The stock DME hardware is capable, the reason it can't do MAF is due to the transfer function code in the DME. A guy on rennlist (rogue_ant) has written his own mods to the code to allow true MAF with the stock DME. As long as the MAF gives a 0-5V output then the DME can read the signal. He says he will release the code for free in due time. I thought this is what Vitesse had done with his MAF.
Vitesse MAF
- Thread starter DivineE
- Start date
Diver944
Active member
No, not yet. It's on the list at some point, but more power is not needed at the moment. The next priority is going to be suspension and then possibly lightweight. Getting back to the Vitesse MAF, it is only really necessary to get the piggyback if you want to get the last possible inch out of the tuning or you have a very non standard spec. The base chips that John provides are pretty much on the ball (I would say 95% perfect) especially if you have one of his own turbochargers and a 2.5 engine. The base map he wrote for my 3.2L and BB turbocharger was near as damnit perfect as soon as I fitted it, just slighty rich, so very safe. The beauty is it doesn't care what boost level you select, the AFR stays the same. I have a Greddy electronic boost controller and I set it to either 12, 15 or 18psi at the touch of a button. The MAF keeps the AFR the same across all three [
]
Suffolk944
Moderator
As I understand it the Vitesse is indeed a "true" MAF rather than say Scivision's which as Scott explains converts a signal to allow the cars standard ECU to read it. Oh and the standard Vitesse item is brilliant even without the Vflex which is really an icing on the cake addition for highly modified cars - if I ever get to the stage of a new turbo then I may consider it, but not until then .
TTM
Well-known member
My observation is that the system feels like it cannot keep up with how fast the turbo spools. Either I could get an AFR that would be fine when avoiding having the turbo spool as quick as it can or I would have to add fuel everywhere in the fuel map when fully exploiting the turbo and as a result it would run too rich on partial load/boost - at any manifold pressure value in the [-5; ~8psi] bracket it always ran rich, never leaner than ~11, until I used a trick John told me to eventually get a correct AFR in this pressure range. The system feels like it doesn't self-adjust in function of manifold pressure and will assume boost pressure in function of what the MAF measures. This means either there is some turbulence in the intake which makes the MAF reads a flow that doesn't correspond to the flow value calibrated into the software or there is something wrong in the wiring, which makes sense since I've had the "lean spike" problem since day one.
I think I understand what you are saying. Because the DME must richen the mixture to say 11 when on boost without knowing the actuall manifold pressure it must use some other method to work out when the car is on boost. Vitesse is using some load calculation to decide when the car is coming on boost, like above a certain mass flow rate at a given RPM means the car is on boost. With your 3.0 litre engine the Vitesse MAF thinks you are on boost when in fact you are not so you run rich. Did it run rich on very low load throttle or just kind of high partial throttle where it might have thought a 2.5 engine was boosting?
TTM
Well-known member
Yes this is what I'm saying, AFR has always been coherent below -5psi (= vacuum higher than 5psi) and used to suddenly turn to 11 as soon as manifold pressure reached -5psi. I suppose Vitesse provide an (overly) safe configuration when their customers use special turbos with unpublished characteristics, which is my case. PS : apologies to rickware, I didn't mean to bring another thread on my past tuning misfortunes.
I don't think its got anything to do with the turbo, more like the engine capacity. Your's is a 3.0 litre isn't it? I think that your engine sucks enough air off boost for the vitesse to think that its a 2.5 litre on boost which is why its injecting rich. He probably has a pretty good idea when a 2.5 litre is boosting but your engine is none standard so the MAF does't really know how much air it can suck off boost. Ask John how he works out when the engine is on boost and thus when to start injecting rich. This is easy at WOT as you just set the AFR to 11/12 immediately but at partial throttle there is a bit of a grey area as you don't want to run rich all the time.
TTM
Well-known member
Since engine capacity is a factor taken into consideration when ordering the kit, I would think it matters in how the maps are made and I'm confident the base maps I am running are specifically for a 3.0 and not a 2.5. How should the turbo not matter when the engine is running off boost? The turbo still sits between the MAF and the exhaust gases, and the shape of the blades surely have an impact on flow regardless of vacuum/boost?
The point is that with a MAF you know the air mass going into the engine so you can inject to whatever AFR you want. It doesn't matter whats inbetween the MAF and the engine as you know the mass of air going into the cylinders. The descision is what AFR you want to inject to. Clearly the MAF is making the descision to inject to 11AFR which is what you want to run while on boost to keep the temps down. This suggests the MAF thinks the car is on boost when in fact it isn't.
