Running lean at higher boost (was not last week)
1 Attachment(s)
Hello,
I have a MSPNP9093 (with the cap upgrade) and 550 injectors. I loaded the HiRes 10g code last week and everything seemed fine (idles well). I loaded my prior base maps and settings and did some short street tuning runs and everything felt good and looked fine on the MLV, at least to me (and I ran the VE analyzer to do some minor tweeks). I drove the car about 60 miles and then did another logged run and noticed what felt like hiccups at boost. The log shows that I am now going lean at higher boost where last week everything seemed fine. I did multiple runs today and yesterday and see the same thing (say around 155kpa and above). I am new to the MS platform and cannot figure out what might have changed between last week and today (I hope it is something simple). I truncated two sample logs that show the difference, if anyone cares to take a look I would be curious to hear what you think? (the car has the base spark map for the MSPNP9093 except around idle and the VE tables are only street tuned using the MLV; dyno tuning is supposed to be this week but I would like to know if something is wrong before I head to the dyno). The first log is what I call the "normal" one and the second is the "lean" one (the "lean" one is a little long but if you jump to the end you will see what I mean). I truncated the logs for ease of viewing but if a longer log helps, just let me know. Thanks. |
the highres code normally leans out a previous good map on 029v.
|
Thanks for responding.
In this case both logs are taken with the hi res code. (I assume there is not a settling period or anything along those lines) |
clueless. way lean now. fuel pump failing?
|
There is no settling period. If you're running leaner then something must be difference. I can't look at your logs now but are your IATs substantially higher? Could a weak fuel pump cause this as well, I don't see why it wouldn't cause issues if it can't keep up with your fueling needs due to age/wear.
Something is different for it to go lean and unless you're trying to tune while the engine temp is changing then there shouldn't be problems. It's possible that you're getting into cells that have not been tuned yet. Look at your fuel table and do some hand smoothing to fix any large jumps that might be present. If you don't hit a cell in a datalog then MLV can't tune for it. |
The cells that are now going lean were sampled in the prior runs and looked good (using hires code). I will take a look at other variables that may be different but nothing jumped out at me (the temperatures were similar for the bulk of each run). The logs look like night and day to me.
I have records on this car going back to 1993 when I bought it (it is a a 91) I'll check to see when/if the fuel pump was changed (I think when the turbo went in the pump was swapped). Thanks again and if there is anything else I should check or information I can provide that will help narrow the possibilities of what might be causing the latest logs, please let me know. If more details about the car's set up might help, here are a few more line items: • MSPNP9093 (DIYAutotune) (MapDaddy MAP sensor, IAT kit, HiRes code) • FM1 turbo package • FM updated turbo housing outlet. • FM hybrid downpipe w/ 1.8 hiflow cat • Custom exhaust (dual, magnaflow muffler, custom install). • Racing Beat aluminum flywheel • FM dual Spal fan kit. • Mobile 1 & Redline • RC Eng 550 injectors • Knocksense • FM fuel rail • Innovate LC-1 wideband O2 sensor and blue gauge kit. • AutoMeter SportComp boost gauge. • FM Aluminum Race Radiator. (Samco silicon radiator hoses) • Hose Techniques silicon vacuum hose kit |
1 Attachment(s)
here chad:
|
Thanks for posting that (I just noted the Compare feature). I was going to note that at the end of the "lean" log I am seeing AFR's of around 17 where on the "normal" run it was about 12 (this at about 170kpa). Increasing monotonically with MAT and RPM's (where on the "normal" log the AFR seems to stay relatively flat.
You guys have a much better feel than I do for how the logs should look, I appreciate you taking a look. I will have the car checked out tomorrow hopefully if I can get squeezed in (as I was about to drive 500 miles to get it dyno tuned and if there is an issue, I need to get it resolved). Dave |
As the HiRes code was never released with the MM9093 MSPNP (though it can work just fine, we just didn't release it that way so you'll need to make sure you set it up right) -- do you have it easythermed properly for use with the IAT sensor we provided and the stock CLT sensor? The values we used can be found in the .inc files that are a part of the MM9093 MegaTune package.
