This issue has cropped up here and there in (many) other threads. To avoid dragging them off-topic, I thought I'd put up a simple log of my SST's apparent delay when taking off quickly from a standing start, and also my conclusions after reviewing it.

It's the sort of thing that can ruin your day if you try and enter/cross heavy traffic. A gap comes along, you apply a good amount of throttle, and the SST "thinks about it" for what seems like a VERY long time, then gets going. Only once boost builds will your butt get hauled out of harm's way. Pretty scary the first time!

Anyway, I got sick of spouting phrases like, "it seems to pause for a full second", and "seems like an eternity"... so I'm starting to collect some hard data from the OP2.0 standalone logging files.

Here's a perfect example of the behaviour in question... (I pared it down to a few key elements, and got them all being sampled, every cycle)

As luck would have it, my right foot started to twitch at timestamp 90.000.


The part in RED looks to me like the SST performing an open-clutch RPM increase phase... check out the evidence of clutch "grabbing" at 90.825... where the revs drop even though throttle is still wide open.

The part in BLUE is the acceleration phase.


This log tells me that the car didn't actually start to go anywhere until 0.8 seconds after I stamped on the throttle. And by "go anywhere", I mean 2600rpm, a whopping 1 psi boost, and the wheels just starting to roll forward.

That can indeed seem like a long time in some traffic situations.

However...

Looking at this critically, and comparing SST behaviour to my own manual transmission driving, it kind of makes perfect sense...

If I were in a manual vehicle, I'd see that gap in traffic coming, get off the brake, on the throttle to bring the revs up in preparation, but leave my left foot on the clutch. My brain's "go" signal would result in my left foot coming off the clutch, timed to perfection with the gap in traffic... and the car goes.

If I were in an automatic vehicle, I'd simply stamp on the throttle... and the car goes.


But with an SST, it can't "see" the gap coming. I can sure see it, but I have no clutch pedal. The engine revs are at idle... the car is essentially parked. So there's honestly no way to reasonably expect an instant fast take-off from an SST-equipped vehicle - it needs time to get the revs up before it can engage the clutch.

The "missing" 0.8 seconds before an apparent reaction isn't missing at all. The car is reacting... it starts building revs from the moment the pedal goes down. But there's no boost - the car was "parked" right up to that point.


After seeing this from the inside, it's clear now that such situations require a similar approach to a manual... namely, getting the revs up before the "gap" arrives.

The good news is, it's possible to do this with left foot on brake, and then very quick jabs at the throttle. The jabs need to be very brief for the SST to keep from "stalling it up", as auto drivers would say!

Then, when the gap comes, off the brake and steady throttle. My guess is, if the revs are already at or above 2600 (ish), the SST will be able to link up engine to wheels waaay quicker than 0.8 seconds.

I'll go log that tomorrow, and see...

Rich

 

Great post! It quantifies what we've all felt. I'd love to see logs of the stabbing method, and if you hold it at 1000 rpm and gas it as you let off the brake...

 

Having a Blitz throttle controller also helps with the pause when stomping the gas.

 

This is because the sst is trying to guess what you're doing (I'm assuming you're in auto) by looking at your pedal pressure.

It looks at the brake pedal pressure, and looks at throttle pressure. If you are barely touching the brake to where the car wants to creep forward you are engaged. If you push the brake pedal much harder it will completely disengage the clutches and will get this lag.

What works great for me is, always been in manual / sport mode. I lightly touch the brake so I don't go out in the traffic, then when the timing is right I step on the gas back off the gas to about 1/4 throttle, then quickly shift to 2nd gear once I'm going. The shifting to 2nd gear keeps my car from shooting off like a bat out of hell once it finally gets spooled up and 100% engaged.

 

That's interesting seeing data to show what we are feeling. I noticed the same thing too. If the car is standing completely still... foot off the gas and on the brake... there is definetely a pause between when your foot gets on the gas and when it goes.

I find I can get rid of this a bit by blipping the throttle before actually taking off. I don't mean revving all the way upto like 4 or 5K rpm... I just mean lightly tapping the throttle before going.

