MAFless mapping

[jammapic]

Hi peeps,

Just FYI, I've been asked a few times recently to map out the maf sensor on the hdi 90. This has been for reasons ranging from "I'm fitting a big turbo and the pipe work is a ball ache", to "I'm too tight to replace it".

We did it on "the milkman"'s 307 because there's some fault with the wiring and I'd argue the car drives better than it ever has, smoother, more power etc etc.

Cheers,

JP

[Connor]

Was this just to let people know you were offering to do this? Or simply just a statement, or looking for people opinions on the subject? Im in no way trying to be a knob, just cant figure out what youre getting at?

[jammapic]

A bit of both really, just letting people know it can be done if required, and trying to spark a bit of conversation around the subject.

Personally, I wouldn't bother... The maf is there for a reason, but I guess I can see why people would...

[Dum-Dum]

I was about to say I'm not sure I'd bother, when they work they do 2 important jobs, firstly accurate airflow helps with fuel economy and second the IAT sensor is within it and loosing that means a whole correction table that the car will never get to. Can you replace them with a MAP sensor? That would be far more useful (and the 406 ones are only about £20)

[mcglynn]

we used t do this in the mg/rover scene.

the MAF's even OEM bosch ones where naff. They would get ruined very quickly with oiled panel filters (k&n etc) and left you with probems regarding boost not going very high due to low fuelling. We used to bridge two connectors in the plug with a paper clip as a temporary solution if the car wasnt mapped yet, essentially feeding 5v to te 0-5v sensor, and it fuelling max for that set revs etc. The cars when mapped were very much so MAFless maps, due to the problems with the maf.

It made for the car being a bit smokier at times and mpg MAY have suffered, but we still returned 50+ on them when drove right. basically it made the whole cars driveability depend on your drivers wish etc instead of the maf

[Uberderv]

Can you replace them with a MAP sensor?

I wondered that too. Still 0-5v output. Would be an interesting experiment.

[Martin306]

Can you replace them with a MAP sensor?

I wondered that too. Still 0-5v output. Would be an interesting experiment.

Yes and No. pressure and air flow are completely different.

Big turbo, low pressure can shift the same air into your engine as little turbo and high pressure. so an ECU map based on a MAP would be all to hell if you changed your turbo or if boost was different (due to atmospheric changes) etc. You would have "good and bad" days, and nobody wants that

Also, unlike a petrol, there's no throttle position so the airflow can't be calculated based on pressure. Petrol uses MAP + TPS to calculate air flow and it's not accurate.

That's my understanding anyway

[jammapic]

You could.

Many EDC15 based systems allow for MAP based fuelling anyway.

JP

[Dum-Dum]

Can you replace them with a MAP sensor?

I wondered that too. Still 0-5v output. Would be an interesting experiment.

Yes and No. pressure and air flow are completely different.

Big turbo, low pressure can shift the same air into your engine as little turbo and high pressure. so an ECU map based on a MAP would be all to hell if you changed your turbo or if boost was different (due to atmospheric changes) etc. You would have "good and bad" days, and nobody wants that

Also, unlike a petrol, there's no throttle position so the airflow can't be calculated based on pressure. Petrol uses MAP + TPS to calculate air flow and it's not accurate.

That's my understanding anyway

No mate, I think you might be having a bit of a brain fart but HDis definitely have a TPS. You can calculate boost by airflow and rpm into a known cylinder volume or airflow by boost and rpm into a known cylinder volume.

[Poodle]

Diesels don't need to work with the tps to calculate airflow because there's no throttle...

[Ruan]

Hmm, think people are getting confused :/

In a Petrol airflow is calculated with a VE table with regards to MAP vs TPS, their airflow rate is variable with throttle - it's a completely different management principle - diesels don't require a VE table since they aren't throttled by airflow.

EDC15 calculates the amount of fuel to be given with the "smoke" map - a coversion table to dictate how much fuel can be injected at a given airflow rate vs RPM - i.e. so you don't floor it and jam in 120hps worth of fuel at 1100rpm with of boost, it's going to smoke like a motherfucker. When you go to MAP based tuning, you literally swap this map to rather than being based upon airflow rate (from the airflow meter) you change it to be based upon manifold pressure. Some EDC15s have both however, a MAP and a MAF sensor for different situations.

