Firmware Upgrade broke AFR

[McMark]

Apologies if this is the wrong forum. Wasn't sure where to put this exactly.

I have a MicroSquirt (v3.8 firmware) install on a Mercedes C230 engine that I installed and I noticed that the AFR wasn't working. TunerStudio would sit at 12 with occasional jumps to 17 (nothing in between, ONLY those two values). At first I thought I had screwed something up in the wiring (still could be, I suppose) and blown part of the ECU so I installed another MicroSquirt ECU (v3.78) I had on hand which was known good and it did the same thing. I did upgrade the firmware during this process to 3.8. I think it read fine for awhile, and then started exhibiting the same 12 or 17 reading problem. I put the spare ECU back in the car I took it from (which was working fine) and the AFR still doesn't read right. So I walked away from the problem, thinking it was a wiring issue I didn't understand, but didn't have time to beat my head against a wall.

So today I had a completely different car come in. This project has been running fine (bit of a bucking problem, but that's beside the point) and has been daily driven for quite awhile. This person wanted to upgrade to the newest firmware to make sure it wasn't a firmware issue that was causing his bucking problem. The ECU had 3.78 on it and I flashed it to 3.83. He took it out for a test drive to confirm the settings and came back with a log he took along the way. And lo and behold, this ECU now has the same issue with AFR, even after swapping back to 3.78.

Any ideas on this? Could my wiring harness have some 'bug' that works fine with 3.78 and no with 3.8?

[Bernard Fife]

McMark,

Did you set the correct EGO settings (http://www.megamanual.com/mt38.htm#se) and calibrate the AFR table (http://www.megamanual.com/mt38.htm#oa)? You have to do this every time you load new code, it only takes a few seconds.

Lance.

[McMark]

I reset it in the Tools menu -> Calibrate AFR Table... as well as in the Project Prefs and in Fuel Setup -> EGO control. Just went and tried it again.

Does AFR not read if EGO Control is disabled? (mine is disabled)

[McMark]

I was doing a little testing today to see if I could overcome this issue and one of the things I tried while fiddling around in TunerStudio (2.0.8 ) was to recalibrate my thermistors. Well I put values in the Calibrate Thermistor Tables and wrote the changes, but the readings didn't change on the dashboard. So I tried something out there and put in really odd values (something like 0C=10000ohms, 40C=1000ohms, 80C=100ohms, with a 100 Bias Resistor Value) because this should make the readings on the dashboard change dramatically. Well it didn't change at all. My CLT was reading 76 before the change and 76 after. So I tried different values again, and it still read 76. So somehow I'm not able to write to the ECU. I suspect that AFR will work once I can reset the O2 Sensor in the ECU.

[grippo]

You say the same problem occurred with the 3.78 as with the 3.83, so I don't think it's the new code, and I don't think its wiring because as you say changing the coolant or afr tables doesn't affect the gauge outputs as you would expect. So it seems something is wrong in the TS/ MS interface that is preventing the flash burn from working. We should focus on the coolant or air temperature first, because they don't involve files and are much simpler than the ego sensor calibration.

Do you get any error message from TS when you do the cal ? Is the engine running when you do the cal or is it stopped ? Have you tried repeating the coolant or air temp calibration with the 3.78 code ? (Iknow you did it with ego cal, but coolant/ air is a bit different).

[McMark]

I don't get any error messages. When I change the CLT calibration and write to controller, the progress bar moves along the bottom and then the ECU resets. I'm still on 3.83a.

The engine is off when doing this.

I will have my buddy who upgraded to 3.83 had this problem, and went back to 3.78 try recalibrating the CLT.

[McMark]

Same situation on the upgraded then downgraded 3.78 unit. Clt stays the same dispite values input on calibration screen.

[grippo]

I just tested this on the bench with 3.830 code, new download, and selected 2 of the coolant sensors and each of them resulted in a big change in the coolant temperature after I did a "write to controller".

Have you tried different sensors just to see if you get a change ? Do you get a change if you pull the connector off the sensor ?

[McMark]

The sensor changes when you unplug it and IAT and CLT operate while driving (they go up) in a reasonable fashion. They're just not scaled correctly (i.e. my IAT reads 18 today when it's 70+).

Here's a copy of my wiring diagram in case there's a clue in there. Also, I will end up with a spare 'broken' ECU if you end up being interested in checking it out.

[grippo]

I don't think this is a wiring problem or a code or ecu problem. The burn just isn't taking. The reason I asked about other sensors is that if the cal table you downloaded was the same as you already have in the ecu, then you wouldn't see a change, but if you have made drastic changes to the resistor values and then done another" write to controller" and still see no change, then the coolant table in the ecu flash memory is not getting burned with new data.

If you bring up a user menu, say cranking rpm, and just change it a few rpm, then cycle power on the ecu, does the new value come up ?

The only other thing I can think of right now is to turn on "debug Log comms" under the TS Communications menu, then push the "write to controller" button. When the write is done, close down TS and look at the end of the file "TunerStudioAppDebug.txt" I would delete this file before opening TS as it contains a hug amount of past history and it won't hurt anything to delete it. Then it will start with a clean file. This file is 2 directories up from the directory where your ini files reside.

Another thing you could try is to update to the latest TS version and see what that does.