Pseudo Real Time Tuning for E38 / E67

[kingtal0n]

So according to Greg Banish, MBT is Maximum Brake Torque. There wouldn't be anymore than maximum.

Torque is an integral function, occurring over a range of connecting rod and crank angles.
Pushing really hard on the piston for an instant may generate as much torque as pushing on the piston gradually for longer duration.

Thus we are not concerned with torque or max torque while tuning as much as we are concerned with cylinder pressure and the duration of torque integral over which cylinder pressure is applied to the surface of a piston.

For example I can add too much timing and damage a 300tq engine easily, shatter the piston with a sudden spike in pressure with very little torque.
Or I can pull out a ton of timing using the same fuel and add boost pressure to make 600tq with no piston damage, no cylinder pressure spike.
The timing doesn't make the torque; our airflow rate net flux is where torque comes from.


You aren't using the timing to add torque in a performance application. This is mistake #1, adding timing for more torque.
It should not be done this way. Instead, you want to increase cylinder fill & fuel quality, and remove as much timing as possible for the longest duration integral of torque applied to the piston surface which will not break the piston or bend a rod. That is the minimum timing possible.

[Sub8]

Within the industry MBT is defined as minimum timing for best torque. Its not someone's opinion lol!

Pick an rpm and a load. Any rpm or load. Run on an engine dyno at that point steady state. Starting from a retarded timing add timing. The measured torque will increase, the magnitude of the tq increase will slow as you approach MBT then stabilise over a small range and then the tq will start to drop off as you continue to add timing.

On typical production engines MBT timing corresponds strongly with having the 50% mass fraction burned occurring around a crank angle of ~14degs after tdc. For typical bore and stroke geometry 14degs is the most mechical efficient angle to get the most work (torque) from the engine.

The only time you deliberately retard from MBT is if you are knock limited, which in most real world situations you are (ethanol fuels aside) or at idle when you when you want to build a torque reserve to allow the pcm to reject an incoming load ( ac turning on, power steering etc).

Idle, cruise at 30mph or 130mph or running wide open pedal are just visiting different parts of the engine speed vs load grid.

The shape of tq vs timing changes depends on where you are in the speed load grid. At lighter loads the charatetistic can be quite flat; very little change in tq for large changes in timing. Conversely at peak torque the rise in tq can be very steep with small timing changes.

[kingtal0n]

The only time you deliberately retard from MBT is if you are knock limited, which in most real world situations you are (ethanol fuels aside)

Right, real world situations. Example: on the dyno an engine takes more timing than it would in a real world situation. Therefore, you would retard deliberately away from peak torque and peak cylinder pressure on the dyno such that when the vehicle runs in real world situations for 20 years 200k miles the head room provided will account for knock limiting situations due to higher temperature/load so the engine doesn't wind up damaged or wipe it's rod bearing out from an aggressive peak torque dyno tune timing profile.

If EGT is under control there is no downside to a reduced timing in a longevity sense.

For example a 5.3L at 600rwhp ~11* of timing is what I called MBT tuned on dyno and for the real world situation using 93 octane.
If I ran 14* or 15* instead, the 93 isn't knock limited on the dyno, and the torque output is slightly higher (4 to 5% maybe +25lbf-ft total) but this would become a disastrous value for real world traffic situations when the intercooler heat soaks and fuel gets hot from circulating all day, IAT goes up, full noon sunshine surrounded by concrete etc... The engine would become damaged at that setting even though it is well within MBT boundaries for steady state or otherwise dyno tuning.




Additionally A factory ECU may hold peak torque values throughout their 'ideal' ignition maps but they also run aggressive timing retard strategies as well, and typically lower VE than when we modify an engine. The double edged knife mistake that is so common to make here is tuners will disable or tone down knock retard strats while also searching for more torque via timing input.

the magnitude of the tq increase will slow as you approach MBT then stabilise over a small range and then the tq will start to drop off as you continue to add timing.

When steady state tuning continue to add timing beyond MBT will place tremendous stress on the rod bearings especially near TDC where no torque can be applied to the crankshaft due to lack of vector cross product at TDC.

However, on a Dynojet there is no steady state, rate of change of RPM is always positive, the engine continues to accelerate throughout the load procedure. Therefore when adding timing beyond MBT due to the rate increase allowed by the dyno, the torque may increase instead of falling off post MBT due to cylinder pressure spike and modern combustion chamber accommodating, we would see a more jagged curve where the dyno reports the highest peak as the highest torque, making people think more timing = more torque, where it was merely an artifact of data collecting from a perturbed accelerating roller.


Bottom line is this: We decide how much torque an engine will make before putting a combination together, and not go searching for torque on the dyno via timing after the fact.

For example, say I wanted 1000lbf-ft of torque from a 5.3L engine.
We can't take a 300hp engine and simply increase the timing until it makes 1000 torque. Torque doesn't come from timing.
Instead, we choose a compressor to support that kind of torque, 120lb/min or similar.
Then the engine can make 1000 torque with deliberately reduced timing, less than max torque, on a dyno, whether steady state or otherwise. The least amount of timing deliberately away from peak cylinder pressure, because we don't need to squeeze every last ounce of torque out of the engine in that situation.
In other words, we aren't going to make 950torque and deliberately add 1* of timing to gain 50 for a total of 1000, putting the engine into a dangerous situation.
Instead we will flow 1100torque worth of air, and deliberately reduce timing so that it makes on 1000 or 1040 or something, leaving the rest of the timing on the table for when the real world situation causes a shift in the timing:torque which reflects cylinder pressure maximum values.


