|
Title: KUMSRL: Calc <> Value in Binary File? Post by: jpurban on December 13, 2016, 05:13:38 PM Out of my 996TT file...
KUMSRL is 0.001432, but the calc'ed value, as described in the FR, suggests it should be 0.001399 (3.606 Liters / 2578 ). This isn't rounding error - the difference is 15% - since 0.001400 is an available value that is much closer. So, I'm guessing it is an intentional deviation from the FR. Has anyone else noticed this difference in their "application" versus the suggested FR value? Perhaps the factory tuners are including manifold volume (Vs) in addition to cylinder displacement (Vh)? Title: Re: KUMSRL: Calc <> Value in Binary File? Post by: nyet on December 13, 2016, 05:29:00 PM Perhaps the factory tuners are including manifold volume (Vs) in addition to cylinder displacement (Vh)? Doubful. It is used to convert load->airmass and back, which is not related to manifold volume. I'd be very wary of those sigfigs, though, the rounding in WinOLS is done in a very bad way; it is fixed point, not floating point. Multiply your scalar by 100 or even 1000 and see if you get 1.400 or 1.432 Title: Re: KUMSRL: Calc <> Value in Binary File? Post by: jpurban on July 06, 2017, 12:15:09 AM Doubful. It is used to convert load->airmass and back, which is not related to manifold volume. I'd be very wary of those sigfigs, though, the rounding in WinOLS is done in a very bad way; it is fixed point, not floating point. Multiply your scalar by 100 or even 1000 and see if you get 1.400 or 1.432 Nyet, you're a wizard. Thanks for the advice... fucking fixed point math. WinOLS rounded the factor from .0000078125 (1/128,000) to 0.000008. 15% rounding error. Ridiculous. Title: Re: KUMSRL: Calc <> Value in Binary File? Post by: nyet on July 06, 2017, 09:55:34 AM Nyet, you're a wizard. Thanks for the advice... fucking fixed point math. WinOLS rounded the factor from .0000078125 (1/128,000) to 0.000008. 15% rounding error. Ridiculous. Yea, it is unconscionable. Truly lazy programming. Title: Re: KUMSRL: Calc <> Value in Binary File? Post by: STEVEPHILP on July 08, 2017, 03:59:26 AM I've just checked mine. Exactly the same. 0.000008
But in checking, I've discovered that my 2401cc Volvo S60 T5 has the exact same KUMSRL as the 2521 S60R. In effect, it's using kumsrl of (2.521/2578) from the factory... What reason would Volvo do this? Title: Re: KUMSRL: Calc <> Value in Binary File? Post by: nyet on July 08, 2017, 09:20:04 AM I've just checked mine. Exactly the same. 0.000008 Please read the post above about sigfigs. Multiply the scalar of the map by 1000 - the result won't be 0.008 WinOLS rounded the factor from .0000078125 (1/128,000) to 0.000008. 15% rounding error. Ridiculous. Read this again until it makes sense to you :P Title: Re: KUMSRL: Calc <> Value in Binary File? Post by: STEVEPHILP on July 09, 2017, 01:35:33 AM Yes NYE I'm aware of that. I understood the first time.
My point reiterated is thus: The corrected KUMSRL value is exactly the same for the 2.521l R engine and the 2.401l T5 engine. So the 2.521/2578 is translated to the T5 file. It's widely known that the T5 engine map is pretty much a detuned/neutered R map. So I'm asking if there might be a particular reason they did this?? Title: Re: KUMSRL: Calc <> Value in Binary File? Post by: nyet on July 09, 2017, 01:39:31 AM Yes NYE I'm aware of that. I understood the first time. My point reiterated is thus: The corrected KUMSRL value is exactly the same for the 2.521l R engine and the 2.401l T5 engine. So the 2.521/2578 is translated to the T5 file. It's widely known that the T5 engine map is pretty much a detuned/neutered R map. So I'm asking if there might be a particular reason they did this?? Ah. I understand. Likely outright laziness. There are enough other MAF->load and load->MAF corrections going on pretty much all over the place that have to be tuned anyway, so changing KUMSRL isn't really a priority (much like KRKTE seems to be all over the map compared to the theoretical flow of a given an injector, and a million other fueling maps have to be tuned anyway, so why futz with getting KRKTE dead bang on theoretical?). Honestly, within 5-10% is actually more than sufficient, IMO. I could be wrong, though. Title: Re: KUMSRL: Calc <> Value in Binary File? Post by: contrast on July 09, 2017, 05:50:50 AM Volvo engineers were just lazy. Its evident in many other modules/maps too.
|