Piet Posted March 7, 2015 Report Share Posted March 7, 2015 It's not ment for the P-part hack. This method gets the TVC duty cycle in advance as close as possible to the value which is needed to get the right load to match with the requested load. The P-part hack doesn't work with requested load, it doesn't work with the load concept at all, wich in my opinion is a bit of a pity Link to comment Share on other sites More sharing options...
Piet Posted March 7, 2015 Report Share Posted March 7, 2015 EDIT: I saw that there is also a MAF Linearization ADX, that is not qualified. If it is the one from >tjwasiak, shouldn't it be vetted first? Ik looks like it is already been removed? Link to comment Share on other sites More sharing options...
razorx Posted March 7, 2015 Report Share Posted March 7, 2015 Please feel free to use or improve the XLSX. Venderbroeck, if he likes it, can add this to the packet. Link to comment Share on other sites More sharing options...
rkam Posted March 7, 2015 Report Share Posted March 7, 2015 Since it didn't sound like bank switching was related to any car driving situtation I connected my scope to do a little testing. My M4.4 seems to be running the 2.bank about 58 times in 700ms. Approx. 82.857Hz. With nothing to do, it takes about 470 microseconds to go through its routines. Running idle KWP communication the running time is close to 800 microseconds. Idle D2 a bit less. 585 microseconds. Not every cycle. Link to comment Share on other sites More sharing options...
lookforjoe Posted March 7, 2015 Report Share Posted March 7, 2015 Ik looks like it is already been removed? There is a Rev5b_Maf_Linearisation.xdf on the Wikia right now, with no provenance ....... Link to comment Share on other sites More sharing options...
Piet Posted March 7, 2015 Report Share Posted March 7, 2015 Since it didn't sound like bank switching was related to any car driving situtation I connected my scope to do a little testing. My M4.4 seems to be running the 2.bank about 58 times in 700ms. Approx. 82.857Hz. With nothing to do, it takes about 470 microseconds to go through its routines. Running idle KWP communication the running time is close to 800 microseconds. Idle D2 a bit less. 585 microseconds. Not every cycle. Interesting, although nearly a 1/2 sec (470 micsec) seems a bit (to) long for a full main loop pass. Where did you measure with your scope how often the controller switches to the upper bank? At P5.7? During a main loop the controller switches at least 80 times to the upper bank and back to the lower bank. And that's without interrupts which do also occur. There are 60-2 teeth on the flywheel for the RPM sensor and if i'am not mistaken an external interrupt is triggered not only on the positive edge but also on the negative edge of the RPM sensor signal. At least ....that is the case with earlier motronics, i' am not quite sure if it is the case with M4 as well. But there will be at least 60-2 external triggers per crankrotation (maybe even 120-4) each causing a bankswitch when the proceesor is in the other bank as where the interrupt routine is at the moment the interrupt occurs. Link to comment Share on other sites More sharing options...
lookforjoe Posted March 7, 2015 Report Share Posted March 7, 2015 Software AC mod for M43 cars with a M44 brain implant: No hardware alterations are required with this, non in the M44 and also non in the M43 car itself. M43 and M44 will stay interchangeable with this. Added a link to this in the M4.4Wikia FAQ section, under AC Link to comment Share on other sites More sharing options...
rkam Posted March 7, 2015 Report Share Posted March 7, 2015 Measured at P5.7. 470 microseconds is nearly 0.5 milliseconds. Pretty fast I would say ;-) Link to comment Share on other sites More sharing options...
Piet Posted March 7, 2015 Report Share Posted March 7, 2015 There is a Rev5b_Maf_Linearisation.xdf on the Wikia right now, with no provenance ....... I'am not very happy about it either, I have to admit. Together witht Maarten we are working on a modification of the bin for its next revision which is a scalable factor with wich a larger MAF can be easily implemented by simply filling in a conversionfactor appropriate for that MAF through the XDF Measured at P5.7. 470 microseconds is nearly 0.5 milliseconds. Pretty fast I would say ;-) Whaa , off course you're right !! Yes, that would add up. Link to comment Share on other sites More sharing options...
Piet Posted March 7, 2015 Report Share Posted March 7, 2015 Since it didn't sound like bank switching was related to any car driving situtation I connected my scope to do a little testing. My M4.4 seems to be running the 2.bank about 58 times in 700ms. Approx. 82.857Hz. With nothing to do, it takes about 470 microseconds to go through its routines. Running idle KWP communication the running time is close to 800 microseconds. Idle D2 a bit less. 585 microseconds. Not every cycle. For clarity.... what exactly do you mean with "nothing to do"? Is that at idle or prestart when the microcontroller is running in it's (much smaller) prestart loop? Link to comment Share on other sites More sharing options...
rkam Posted March 7, 2015 Report Share Posted March 7, 2015 M4.4 on my desk with nothing connected. Link to comment Share on other sites More sharing options...
Piet Posted March 7, 2015 Report Share Posted March 7, 2015 Ah Ok. It's running in its much smaller prestart loop then, which is a whole different story. In the prestart loop there is a repetitive lcall for a routine in the upperbank which is lcalled only about 13 times instead of the repetitive routine in the running loop which is lcalled round 80 times. Also there are no interrupts caused by the RPM sensor in the prestart loop which could cause bankswitching. On the other hand is this prestart loop much smaller then the main running loop with 28 lcalls against about 160 lcalls Link to comment Share on other sites More sharing options...
Piet Posted March 7, 2015 Report Share Posted March 7, 2015 Just curious... what are you planning to do with this information about bankswitching? Just scientific interest or do you you have something in mind? Link to comment Share on other sites More sharing options...
rkam Posted March 7, 2015 Report Share Posted March 7, 2015 I don't have patience to work out how the whole disassembled code works, so I just picked a subject and poked into it. Sometimes I find gold, sometimes nothing. There is no plan behind it. Although I don't see how it could, it would be nice if a method to read the first 32kB magically revealed itself. Previously when I have found something interesting, I have let other people make something useful out of it. (Like M4.4 internal ROM -> Dilemma -> Flashing) Sadly quite a few times this has led to people continuing the work for commercial use, which often never even gets to the market. There are tendencies in topics of this thread also, but hopefully information will get out in the open. Very few people will bother to do the required work to get from proof-of-concept stage to something user-friendly for free. Link to comment Share on other sites More sharing options...
Piet Posted March 7, 2015 Report Share Posted March 7, 2015 Yes,I know You stood at the cradle of this whole adventure. That's exactly why I was curious what was behind your interest in mapswitching. You are right, many times it starts with investing things just out of curiosity and while doing this usefull things just emerge. Concerning keeping mods freeware I think you are right.But some mods are suited to go public right away but with others some care must be taken. Also we have to be carefull that things doesn't become scattered. That's why it's a good thing to have a supervisor who collect all mods to put together for a new version. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.