Discussion and questions related to the course Practical Dyno Tuning
Have a supposedly well-tuned M1 map for a high performance NA BMW S54 engine. Two things about the map strike me as very odd.
First, the fuel (i.e., engine efficiency) table has a number of cells in which efficiency exceeds 100%. In fact, every cell above idle at 100kpa MAP shows 100% efficiency or greater. Does this make any sense on a NA engine?
Second, the idle mass flow feed forward table shows DECLINING numbers as RPM increases. Does this make any sense to you?
Appreciate any guidance here.
It's not unusual for a good N/A engine to exceed 100% VE. You do need to understand however that VE-based ECUs such as the M1 are only going to be as accurate (in terms of the VE numbers) as the rest of the inputs into the ECU. This means it's easy to skew the VE numbers if some of the other parameters such as injector characterisation or fuel properties aren't 100% accurate. In general I wouldn't take too much notice of the absolute values in the VE table but I'd be more worried about how closely you're hitting your fuel mixture aim.
The idle mass flow feed forward would typically have larger numbers at higher idle targets. It may however just be a case where the engine hasn't been tuned to target the higher idle speeds so the mass flow values aren't relevant.
Well designed intake and exhaust system can push the volumetric efficiency to over 100% for normally aspirated engines. This is pretty typical of M1 VE tables actually.
On the idle mass flow feed forward table, perhaps that is what is required to prevent the engine "hanging" at a high RPM and not dropping when the ultimate desired RPM is less. So that table setup may be required to allow the engine to quickly return to idle when the pedal position returns to 0%. The idle system will compensate for less air by advancing the ignition, and can still hit the target RPM.
Are you experiencing a problem with idle control?
-- Well, I took a break while typing this, and I see Andre has provided similar info....
Thanks to you both for these helpful responses. The top end of this VE table (viz., 8500 rpm and 100kpa) shows 189% VE! But I get that the absolute values really do not matter, but rather, hitting the correct Lambda. Worst case scenario with these numbers in the efficiency table is that the engine runs extremely rich in the upper load cells, and better to start there and correct if need be than start with too low a number.
Have not yet actually run the engine with this map, so can't say it doesn't work. I was just looking it over and comparing it with others I have seen and noted these two seemingly unique aspects.
OK, with values that high, I would suspect the either the injector characterization or fuel pressure sensor/defualt values are not correct, and this has been compensated in the VE table.
Injector characterization is now hard-coded by Motec in my package on an injector by injector basis (meaning one selects the characterization by injector model number), so it would seem difficult for the characterization to be wrong. Same with fuel pressure, viz., 72.5 psi (5 bar). The engine does run at .85 Lambda in most load cells, so perhaps that is the cause?
Are you sure the correct injector has been selected? Selecting "Manual" will make the current setting available so you can compare. Is the Fuel Pressure sensor type correctly set (absolute, vs. gauge pressure). These differences can lead to "baked" in values that will still run the car, but aren't actually correct. Fixing the root cause will often mean that you need to revist the VE table.
One test is to change the Fuel Mixture Aim (by say 10%), and see if it hits that target (with no change in closed loop trim if that is active). If not, you know that some parameter is not correct.
Well, I will be running different injectors, but the correct injectors were selected for the car on which the tune was created. I will be running a gauge pressure sensor and so was the car on which the tune was created. I guess I understood one of the benefits of the VE-based fuel model in the M1Tune software was that changing injectors or sensors would not change the VE table--that the M1 would make the requisite calculations to achieve the specified Lambda once the appropriate injectors and sensors were selected. Put differently, I understood the VE table to be engine-specific and that once tuned, changes to injectors, etc. would not require that it be revisited. The engine on which this tune was created was the same model (BMW S54) as mine and with comparable modifications (e.g., cams and higher compression).
I think my best course of action is to re-tune the entire VE table once I have made the requisite changes for injectors, sensors, etc. With the numbers currently populating the VE table, it should run VERY rich, which should make it a VERY safe tune from which to start.
Thanks again for the help!
VE numbers of 189% on a NA engine are pointing to there being incorrect data setup in the M1. Well tuned NA race engines typically will peak around ~120% at Peak torque, and then taper off higher in the engine speed range. The VE table shape should also have a similar shape to the engines torque curve, if it doesn't then the M1 is incorrectly calculating the VE.
I find the Complete Fuel Calculations worksheet in the Monitor workbook to be a very useful tool for diagnosing these sorts of issues, as a bad bit of data tends to show up very quickly.