 

How?

 

That log was taken in NORMAL mode, manual select. I've not felt any difference in the various modes at all - SPORT/manual "feels" the same.

Of course, now I can just log all the different modes, and find out for sure!

Rich

 

On the Evo X, the torque tables, clamping speeds and rev matching are effected by the mode. I'm not sure about the Ralliart.

 

I get the same delay, and subconsciously I have noticed myself coping with it by 'blipping' the throttle in this fashion:

Stopped waiting for an open space:

1. Blip the throttle quickly when you see the space
2. Cover (hover over) the brake
3. If the gap is now infront of you, gas into it quickly - if circumstances have changed, brake and repeat the process when the next gap is seen

From the outside perspective your car may move very slightly or not at all for the first half-second to second after the blip, however it helps to prepare the car to move forward by engaging the clutches

 

I've noticed the same type of effect. The delay is much less pronounced when you blip the throttle so it's essentially already jumping to the low end of the boost before you go. Without putting the full strain of launch control.

The car definitely can react weird if you make sudden changes as it's almost reacting too much on the low end. Same type of effect with the stuttering on rolling stops in 1st/2nd. It's expecting you to go, but when you only blip the throttle it goes from all out, to full off in a fraction of a second. Takes some finesse to smooth out those reactions.

 

all this was done in normal mode i take it?

normal mode cuts throttle input like a *****.

the sst closes the throttle plate and pulls wgdc's

some of the "lag" you feel can be tuned out.

 

While the first exploratory log was done in NORMAL mode, the standing-start behaviour doesn't "seem" to be different in SPORT mode. Having said that, I'm trying to rely on objective stuff in this thread... not my seat-of-the-pants impressions. I'll repeat the exercise in a variety of SST modes and situations tonight, and post an update.

Would SST throttle trickery not show up in logged TPS values? Seems to go to 87% and stay there.

Dunno about WGDC though - wasn't logging that...

Rich

 

@IndyEvo66:

I noticed you drive an MR... which jogged my memory about this from Bryan:

https://www.evolutionm.net/forums/ev...-mr-vs-ra.html

For the MR, I see what you mean...

Rich

 

Okay, I did four "standing start" tests back-to-back. Normal vs. Sport, and also ASC on vs off.

Test #1: NORMAL mode, ASC on.
Test #2: SPORT mode, ASC on.
Test #3: NORMAL mode, ASC off.
Test #4: SPORT mode, ASC off.

I didn't bother testing "auto" versus "manual" gear selection mode because there's no gear selection going on. If that makes any difference, then shoot the Mitsubishi TCU programmers, hehe.


The results were totally, 100% identical in every test. There isn't any difference from SST mode selection when it comes to a standing start take-off.

On my RA, the SST will take exactly three quarters of a second to bring revs to 2500rpm, at which point it will engage the clutch.

It's like clockwork. Actually, inside, I bet the SST is exactly like clockwork. :P


The "start" time was measured from the first twitch of the logged TPS value.

The "end" time was measured from the first observed RPM value that was no longer increasing. In 3 out of 4 tests, this was exactly 2500rpm.


In each test, the difference was between 0.72 and 0.76 seconds (logging trimmed right back to get a line every 0.04 sec).


The speed value I'm logging doesn't react very fast - it only seems to get updated by the ECU every 0.75 seconds - so I can't say exactly when the car started to move on each test.

However, on one test, I was fortunate to see a speed of 2km/h logged a fraction of a second after the obvious clutch "grab".

So it's fair to say the car starts moving off about 0.8 seconds after you stomp on it... albeit in a relaxed, Sunday-driver kind of fashion.


The Evo X gets Launch Control.

The Ralliart gets Out-To-Lunch Control.


The next logging test will be of the "quick throttle blip" approach, which gets the revs up before you actually want to get moving. Watch this space.


Rich

 

interesting find; I wonder if this is also the case in other non-mitsu tc-sst like the gt-r's. If not then we could start looking in that direction for solutions