I'm not a 100% sure what you mean in your last post, but I will try to clarify my understanding. 'True MAF' is really just saying we are measuring air mass directly. The standard DME works with air mass but it calculates it from air flow, air temp and probably atmospheric air pressure rather than measuring it directly. Once you know air mass you must still make the descision on what AFR you want. This might be down to cylinder head temperature, throttle position, air temp, exhaust gas temp or boost pressure. If you want to adjust AFR based on these parameters you must either measure them or guess at what they might be. As the Vitesse MAF and standard DME don't use measured boost to adjust the AFR they must guess at what boost might be.
sawood12
New member
My understanding is taht the stock ECU cannot manage the realtime temperature calculations required for MAF calcs. The DME is only contains reference information which is looked up by the ECU hardware so you are limited to what you can achieve through clever re-programming of the ECU chips. The stock ECU does not have the capability of sampling temperature in real time. The best you can do is put some logic in the MAF sensor itself but all you are doing is mimicking AFM. You need an extra input into the ECU that samples temperature in real time and then the relevant map can be burned onto the ECU chip. A piggy back computer obviously provides the extra functionality that is required to provide realtime MAF. Also the Vitesse kit also gives you semi-sequential fuel injection which is something the stock ECU simply cannot provide. Not sure what Rogue Ant has done - I thought he was fiddling about with a cheap standalone ECU kit similar to Megasquirt - MAF Translator I think it was called - has he started another project on the stock ECU? If it was as easy as more clever programming on the ECU chip then I can't see why there are so few MAF kits that provide True MAF. This isn't rocket science afteral. It is hardly cutting edge technology.
You don't need air temp for MAF, the MAF sensor compensates for air temperature already. The stock DME talks in terms of AFR I think, I have the stock programs and some software to edit them and all the fuel maps have values in AFR. The DME is looking up what AFR it should be injecting too and then working out how much fuel to inject. The software is not really any different for MAF other than you don't need to do temperature/pressure compensation calculations to work out air mass which actually reduces load on the microprocessor. You simply calculate/lookup the air mass from the MAF sensor voltage, lookup the AFR using the same fuel maps as before and then calculate/lookup the injector pulse width in the same way as before. With the AFM before you have air mass you must calculate/lookup air volume, then air density using temp/atmospheric pressure and then you have air mass to use with an AFR value to calculate IPW. The stock DME can sample temp in real time, the temp sensor is hooked up to an 8 way multiplexing ADC which the microprocessor can access at will, it is simply a matter of coding the DME to do so. Its on the same input chip as the AFM for example. The reason our DME cannot be easily adapted to use a MAF sensor is as follows: The AFM produces a linear (ish) voltage signal in relation to air flow, the DME computer uses a transfer function (think a2x+bx+c) to map from the voltage value to the actual air flow measured by the sensor. A MAF sensor produces an exponential type curve of Mass vs voltage, this type of curve cannot be fitted to with the transfer function on the DME. This makes it hard to use a MAF sensor on our DME. However if you can code a different way of translating from the MAF output voltage to actual mass flow then this problem is solved, I believe rogue_ant uses a lookup table (map) and interpolation to do this. This is no easy task as it involves reverse engineering the DME code, this is not somthing your average guy is going to be capable of. It takes a lot of coding knowledge and the right motivation. Also many of the DME's only have part of the code on the EPROM which limits the bits you can access. As far as semi sequential injection goes, I'm not sure how he does this but the injectors on the 944 batch fire as they are all driven from the same output. There is simply no way to fire one injector and not the others. Does he run the injectors from the piggyback? Rogue_ant has made this custom software for MAF but also uses somthing called ostrich which is an EPROM emulator. Basically u can pretend to be the EPROM with your computer and change all the values in the maps on the fly.
Diver944
Active member
Let's not forget that my 3.2L engine with a very quick spooling ball bearing SPS Turbo has no such problem. I can't begin to understand all the techie talk recently on this thread but I thought we'd established a while ago that Thom has some unknown problem causing his spikes. The Vitesse itself copes with whatever air you throw at it (unless the unit itself actually has a fault)
Fitted all the rest of the kit today, map sensor, piggy back and switch vflex to image #4 which is a less aggresive timing setup. Car works like a treat but now I am wanting an expert to check my tuning to make sure the fuel air ratio's are correct across the rev range. Any body suggest anyone in the surrey area who might have experience with the vitesse computer software that comes with this?
Posts made and opinions expressed are those of the individual forum members
Use of the Forum is subject to the Terms and Conditions
