Sequential Paddle Downshift Blip (no clutch)

[beetos]

Setting up a GR Yaris with sequential, all is good but downshifts are very harsh at high rpm. Tried closed loop on barrel%, request ends cut like in the syvecs help video, open loop etc. its downshifts nice at low rpm but very harsh at high. Open loop with longer blip times seem best.

TPs bliptarget multiplier changes don’t seem to help as while in gear the rpm will never be free to reach postdownshift rpm and bliptarget multiplier drives the target too high. Not sure if that’s mostly for twin clutch cars.

Need to yet map the engine predictive time time table but I would be interested what blip set up people are using who have nice high rpm downshifts on a sequential to share their successful set ups for info while I continue to experiment.

[stevieturbo]

first you said upshifts are harsh, and then downshifts ?

For most part for downshifts, it'll either not open the throttle enough, or open it too much. Start somewhere and work one direction only to see how it feels.

Does logging look like the blips are starting at a correct time ?

[beetos]

Apologies, meant downshifts. Up shifts are great. I’ll post up logs but blips look like they are coming on at the right time.

[beetos]

logs of downshifting from 6 to 1.
This is closed loop. No other parameters ending blip. Rev limit active on blip.

[stevieturbo]

Not sure what's going on there.

Your gear cut input voltage is a very curvy line, as if someone is either leaning or very gently and slow with the shifter ?
It never looks like it is at rest, then a sharp precise shift action made ?

Also throttle, it looks like then throttle is a little out of sync with the DBW target. Nothing huge though. But blips do seem quite small ? Have you tried a larger opening target ?

Also, you've quite a lot of fuel cut and ign retard during the downshift. Is that really necessary ? Like literally after the blip, you're effectively killing the engine. That can hardly feel smooth ?

Even between gears you have a lot of ignition retard going on, although you are not logging the reasons behind any torque reductions to see why it is doing it.

[beetos]

Hi, Torque reductions during blip are coming from post downshift rev limit during blip active, after down shift they are ALS settings for burble. I’ll try turn off ALS to see if it makes a difference.

Blip targets are small but feel smoother with smaller targets as If I try more blip target, blips get harsh and you get even more torque reductions because it drives the rpm harder into the limiter that sits at post downshift rpm, locked in at blip request. It’s getting a bit better the more I experiment. Still trying to fully understand the bliptarget modifier table ( error between rpm and post downshift rpm target). With a twin clutch or clutched down shift I get it as engine can spin independently on drive train, but a sequential that is either in one gear or the next, wondering if that correction is negatively impacting blip target?

Noted on the blip input voltage, that would be my lazy driving.

[beetos]

One thing I realised through testing is roll on rate has almost no effect, right from 0.01% to 100%.

Blip targetMult has a massive impact on how fast the blip comes in and has a large impact on smoothness of downshift.

Continuing the testing…..

[beetos]

So the finding is the DbW on the GR yaris simply can not respond fast enough above a certain DbW blipTarget from closed throttle.

I can get my sequential box downshifts nice and smooth with no clutch up to around 3000 rpm, which demands about 35% tps depending on gear. above 3000 it gets harsh as shit. Logs show the DbW is still ramping up to try and meet demands of anything over ~40% TPS before the lower gear engages. No matter how hard how try and drive the TPS using bliptarget multiplier or higher base blipTarg values, same outcome, lazy DbW.

The base map for syvecs does need a bit of work to get the DbW PID to behave properly and that needs to be sorted first to stop it overshooting in both positive and negative directions, but with the DbW motor capped at 80%, logs show its driving the motor as hard as possible so there is nothing left in DbW PID calibration to be had.

Solution is either using the pre-blip feature or an ALS cal that holds the TPS open. I have tested with the preblip feature at present set around 25%. At RPMs over 3000, this allows the TPS to be close enough to the required blipTarget at higher engine speed (~40%+) before the blip request comes in, and now I can get clean downshifts with reduced shunting. I'll probably migrate to an ALS cal that I can select at the track when I need high overrun TPS for the blipper and also use it for antilag at the same time.

I think because on such a heavily modified small engine set up like my Yaris (1.6L 3cyl), the stock throttle body being so tiny needs a lot of opening angle for decent engine acceleration transient as demanded during blipping a sequential at high RPM, so the small DbW TPS motor struggles. Longer term I'm probably going to drop in a bigger TPS.

[stevieturbo]

And why capped at 80% ?

Have you tried at say 95% ?

What size is the actual blade ?

[beetos]

Blades 49mm

Syvecs base map caped at 75, I added 5%. Don’t want to burn it out. Anyway, not seeing much tpsDTR improvement anyway by pushing boundary of duty limits so blade is going as fast as its capable at 80%