Crashes when put into drive

[1991blazer]

Hi all,
I have built a megasquirt a few years ago and still run it in my Volkswagen, now to dive into the MegaShift for my Blazer.
Just built my MegaShift and things seem to be working OK until I did a running test.
Application: 1994 4L80E running no-CAN
Problem: As soon as any manual gear is selected the whole Megashift crashes and I lose all communication with the computer. Loading code goes well as well as changing parameters. Input Tach works, Load works. Rechecked all wiring and loaded code a few times. Running the latest code 4135.

Any Ideas?

[Bernard Fife]

Datalog, MSQ?

[1991blazer]

Found my problem. Got to making assumptions about 12V power. I had the brake and 12V switch swapped. So of course every time I hit the brake to shift into gear it shuts off! Took me starting from scratch.

I have another problem though. GP12 circuit for switchB was not reading, found that I put in the diode R30 in backwards. Fixed that and now Amp seal 6 is only outputting 3.77 volts instead of 5V, so in Tunerstudio it reads always low. I may have damaged a trace removing R30, is there a schematic of this circuit somewhere?

[Bernard Fife]

1991blazer,

There are schematics here: http://www.msgpio.com/manuals/gpiohw.htm and more details on the input circuits here: http://www.msgpio.com/manuals/gpi.htm

Lance.

[1991blazer]

Thanks.

Found the schematics and could not figure out what was going on, so i moved the circuit over to EGT2 and everything is the same. Output of 3.7V instead of 5V, if I disconnect the jumper on AD1 it goes to 5V. Is there something wrong with the processor? Or can the software be changed for this one input to switch with a lower high voltage? In tunderstudio I get a voltage of .098V on input SwitchC with the jumper on and with nothing connected to the correct output wire on the ampseal. It just seems odd that I can move the circuit and the jumpers from GP12 to EGT2 and get the same results. Any ideas?

[Bernard Fife]

1991blazer,

The digital voltage threshold level for the ADC pins is a user setting and you can set it however you like. You can read about that here: http://www.msgpio.com/manuals/mshift/V41tune.html#il

For the non-ADC inputs, the voltage levels are:
- less than 1.75 Volts is logic low,
- more than 3.25 Volts is logic high.

It is not impossible that the pin was damaged when the trace was burnt.

I plan to add more option for a few more digital input pins in the next beta code, so that might work for you.

Lance.

[1991blazer]

OK. Now this is getting weird. There are times when I can shift just fine all the way to 4th in auto mode, then I will get nothing but 1st even though target gear is 4.

Here is my MSQ two data logs. One is in the selector in 4th and the other is in 1st only. The target gear really jumps around when in 1st only. Just does not make any sense to me.

running the latest code 4.135

[1991blazer]

Ok, found the problem with the logic sensing the wrong gear, had two gears with the same output. So I think the root of my problems comes down to InputB. When I drive for awhile it will stop shifting into 4th, so I pull over place it into Park and put it into driver again, it it won't go over 2nd and the inputs are frozen on the bottom of TunerStudio. If I pull over stop the engine and unplug the GPIO from the truck and plug it back in everything works great for a period of time. So I turned on the gauge for InputB on the screen and it reads right until 3rd or 4th gear. It is going from 0.00V to 5V back and forth. Here is a datalog of what is happening. This is driving in Drive and no manual input. Watch InputB.

[Bernard Fife]

1991blazer,

I tried your MSQ, the input2 voltages are stable for me (that doesn't rule out code problems as a possible cause, but it certainly makes them less likely).

There can be a number of reasons why an auto shift won't happen, even if the target gear isn't the current gear. This can include rev limiting (so make sure you have allowed enough rpm range for this to happen when shifting to high gear, which can often induce very low after-shift rpms).

If the gear lever input is ambiguous, then the shift definitely won't happen - the code really needs to know what gear the lever is in. When it doesn't know, the code does nothing at all (which is why a shift won't happen). The hardware shouldn't ground the input pins, only the shift lever mechanism should do that (unless there is an error in the build). So fluctuating from 0.00V to 5.00V suggests a hardware problem that you need to find. If this was mine, I would disconnect the pin entirely, and see what side of the connector (GPIO or shift lever) is doing the intermittent grounding. Then I would work backwards from the pin towards the processor (using the schematics) to see where the problem lies.

Lance.

[1991blazer]

Lance,

Thanks for the help.I will trouble shoot a little more. The very last datalog shows the inputB going into a square wave pattern above 45mph or so. Just seems odd that it will work and just unplugging the ECM makes it work until the next shutdown.