Its not at all, but its accepted because we are screwed for other options guess the venors have us right where they want us
Its not at all, but its accepted because we are screwed for other options guess the venors have us right where they want us
I am unsure how 5Digits is planning to do this. One idea is to use a fast modern processor with a 68XXXXXX emulator on board. The main processor would collect data from the outside world and could indeed adapt fuzzy logic(for example) for self calibration like 1BADVAN states the result will be sent into the input of the 68XXXXX emulator and the output handed back to the main processor which will then send out signals to the injector drivers etc. Another would to create a 68XXXXX soft processor in an FPGA but would work similar to the above. This could be a small daughter board that plugs into the existing processor slot. Or use two daughter board one into the EPROM slot and the other into the processor slot.
Just a hacker's thoughts
Regards,
Miles
DD '87 Sundance T1, SLH with rear disks
'87 CSX #432 2.5 CB TII, SLH
I am definitely keeping an eye on this thread. I'm interested in the software, but I am also interested in seeing where this goes with the vendors. I am looking to spend some $$$ this winter on the new engine/trans combo going in the Scamp. How this plays out may seriously affect who I buy parts from...
89 Voyager LE, 2.5T2 - rest in peace
87 Charger Shelby T2 (2.4 conversion in process)
this is what i would imagine also.
5digits - before this dies let me ask a question about the code. is there anything special about the 11mS loop? was that just to ensure that the injectors could be fired fast enough to support up to 5500+ rpm? there are a couple places where the code loops to fill time if needed. also the code uses various timers for calculations. if would seem that we could upgrade to a 68hc12 chip, e.g., and benefit from it's capabilities/speed while maintaining code compatability - my concern would be what the increased processing speed would do to those calculations. maybe it wouldn't be an issue?
89 Voyager LE, 2.5T2 - rest in peace
87 Charger Shelby T2 (2.4 conversion in process)
well, I'm staying tuned just incase, I just found this thread myself...
but I am a little confused, I understand that "this" "should" be "given" to the community, and cals that have been created using it could be "shared" for free, but I am more than willing to pay (within reason) for a custom cal for someone to create for myself (within a "reasonable cost, and once I get 1 of my 7 turbo dodges back on the road), especially with boostbutton about to take a vacation (forever or?). I cant imagine someone doing cals for free unless it was a friendly gesture.
I hope something comes of this though, I still have a passion for these cars, although it has dwindled with all the many troubles ive gone through on multiple fronts (stolen cars and parts, broken cars and parts, personal issues).
and no disrespect from me at all, im very greatfull of the work you have done for these vehicles and this community...
Last edited by OmniLuvr; 11-16-2016 at 08:52 PM.
I don't quite understand why you're looking for an oath from the vendors that they won't use this stuff. If it's available free and public, they'd have no incentive to.
I don't want to be THAT guy, but it kind of sounds like you don't want to devote the time for this project and looked for an excuse not to. None of us should expect you to put in all that time on a project for a car community you're getting out of. It is however just a little mean to dangle that carrot and then take it away while blaming someone else.
Very true.
This 'code' has evolved and is significantly different.
Some tables no longer exist, others added, along with the calibration area being completely different.
Most importantly, there significant change to the software with regards to managing spark, fuel, boost, diagnostics and other areas.
It corrects some short-comings within the existing code and benefits from some enhancements throughout.
In time, it'll work out.
The greatest benefit going to which ECU platform? SMEC, SBEC, SBECII?
If it all even happens can us SMEC cars have some of those improvements you mention?
I know from messing with the flashable smec and the turbonator cals using MPTune, it looks like there is alot in there that could potentially be cleaned out. But, its not something that will likely get done without you.
Wayne H.
'91 Dodge Spirit ES 2.5L turbo 5spd
'05 PT GT 2.4T HO autostick (RIP)
'89 Plymouth Acclaim 2.5L turbo auto, "Slugmobile" yes, THE Slugmobile!
'89 Dodge Caravan SE 2.5L turbo auto, "Mean Mini" yes, Gus' Mean Mini! (Current best 11.699 @ 114.43 mph! - Oct 15th, 2022 Cecil County Dragway, MD)
MeanMini dragracing videos: https://www.youtube.com/playlist?lis...URZLB1RxGYF6vw
and other cars, trucks and motorcycles
https://www.youtube.com/user/SlugmobileMeanMini
5DIGITS
Going away present
Post time:
11-14-2016, 06:39 AM
glhs727
Last Activity:
11-03-2016 09:06
http://www.turbo-mopar.com/forums/me...hp?148-glhs727
Anyone email or call her yet?
What do you mean? I use LM and Turbonator and my car seems to do fuel and spark calculation just fine at 6700rpm.First - I tested the production code for the 1987 and then the 1990.
The 1987 no longer properly calculated fuel/spark beyond 5400 - which means anything done to the spark and fuel curves beyond that RPM are never realized.... 1985 and 1986 are worse (4800rpm) - so lets move on.
Harry,
5DIGITS mentions
The JSR allow the code to immediately run an important task and then return to the main loop.This is how the engine is able to run faster than the loop time in the real world.Please keep in mind, these tests are being performed on 'production' level code WITHOUT any JSR (jump to sub-routine) interruptions.
Last edited by chromguy; 11-17-2016 at 09:18 AM.
Regards,
Miles
DD '87 Sundance T1, SLH with rear disks
'87 CSX #432 2.5 CB TII, SLH
OK I will try to explain. The code does tons of calculations/tasks so if we run all of these one by one, it takes 22 or 11msec(loop time) to run and the code. After complete it starts back at the top
If we have a SMEC ECU its loop time is 11ms, so if left as is no decision can happen faster than 11mS. For example, let's say we want spark to be accurate to within one degree. So we would have 360degrees X 11ms or 4 second per Revolution. Well that will not work!
