|
nyet
|
 |
« Reply #45 on: November 02, 2015, 11:53:11 PM »
|
|
|
I'm interested in knowing how the MAF airflow is translated into the UPM % value.
No clue what you mean by "UPM %", but: http://s4wiki.com/wiki/LoadAnd if you "fix" the load to be correct, the ECU won't know WTF to do under boost. Again, for the umpteenth time, you're using an ECU designed for NA, not FI. |
|
|
|
« Last Edit: November 02, 2015, 11:56:04 PM by nyet »
|
Logged
|
|
|
|
hackish
Full Member
Karma: +1/-1
 Offline
Posts: 56
|
|
« Reply #46 on: November 03, 2015, 11:40:45 AM »
|
|
|
UPM was airflow per minute or the axis used on most of the tables. In my case the concern is mostly just the ignition table. I've done some research and found that the car has a VR6 MAF and "green giant" 440cc injectors installed. I expect this explains why the car never goes above 50% load yet the fuel is perfect.
I'm thankful for the help people have been providing but a simple answer "won't work" explains nothing to me. At present the car runs, starts, idles and drives reasonably well considering. The load numbers do need to be corrected so the timing map can be used properly but I'm not seeing anything here that is so critically wrong.
Why will the NA firmware not run boost?
|
|
|
|
|
Logged
|
|
|
|
|
nyet
|
|
« Reply #47 on: November 03, 2015, 11:42:13 AM »
|
|
|
Why will the NA firmware not run boost?
Itemize every map that uses load as an input. Now, rescale the axis for each one, AND its corresponding data. Now you'd also better pray that the load scaling used doesn't hit a representation max (like the pressure limit in ME7). Of course, you can then go and scale EVERY single piece of load arithmetic in the system to give you headroom there too. You can make anything work, that isn't the question. The rabbit hole goes pretty deep. Every hack thinks 'well this is good enough' until something breaks or somebody gets frustrated with shitty drivablility or codes that keep popping up. |
|
|
|
« Last Edit: November 03, 2015, 11:44:57 AM by nyet »
|
Logged
|
|
|
|
hackish
Full Member
Karma: +1/-1
Offline
Posts: 56
|
|
« Reply #48 on: November 04, 2015, 09:16:28 AM »
|
|
|
Ok, I understand. I know that a number of NA cars have items that actively prevent you from going into boost so I needed to make sure it would not suffer from that.
The car arrived running like crap and it blew up an engine before it came to me. While I'd love to do a perfect job this project is being done as a bit of a favour and to get some bosch experience. I pulled a bunch of timing out of it to stop it from detonating and the guy already thinks it's amazing. With limited resources and no budget to replace sensors, ecu and wiring for a boosted ecu there are not many options. Hope this explains why I didn't go ahead and swap it over to a 1.8t. I'm not going to pay money to work on a guys car.
I'm going to rescale the injector and MAF table to gain the amount of load I need. If the torque system gets in the way too much then I'll strike a balance between the extra 30% power it needs by underscaling them slightly.
With the MAP/Injectors scaled correctly can anyone point out which tables I'm going to need to adjust for the torque system? In other ECUs I'm used to having a gas pedal (APS) versus torque or airflow type table that attempts to achieve the desired airflow based on the APS position. I've been reading the functions descriptions documents but it feels like there should be a strategy document like seimens does to better explain each module and how it operates.
|
|
|
|
|
Logged
|
|
|
|
|
nyet
|
|
« Reply #49 on: November 04, 2015, 11:21:41 AM »
|
|
|
to get some bosch experience.
This is not a good way to get bosch experience for a noob; the best route would be to start with basic stage 1 tune of a car with an FI system in it stock. there should be a strategy document like seimens does to better explain each module and how it operates.
There is, but unfortunately it is probably more general than you want. http://files.s4wiki.com/docs/Audi_S4_engine_english.pdfhttp://nyet.org/cars/info/ME7sw.pdf |
|
|
|
« Last Edit: November 04, 2015, 11:24:55 AM by nyet »
|
Logged
|
|
|
|
|
adam-
|
|
« Reply #50 on: November 04, 2015, 01:28:19 PM »
|
|
|
This is not a good way to get bosch experience for a noob.