It's all down to accuracy, the HDi90 in stock trim really doesn't need much other than a MAF sensor, it's a pretty simple engine that doesn't require that much boost, it's also run at a pretty lean AFR most of the time to prevent smoke. For better control, airflow meters ARE better, it's just that they go wrong more often and can be a pain to map.

[jammapic]

Been doing some serious work on this tonight.

I believe I've done the basics of the mapping required to get a map sensor based smoke map on the dw10 motor by using the maf sensor wiring.

It'll need some Fettling to get right once on a car, but it should work.

Could be useful for running multiple (or infinitely variable) tunes on the car on the road. I.e. Upping the boost will automatically up the fuelling.

JP

[DeeTurbo]

A bit off topic here but is it possible to control the wastegate with ECU via the EGR vacuum solenoid, is it worth doing?

[Poodle]

You'd need a vacuum actuated wastegate, but otherwise yes. Not worth doing imo, unless you're going for something other than power.

[DeeTurbo]

What about if you were going VNT?

[Midnightclub]

It is possible to have some form of VNT control, but it was never massively successful i believe..

[Poodle]

That'd be fine if you could find someone capable of mapping it properly.

[Ruan]

It's not that the ECU can't do it, it's knowing the right maps to adjust and knowing a little about PID control to be able to set up the control maps to hold the boost levels stable.

Doing what I did with the Arduino is almost exactly the same in terms of effect, only difference is that you're not using the main ECU to do it, you're using another ECU standalone to do - far as teh ECU goes, it's just the same as how a HDi90 is set up.

[jammapic]

It's been done a couple of times I believe.

The HDi 90 ECU has ALL of the correct maps to run either a wastegate turbo or VNT turbo, both on ECU control. It needs a couple of flags setting to tell it to do so, the right hardware putting in place - namely a map sensor and a solenoid valve, and it needs the correct maps optimising to run correctly without bring up the EML etc.

I had my old HDi running on ECU control with the wastegate turbo - but I missed a couple of limiters and it kept dropping into limp mode. I have subsequently now found the right bits and will get round to having a go at making it work properly at some point on the test car.

Worth noting though, like poodle says, it won't gain you any power per se, it will however give better drivability, probably reduce fuel consumption a bit (reduce emp a bit by only boosting when it's needed) and allow more granular control over what boost is needed when.

As above, the other thing I am working on at the moment is replacing the maf sensor with a MAP sensor for smoke control.

In an ideal world, I'd like to have the car running on 2 map sensors (maybe one if I can work out how... Ruan??) to control both the smoke and turbo. You'd then have a car with totally customisable boost levels in the mapping, and you could set a target AFR based on boost...

JP

[Ruan]

That's exactly how the EDC15 VW PDs do it when you unplug the MAF sensor, they have VNT turbos, so therefore have MAP and MAF sensors, but use the MAP sensor when the MAF sensor is in unreadable - they do MAFless tunes also - I need to look into how they do it, that'd give us a good idea.

Though you said IIRC you'd changed the wiring on the MAF to accept the 0-5v MAP sensor - did you install the MAP sensor in the place that the 110s have it? There should be some switches to enable MAP Smoke Limitation *and* use it for boost control at the same time.

Should only need 1 MAP sensor then.

"VNT" control isn't a something you enable in the ECU like most people assume, you use all the PID control maps that the 110HDis use, but you have to completely re-do them - it's the same thing, but just controlling something totally different, thankfully there's lots of EDC15 ECUs out there that use VNT turbos that we can yoink calibrations from - at least for a starter point.

[jammapic]

Yeah, in a nutshell you have to flip the calibration backwards for VNT rather than wastegate - plus the response is completely different.

You can do MAF deletes on VAG stuff and use MAP for smoke and boost control - I've done it, it's literally a switch and it reads a whole seperate smoke map and ignores the maf table.... On an early VAG EDC15, when you unplug the MAF sensor it does the same as ours do - reads a default nominal value, on the later ones it reads the map based smoke map as above. I expect it temporarily "flips the switch" but, they use EDC15P - which has a bunch of other features which C2 doesn't. One of which is, I believe, to be this. Maybe I'm not looking hard enough - I'll keep hunting.