To put this one more way, everything in a performance sense is a waste. Rich a/f ratios waste fuel. Intercooler throw away heat energy, reduce power. Water injection and alcohol fuels pull energy out, reduce efficiency all else equal. Use of heavy transmission parts such as 4l80e are unnecessary for economy, its a huge waste of energy and fuel.
Everything is a waste for the sake of safety.
And when it comes to timing, especially for 1000torque setups, the timing decision is no different. We will use deliberately reduced timing numbers which throw away potential energy and waste fuel, just like everything else, because this will help ensure longevity and reliability. If the setup was for 1000torque then it doesn't matter how much power we throw away to get it down to 1000 from whereever it started... the timing isn't needed to get it there. I hope that makes sense.

[Alvin]

Alvin and Michael, this is a side note but I thought I would ask. What kind of fan set up are you using on your dyno?

I know steady state tuning taxes the vehicles cooling system and the dyno's fan system but I am curious what you are using or have used and what you like for doing tuning like this. Multiple sessions are required for sure especially at the top end of the map.

Just imagine what it would take to even make a 25mph wind over an entire car.. There aren't fans big enough to do this. It's not just radiator but headers exhaust rear end. the whole thing.

[cadillactech]

Just imagine what it would take to even make a 25mph wind over an entire car.. There aren't fans big enough to do this. It's not just radiator but headers exhaust rear end. the whole thing.

I get that for sure, just wanted to know what fans you guys were using, buying a dyno in a month or so and choosing a fan is my next purchase.

[GHuggins]

Alright, just to throw another curve ball in the mix. When I am on a dyno and tuning in timing (of course it's been awhile tuning for a shop with a dyno) - it's not necessarily optimal, however it is safer for the engine - as I am tuning in the VE model I set the main timing table a little high - allow some safety with the low octane table to be lower then the high just like oe then set the knock settings up just barely over OE... As I'm tuning VE under steady state conditions, I can watch knock voltages and what timing is being pulled. This tells me if there's more room in the knock parameters and where timing may need to be pulled. I then do some 3/4 throttle pulls and wot pulls to do final timing dial in all with vehicle load simulation activated on the dyno during the pulls. That uses the eddy brake to load the vehicle to "simulate" wind and road resistance. I even do final wot dial in with the same rolling resistance profile.

As pointed out, you need at minimum 3 fans in front, possibly 2 blowing under the car around the fuel tank "it will boil the fuel in the tank just from exhaust heat" and a big fan blowing on the eddy brake while watching it's temps - don't go over 400...

As for temp swings - a lot of that is in the tune too. Inj tip, iat, intake valve and so on to keep fueling in line and to pull timing for safety.

[Sub8]

Alright, just to throw another curve ball in the mix. When I am on a dyno and tuning in timing (of course it's been awhile tuning for a shop with a dyno) - it's not necessarily optimal, however it is safer for the engine - as I am tuning in the VE model I set the main timing table a little high - allow some safety with the low octane table to be lower then the high just like oe then set the knock settings up just barely over OE... As I'm tuning VE under steady state conditions, I can watch knock voltages and what timing is being pulled. This tells me if there's more room in the knock parameters and where timing may need to be pulled. I then do some 3/4 throttle pulls and wot pulls to do final timing dial in all with vehicle load simulation activated on the dyno during the pulls. That uses the eddy brake to load the vehicle to "simulate" wind and road resistance. I even do final wot dial in with the same rolling resistance profile.

As pointed out, you need at minimum 3 fans in front, possibly 2 blowing under the car around the fuel tank "it will boil the fuel in the tank just from exhaust heat" and a big fan blowing on the eddy brake while watching it's temps - don't go over 400...

As for temp swings - a lot of that is in the tune too. Inj tip, iat, intake valve and so on to keep fueling in line and to pull timing for safety.

All true!

I guess I was more trying to help anyone interested in what the underlying physics for timing looks like.

Running the engine on dyno allows you to control the combustion parameters (intake air temp, coolant and oil temp) and to allow the combustion chamber temps to stabilise. So on this regard the engine dyno gives effectively a worst case operating condition for spark - the head and particularly exhaust valves all reach their peak temps, exhaust back pressure stabilises. So actually for a like for like condition for intake air temps etc dyno derived spark timing will be conservative in the vehicle.

But we must remember that the control strategy is built on blocks of info. For OEM the fundamental is an MBT spark map that gives optimal engine torque and a steady state exhaust gas temperature map that supports either LBT or component protection enrichment above this.

Then you typically build BLD spark maps. This is borderline spark timing. For this you use market fuel and this gives you the spark timing that the engine will tolerate with some minimal predefined knocking - measured using cylinder pressure transducers. So typically on an NA engine most of the part throttle map will achieve MBT. Typically you start to be knock limited at high load as you approach peak torque and then you may be able to start to approach MBT again at higher RPM. This is why just tuning for knock on an NA engine doesn't always give great results. On a turbo application tuning for knock is much more effective over the whole map.

You then set up your temperature modifiers for spark and fuel based on intake air temp, coolant and oil temp etc.

This in theory gives you the optimal spark timing under the whole set of real world operating conditions.

Then you calibrate and turn on your knock control. If you've done all of the above correctly knock control should only ever be adding/ subtracting 1deg of spark.

The difference between the optimal MBT spark map and the actual delivered spark is then also used in the PCM torque model calculation.

All modern control systems work in the torque domain:

Throttle pedal position and rpm give a torque request to the pcm. This is then split into an ignition and air request that is delivered to the combustion chamber. The pcm then calculates actual current torque based on measured airflow and timing.

All great in theory of course - but what we do in the aftermarket is basically massage around the edges of the entire control system. But it's still useful to understand the underlying theory. We also have to bear in mind that even on a long derived aftermarket interface such as HPT we are only seeing probably 5% of the actual pcm calibration functions and datasets.