EcuTek SD problems in the cold
 

I've been running on a tune I made for the past two weeks, and it's been running great so far until about 2 days ago. I run hybrid speed density and at 4000rpm it switches to SD mode. In the past few days it's gotten colder and suddenly this weird issue cropped up. The tune worked fine from 70F temps to 40F temps. It was 20C the other day and as I was driving, not even under major load, as soon as it hit 4000rpm I saw my AFRs shoot to as lean as the sensor would read and the engine would basically lose virtually all power. I've made sure my IAT compensations are set correctly. From a log, I can see the calculated mass air flow suddenly drop off once it goes SD mode, so something's happening with airflow calculation. Perhaps a bug? I'd have to guess something at 0C. Anyone else run into this?

I read about Ecutek having issues with SD in cold climates on here. If you have a read of the @DeliciousTuning thread you will find some info on it.

Link to thread. I couldn't find it quickly but I am fairly sure it is in the last 10 pages.
http://www.ft86club.com/forums/showthread.php?t=52200

Which ROM? Is it set to use the MAF IAT sensor?

The quick flash of power loss bug I've always had, I attributed that to the changeover. We don't have the hybrid blend feature of the COBB SD maps. Not sure why that happens. I've taken a look at the datalog, and I notice something interesting. Once it goes speed density (mine's set at exactly 4k) you can see the calculated mass airflow drop like a rock. The first dip is easiest to see, the mass air drops before the MAP drops as a result of me letting off. It's the little step before I lift up on the gas since I don't want to blow up my engine. I have a theory about what's causing it, and am going to try something to fix it after work.

http://i100.photobucket.com/albums/m...pse8adae82.png

I'm pretty sure we have this one identified.

RR7 based tunes are having this hiccup under 0C intake temps. So once they go negative fuel cut ensues. Not sure exactly why, but what we'll do for our customers is enable a full MAF mode for those times where it is that frigid. For some of the higher powered folks this may mean capping the revs to 6000 rpm or wherever the MAF maxxes out.

Likely we didn't have this issue last winter due to using RR5.

Why not roll back to RR6 or 5?

We just got the fix here today from EcuTek, we will issue the patch in the next couple of days as we have to re-write the custom maps for this version.

Is this something that can be pushed out to retail customers? Those who have the RaceROM upgrade will need this too.

I never quite understood the point of releasing tuning that does not address all temp ranges. Almost unacceptable really.

I'm under the impression it's a software bug that they didn't notice. Judging by the datalogs it looks like anything below 0C rolls over to the top of the scale instead of into the negative scale. Case of programming error. I'll check tonight if there are any RR updates.

I may have come up with a workaround which should work during the interim until we get the official fix, as I don't think the retail customers have received it yet (at least I haven't).

Originally I had planned to mess with the IAT sensor scaling but that'd produce issues with anything else that relied on the IAT/CAT values. I noticed that RaceROM custom maps were able to replace the SD CAT value, so I decided to just add 20 degrees to the SD Temp Compensation table IAT values. Then I created a custom map telling the ECU to lookup the +20 value based on actual CAT. This gets around SD using a negative number for calculation since they're all positive technically. By using a custom map this only modifies what speed density calculates airflow with and leaves anything else that relies on IAT alone.

It's worked for me down to 26F (-3C) so I'm seeing it as a potential fix. My AFRs were where they should be. I'm not sure about what function the SD Calibration Temperature does, I run a JDL kit and it's blow through MAF so IAT is CAT. If anyone else wants to try it out I've pasted the tables in a screenshot.

http://i100.photobucket.com/albums/m...psabdf85d7.png

Granted these numbers will work down until IAT actually drops below -20C or -4F. Scale accordingly if you live somewhere exceptionally cold.

EcuTek SD problems in the cold
 

I guess I should've posted here when these issues first cropped up in October. Unfortunately it was difficult for me to replicate until it got colder as it's been unusually warm this fall and winter. The issue finally cropped back up earlier this week and I was able to get a log. I sent it of @ptuning for Toan to review. He promptly saw what the problem was and got a revised map to me that has rectified the issue. He noticed there was a bug within ECUTek that was popping up when transitioning to SD in frigid temperatures.

"Happy New Year to you!

So, I looked at your logs and I'm thinking that there may be a small SD bug
in your current RaceROM file. ECUTek has since released a new RRFF. Anyhow,
I've updated your tune to use the latest 16325 ECUTek RaceROM Feature File.
Let's go ahead and see if you can replicate the problem."

I'm not 100% what he did to resolve it, I asked but he said he's working with ECUTek and will get back to me once they resolve the issue with their software.

When I first reported it back in October/November and got a log I wasn't logging the MAF parameter needed for him to see what was going on. He thought it was an intercooler coupler that came loose or a faulty ignition coil. I assured him that it only occurred when it was freezing. He asked that I log again with the MAF parameter but of course it warmed up and I was unable to replicate it and the logs looked good so that left me scratching my head as to why it was until this week.

I know why he wanted the MAF Sensor parameter haha, because he probably saw the same dip as I did. I logged MAF parameter and you could see that held constant while the calculated MAF from SD dropped like a rock. I'm pretty certain now it's caused by a software bug transitioning into negative numbers. Incorrectly coded it would roll over to the top as opposed to scaling further down, I would assume. Just took digital logic design this semester :bonk:. So far this little fix works at 20F (-7C) + 20 so I'm thinking it works as long as SD uses a positive number for CAT. It's a workaround until we get an official fix, but it works so I'm okay with it for now.

The latest V7 RRFF available to retail is svn14599 (other than some specific versions that have a later number). Nothing in the 15xxx.

What you did works fine with that custom map I think, the only downside is incorrect fueling calculations for the temperature, but it will be a small difference and certainly words better than a full on fuel cut.

You'll likely be seeing an update to the RR features soon, there were some other minor bugs that were fixed as well. Ideally the RR feature files are not released any more often than necessary (i.e. critical bug fixes) so as to prevent having to redo all the files. You gotta keep in mind that EcuTek primary income comes from tuning and not individual shops, and the last thing we need to see is a constant stream of half useless updates that create more work for little to no reason. Makes much more sense to get a bulk release so we can update the library of maps when there is actually something worth revising for.