I have chopped a maf sensor off, and soldered directly onto some wires. It meant re-jigging the smoke map MAF axis, but it should work. Haven't had time to test it properly yet.

At the end of the day, if we have to have 2 map sensors, so be it... but it would be nice to only use 1!!

JP

[pro_steve]

In my opinion MAF is important, the ECU makes small timing corrections based on air mass and temperature, you'll probably notice that MPG is better with a (fully working) MAF.
Sometimes It can be a hassle for example when broken or when flowing above the readable range when you've got a big turbo.

However, MAF based smoke limiter will always be a lot more precise than MAP based smoke limiter, I think they only really got popular because of the VAG boys who like to fit tje 3 Bar sensor and then use MAP limiting.

The 2.2 HDI has a much bigger MAF, I'm going to try and calibrate this into a 306 with a K03s, hopefully this should then allow for precise control of AFR after 4K RPM where the stock MAF just can't cope. Likewise on my 1.6 HDI, the MAF can't read high enough for even 150 BHP worth of air, this kind of sucks because I either give a lambda ratio of less than 1, or I have to calibrate in a new sensor to keep it smoke free. It all depends on what you're trying to acheive I suppose, on a stage 3 then yes it's a good plan to remove complexity, on a stage 1 it's best to fix the fault rather than masking it surely?

[jammapic]

Agree completely Steve.... but some people are too tight to replace sensors :-p

[pro_steve]

But the problem with masking the fault is the car will always be smoky/bad on fuel, when they ask who tuned it it comes back to you doesn't it. I've done it a few times when setting up a car with a bad MAF, but I do it so that once a working MAF is plugged in it over rides the setting, probably be easier if I kep a pile of MAFs in stock?

[Mr Whippy]

But the problem with masking the fault is the car will always be smoky/bad on fuel, when they ask who tuned it it comes back to you doesn't it. I've done it a few times when setting up a car with a bad MAF, but I do it so that once a working MAF is plugged in it over rides the setting, probably be easier if I kep a pile of MAFs in stock?

I agree Steve.

If you can't afford a MAF, or just go to a scrapper and pick up a few cheap and test them out (95% of the time MAF are fine over the hundreds of cars I've seen) you are not gonna be able to afford a MAP sensor and the time/effort to do that, AND then have the appropriate remapping changed which is a faff, to suit the mods.


MAF is only really present because of EGR, EGR needs to get AFR really accurate to keep emissions perfect.
So you could argue MAF is the best solution all round. The problem is sizing at big power. But the standard MAF in theory is calibrated (that means it meters pretty accurately) out to 180bhp... even with some non-perfect metering it'll be fine to 240bhp purely on the cross-sectional considerations of restriction.



I'm fairly certain there is a MAP in the ECU that is what is used when the MAF is un-plugged, that is basically a Look Up Table for MAF values when they are not provided. To run without a MAF is basically running off another MAP that guesses what MAF value will be present at certain RPM and throttle inputs.


Like Steve says, you are 100% better with the MAF present than not because it is measuring all sorts of variables and real air mass, vs guesses mapped into a LookUpTable (LUT).


As per boost adjustments impacting fuelling, you can do that too with MAF. More boost = more MAF flow = more allowed fuelling.
You can punt around with 0.9bar boost and about 130bhp, then as you turn the MBC up to 1.3bar you get about 155bhp. Weeee!
If the smoke map is populated properly with good values that is how it should work all the time surely!?

That is if you don't use the boost control on the ECU, which is as much a pain in the arse as anything.
Remember that it's values are hard-set (calibrated) to the smoke map, and advance maps and everything else. It's a circular tuning affair. Changing boost request low down will impact restriction on the turbine, lean out the AFR, change the ideal advance point in steady state etc etc...

I guess if you went really wrong you could end up with a smoke map set such that the requested boost might never be met, or not met so easily, because the AFR allowed via the smoke map isn't enough to get turbine energy up to achieve the required boost.



