4l80e downshift with no reason

[RICH454MC]

Hello. I currently have a 4l80e setup that is driving me insane. I'll be driving along in 4th gear on the highway and all of the sudden, BAMM. it downshifts to 3rd. if I don't do anything it will stay in third. If I put the shifter in neutral and then back in 4th, it will drop to 1st (sucks when your doing 70 mph) and then 10 seconds later it will jump to 2nd, 3rd, then 4th and then will act normal. I viewed the log and I do notice that just before the shift, the in2adc jumps from 0 to 0.049 v. this shouldn't have any effect because i have my digital voltage threshhold set to 2 volts. All the other values show normal. I have also noticed this happens a lot when i am driving the car hard. Any ideas?

[Bernard Fife]

MSQ and datalog?

[RICH454MC]

sorry. here you go. the downshift happens at 194.9 seconds

2013-06-08_10.10.46.msl

194.9

(508.1 KiB) Downloaded 760 times

CurrentTune.msq

(44.79 KiB) Downloaded 987 times

[Bernard Fife]

Rich,

The problem is caused by the short term load in your datalog. At one point it goes from 28.0 to 6552.6 (65526 in internal units) which looks a LOT like a 'roll-under' (65535 is the maximum that can be put into an 'unsigned int', but if you try to assign a signed int negative value of -9, for example, to an unsigned int, you get 65526, which TunerStudio would report as 6552.6 kPa). The code then averages it down over time, but the table value used is the modulus value (the remainder after dividing by the maximum index for the load bins). The result is odd shift behaviour.

This should be a straightforward fix, but I do need to make sure to accommodate all code variants. So what code (if any) are you using over CAN?

Lance.

[Bernard Fife]

In the mean-time, here is some test code (v4.125a) that *should/might/could possibly* solve the signed/unsigned issue:

Monitor_4125a.abs.s19

Uses 4.125 INI from here: http://www.msgpio.com/manuals/mshift/files/ini/GPIO_MShift_4125.ini

(70.94 KiB) Downloaded 701 times

Lance.

[RICH454MC]

Hi Lance. I am not using can . I have a map sensor to calculate load

[Bernard Fife]

Okay, thanks, that will help narrow down the problem. Lance.

[RICH454MC]

Hi Lance. I was going through my datalog and I can't seem to find where the load value jumps up to 6552.6. Also I found something else that is weird. My rpm is inacurate. It is reading aprox 25% higher then it should. I managed to temporarily fix this by setting the pulses/revolution to 40 instead of 31. the problem now is that the rpm reading freezes at around 200 rpm when in gear at idle. when i take off at the light it starts reading normally. but if i'm in neutral it reads fine.

[Bernard Fife]

RICH454MC,

It's possible that mslvv didn't parse the log correctly, I will have a look.

What is the rpm low relative to? (I.e. how do you know it is that amount low?) Have you experimented with the input mask value? ( http://www.msgpio.com/manuals/mshift/V41tune.html#gn ). I assume this an input shaft sensor (i.e. after the converter), not a real tach signal? The ISS may not be properly set up to go to zero rpm (when the changes were made to the VSS/ISS filtering a little while ago), I will have to have a look (and will do that shortly).

Lance.

[RICH454MC]

I am using the ISS as a tach signal. I Just realized that the input shaft does not turn when stopped in gear. (my bad) Any ways, I am comparing the tach signal to my autometer tach which has been tested to be accurate. I have been playing around with the shifting some more and can now duplicate this gear change glitch very easily. It seams to happen 80% of the time, if I go to full throttle for any more then 5 seconds while going above 50 mph and then let off to slow down.