Awesome. Tomarrow I have to go clean up the lab-it's a wreck. John will not be happy....
I also have to figure out a new laptop situation, as I'm not going to continue trying to work with a laptop that can't keep the USB ports going. The FTDI cable fails to start 9 times out of 10 or more, but plugging the cable into an alternate laptop has it working every single time. The issue seems to be with my laptop, not the cable.
So it doesn't seem to matter at all what cable, what computer, which PCM, or if it's in the truck or on the bench harness-I can not talk to ANY PCM anymore.
I don't even know where to start now-it won't talk to any of the four P66 units, nor the 8062 that I pulled out of the truck. NONE of them talk with either the original FTDI adaptor, a modified Prolific adaptor, OR the new shiny CP2102 adaptor I built from Silabs parts. All of them will do 8192 baud, (yes, the prolific will work but it takes a registry entry) and the CP2102 can *actually* talk at 8192...not just "close'nuff", like the FTDI and Prolific parts.
I've used three computers: My desktop, which is what I initially got things going with, my laptop, which worked for a while, but then I thought was dying, and a borrowed laptop. I've used three cables, outlined above. I've used five different PCMs. I've powered the four P66 computers from both the bench supply, and the actual truck harness.
What can I do to get these things talking again???
All the cable/computer/PCM combinations show exactly the same symptoms. I get DIC data coming in, I can see it on the CP2102's Tx/Rx LED indicators and on the scope. I can't *read* the DIC data in anything. All I get is constant checksum errors, strings of C0 or C2 or C4 and bad checksum errors, any time I try to connect.
OK. I've now got Scan94/95 working on one of the PCMs, in the test harness. Tunerpro flat doesn't work, Winflash doesn't work and claims "Invalid response from PCM" when trying to connect to it, even when Scan94/95 works fine.
This is with the same cable that bricked two other pcm's and wouldn't talk to anything else, so I'm hesitant to use it to flash anything.
Here's the odd part: This is a PCM for a 95 Oldsmobile Achieva (N body). Service Number is 16196397, part number is 16218404, BPJK. This one isn't in your list of BINs, Robert, so if I can ever manage to get a read from it I'll post it up here for you.
Some miracle has changed-I got the Winflash software to talk to the PCM and retrieve the bin. 10 times in a row, no less.
I am wondering if it's related to the PCM voltage. I am going to try tunerpro next, and then try various voltages and see if I have a problem somewhere along those lines. Pretty sure I don't, but at this point it's got to be SOMETHING.
Anyway, this is the stock bin for the 95 Achieva 3.1L V6, 16218404, BPJK
OK, so now I have blindly stumbled into a mostly OK tune. It will not idle or return to idle, but that's mostly the incredible blower surge...rummmmp, rummmmp, rummmmp, foot off the pedal and it stalls.
Not a computer on earth that'll suppress a blower surge out of the box.
Maybe there is a bit of idle valve tuning I can do to command a fixed idle valve position?
EDIT:
Thinking out of my ass up there. Blower surge is a lean/rich alternating condition, not an idle control issue...Also, the ridiculously powerful surging rubbed the hood in just the right spot to show me where to clearance it some more.
Anyway, the conditions are currently:
Will not sustain idle without help-needs a tiny amount of throttle to attempt an idle. Without this, will choke out. Too rich.
With small amount of throttle, will idle-the surging lean-rich-lean-rich-lean-rich condition hunts for air, not fuel. If I give it more throttle, and thus more air, even a tiny bit-it evens out an runs acceptably, at 1200-1400RPM. I haven't tried a mode-4 command yet because I'm still on the hunt for the ECU killing issue.
I haven't tried any revs over 2500-3000RPM yet, mainly due to not having the wideband fitted to the collector pipe yet. I also don't have any datalogging going, due to the ECU-killer...so it looks like some educated guesses for the idle issue, and then start hunting the voltage problem that is messing with the datastream. (and replace the dying laptop, it's time I got a new one anyway...this one is only a first-gen "Centrino")
Last edited by Xnke; 01-14-2017 at 02:30 AM.
the IAC position when in IAC-open loop is controlled via a few tables that could be messed with, but they're not in effect with a closed throttle other than cranking, immediately post-cranking and after immediately throttle closure until the closed-loop IAC control kicks in to maintain target idle speed.
it's not necessarily the greatest workaround, but moving the commanded idle speed tables around to get into the 1200 range SHOULD allow you to keep your foot off the throttle long enough and give attention elsewhere until you can get the fueling and spark roughed in. then you should be able to lower the idle to whatever speed you're looking for, or at least play with the IAC control settings to get it to react the way it needs to.
1995 Chevrolet Monte Carlo LS 3100 + 4T60E
Getting the fuel and spark roughed in off-idle is no big thing-it's already very close in the N/A form. (Still haven't got the boost control acutator fitted) Right now, this is just a moderately strangled N/A engine...