In the end I think unless you are gonna fit a 200bhp+ engine with a VNT turbo, then keeping everything standard HDi90 with a working MAF is perfectly tune-able and fast and good enough.

MAF-less is do-able but just unplugging it and tweaking the appropriate map is 'ok' if you are THAT tight for cash (for a £15 spare from a scrapper)

MAP seems a backwards step for me. It's slight advantages are outweighed by a complete faff of work. If you want mega power just use a bigger MAF. Iirc some C5's use a MAF that will meter accurately out to 240bhp or so of air, and they are a fair bit fatter too, probably 280bhp theoretical air flow limit.

Dave

[pro_steve]

Dave, I agree with you on most of your points, except that MAFs usually work; I would say 50% of the cars that come to me have a MAF problem. If you know what the default MAF table values are it's quite easy to spot, they are also different on a 306 or 206, so a 306 would be down on power and a 206 you wouldn't notice it.

I'm going to try and calibrate one of these in to Ian's car with the k03s hybrid http://www.ebay.co.uk/itm/Citroen-C5-2-0...2a34447df0

Think I will buy a new VDO unit so that we know it's not got a problem with it, the cheap units tend to cause problems, I think it's mainly because they read the wrong air flow, if EGR is still connected then the EGR will open to reduce air flow and make the car chug etc etc.

Bottom line of it, only get rid of the MAF if you're above a stage 2 and that's if it's giving you massive problems.

I've got a strange feeling MAF readings are affected by fitting a cone filter, the air doesn't flow in a laminar manner as it goes through the sensor and I don't think it reads properly. I'm going to fit a 90 silicone hose between the cone and MAF on my brothers car to see if it changes low down response (don't shoot us for having a cone filter in there, it's so that we can fit other parts into the tight engine bay! haha).

JP, you asked us to get some conversation going on the matter, so hope you found this interesting?

[Uberderv]

I thought the grid on the Maf inlet was to create laminar flow........ Some good info here chaps

[pro_steve]

Yes but I have worked with wind tunnels before and the the 'grid' would be a lot deeper, obviously the original pipe work takes it through a series of pipes and makes the flow nice and laminar. Similar to the intake for a fighter plane, we have to try and prevent flow separation and encourage laminar flow by changing the design of the intakes.

I think with a 45 or 90 bend there will be enough pipe length to bring the flow laminar rather than sucking from the edges of the cone filter. Will test when I get a chance and if there is a suitable sized piece of silicone laying around... I think Mr postman is bringing some 45mm but I'm not sure what the MAF diameter is on the 1.4 HDI so it will be pure luck if I have anything to go on there.

[jammapic]

JP, you asked us to get some conversation going on the matter, so hope you found this interesting?

It makes good reading for sure. It's stuff I had kinda already worked out, but very interesting to see about the cone filter and air flow issues....

There's no replacement for theory and practical testing... at the end of the day, you can really only work out how these things behave when you test them.... that for me, is the fun bit!

Back to the original point I made though, in some circumstances it's beneficial... "The Milkman's" 307 is driving better, with more power, than it ever has on a MAFless map.

I don't disagree at all though that ideally, you'd have a fully working, decent quality MAF in there. After all, it's there for a reason....

JP

[Ruan]

Totally agree Dave... Though surely we're not having issues with recalibrating MAF sensors? The 2.2HDi comes with a 1000kg/h sensor - surely we can nick the linearisation map off it and extrapolate from that - or is there other things that are referencing off the MAF Linearisation axis?

MAF is such a better way of doing things from a tuning perspective - as far as I can see the only reason you'd want not to is if you can't work out AFR and understand how the engines VE changes. Only other reason for a MAP sensor that I can see is to notice if a boost pipe comes off and stop people billowing smoke... Almost an expected boost vs MAF - again, I think the EDC15p TDis do this.

But obviously when you start introducing PID control for VNTs etc, you need one for feedback - though I'm looking at the possibility of rather than using an Arduino - which although is a nice piece of kit, it's just not designed for the harsh environment of a car... Using instead a proper PLC that can do PID - that's what they're designed to do and are of much better quality - probably a bit more expensive, but rather than having the associated issues of not knowing what references what inside the ECU, be able to concentrate on making the ECU purely deal with injection data.