nAst1: Progress and Concepts Thread

[SappySE107]

There is a throttle follower steps constant that will change how fast the RPMs drop between shifts for manual transmission. Tables Maximum vs rpm and vs mph will control the drivability in gear but at $0DE3 in the stock A1 code there is a manual trans throttle follower.

0DE3 10 16 KISMANOF 16 STEPS THROTTLE FOLLOWER OFFSET FOR A MANUAL VEHICLE, IF THE
VEHICLE IS MOVING.

[Fieroghini]

My constant is is set to 6. A1 is set to 16? I am using the dohc bin. You mentioned driveability in iac vs rpm and mph? What kind of issues are you talking about? Can I just start lowering these to keep the iac transitions smooter? I'm not super concerned with fuel economy. Just going for driveability first and I just switched from auto to manual and this thing is a bitch to shift. The spec 3+ clutch probably doesn't help much either.

[Superdave]

Just a random thought.. how hard would it be to add in something like the VE adder table, but have it act as a PE mode VE adder? or maybe a PE mode RPM vs VE multiplier table?


I finally datalogged mine today to try and fix the surging. What i found is basically while cruising around at 2K RPM, 20% TPS i'm in the 90Kpa range on the VE tables. Of course i need the 90-100Kpa range for WOT use so if i lean them out for cruising it's lean city launching at the track.


Ben, that throttle follower offset also works in auto mode. I adjusted mine this morning down to 4 steps so i didn't have to ride the brakes so hard pulling up to a stop.

[robertisaar]

couple of things i would consider playing with:

F129 table
the KISMP decay stuff



and Dave... you may need to go alpha-N.

i'm not sure how well a more or less PE only VE mod would work.... i can work up some beta code fairly quickly, but i'm just having a hard time visualizing a way to get this to work smoothly. the first thing that really comes to mind would be a table that modifies the VE calculation based off of TPS%. essentially a blend of alpha-N and speed-density.

anyways, PM or post whatever ideas you have, i'll see what i can come up with.

[Superdave]

Alpha-N keeps coming up, i may just have to bite the bullet and do it.

I'm going to do a ton of datalogging at the track this weekend, plus too and from highway logs. I'll post up some of them.

[robertisaar]

Alpha-N keeps coming up, i may just have to bite the bullet and do it.

I'm going to do a ton of datalogging at the track this weekend, plus too and from highway logs. I'll post up some of them.

i do have the features documented and enabled in the current nAst1 release, should you want to play with it. being at 90kPa with only 20% throttle is going to be rough with S-D. the Monte regularly sees 1600RPM, 85kPa at 23% throttle when in 4th with the TCC locked, but that's really the only situation where i'll ever be in that situation. if i go WOT, the converter flashes right past 1600 and sits around 2200, so i definitely can't test it on-road myself.

[Superdave]

Mine seems to flash to around 3600, i'm kinda thinking that i can get away with leaning out the table under 3K just because of that. Needs quite a bit of testing first though.

[robertisaar]

so.... me with things churning in my head. generally not a bad thing, but i tend to make things that aren't complex far more complicated than they need to be...

so i put some brain cells to use and came up with what i think are some reasonable ideas....

we all know about the cars with paddle shifters on them right? well...... had a thought that should at least work with 4T60Es. stick shifter in 4th, mount paddles or some other kind of momentary switches, only need 2 circuits, use it to signal the ECM to force an upshift/downshift. i could probably piggyback most of the code on the existing "manual valvebody" 4T60E code i threw together already, except instead of decoding the PRNDL combination circuits and forcing whatever gear combination comes out, command first gear and then only shift and always shift based on the paddles/switches... Reverse uses the 1st gear shift solenoid combination, so that won't be hampered unless you bump the shifter from 4 to R without downshifting via the paddles first.

i'm THINKING.... that for cars like this, the 2nd gear start switch definitely won't be needed anymore, since you could command a startout in 4th gear if you wanted, so i would setup an option that would be checked during the normal 2nd gear start request routine, if the "paddle shift" option were selected, and the 2nd gear start switch were in the "request" position, to bypass the 2nd gear start code, to do the paddle shift algorithm. if in the non-request position, have the shifter operate normally(no paddle requests). that way you could switch the car on the fly and have it working however you want it to within 1/80th of a second. TCC logic could probably be left alone entirely and have it function however it would normally.



and for the row your own side of the coin..... a similar concept, in that a pair of switches would be used, but wildly different operation.

not sure where to mount them to be effective, but an upshift and downshift anticipate switch to tell the PCM what you're planning on doing. example being on the track, know a downshift is going to happen, tap the downshift button before clutching, ECM calculates the gear you're in based on the N/V ratio, calculates the gear you'll be downshifting into and then waits for the clutch switch to indicate that the pedal is being moved. when it does, it takes control of the IAC to move the engine speed around while you're shifting, so that by the time you let off on the clutch, you won't have to attempt to manually rev match the engine and transmission speeds via throttle. that's the theory for that one anyways.

and for the upshift, same concept. press button, calc gear, when the clutch goes in, move IAC around(assuming the throttle was let off on, otherwise no amount of IAC movement is going to control engine speed at WOT), match engine and trans RPM by the time the clutch is enagaged again. assuming anything other than roughly 0% throttle during a shift, it would be interesting to actually have a dynamic rev limiter that changes based on where the engine speed will be in the next gear...

