Why is it that some people knock off a few degrees just before the torque peak and during?
Safety margin.
If the vehicle is approaching any sort of limit - thermal, octane, what the internals can handle, the owner/operator is obviously dumb as shit or lacks common sense - you pull timing for safety. If the vehicle will be used in an abusive environment like roadracing then you really have to worry about shit like that, drag racing not so much because they are only in the pedal for a short period.
Also, some engines that are biased for low end and midrange powerband, like the traditional domestic small block V8 with a stock-ish cam meant to make 300 ft/lbs right off idle and torque peak at 2900 rpms, or horribly mis-built motors like my D16Y5 Type-R with way too much CR/stock cam/pumpgas, will continue to make power as you add in more timing until well after audible detonation has set in.
You stumble upon a slut pile of RB's JD?
Pretty much. This RB20's been sitting around for a minute because it's a bit of a wiring calamity and I've not had the patience for it. We managed to get rid of the one I was talking about to you yesterday, guy traded it for an 04 or 05 IS300 with a .60 trim T3/T04E on a Greddy piggyback, so of course it doesn't run right and my next paycheck is ensured. However, bossman is trading around for an R33 GTR SpecV and part of the deal is he has to pick up ALL of this guy's Nissan parts so a spare RB25 swap comes with it.
You forgot about the effect engine speed has on this though. If at a given density you have half the time to complete the burn, you need to start sooner( considering ignition delay and burn speed vs RPM) Which is the main reason MBT timing goes up with RPM after torque peak.
No. Burn rate for iso-octane based fuels remains proportional to engine speed, this is plainly stated in several engineering literatures. Some engines like more timing as the RPMS go up (before as well as after torque peak) like SR20 and KA24, but I've had no experience with that happening on a dyno as being anything that happens more often than very rarely. I hear people saying it, but they always seemed to have a dynojet so I just shrug and do something else more interesting than argue with them.
I guess I'm just used to looking at peoples DSM tunes then, almost all of them I see ramp up timing slowly as RPM rises. Usually its knock limit that they are up against, not sure if thats part of it.
Do you have a link to any of that engineering stuff. Sounds like something I'd like to read a bit of, always trying to learn.
MAF + a preference for internally gated turbos can result in weird ideas concerning how engines work. If you are dealing with a setup that tapers boost past torque peak, and is dealing with timing in terms of horsepower (airmass is directly related to horsepower), then it very much looks like the car wants more timing in the higher rpms.
Let me step back a bit from the concept at hand and give you a frame of reference as to how different shit works, and how that forms the way people "think" a thing works. It's pretty interesting, actually, as it makes some designs that start out looking fucking stupid turn into something that makes sense.
What I cut my teeth on, Hondas, you can view fuel maps in 2D format. Speed-density tunerboi adage, "fuel follows torque," and if there's nothing wrong with the vehicle your 2D fuel curve looks like your dyno torque curve. It is common practice to refer to 2D fuel maps while tuning as a sanity check, and people start thinking in these terms. With eCtune I commonly monitor AFR and MAP plotted out across an RPM axis in the same window that with a keystroke hosts the 2D fuel maps. With speed density eyes, boost is flat across a targeted scalar and timing is flat. When I tuned Robb's Integra with the internally gated 16G it would hit 24+ psi and taper to 12-14 with the boost controller on nig nog setting, and I chased his tune for that one boost setting across eight pressure scalars - at no point in time did I view it as requiring more timing past the (3800 rpm) torque peak... although it "did" due to boost taper.
MAF guys deal directly with airmass. Airmass is directly proportional to horsepower, ASSuming correct ignition timing. MAF guys could also view things as fuel follows torque as easily as speed density guys, but when dealing directly with airmass they have no reason to and therefore do not. The manner in which a MAF tune wanders across the scalars - or not depending on wastegate dynamics or
appears not to due to a lack of resolution in the mappings - shapes the way a DSM or Ford or Nissan tuner thinks an engine operates. Dennis (DmC) is a couple Fords deep now, and has a similar Honda background, he'll attest to the number of "boost scalars" above 1.0 load axis on a 750 whp Ford aren't enough to correctly tune a 225 whp turbo Honda. On those same Fords the ignition map would look like timing goes up (slightly) with RPM, but not really. DSMs also have a similar lack of resolution, although not as pronounced, not that the average DSM tuner who uses SLIDERBARZZZ knows this or in truth has any real grasp of what's going on.
I've by no means tuned everything I need to in order to learn everything I feel I need to, but I have been through a LOT of stuff in the last two years and my perceptions have grown a lot. On top of the OEM ECUs mentioned previously, GM (MAF + MAP) and Chrysler (MAP expressed in terms of airmass a la MAF lolololol) are interesting things to poke at, as well as a rainbow assortment of standalones.
Also, keep in mind that the premise of a turbo 4 cyl is not to make huge peak torque, but to maintain whatever torque it makes as long as possible and then let high rpms spin it into higher horsepower. Given a traditional V8 engine that makes big midrange and then falls off suddenly, yes, the amount of airmass entering the engine falls off abruptly and would seem to require additional ignition timing. I've tuned slightly less than a dozen mod motor Fords and one SBF, none on a dyno, but this might be something Dennis or 79fairmont could chime in on. Something with exaggerated torque that falls off suddenly, like some of the redneck built 347 strokers that make 350 wtq right off idle and can't break 300 whp, that the owner thinks is fast because it wants to rip the tires free when trying to pull away gently from a traffic light, even though it runs a 14 at the strip.
As far as the particular reference in re: rate of burn for iso-octane based fuels, it's somewhere in the first 15-20 pages of the second volume of Taylor's ICE primer, a pretty vanilla read although his bibliography is a rewarding reading list that you can spend a lifetime trying to plow through.