But, it's important to tune the bypassed mode of the super because you never know when you'll have a failure and need to rely on the non-boosted engine. Also, the super is bypassed at light load to about 2800RPM, anyway.
Fueling under boost is "rich safe" at the moment...probably Scrooge McDuck rich.
If I can get idle and off-idle to behave, the rest comes very quickly. AND it helps to avoid a lot of pitfalls that come from trying to tune mid-range and high-speed first-namely making timing adjustments and wide swath fueling adjustments that have to be undone once you get back to idle-lots of interlaced tables and all to screw up.
Tomarrow is a free ride to the parts yard, so hopefully I'll pull a few trim panels, some wiring harness bits, and maybe if I'm lucky an evaporator temperature sensor and control box.
I ran into the issue of patched/unpatched code again today-the code I pulled from the truck came up as "indeterminate" again for the Rev3v1.01 patch, even though I've patched it right before I sent it and it claimed it was clean. I wonder if it's not getting mangled "in the box"?
I'm baffled as to how sections of the patches are being undone. do you have a copy of the BIN where it was unset again, I'm curious to see if there is a pattern as to what is getting unset.
the BIN as read out of the PCM should be identical to what is being sent... it's not like the LT1 where some of the BIN area is actually occupied by RAM/registers.
1995 Chevrolet Monte Carlo LS 3100 + 4T60E
I don't have a copy at the moment-it'll happen again I'm sure.
I've ordered another FTDI 5v TTL cable and a new laptop. This insanity should stop soon enough...one way or another.
I pulled the flash chips off the boards to fit sockets on the bricked ECM's, and the flash chips for the E-side are no good anymore. They all have the first bit (address 0x0000) stuck at "0x80", and will not erase or flash to a different value.
I've got more flash chips coming so those ECM's will get repaired. Do you want a socketed ECM for code development, Robert?
And it won't connect again. It's as if it's not even connected to anything-but the ECM hasn't even been turned off since the last successful attempt. The PC hasn't been turned off. The setup hasn't been disturbed at all.
Same C0, or C2, or C4 data being sent over and over and over. Same checksum errors. There's no point in trying to troubleshoot this if there isn't a consistant problem.
Looks like it's time to give up on the FTDI stuff. I pulled out the Silabs CP2102 3.3 volt adaptor and built a 3.3 to 5v level converter for it, and it's now the only thing I'll use...because while it does not work out of the box, once configured, it will generate a *real* 8191 baud signal...not 8192, but far closer than anything else I've found/researched.
The Silabs chips work flawlessly on the bench so far, with both surviving ECM's (still waiting on enough chips to fix the others) and it works with my laptop as well. Bad cable, lots of conflicting issues...but now the new cable, while ugly, does work.
However, Tunerpro is still flickering data like mad-Scan94/95 datalogs are clean and have very little to no noise. Tunerpro logs are even worse now than they were before, but no data errors now. 4-5hz update rate, but so far, random garbage data. I'm going to see if it persists in the truck as well.
Here's a video of what happens when I try to get a datalog. I can flash a bin, read a bin, get data, send data, but I can not get "live data" or any kind of running data.
Not sure why.
https://www.youtube.com/watch?featur...&v=Fl6vuO6qLj4
Robert, I may have some bad news.
Removing the 2 bar patch code lets the engine run, instead of sputter and surge. I pulled the 1 bar map sensor off the truck I drove to the shop, cobbled an adaptor, and hooked it up with a 1 bar sensor. Unpatched the code, copied the stock 3.1L tables into the tune and multiplied by 1.1, so as to eliminate my tuning efforts, and fired the engine up to a steady, smooth idle and nice throttle response. At this point I called it a night and came home.
Clearly, this was all with the supercharger in bypass mode, so no boost. I'll do some more testing when I wake up, I've got a spare 1bar sensor here.
The plan for when I wake up:
-re-fit 1 bar sensor, and start engine with unpatched code. It should run.
-Patch code, and deselect 2/3 bar support, set the scalars to 1 bar settings. Find out if this runs exactly like stock (it should, I would think)
I'll report back with what happens.
Doesn't work with the 1 bar sensor, once patched-but I think that was suspected behavior? I'm still not ready to suspect the patched code, but I'm running out of leads to chase down.
Anyway, I'm going through the wiring harness with a fine-toothed comb, getting every input/output functioning as best I can figure it should, and improving the grounding anywhere possible. I am hoping to get a magic ground system that makes the communication link while running start working...from there it should be trivial to fix whatever problem I'm having. Lots of time crammed up under the dash with meter probes and wiring charts...
here's a datalog from today-Notice what i noticed after going through all the wiring again?
Check out the Ignition Voltage...At 8.5 volts on the bench, comms drop out and the same thing happens-but with a stock bin and 1 bar sensor, the engine snaps up and runs good enough to get the alternator going-and the voltage comes back up, I guess.
Time to look at my supply lines and battery again, I guess!
Posting Permissions
Reply With Quote
Bookmarks