwell values to/from milliseconds.
This is the stock table.
This is what I've been running for the past few months.
This is the stock table.
Code:
1000 2000 3000 4000 5000 6000 7000 8000 9000 6.50 7.58 6.00 4.75 3.56 2.85 2.38 2.04 1.78 1.58 7.75 6.75 5.33 4.67 3.56 2.85 2.38 2.04 1.78 1.58 9 6.08 4.83 4.22 3.56 2.85 2.38 2.04 1.78 1.58 10.25 5.58 4.42 3.86 3.52 2.85 2.38 2.04 1.78 1.58 11.5 5.17 4.08 3.58 3.27 2.85 2.38 2.04 1.78 1.58 12.75 4.83 3.83 3.36 3.04 2.83 2.38 2.04 1.78 1.58 14 4.42 3.58 3.14 2.85 2.65 2.38 2.04 1.78 1.58 15.25 3.75 3.38 2.97 2.71 2.50 2.36 2.04 1.78 1.58 16.5 3.25 3.13 2.81 2.56 2.38 2.24 2.04 1.78 1.58
Code:
1000 2000 3000 4000 5000 6000 7000 8000 9000 6.50 21.09 16.08 12.34 11.02 9.00 7.00 5.55 4.50 3.50 7.75 15.39 12.61 10.07 9.03 8.00 7.00 5.55 4.50 3.50 9 9.72 9.15 7.81 7.04 6.30 6.30 5.55 4.50 3.50 10.25 7.51 7.23 6.55 6.12 5.55 5.55 4.86 4.50 3.50 11.5 5.71 5.71 5.64 5.40 5.00 5.00 4.17 3.84 3.50 12.75 4.73 4.73 4.73 4.68 4.46 4.46 3.80 3.65 3.50 14 4.16 4.16 4.16 4.16 4.11 4.11 3.50 3.35 3.20 15.25 3.83 3.83 3.83 3.83 3.81 3.81 3.30 3.11 2.91 16.5 3.50 3.50 3.50 3.50 3.48 3.48 2.92 2.77 2.61
11-02-2012, 06:58 PM
#73
Hey I just wanted to thank everyone who contributed to this thread, it's been very useful! I've successfully upped my idle to 875, lowered my fan turn on temperatures, increased omp, and have hidden my cat cel. So far no more a/c vibrations due to the upped idle, and my coolant temperatures don't go above 190, it's awesome!
11-04-2012, 10:33 AM
#74
Because this is open loop flow range so there is no trim to adjust for. Whatever the true flow is doesn't matter, only that you adjust the tables to provide proper AFR for it. If you want to go **** and try to make commanded and actual AFR line up you can devote a lifetime to chasing your tail with playing there.
In my experience using 1-12 for idle and 12-80 for cruise is all you need to be concerned with for practical NA trim adjustment.
.
In my experience using 1-12 for idle and 12-80 for cruise is all you need to be concerned with for practical NA trim adjustment.
.
Last edited by TeamRX8; 11-04-2012 at 10:39 AM.
11-05-2012, 09:44 AM
#75
^ On the other hand, scaling the MAF table above 80ish gm/sec based on the highest level CL trim doesn't do any harm either. This is especially so if one plans to mess with the main a/f tables to get the desired values. In my limited experience, I find that scaling the higher values as well, results in OL a/f measured values closer to the table values.
I agree that it matters not much how it's done if one achieves the desired measured a/f in the end.
However, playing with other stuff now has me running within +/- 2% of the a/f given in the main a/f table over most of the rpm vs load range in OL. Getting it right for WFO is pretty easy. Filling in the rest of the table is time-consuming and tricky for the low rpm or low load side. I need to improve my data-taking techniques and perhaps automate the corrections, a la Kane, I suppose.
It takes time though, and winter is coming. I also need to find another test route since certain ppl seem to be upset with full-throttle runs with a rather open exhaust.
I agree that it matters not much how it's done if one achieves the desired measured a/f in the end.
However, playing with other stuff now has me running within +/- 2% of the a/f given in the main a/f table over most of the rpm vs load range in OL. Getting it right for WFO is pretty easy. Filling in the rest of the table is time-consuming and tricky for the low rpm or low load side. I need to improve my data-taking techniques and perhaps automate the corrections, a la Kane, I suppose.
It takes time though, and winter is coming. I also need to find another test route since certain ppl seem to be upset with full-throttle runs with a rather open exhaust.