I loaded up 4138 (have some miles on this version) and did not use the TS saved MSQ load, instead sticking what was in MSShift. Then used TS to connect directly to MSShift, it was communicating nicely. I got my stop light switch working again (as seen on TS). So then I progressed to my main project, which is MS3Pro integrated with MShift. Again, forgoing the TS MSQ version. Again, success, CAN worked nicely, MShift was seeing TPS and RPM via CAN, and Eng Oil Press was going in the other direction to MS3Pro, over CAN. Then I loaded a week-old version of MShift MSQ, and it still worked. I then went out and bought a lottery ticket (joke). This was all done via "Old Faithful" BlueTooth (COM16). Next, I tried RS232 into MS3Pro, via my USB adapter that I use to talk to MSShift (COM1). That acted up like the built-in USB chip set was doing yesterday (COM10). It would connect for 2 secs, then gray screen for 2 secs, and repeat. Like yesterday, I pulled the MSShift fuse, and then the connection to MSPro stablized.
So... Then I re-installed the fuse, cycled power, and for some unknown reason, I tried File -> Recent Project, selecting same project I was in, and BINGO it worked! I immediately switched to the built-in USB cable (COM10), and again it was acting up, until I did the Recent Project trick.
Ok... So can someone smarter than me explain what's going on?
Andrew, give this a try and see if you get the same results.
I know it is functioning...it does make a change, it is just somehow links to the wrong calibration?? That part of it is beyond my abilities.
I was able to get CAN VSS to work. I think I might have something going on with the OSS signal, so I took what is actually my VSS2 (tested to work perfect) and changed it to VSS1. This is a front wheel speed sensor, so its not ideal, but I figured it would do the job for the testing I was trying to do. I had to load some old code where outpc settings allowed for greater than 255, but once I got that going the VSS on MS3 and GPIO didn't match up. They would accelerate together but GPIO was half the speed. No matter what I tried nothing changed, so instead, since my main objective was to test the dropout issue, I lied to MS3 and told it the wheel speed sensor had 24 teeth instead of its actual 48 teeth. This got me an MS3 signal that was 2x as fast, but got the GPIO the correct speed. Finally after all of that, I was able to do some testing again. I drove the truck around for a while, but never did the connection drop out. I did have some odd things with shifting, I think I am having some issues with the -99 on manual gear, but I need to measure the outputs on the range sensor to see if maybe my settings are off some, however, at least with this old code, it would drop out and not have a target gear and require I stop, move the shifter to park or neutral then back to drive to try again. Is there a way to make it less sensitive to small voltage changes? I was adjusting the voltage dead band, but it didn't seem to be changing much...I'll play with it more today. I need to work out my VSS signal issue, then I will likely hook it back up to GPIO as the CAN VSS is pretty laggy.
Andrew and dave. This is exactly whats happens to my MS3Pro and Megashift combo. Version below 413x have no issues with CAN dropout. I'm following Lance instruction to contact MS3 developer once I have a fresh log.
Here's what happens to Megashift when connected to TS using MS3pro USB:
1. Froze after some time. I think this have been cured with a fresh msq.
2. Can communication error and TS turned to grey. I have to restart the car and or plug and unplug usb cable and or restart TS. Same like Dave and Andrew.
3. The last code I tried until now is 4144b. its very stable if connected directly to TS and the worst of all is connected to TS via MS3Pro. I did get a message that a controller is found but appeared to be no firmware. This is when the car keep sticking to first gear and no VSS reading ( VSS connected directly to GPIO ) . I have to restart the car and TS a few times and plug unplug USB cable. This is the last time I connected Megashift to TS via MS3pro. Haven't have a courage to try again. Hehehe.
I think the problem only happens if I connected Megashift to TS via MS3pro CAN. Other wise this combo works like a charm. Have been driving the car for almost 500KMs without any issues.
Here is my thread in megashift forum: viewtopic.php?f=4&t=823.
I just read msextra forums. I haven't made any change since I wrote here last time. My car runs fine with MS3pro and Mshift 4.144b and without TS, until a few days ago it acted strange. It shift to neutral when I select 2nd gear on the lever. This happens without any laptop connected to it. And trying to change the shift output in TS didn't fix anything. So I'm back to 4.126 right now since I need the car for daily use and everything back to normal.
By the way I need some help clearing noise from my VSS signal. VSS sensor is a reed switch with single tooth per output shaft rotation. I have made a debounce circuit from arduino forum which help a lot in stabilizing my vss reading. But GPIO still reads 5,1 kph when the car not moving. I borrowed an osciloscope and found the signal to PT0 is jagged. Can this cause GPIO reads false speed? I need the transmision to shift to 2nd at about 5-6 kph due to low gear ratio.
PS: I experienced late shifting too if I change the top ladder of "2nd" to "1st". It will shift around 19 kph, this is without selecting "enforce rpm check". I think this is because the pointer reads 4 square and makes an average value. Maybe if the firmware can be modified so that GPIO only reads that one square under the pointer instead of 4, the shift point can be more exact. Lance, please correct me if I'm wrong. Thanks
- Shift Table.png (21.93 KiB) Viewed 14265 times
- Current signal To PT0
- IMG_20140531_220933.jpg (248.84 KiB) Viewed 14265 times
- Debounce circuit from Arduino Forum
- debounce2.gif (3.66 KiB) Viewed 14265 times