What MAF Sensor

[barks944]

So I'm going to order an wideband AFR meter to help me do a MAF conversion on my car. The one I'm looking at is made by innovate motor sport and along with a dash mounted gauge gives you a 0-5v output for use with a data logger. I've got the basic design for a signal converter to convert both the Air Mass output and the temperature sensor output to the range the DME expects. What I'm not sure about is what MAF sensor to use. Most of the kits ive seen use a bosch sensor of some sort, does anyone know the exact model? I'm not set on a Bosch one, any appropriate sensor will do.

Tom

 

[Fen]

I believe along the way I've picked up that a Hitachi one is much more reliable than the Bosch ones which fail for a passtime in pretty much every car they are fitted to.

BTW I can recommend the Innovate. I didn't fit the gauge but I ran the wideband with the big box thing that datalogs.

 

[barks944]

Do you think its acceptable to tune a car based on the data from one of these wideband sensors? From what I know if the AFR is right then its properly tuned. I guess then its down to how accurate the sensor is.

 

[Diver944]

As long as you know it's accurate and calibrated when you do the tuning then yes, that's how I tuned mine. It is a good idea to get it double checked on a professional rolling road afterwards, just in case yours is reading lean (or rich) across the board.

I used the Dynojet Wideband Commander in mine. It does the same job as the Innovate but comes with a nice analogue gauge that is much easier to read on the move than a changing display of numbers

http://www.dynojet.co.uk/wideband/index.htm

 

[barks944]

That's good to know. How did you tune yours? Did you have a device to re-flash the chip on the 944's DME or have you also done a MAF conversion?

 

[Fen]

From experience of trying it, categorically no.

It sounds great in theory (and my EMS "self-tunes" which should make it easier), but in practice you need to target every cell in the matrix which means every RPM increment at every load level your EMS supports and that is virtually impossible on the road, especially when you consider it has to be done in a high enough gear to actually control the load and revs.

My Turbo was mapped on the road by someone who knows the EMS Ihave very well, using his own wideband sensor in the tailpipe and det cans to listen for pre-ignition, which had to be done to the redline and it 4th gear, which is north of 140mph in a Turbo (I was driving). That is not something I'd happily sign-up to do again especially as it was wet that day. While it ran OK it made another 19% more power after being set up on a dyno and the only difference in the configuration was a change from OE ignition to coil on plug CDI - which would account for some of the improvement, but not close to 20% as all it brought to the party was a fatter spark, not any advantage in timing.

Another thing to consider is ignition as the wideband is only going to help you with fuelling and that is only half the equation. You also need to tune the ignition timing to optimise the burn of your optimised fuel mixture.

 

[sawood12]

Most of the plug and play kits you get, which are worth getting, and are truly plug and play use a combination of coded DME & KLR chips that will get you 90% there and customised software which hooks into a piggy back computer which allows you to tweak the maps in real time as you drive to get that last 10%. From what I can gather simple MAF kits that are no more than MAF to AFR signal converters don't have a very good reputation which is why the best (and most expensive) plug & play kits use piggy back computers which provide you with a true MAF signal. It's pretty much like a Standalone ECU (though very specialised and limited) which hooks into the stock ECU and proides another level of intelligence which enables you to achive full MAF metering and hence all the benefits of MAF. Also with simple signal conditioning solutions you are limited to the 'clock speed' of the stock ECU - which isn't very fast by modern standards so you are not getting a decent sampling rate and therefore not really getting huge benefits over the mapped AFM. Piggy backs allow you to use much faster sampling rates and therefore take far more advantage of the increased sensitivity of a MAF.

Also there is evidence to suggest that you are not really getting much more benefit from a MAF until you get upto 300bhp or more. Below that there isn't much wrong with the AFM assuming the rest of your engine is working OK. In fact there is a chap on Rennlist who got very good HP (well into the 300bhp's I think) using an AFM with spot on AFR's across the rev range. I fancy the Vitesse MAF kit but not just for the MAF sensor itself, but for other benefits it offers including semi-sequencial fuel injection, use of large injectors, ignition timing based on MAP, utilising a wideband sensor for closed loop AFR control, utilising multi fuel maps and I believe his latest development is boost control. The fact the air sensor is a MAF is an added bonus on top of all the other features.

Again I don't know all the in's and out's, but am going off what i've read about others' experiences. The route you are going down is not a well trodden one, maybe for very good reasons - but good luck! From your other post it sounds like you've got the right background to be tackling such a project.

