TB993tt

When the temperatures drop we all know our turbos are faster, cooler air = denser air, more oxygen and bigger bang giving more hp.

There have been various debates where people claim their tts give MORE boost when the temperature goes down, I have always observed (on an accurate boost gauge) LESS boost when the temperature falls and MORE boost when the temperature rises.

On the older non Motronic boost control engines it is probably accurate to say that lower temps will give more power but on our Motronically controlled cars it is not quite as straight forward.

I recently purchased a Bosch Hammer which allows one to observe the various inputs which the ECU sees and uses to determine how much boost/timing to allow.

The ECU uses the mass of the air predominantly (combined with ambient pressure and the important various temperature inputs) to determine how much boost to allow. The ECU sees this in kg/hour.
It is amzing to watch this "air mass" number vary according to atmospheric conditions. The only easy way I can compare is at idle speed of ~960rpm - today the air mass is ~38kg/hour with an intake temp of 12degC and air pressure 985mbar.
On other days, with higher air temp and different pressure the air mass is at 30kg/hr at the same 960rpm.

This is a 20% variation in air mass !!!

Thos correlates exactly with my observation of less boost at peak power on a high air mass day - usually a full 0.2bar less than a warmer day and in the summer the boost can be as much as 0.3bar higher (like for like) as the ECU winds it up to hit the air mass targets which the ECU has determined.

I am unsure exactly how the pressure reading fits into the calculation -maybe someone more knowledgable can help ?

So your Motronic tt will be faster in cooler air since your intake temps will remain below the ~32DegC level (at which the ECU retards your timing) - but we do not have the "advantage" of the older systems which could utilise their over rich mixtures and burn more of the fuel as it mixed with the oxygen rich air.

This also answers the question as to why the "tuners" who attempt to run fixed boost on the 993tt engine come unstuck when the temps drop - their wizards can map the fuelling on the chassis dyno for a certain set of conditions and a certain base air mass level - but when the base air mass increases past a certain point there is not enough fuelling to cope with the volume of air which the fixed boost is still pumping into the motor and the ECU cannot lower the boost so it resorts to cutting the spark - this has been experienced by all the fixed boost guys I know of who run the cars in varying climates like we have in the UK.........

Red rooster

DIN70020 is used to correct engine power for differing atmospheric conditions .

The correction factor for pressure is P/P0 where P0 is 1013.25 mBar .

So for 985 mBar the CF is 985/1013.25 = 0.97

So at 985 mBar the engine , with no alteration to boost etc ,makes 3% less power.

For temperature the CF is the square root of ( T0/T ) , where T0 is 293 degrees Kelvin . To convert degrees C to K just add 273 .

So 12 degrees C = 285 K

Square root of 293/285 = 1.014

So at 12 C atmospheric the engine will make 1.4% more power if nothing changes .

All that is taking account of atmospheric conditions which give guidlines as to what the DME is doing !

Happy new year

Geoff

Jean

The way I understand the question raised by TB is that why would Motronic lower boost in cold days and increase it in hot days.

My interpretation of that is that the Motronic is supposed to be adjusting for air density as it controls the engine for mass flow. My opinion is that what triggers the boost drop at lower temps is the combination between denser air and turbo flow capability, not strictly the Motronic, which controls boost based on a pwm output (pulse width modulator) vs.RPM/throttle position. I don't think Motronic "controls" boost per se.

So my guess is the following: If the tuning is set at the limit of the turbos' capability at DIN standard and under very closely controlled temperature and pressure conditions (can be done on an expensive temp controlled engine dyno only)....therefore on cold days when the air is denser than the tuned for parameters, you will run out of turbo sooner because they are cfm based, not mass, this translates into lower boost levels, but not performance levels as timing compensates.

Then, on hot days, the air is less dense, and since the turbo can move the same amount of air, it will use more pressure to get it there, thus, more boost and nearly the same power level, for all different conditions.
This will of course require very stretched control of Motronic and one that very few can master.

One last point is that I have never had any issues with spikes or other on my fixed boost setup on Motronic, the coldest I have tried it was in Germany, with temps in the 15C, certainly not very low, and the hottest was around 50C.

Red rooster

Jean,
The 993tt Motronic is a mass based system .Mass measurement automatically takes care of intake air temperature/pressure .
At low intake temperatures , with higher air mass, a smaller volume of air is needed to acheive a level of intake mass so the Motronic will lower the boost pressure to that required. Visa versa for higher intake temperatures. Same situation for ambient pressure .
993tt Motronic has a few more control functions and safety limits as you would expect.
This system allows a pretty well consistent bhp level regardless of ambient conditions and the small resulting boost level variations provided that the supporting hardware -turbos/injectors are not running dangerously close to their maximum limits .

A motor where an increase of 5% in airflow due to temperature rise, causes flow problems ,certainly does have turbos sailing close to the limit !


I would expect a fixed boost level to have more bhp varience with temperature/pressure.


