StealthTL

iTrader: (1)

A fascinating post on the electronic throttle and its relationship to your pedal inputs -
Copied 'word-for-word' from the UK RX-8 forum, and posted by Fabrice, (forum name Rasputin) who lives in Belgium and is the European dealer for 'Hymee' accessories.

(Fabrice, we would love to learn more, and access the graphs you refer to!)

Posted: Mon Jun 13, 2005 11:08 am Post subject: Electronic Throttle Body operation and calibration

Operation

Until recently, the throttle body that controls the air mass going into the cylinders (or rotors...) on a petrom engine, and therefore, the engine load (= torque), was directly linked to the accelerator pedal with a cable.
There was then a univocal relationship between the pedal position and the throttle opening, for all operating conditions. It was not necessarily linear as the progression could be affected by rounded levers to give more (= sproty feel) or less (= better driveability at lower speed) aggressiveness to the throttle body control.
Recently, theat system was replaced by electronics. The mechanical link disappeared in favour of an electrical wiring. The Electronic Throttle Body Control system contains :
- a pedal box with 2 position sensors ;
- an electronic control unit, nowadays integrated into the Engine Control Unit (as in our RX8) ;
- a motor-driven throttle body connected to 2 position sensors.

Its operation can be summarised so :
1) The Pedal Position Sensors (PPS) measure pedal position. There are 2 sensors to offer redundancy to :
- cancel any signal coming from a defective sensor ;
- ensure the car remains operative after a sensor failure.
2) The 2 PPS signals (PPS_A and PPS_B) are processed to obtain a single value : PPS.
3) The electronic control unit affect a requested throttle position value (TPS_DES) corresponding to the input PPS, according to the operating conditions (mapping).
4) Then, the control unit controls the motor driven throttle in closed loop in order to achieve the res-quested TPS-DES value. To do so, the system relies on the throttle position sensors (TPS) feeding TPS_A and TPS_B to the control unit that processes them into a single value, TPS. The control unit will make sure the TPS value stays as closely as possible to the TPS-DES value.



TPS/PPS relationship engine not running.

The data was collected with my Hymee sCANalyser Live OBD tool.

Reminder :
TPS : Throttle Position Sensor (also value in %)
PPS : Pedal Position Sensor (also value in %)

Tip-in = pressing the accelerator pedal
Tip-out = releasing the accelerator pedal

As you can see on the attache graph, the relationship is far from linear. Indeed, the throttle stays at his minimum crank value (8.2%) until you reach 50% of pedal travel. Comme vous pouvez le constater, la relation est loin d'être linéaire. Then, the system rapidly reaches 100% TPS before the driver reaches 70% pedal travel.

This mapping ensures the throttle remains at 8.2% (optimum for engine start) even if the driver accidently and sligthly apply some pressure on the pedal, in order to prevent overfuelling and, potentially, flooding of the engine.
In case of flooding, as the manual tell you, one must fully press the accelerator pedal in order to achieve 100% TPS. At this maximum throttle opening, the fuel map actually completely cuts out fuel in order to cure any flodding condition with induction air. To make sure one achieves 100% TPS even with a weak leg (before one's morning coffee) or a misplaced carpet, this wide open throlltle (WOT) condition is obtained with only 70% PPS. Failure to obtain 100% TPS would end up in adding massive quantities of fuel in the engine, aggravating the flooding condition.

Its seems there is some sort of hysteresis in the system. However, as I don't see the need for it, I guess it might be caused by some dashpot or timer strategy in the system.


TPS/PPS relationship engine running

The first graph contains all the data points recorded with my Hymee sCANalyser during normal driving, and contains city and motorway driving conditions.

There is a high concentration of data points in the low TPS/PPS area as that's where most of our commute driving takes place.

An important dispersion is visible in the area of low PPS. This looked suspicious to me and I decided to push the investigation a bit further by sorting the data points by gear (only 5 gears on mine I'm afraid).
The second graph seems to confirm a trend showing different throttle progression with regards to the gears. In 2nd, it looks like higher value of PPS are needed to obtain throttle movement. In 5th, the throttle almost opens immediately. Data points corresponding to the 1st gear are not clearly visible (only a few data points correspond in that gear, as traffic was flowing) but they are roughly in the same area as 2nd gear data points. Data poit in 3rd and 4th are roughly situated together somewhere between the 2nd and 5th data points.

It seems several ETB progression maps (it could just be a different gain, not full map) are available as a function of gear, or engine speed, as there is a relantionship between them. Please note that a gear selector sensor is not needed for the ECU/PCM to know in which gear the car is running. It just needs to divide engine speed by vehicle speed. That's what I did to determine the gears in my data set. Indeed, I already suspected previously, just by driving the car, that such a difference in progression was built into the system.

Going back to the graphs, the progression at low PPS is smooth in 2nd gear in order to give the driver a better feeling and control on the engine torque for a more confortable drive, or better control on slippery surfaces (snow).
As soon as the pedal position reaches 85%, you have WOT! No need to try to smash the pedal through the firewall to go faster . Moreover, at low engine speeds, very high loads are reached with small throttle opening as relatively low air delivery is sufficient at these speeds. But we'll discuss that somewhere else.