Some sensors don't need to be read as often such as Air temp, coolant temp, Map etc as they are not going to change significantly over 11msec or so. For example, spark control, injector firing and Hall effect sensor are tres important so we interrupt the main program to work on these important tasks. This is the interrupt JSR
Pseudo code
measure coolant temp
other tasks
more tasks
If pin X on microprocessor changes to 5V (HIGH) run sub rountine "priorty"
lots more code
End repeat loop
The above would be the main loop. For example, when the Reference pulse on the HEP starts, this transition is sensed on the micro on a special pin of the chip (called interrupt pin) this will cause the micro to jump to execute the sub routine code shown below, When complete it returns to the main loop. Note the interrupt JSR will occur at any time in the main loop. I placed it above as a place holder.
here is our sub routine
sub routine "priority"
run tasks to fire spark
end sub routine and return to main loop.
I hope this helps a wee bit
Last edited by chromguy; 11-18-2016 at 03:22 PM.
Regards,
Miles
DD '87 Sundance T1, SLH with rear disks
'87 CSX #432 2.5 CB TII, SLH
Not problem and thanks for posting this.
Think of a 'JSR' as an interruption.
As an analogy, you're in the middle of a conversation and someone taps you on the shoulder and interrupts your conversation and train of thought.
You begin a second discussion while forgetting where you left off on the first one, so it never gets completed.
This is what a 'JSR' (jump to subroutine) does while the program is running.
It jumps to another area of software skipping everything currently being processed, only to return on the next program loop.
The program loop (like reading text - left to right and top to bottom) cannot go back in time and only returns to its original location when it completes the next cycle and comes back around on the subsequent pass through the code.
Do this enough times and the loop time (the time needed to process the software and data one time) becomes longer and longer.
How does this impact processing, engine function and why does it matter?
Engine software is 'interrupt' driven... this means that the engine RPM dictates when critical parameters need to be calculated.
If the code is in the middle of determining whether the cruise control speed needs to be updated and an 'interrupt' occurs because the next engine revolution has occurred, it stops what it's doing and starts over to calculate information considered more critical - like fuel and spark.
BUT....
If those routines (sections of software) now include interruptions, the potential for their calculations to be completed becomes compromised.
This results in fuel, spark, boost and other routines not being completed and no updates occur as a result.
In english, what's calibrated for fuel and spark delivery at high RPM is not updated beyond a given RPM - calibration accuracy has now been jeopardized.
A great example is this...
At 6000 RPM, a 4cylinder sequential fuel injection engine has only 20ms (milliseconds) available between injector pulses.
IF the injector requires more than 20ms at or above that RPM, the injector is considered now FULL ON because the pulse was not completed before the next pulse is initiated.
I hope this helped and by all means ask if further information is desired.
There is a science to getting this right and it's not easy from a programming stand-point.
Another post mentioned 'dangling a carrot and using the vendors cooperation as an excuse'.
Maybe a better way to express what I'm attempting to offer for free is as follows:
If someone dedicated numerous years to develop software that further refined calibration accuracy and was willing to offer it for free, anyone seeking profit for what they had no part in developing seems a bit out of place. If there was/is no interest in profiting from these efforts, we'd already be well on our way to something new. Until then, the carrot and numerous other vegetables exist but whether it's realized is out of my hands and decided within the community and the intentions of those involved.
Thanks for this post...
Although fuel and spark are being delivered, they are no longer being updated as a function of their calibrated values.
This means, their preceding calibrated values are being used based on when the software became exhausted - what was calibrated for fuel and spark at 5400-5800 is simply used and stuck in time, beyond those given RPM's.
the words ,just a secondary routine, come to mind lol
in a mass production 7/70 , emmissions , fuel regulated pre determined power level application , make as much power and go as fast as possiable are kinda likely to be a "good enough " sub routine , go to program .. I'd think
not knowin' nuttin ...but a sub routine that's, good enough , when putting back up the return road would be good if the primary intent was go as fast as you can - you could devote more computer to going as fast as you can
Well, I for one hope that this development continues because it seems to me that being able to make power ABOVE ~5400 rpm is one major thing this community rarely succeeds at. Not an issue for most builds but certainly the main reason you don't see 'big numbers' coming from our turbo mopars very often.
I don't want to derail the thread but i also have to wonder how much timing variance is introduced by the timing belt's movement and the I-shaft>Distributor gear interface at those rpms. It begs the question of what degree of accuracy is even possible without going to a crank trigger setup.
Dont push the red button.You hear me?
Or adding another pulley on the right side? ...to stop the flapping of the belt.but i also have to wonder how much timing variance is introduced by the timing belt's movement and the I-shaft>Distributor gear interface at those rpms. It begs the question of what degree of accuracy is even possible without going to a crank trigger setup.
Great stuff for great cars! Poly engine mounts and bushings at: http://www.polybushings.com
Oh? There are several options from buying canned tunes or tuning yourself (with more than a few tuning packages which is a godsend, try edits in HEX) to affordable fuel injection packages ranging from Megasquirt, FAST, Holley, etc.
Much better than the old days when you had to dink around with stock cals with diodes / cutout raisers / pressure regulators / extra injectors / rising rate regulators / add on modules and injectors, etc.
I can understand 5digits reluctance... but i'd not throw out the statement "vendors have us right where they want us" which does a disservice to both the vendors and those who freely contribute to the shelby-dodge community.
Gary Donovan
Working on clearing the decks.