All the best

Geoff

TB993tt

I'm a bit confused - where is this 5% number from

Jean
I agree with your reasoning but I think there is another dimension to it which I think you didn't make clear:
The way Porsche tunes the engine (and the way RS does it ) is to get maximum torque as early in the rev range as possible and then use the boost control (as well as the other controls) to keep the torque as high as possible for as long as possible. The act of winding the turbos up to their maximum as quickly as possible brings the "overboost" situation on Motronically controlled boost cars wherebye the ECU boosts until knock is detected (peak torque) then the boost follows the knock curve all the way to the limiter -since THE major factor in determining where the knock curve is is the heat of the boost air -then your reasoning about the "limits of the turbos capability" is spot on (obviously the limits of all the rest of the componentry come into play big time also).

people tend to forget (particularly on cough.....6thing) that a boost gauge showing engine 1 at 1.2 bar @ 6000rpm may be producing less power than identical engine 2 showing 1bar @6000rpm because engine 1 is blowing hotter air and not producing the power........compared to the correctly tuned engine 2.

Felix

This thread is getting near to my question of a few weeks ago where I queried if it was fair or correct to apply ambient air pressure corrections to a turbo engine. (Opps, hit send too soon.) Some of the above would seem to imply pressure correction is done by the Motronic system and therefore isn't subsequently required when calculating power outputs.

Red rooster

TB,
The 5% number is an informed guess at mass variation due to atmospheric condition variation ,which is where this all started ??

No way do Porsche map the 993tt DME to rely on knock as a control function !
Knock is a safety /fuel quality system .
Boost control as a function of mass flow along with all the other control functions is clearly defined.

Sorry if that sounds a bit heavy but I couldnt let that pass.

All the best

Geoff

TB993tt

So you are you saying my observation of the 38kg/hr at 960rpm vs 30kg/hr under different conditions only represents 5% and not the ~20% it appears ?
Don't worry about "sounding heavy" if you know what you are talking about, we need to hear it
I understood that part of the "learnability" of our Motronics was constantly testing timing/boost until knock was detected - you are saying this is wrong ?
How else can the system optimise its performance -unless it has maps for every possible atmo condition ?
I realise that there are set limits in terms of temperature which limit timing and boost in a set way but still thought the "knock chasing" was valid - it certainly looks like what happens on my motors also......
Edit
Just to clarify - the Porsche mapping I have no doubt is done to fit all conditions where the cars are sold and I'm sure the knock limits will be known and the engine mapped to stay away from them - I guess I am talking about ECUs which have been "tampered" with
Kevin M - any input from you - you've been messing around with these things for a while now

Red rooster

TB,
Idle is probably not a good place to access mass variation under load . Some quick summs show that 8% would need some pretty dramatic atmospheric conditions !

When knock is detected a relatively major retard is applied , the amount of retard is set by the level of knock but is enough to easily eliminate knock.
The retard is then slowly reduced until knock is detected again or the next knock event occurs.
If you tried to use this as a control parameter the likely result would be a sawtooth timing characteristic accompanied by relatively poor performance !



Yes, all atmospheric conditions are mapped along with all the other parameters that are needed.
A modified DME should have all the necessary maps recalibrated so that the desired performance is delivered , using the correct fuel quality , without knock issues.

Knock is the sound of uncontrolled explosions in the combustion chamber and is well known to be a major cause of engine destruction !!
For that reason it is to be avoided at all costs .
If your engine has knock issues I would get that fixed pretty quick !

All the best

Geoff

Felix

The Service Information Technik book describes the knock control as follows:

If a knock condition is detected the ignition timing of the respective cylinder is retarded by 2.25 degrees. If the knock condition persists, ignition timing is retarded to a maximum of 12 degrees in 2.25 degree steps per knock signal.

If the maximum retard setting of 12 degrees is not sufficient, the boost pressure controller reduces boost to reduce the danger of knock during combustion. If no knock condition is detected the ignition timing is readvanced incrementally to its optimum setting, i.e. the programmed setting, in a time dependent manner.

Thus it would seem the timing is only advanced back to the baseline; the ECU doesn't go beyond that baseline by finding the point where knock starts and backing off slightly from there.

Red rooster

Felix,
Couldnt have put it better ! Now why didnt I think to read that !!

The emphasis is entirely on saving the engine. The reduction in retard time is measured in seconds ! That could really spoil a 60-130 run !!!

All the best

Geoff

TB993tt

I presumed that the retard time and boost reduction folowed by increasing back up to optimal was measured in micro seconds NOT full seconds on these systems - that was the whole point and why the program could "test" for knock.
If it takes full seconds then of course you are correct but that is not what I have witnessed.
A few years ago when I was tooling about with some K16 hybrids using RS programming we messed around with wastegates and the thing was overboosting stupidly at 4000rpm up to ~1.5bar worth of hot air, it was clearly producing knock and the ECU allowed it to momentarily hit the 1.5bar then dragged it straight back down (presumably with the timing alteration as described) -this did NOT take "seconds" merely a fraction of a second.

How do you think/know it takes seconds ?

TIA

Felix

The next para says:

In the case of an engine speed of 2000 rpm, for example, eight operating cycles are required, and at 6000 rpm, 25 cycles are required until the programmed map is attained again.

I can't quite get my head around the difference in engine revolutions required unless this is because the process takes a fixed amount of time as opposed to occuring every firing cycle.

TB993tt

Felix
So how long would the ~15 cycles (if its linear) take at 4000rpm - fractions of a second ?

Felix

I think I've found the answer; in my earlier post there's the phrase "in a time dependent manner" used to describe the restoration of the timing.

Assuming an engine cycle is 2 revolutions it takes 16 revolutions at 2000 rpm to restore the timing; this is less than 0.5 seconds. Pulling the timing in 2.25 degree increments is almost certainly quicker still.