and with this system, if you never press the buttons, the PCM will never care and it will act like it's non-existant and normal shifting behavior will happen. of course, this can have street benefits as well, not just track, do some slow shifting with no throttle and still rev match on up and down shifts... especially downshifts.

the only problem i'm having a hard time wrapping my head around is a way to signal to the PCM that you want to skip a gear entirely... maybe double tap the up/downshift buttons? hold the buttons for a long enough time? not sure how i would want to try and account for it yet. or how to cancel a shift anticipate? i figure i can throw a few second timer into it and if a shift isn't made in that time to cancel the selection, but how to cancel before the window elapses?







anyways, this idea has been going through my head for a while, figured it would help out in the realm of both drivability and fun, which are always good things to shoot for in any vehicle. i really need to repair the testbench to get this kind of stuff pounded out though... killed a few pots by resoldering them too many times, could only simulate MAP being between 60-300kPa(yes, boost code was being tested ), couldn't get the TPS to go to an idle position, stuff like that. everything else is working though, managed to calibrate a few instrument clusters (bad tachs and speedos) in the process, so the bench has already paid for itself and then some.

[robertisaar]

2 somewhat small but potentially important updates...

new tunerpro version has just been released, a "patch" item is now natively supported in the XDF. this should be quite useful for me, at least.

and some of the smart guys in austrailia with their commodores have implimented a large capacity(entire BINs, not just 4K) NVRAM module into their ECMs in place of a PROM and developed a windows program that can read/write the entire program area of the NVRAM. essentially it allows them to update their tunes over the ALDL connector, some of which are done in realtime with the engine running... i'm currently looking into how difficult this would be to support. it doesn't seem too difficult, since i would be able to use some of the space on the 2K of SRAM that a 9396 provides that i can't possibly fill up otherwise... it would get used for temporary storage of data that gets sent to the NVRAM module and changed in realtime. it also allows for only specific portions of the calibration to be updated, if need be, so it is in every sense of the word, emulation.

the only potential issue is that the NVRAM modules need an initial calibration written to them that has the code in them that allows the calibration to be updated over the ALDL to already be burned in before attempting to install into the ECM. so, minimum cost would be something like the MCUmall Willem units to do an initial program on the module, then just update over ALDL, unless the battery in the NVRAM dies(replace NVRAM) or the calibration gets corrupted(reflash in programmer).



in other news: i'm the world's worst procrastinator..... my testbench is still sitting, half-disassembled, waiting for me to solder in new potentiometers.

[robertisaar]

assembled the bench, only issue was that i connected the terminals for the pot backwards, so now instead of rotating it right to increase MAP, i rotate left...

but now that the diffiuclt portion is dealt with, now i just need another MEMCAL since i loaned out mine. otherwise, knock counts are continuously run up at a rate of 160Hz(if i was watching it correctly), not something that is normally experienced with a real engine.

[robertisaar]

p-p-p-progress????


finally got the motivation to test a few things without a MEMCAL, above 1BAR MAP sensors are now functional, as is the 3 step rev limiter.

notes/quirks: for some reason, the BPW delivered at any kPa under boost using a 3BAR is not the same as the BPW delivered with a true 1BAR in the same situation.... it's usually off by 10-15%, and i'm not 100% certain on why.... i have a few thoughts on where to look first, including for some reason using a different base BIN between my boosted and non-boosted tests, why i did that, i don't know...

3 step is working perfectly, from what i can tell on the bench, just have to be certain to setup the switch open/close options correctly... i think i inserted some tips in each item to help.

anyways, pending a few more tests and assuming i don't find anymore bugs, release will be soon.... couple of hours to a couple of days, depending on how confident i am in it. 3BAR REQUIRES the extra RAM present in a 9396, 3 step does not.

[pocket-rocket]

Very nice update, Robert. I figured out what was confusing me about messing with the tune on my car some time back... Ben always said the fueling was spot on in the 92 so that's why I never saw what I thought I would see scanning it since it was already ideal, even with a stock tune. Hoping once I get this other engine in after installing new gaskets, arp rod bolts and clevite 77 bearings it will be off so I can get some tuning time under my belt.

Sent from my HTC One X using Tapatalk 2

[robertisaar]

oddly, GM did do a few things right with a couple of tunes....



also, figured out my BPW disparity... changed my comparison tunes so that the only thing different between them was that one used a 3BAR and one used a 1BAR and now the BPW issue is resolved. now to disable barometric updates when using a 3BAR..... since that's exactly what GM did with all of it's factory boosted stuff. or at least only allow it under a certain RPM(below the boost threshold).

[Superdave]

the P/N limiter as a 2 step does work quite well, kinda fun on the street too.

[robertisaar]

well, assuming you want to redo the wiring, you can now use 7727 pin D22 (7730 pin D14) for a 2/3 step rev limiter instead of tricking the PCM thinking you're in P/N instead of D. i don't THINK A1 has trans abuse prevention code in it(been a while since i've looked), but if you were to do it that way with DF, it can cut fuel temporarily as the trans "shifts". of course, it could always be tuned out, but if that code were present in A1 and it weren't defined in the XDF, it could cause lots of hair removal in the process of figuring out why it happened.

also: 4 circuit PRNDL decoding has proven to be more problematic than i thought, but i believe i'm on my final fix of it...... most of the issues have been due to GM leaving in legacy support for older stuff, like a/c pressure switches instead of the pressure sensor, stuff like that. i could have done a full PRNDL decode with only 3 circuits, but i'm using all 4 as used with the 4T60E for now.