I'm not sure which MAF is suitable. Maybe if you look to current cars with similar engine capacity and HP then i'm sure they'll be suitable (Ford Focus ST comes to mind). The SciVision kit uses a Bosch one.

 

[barks944]

I think Ill grab a couple of cheap MAF sensors from a local scrappy and see how I get on. Anyone know of an equation which would model how the output of a hot wire MAF sensor would change with temperature?

 

[Diver944]

ORIGINAL: barks944

That's good to know. How did you tune yours?

Mine has a piggyback ECU from Vitesse which came with a fuel map that was pretty close for my engine. Then it was simply a case of spending some time on the road tweaking each cell in the RPM range to deliver the correct fuel under WOT, part throttle and also at idle. I simply drove the car with my Wideband recording the AFR and then adjusted the piggyback whilst pulled over in a layby.

 

[barks944]

Yeah this is essentially what I want to do, you might know the answer to these questions. Does the DME ignore the AFM signal at full throttle, is it only at full throttles it does this or at high air flow too? Do you know how the car alters the injection at different temperatures, does it do some calculation of can you have a map for a set temp range or somthing.

 

[Fen]

DME ignores AFM at full throttle, I'm not sure about other high air-flow conditions, but logically if the flow is enough to open it fully then the signal it is sending has no resolution so isn't useful.

Trusted posters have previously stated that the DME records a temperature at startup but does not dynamically measure temperature at all. That alone highlights to me that it is an archaic system and not one worth investing an awful lot in before it is replaced with something better.

 

[barks944]

Im wondering if I should output a constant voltage ratio for the temperature to the DME and do all the compensation for temperature in the MAF sensor. That way I can adjust it dynamically while the engine is running and it takes all the guesswork out of trying to alter the MAF signal by temperature only to have the DME re-adjust it, I can imagine it being a pain to tune in that case.

 

[Fen]

I seem to recall that a MAF has to compensate for temperature as part of its core design, or have I made that up in my head?

 

[barks944]

No thats correct. essentially you convert the maf signal to AF before sending it to the dme. You could also do some temp compensation by including some temp variance in the signal from the MAF. However if you allow the dme to do its own temp compensation as well it would likely to be very hard to tune as you would have two systems trying to do the same thing. if you fixed the temp signal to the dme you would fix its temp compensation to a constant amount effectively disabling it. However it occurs to me that the dme would use the temp compensation when at full throttle when then afm is being ignored. This would result in incorrect injection at full throttle despite any compensation done by the maf converter.

 

[Fen]

Understood. I think, given the DME only has start-up temp available to it as implemented by the factory, that the WOT situation you describe does exist, but it will just revert to stock behaviour. If you force the DME to see always the same temperature I don't think that will change much, so long as you fool it into seeing a reasonable figure.

 

[sawood12]

I think the DME ignores the AFM signal from about 3/4 throttle through to WOT and relys on RPM and throttle position mapping.

 

[barks944]

I wonder how much the amount of fuel injected varies with temperature at full throttle. I guess it should vary by w/e percentage the density of air changes over that range of temperature.

Temp C Density Kg/M3
−25 1.423
−20 1.395
−15 1.368
−10 1.342
−5 1.317
0 1.292
5 1.269
10 1.247
15 1.225
20 1.204
25 1.184
30 1.165

I guess from that it could be a reasonable difference over that range.

 

[barks944]

ORIGINAL: sawood12

I think the DME ignores the AFM signal from about 3/4 throttle through to WOT and relys on RPM and throttle position mapping.

I thought the 944 had an on/off switch for full/other throttle rather than a angular measurement of throttle position?

 

[sawood12]

Just double checked in the workshop manual and can confirm:-

There are three pins on the KLR which make up the throttle position/angle input: pins 21, 22, and 23

The RPM sensor is pin 27 on the DME.

Pin 2 on the DME is the 'Throttle switch idle contact' which is what we commonly refer to as the Throttle position switch, which is the microswitch mounted on the throttlebody.

Not sure where the actual device is that inputs into KLR pins 21,22 and 23.

 

[barks944]

I got this from FR Wilks Site: http://www.the944.com/connector.htm

Looks to me like it uses a switch for closed throttle and wide open throttle. Pin 22 is the AFM temp input from what I can see. What is the KLR?