Also-- where is your IAT sensor located? This isn't normal behavior, something is either installed or configured out of whack here... let's figure it out before something bad happens! |
Hi Jerry,
I started with the files/installation provided with your 02/12/08 installation .exe. I then just used the process listed here: http://ozmx5.com/viewtopic.php?f=4&t=19 (I did not easytherm it since those settings seemed to be included). Using the 9093 (No AFM 10G files). The sensor is at the TB input (same as in the manual). I then loaded the stock MSPNP9093 .msq and adjusted the settings by hand for my configuration (my prior VE maps, my MAP sensor, my idle settings). I then took some logs and did some adjustments using the VE analyzer. (the adjustments were not large) The car ran GREAT at first (the "normal" log was taken on that same code, >10psi boost with no issue, the car hauled a$$). Is it possible that something got corrupted, is there a way to tell? Dave (I am going to give the car a going over as well and check to see that nothing is now hanging loose) I have longer logs as well if that might be useful. |
Max MAT 215°F?
|
There seem to be communication glitches in the log. I assumed the odd max values were due to that. (I also assumed that was a separate issue but you guys are the experts)
|
1 Attachment(s)
I'd be pretty concerned about what happens at 310 seconds. Don't know if you have a wiring issue, or if it's internal, or what.
|
This is a communications error and not an ECU issue. I've gone over the "spikes" with aseer before, it seems like a minor computer hiccup.
|
Originally Posted by aseer
(Post 252117)
Hi Jerry,
I started with the files/installation provided with your 02/12/08 installation .exe. I then just used the process listed here: http://ozmx5.com/viewtopic.php?f=4&t=19 (I did not easytherm it since those settings seemed to be included). Using the 9093 (No AFM 10G files). The sensor is at the TB input (same as in the manual). I then loaded the stock MSPNP9093 .msq and adjusted the settings by hand for my configuration (my prior VE maps, my MAP sensor, my idle settings). I then took some logs and did some adjustments using the VE analyzer. (the adjustments were not large) The car ran GREAT at first (the "normal" log was taken on that same code, >10psi boost with no issue, the car hauled a$$). Is it possible that something got corrupted, is there a way to tell? Dave (I am going to give the car a going over as well and check to see that nothing is now hanging loose) I have longer logs as well if that might be useful. I haven't taken a look at those files just yet though I can. But even if they're right it's difficult to say if they are all in place right (both the firmware flashed properly and the .inc files placed properly). If the IAT sensor calibration is off at all then you can get wildly different AFR's in the same load cells with different intake temps. It's easy to simulate on the dyno if you've got this going on... do some steady state loading in a particular cell with the hood open and a big fan on and datalog it, that's your low IAT reading. Then close the hood and repeat the same exact test. That's your high IAT reading. If you're seeing wild AFR variation in the same cell (RPM and MAP) with IAT being the only difference then the IAT calibration is likely the issue. Here are the readings we used: IAT Sensor: ; Input Data: Temp, degF Resistance ; 48 7000 ; 87 1930 ; 146 560 CLT Sensor: ; Input Data: Temp, degF Resistance ; -4 16150 ; 104 1150 ; 176 330 You can compare the values to what your easythermed values currently are in your projects \mtcfg\ folder in the matfactor.inc and thermfactor.inc right at the top of the files. That is 'assuming' the .inc files you're using match the calibration in the firmware you loaded. |
Smart thinking, I was going to suggest a change in the AIT corrections, but that wouldn't make sense; an improper resistance curve does.
|
Ok, I have some work to do. (I thought my runs were at similar temperatures, or close to each other but I can check the values you suggested). I did pull the CLT sensor at the back of the engine block and put it in very hot water and then did a side by side with a calibrated temperature probe. Those numbers were very closely matched (what the MSPNP read and the probe).
