AA80E 8-speed

[Jim Blackwood]

Just some pictures to liven things up a little:

[Bernard Fife]

Jim,

Nice pictures!

here's what I think I need:

1) PT6 and AD4 ports switched. AD4 is the input for Paddle Upshift from VR2. PT6 drives solenoid SR:Shift Trigger through EGT2 and an off board driver. If AD4 could work as an output and PT6 could work as an input I suppose maybe I could try that, but I think we already looked at it.

2) Configurable trigger pulse controls for PT6 to set delay and pulse width. I think that should be enough to start testing.

BUT. The existence of a shift trigger suggests that the shifting logic may require the big solenoids to be preset to the newly selected gear just prior to triggering the shift. I do not know if this logic is shared with any of the smaller Aisin gearboxes, however a GM press release mentions a "Shift Trigger" in connection with the new GM 8 speed.

for the delay, I assume you are referencing this:

what needs to be done to allow the output of up to a 250ms pulse on PT6? From my notes, this drives solenoid SR which is used as a shift "Trigger". From the factory literature, a fault code is generated if this output stays either high or low, putting the controller into a limp mode. We also see from the truth table that this solenoid is used during every shift. Therefore logic dictates that it must be a pulse. The transmission is capable mechanically of shifting in as little as 50ms so the window for the pulse should be something in that vicinity, though possibly the required pulse is much less. However it is unlikely that another shift could be triggered in much under 1/4 second and if a longer pulse works hydraulically it could possibly improve the reliability of the shifts. I would prefer to start out with a wider range and decrease it once we know what works if that's OK with you.

Thanks for that summary, it's been a while and it's great to have it clearly and concisely written down. I am slowly getting my test bench back together (more slowly than I would have imagined!) so I will get working on this soon. I will likely have more questions about the trigger pulse once I look into the code required to implement it (it *should* be straightforward). In fact, in theory it could be done by the spare port logic as a 'clutch' output.

[Jim Blackwood]

Thanks Barney, your assumptions are correct, and it was my intention to periodically review and restate. Good to know that effort has been beneficial.

I hope to finish up the last of the fuel lines tomorrow, at which point I think the car should be ready for testing at any time, so once life let's you get back to it we can rock on.

You couldn't possibly know how much I look forward to seeing what this car can do.

Jim

[Bernard Fife]

Jim,

We have found something in the code wrt PT6/AD4 that wasn't correctly implemented, so this particular issue is fixed in the attached 5.290 code:

Monitor_5290.abs.s19

(92.02 KiB) Downloaded 453 times

GPIO_MShift_5290.ini

(314.3 KiB) Downloaded 463 times

This code appears to load and work at first glance, but it hasn't been properly tested and there are likely more issues to be found. We post it in case you want to do some testing yourself (on a bench) as it has added diagnostic capabilities - so a log would be very useful if combined with the MSQ used to generate the log.

[Jim Blackwood]

Thanks. I don't have a bench to test on but I do have the car on a lift. Still need to install or jumper the inertia switch before I can check the fuel system for leaks and fire it up, but that's next on my to-do list.

However, I don't need the fuel pump to apply power and make sure nothing is getting hot, which should allow me to move on to the CAN bus configuration. So I'll go ahead with that and let you know as soon as I have some results.

Jim

[Jim Blackwood]

It'll be a few more days before I can check operation. Turns out I had the flyback diodes on my two off board TIP120 drivers installed backwards and had to order some more diodes. I'm going with the 1N4744 15v zeners this time. Don't know what I was thinking, must have forgotten they were zeners or something. But that had something to do with them getting hot. I was pulling 2 amps through the one on the shift trigger driver with the transmission unplugged. A bit much for a 1 watt diode, but better it than the driver. But it raises a question in my mind as to why the shift trigger driver was conducting in the first place, since I'm working from the premise that it will output a pulse when the time comes for the transmission to change gears. That being the case, it should not be conducting upon application of power with the selector in Park.

In the meantime, I should be able to do some testing with the engine off, key on.

Jim

[Jim Blackwood]

I removed the diode from the off board driver, and then because there is no flyback protection, unplugged the transmission. Powered up and connected TS and started a datalog to record initial conditions. Two problems were apparent. One of the paddle shifter inputs does not do anything, and the driver for the trigger solenoid is always on. This seems to go back to the problem of the inputs/outputs for PT6 and AD4 needing to be switched in software but perhaps you can find something I am overlooking.

I'm still learning to use the datalog viewer so bear with me, it seems I need a Y axis input from the Megashift CAN bus. So I will work on the CAN settings while I am waiting for the new diodes. I should have had that done by now anyway.

I have attached the last msq and a short datalog. Don't know if you can view it without some sort of Y axis from m-squirt but here it is.

You may notice erratic input signals on the paddle shifters, with sometimes both high at the same time and multiple shifting. I'm having issues with the touch sensor board and have found a different touch sensor that I want to try instead, as I have gotten no support in modifying this one to work for my application. But until that is sorted out I'm stuck with this one, which has become overly sensitive and has cross interference between the two. The new ones will be separate and tunable.

Jim

[Jim Blackwood]

A few more photos. I considered not mentioning my mistake on the flyback diodes but if someone went by the photos on page 11 they'd be wrong. So here are some new ones. Also, some shots of the tranny in the car which show the connections on both sides, a bit of the shifter detail if you look closely. (Or not, limit 3 photos) I think you might be able to see the edge of the big rubber drive donut this tranny uses also. But the main thing is how much room there is around the transmission. Plenty of clearance everywhere you look, but not excessive anywhere. It's a big bulky transmission but packaging is quite good.

As usual, I'm still pushing the boundaries of possible errors in doing a conversion. This time with the CAN bus. Pretty weird results, maybe someone has seen something similar.

So I went into project properties and followed the video and instructions pretty well I thought, for a simple 2 device connection, the MShift and the MS-II, recycled power on everything and that's when it all went to h3ll. I got this selector box on the top right that said CAN1 and AA80E as the choices and TS apparently was switching between the two at like maybe 1/2 second on each one, and one appeared to be connected to the MShift and the other was offline. Just flashing back and forth like that. (at the time, diodes were out and transmission was disconnected)

What's up with that?

I tried loading an earlier msq from before the CAN changes and now I've lost communication with the MShift. Bugger me.

Anybody got any ideas what to do?

Jim

[Rukavina]

Not for nothing I had a problem with can bus cycling and it was a discrepancy with 2 different places to enable can coms. Look back thru my posts and you will see it. Not sure if that's your solution but it drove me crazy for a bit

[Rukavina]

May 2016. Tunerstudio forum