Personally, what he's achieved so far is well beyond that of a noob, and his understanding of terms and what he's doing seems to be a little underestimated. I really think for what he's done so far deserves a pat on the back; and I don't think he's "that" nooby. It's been said time and time again that this isn't the right method for doing what he's doing, we get that. I'd try and rescale the injectors and MAF, but I think you're going to hit a limit, so while I'm all in for trying, I don't think it's gonna work. Try it first and see how you get on, and document what you're doing and how you're finding it. If it doesn't work, stop spending time on it, hand it back and get on with more important things in your shop.  |
|
|
|
|
Logged
|
|
|
|
|
vwaudiguy
|
|
« Reply #51 on: November 04, 2015, 01:52:32 PM »
|
|
|
Personally, what he's achieved so far is well beyond that of a noob, and his understanding of terms and what he's doing seems to be a little underestimated. I really think for what he's done so far deserves a pat on the back; and I don't think he's "that" nooby. It's been said time and time again that this isn't the right method for doing what he's doing, we get that. I'd try and rescale the injectors and MAF, but I think you're going to hit a limit, so while I'm all in for trying, I don't think it's gonna work. Try it first and see how you get on, and document what you're doing and how you're finding it. If it doesn't work, stop spending time on it, hand it back and get on with more important things in your shop. +1 |
|
|
|
|
Logged
|
"If you have a chinese turbo, that you are worried is going to blow up when you floor it, then LOL."
|
|
|
hackish
Full Member
Karma: +1/-1
Offline
Posts: 56
|
|
« Reply #52 on: November 05, 2015, 01:20:44 AM »
|
|
|
Personally, what he's achieved so far is well beyond that of a noob, and his understanding of terms and what he's doing seems to be a little underestimated.
I really think for what he's done so far deserves a pat on the back; and I don't think he's "that" nooby. It's been said time and time again that this isn't the right method for doing what he's doing, we get that. I don't like to go on a site and talk about what I know and what I've done in the past. It is far from my first time at the rodeo and I've worked on ECUs and tuning professionally for over 15 years. I've written firmware for standalone computers and I've worked on OEM software so I have a pretty good idea how most strategies work and I've seen Bosch is pretty similar to siemens which I know a thing or two about. Back on topic. I tried following the instructions for rescaling the MAF. Google told me that the stock size was 65mm and that the VR6 was 73mm. So I tried a 26% gain. I also hacked up a real VR6 file but was told that copying the MAF table from that would not work. I wasn't very successful at finding the stock size injectors to scale the injector size so after a few tries I managed to get about 10%. The load is now maxing out at 70% so it's going in the right direction. The car idles with a -15% o2 correction but driving it's more like -2% so I think I will go find a file from a car that had the green giants from factory and steal the latency table. For the MAF I'm just going to scale it another 10% to expand the range of the load value. The driveability has gotten a bit better too. Given that this thing is probably going to add only 30% more power I don't think a maximum of 100-110% load is going to have any significant downsides in driveability. I also invested some time and wrote a tool to decode my collection of A2L/.hex files and find a file that is better matched so in a few hours I managed to buid a competent mappack. I have had no success with me7logger or visualme7logger. I've been logging the blocks using eurodyne. I also hacked up a perl script to process the logs. The output looks a bit like this:
./stft.pl *.csv
Processing 1445815696.csv
Processing 1445821450.csv
0.00 500.00 750.00 1000.00 1250.00 1500.00 1750.00 2000.00 2250.00 2500.00 3000.00 3500.00 4000.00 5000.00 5500.00 6000.00 7000.00
0.00 ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------
32.00 ------ ------ 5.08 6.58 7.75 1.90 7.91 8.24 5.64 5.88 3.70 -3.71 -3.60 -1.17 ------ ------ ------
48.00 ------ -11.59 -0.99 1.34 1.66 14.73 -0.72 -0.98 1.16 3.35 0.43 1.55 -0.78 ------ ------ ------ ------
64.00 ------ ------ 0.49 2.68 -1.70 6.88 1.60 -3.84 -0.97 2.41 0.09 2.32 -0.59 ------ ------ ------ ------
80.00 ------ ------ -0.48 1.35 -2.34 -0.40 3.50 0.09 -0.66 -0.17 1.51 2.70 -0.39 ------ ------ ------ ------
96.00 ------ ------ -1.51 0.29 -0.64 -0.61 2.95 -1.57 0.92 -0.06 0.31 -0.29 ------ ------ ------ ------ ------
128.00 ------ ------ 0.20 -0.72 -0.69 -1.58 -2.42 -2.86 2.37 0.72 6.79 8.12 ------ ------ ------ ------ ------
160.00 ------ ------ -2.93 -0.78 -1.19 -2.28 -3.91 0.08 6.02 -0.18 8.74 14.71 ------ ------ ------ ------ ------
192.00 ------ ------ ------ 12.89 -2.34 -3.16 8.05 6.28 5.00 -0.09 -2.00 ------ ------ ------ ------ ------ ------
224.00 ------ ------ 13.28 ------ ------ -1.95 1.09 3.20 1.80 3.61 -2.93 ------ ------ ------ ------ ------ ------
256.00 ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------
288.00 ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------
Excuse the fact this example is not actually from a ME7 but you get the idea. |
|
|
|
|
Logged
|
|
|
|
|
ddillenger
|
|
« Reply #53 on: November 05, 2015, 01:32:04 AM »
|
|
|
2.0989 1.3202 0.9174 0.6694 0.4987
8, 10, 12, 14, 16 volts. I just rescaled the axis with these values (from 6, 8, 10, 12, 16) for obvious reasons.