As you probably noticed, there is a relatively high disparity of some data points. I believe it is due to the relatively slow response of the motor-driven throttle to very swift pedal tip-ins or tip-outs. There is one data point situated at PPS = 0% and TPS = 28% which can definitely corresponds to a very quick tip-out : the pedal already is at 0% while the motor is still working at closing the throttle.

Anothe point : TPS never goes under 2%, even in the most violent overrun. More info on this later.

Rasputin
Club Member

Joined: 18 Mar 2004
Posts: 258
Location: Belgium
Posted: Mon Jun 13, 2005 11:09 am Post subject:

Fabrice

rotarygod

I don't expect the throttle to open necessarily the same amount as the amount of pedal I am applying. Remember that the throttlebody doesn't always have to open all the way to provide the engine with all the air it needs. At low rpm's, the throttlebody may only need to be opened half way to flow as much air as the engine can possibly use at that rpm. This will vary slightly with load. The ecu needs to control this due to the induction system these cars have. As more intake runners are added to the system as rpm's rise, the throttlebody's percentage of airflow changes in relation to it's position. This is all to keep intake velocity high. Yes we could set it up to run off of a cable operated throttlebody but you'd find the car a little sensitie to throttle inputs at lower rpm's and much less smooth depending on the load and rpm. The RX-7's uses a quasi-progressive mechanical 3 plate throttlebody to try to do something very close to this. I actually did some extensive work in the past and ended up using a large 75mm single plate throttlebody on an RX-7. It was very tipsy and sensitive. I'd give it half throttle at low rpm's and I may have well jsut floored it since the engine was breathing in all it could. Top end was great though but low end power suffered. It would make sense to vary the large throttlebody opening based on load and rpm. While not mandatory, it would help aid the smoothness of the engine for street use. The current intake could only be done successfully with an electronic throttlebody.

Rotarian_SC

Have the reflashes modified the throttlebody/pedal position maps, because I noticed after the M my car seemed to drive a lot smoother, especially in first and second gear at low speeds it was much less jumpy.

rotarygod

I'm not sure. They did change the fuel/ignition maps at different spots and that would be noticable.

Rasputin

They certainly did not change the steady state fuelling (runs in closed loop stoich) or spark (MBT and/or emission related) at low speed low load .
But transient parameters as spark advance / retard in tip-in / tip-out can make a difference, as some gain changes in the Electronic Throttle Body (ETB) control strategy.

That'sexactly what I say in the text, if you care to read it carefully.
Actually, at 2000 rpm, you can get 70% load with a mere 15 to 20% TPS (throttle opening).


StealthTL,
Thanks for your support.
As long as the thread your refer to is in the public area of the UK forum, I'll keep it that way. If it were moved, I'd post the full thread here.
The thread is visible here : http://rx8ownersclub.co.uk/forum/viewtopic.php?t=7864
My next study will be on the open/close loop fuelling and AFR table.

rotarygod

From which flash to which flash are you referring?

olddragger

iTrader: (3)

I may be just totally off here but I wonder if this contributes to the lift throttle surge some of us notice on the track?
Olddragger

Rasputin

None in particular RG, I was just making a general comment, not Flash or even RX8 specific. But I guess that if Mazda changed the fuel map, as some have measured on this sforum, it will only be in open loop operation, not in low speed and low load steady state close loop stoich conditions.
The only change in that area of the mapping (low speed/ low load) I can remember of, was on 2.0L Mondeo for the South African market. In the field, it suffered a bad case of shuffle at a given engine speed and very low load (in overrun actually) that got through the calibration validation tests. The engineer in charge of this issue solved it just by increasing spark advance at that particular operating point (speed/load) to stabilise combustion.

Fabrice

rotarygod

The reason I asked is because on Brillo's car when he had the M flash, we had driven around alot and I payed attention to the timing advance at different loads and rpms. Remember I'm only talking about normal daily driven rpms and cruising speeds, not full throttle high rpm runs. With the M flash I saw a max total advance of 38 degrees. After he got the N flash we went back and did everything all over again and now we saw a max total advance of 41 degrees which is admittedly higher than I've ever seen on a rotary. All of this was at light throttle below 5000 rpm.

Spin9k

I think you get sometimes a better feeling of acceleration by only giving 80% or so to the floor in the 8. When I do that I get the nice push with 3 distinct 'bumps' in acceleration. With simple WOT, they are not so evident. My resident rotary god says that w/WOT, the ECU thinks you are potentially on your way to abuse the CAT, and so it turns on a pig rich fuel map for open loop, rather than the nice (sort of) leaned out closed loop mode A/F ratio.

Therefore the blaaaaaa result at WOT.

zoom44

i agree that "leading" the throttle gives a better "feel" but i dont know if it is actually better

Rasputin

That's only 3 degrees. Couldn't it just be down to slightly lower IAT and/or ECT?

F

Rotarian_SC

I'd agree with this too, and I've heard that previous Mazdas have run faster at SLTWOT.