I'll switch over and start checking the IAT and CLT settings you suggested. (The process for loading the 10G HiRes code that used the 9093 settings seemed very simple but I could have missed something) And again, thank you to all who took a look. (btw, the idle with the HiRes code is great on my car, I really like the improvement) |
there's a chance you could have flashed one file and didn't overwrite the file your CPU reads off. so you'd see the same exact scale on your laptop but the MS would be reading a way different value.
|
Originally Posted by Braineack
(Post 252207)
there's a chance you could have flashed one file and didn't overwrite the file your CPU reads off. so you'd see the same exact scale on your laptop but the MS would be reading a way different value.
http://ozmx5.com/viewtopic.php?f=4&t=13 I will go through the temperature settings and learn EasyTherm (once it finishes installing). |
it's possible you flashed the "with afm" file, which would be the only cause of this. or if sabotuer created the images incorrectly.
|
Originally Posted by Braineack
(Post 252217)
it's possible you flashed the "with afm" file, which would be the only cause of this. or if sabotuer created the images incorrectly.
I don't know how to tell if his image is correct (though I assume that if it was wrong other people would have seen my issue too - are there other folk running Sabotuer's 9093 NoAFM HiRes image?). |
I have been reading and thought I would try EasyTherm on the HiRes 10g fw (just to see if anything changes). Just so I understand
If I am wrong and there is a discussion/faq/manual that covers this feel free to point me to it and I will keep reading. |
well you wouldn't need his images. you'd download the 10g code. plug in the resistance values, and then create your own .s19 image to flash.
|
2 Attachment(s)
Originally Posted by Braineack
(Post 252377)
well you wouldn't need his images. you'd download the 10g code. plug in the resistance values, and then create your own .s19 image to flash.
Is that different than taking KeithG's 10G s19 file and running Easytherm on it and putting it back into a copy of my 9093 directory and flashing? This is the first time I have had to modify the image and I appreciate you helping me get it right. Regards, Dave (before I saw your note I tried the latter method (EasyTherm) and saw a huge difference in performance at boost, but the run is relatively short and today is a bit cooler than yesterday - but it did not run lean at all up to >10psi, I stopped when I realized that I was now running way too rich and needed to tune - I also realized I needed to better understand how the image is generated to be sure I got everything correct) |
1 Attachment(s)
I took another run on the new files and I don't see anything like the prior behavior. It is cooler out but the MAT is not that different. I will go back to the original HiRes 10G files from the other thread and see if the problem comes back but right now I am not running lean at all at high boost.
|
1 Attachment(s)
So I ran today as well with warmer temps (same as when I was first noticed the issue). Ran to higher boost (>11.5 psi or so) and saw no issues (does not run lean and nothing like the issues we were discussing). The only thing I did was to Easytherm the file I had been using (the one posted here for 9093NoAFM https://www.miataturbo.net/forum/showthread.php?t=17901) and then reload everything. Dunno what made the difference (I had reloaded the .msq which did not help before) but I hope the problem does not come back.
Thanks to all for taking a look and commenting. If it holds up for a couple of days I am off to the dyno. |
Thanks for the PM aseer. I was running the same firmware image on my car (JRSC) without the AFM and I didn't appear to have this problem. I was careful when creating those images but there's a chance I have entered the wrong value somewhere.
I think to be safe I'll publish all the values used when running EasyTherm (along with a screenshot) for each version of the firmware, then everyone can check these against their inc files. Cheers |
the problem returns
So I had a week of good running (including a drive from Nor Cal to So Cal for dyno tuning, and back - no issues at all with it going lean). Unfortunately the problem has reoccurred and this time it is intermittent (looks just like before so I won't post the new logs unless someone is curious). Namely, I have runs where I am hitting >10psi with no issues (and the car flies after the tune) and I have a run where the car leans out at high boost just as before and then later that night the car is fine again as well as the morning. I suspect the Easytherm thing was a coincidence (namely I just got lucky ).
Unless anyone has other ideas I am going to pull the dual stage fuel system I have in now and replace it with a single high flow Walbro (stock pump plus the aux pump that came with FM package are what I have now). We'll pressure test first but since the issue is intermittent I am not sure we'll see a drop. Thanks again for all who chimed in (and to DIYAutotune/Matt for taking a look at the logs, those guys are great). |
All times are GMT -4. The time now is 02:47 AM. |
© 2024 MH Sub I, LLC dba Internet Brands