Make sure when you scale MLHFM you take into account the offset (200) that must be subracted from it. This will leave you with negative numbers, representing reverse flow areas.
Realistically, the stock MAF and housing can be used, and will drastically cut down on the amount of work you have to do here.
|
|
|
|
|
Logged
|
Please, ask all questions on the forums! Doing so will ensure the next person with the same issue gets the opportunity to learn from your experience!
Email/Google chat:
DDillenger84(at)gmail(dot)com
Email>PM
|
|
|
wannabee900
Jr. Member
Karma: +0/-0
Offline
Posts: 42
|
|
« Reply #54 on: November 05, 2015, 09:25:13 AM »
|
|
|
Does the stft perl script average values for every load/rpm from a long log file to present that table? If so, would you mind sharing it?
|
|
|
|
|
Logged
|
|
|
|
hackish
Full Member
Karma: +1/-1
Offline
Posts: 56
|
|
« Reply #55 on: November 06, 2015, 11:04:33 AM »
|
|
|
The perl script does average the values. It is a little simpler than the tuning software I designed since it doesn't know anything about standard deviation but it is effective. It also has some stuff in there to exclude ignition values when the engine is knocking and so on. My intent is to share the script once I've cleaned it up and made it easier to configure. It does work on a unix machine so I assume that will include OSX and I've tested it on windows too.
|
|
|
|
|
Logged
|
|
|
|
|
nyet
|
|
« Reply #56 on: November 06, 2015, 11:12:18 AM »
|
|
|
I also hacked up a perl script to process the logs. The output looks a bit like this:
I LOVE IT BTW you might want to exclude accell/decel conditions as well, not sure how much filtering it does, but it should throw out samples where rpm or load is changing... or, even better, generate accel/decel graphs separately! |
|
|
|
« Last Edit: November 06, 2015, 11:14:30 AM by nyet »
|
Logged
|
|
|
|
hackish
Full Member
Karma: +1/-1
Offline
Posts: 56
|
|
« Reply #57 on: November 08, 2015, 10:32:06 AM »
|
|
|
During accel and decel the vehicle shouldn't be in closed loop so it will by default deal with that. The main intent was to make fuel and ignition trimming easier. If I decide that more is needed then I'll just write a module and use real tuning software. This is quick and dirty and does 99% of what I need.
|
|
|
|
|
Logged
|
|
|
|
|
nyet
|
|
« Reply #58 on: November 09, 2015, 10:36:24 AM »
|
|
|
During accel and decel the vehicle shouldn't be in closed loop
not talking about total throttle off or WOT, there are partial throttle cases where you are still closed loop (even narrow band) |
|
|
|
|
Logged
|
|
|
|
hackish
Full Member
Karma: +1/-1
Offline
Posts: 56
|
|
« Reply #59 on: November 10, 2015, 07:38:52 AM »
|
|
|
I'd have to dig through the thousands of pages of EPA standards but I seem to remember reading that in decel fuel cut the vehicle should go into open loop. In fact I think there is even a status bit for it in the mode 1 requests.
For accel enrichment I believe it should be the same deal but not 100% certain. I haven't found a big need for this because accel enrichments usually involve a stab of the throttle and a corresponding jump in load so the settling time passes before the values are used. In any event the script is quick and dirty but quite effective. Any chance you could try to log such a thing to verify? I have not had any luck using me7logger on this ecu.
|
|
|
|
|
Logged
|
|
|
|
|
