maf transfer curve - dual mafs - question

[Yves]

I posted this here, but I'm not sure if this is the right place, so admin, if this needs to be moved, please do.

As some may have read, I'm set on using a dual maf setup on my car because of the ITB crossram manifold that is attached to 2 seperate airboxes. I just laid my hands on a circuit to get the values from each maf averaged and to have a single voltage output to the ecu (with thanks to Don !). In essence this means that if one maf puts out 4,7 V and the other 4,5 V the average output to the ecu would be 4,6 V.

So far so good, but I need to know what the total amount of airflow is through the 2 mafs and relate that to the voltage. That much I can understand.

I have a system coming from Don that incorporates 4 map sensors and I still have my baro and the map on my board.

I know that a normal flowbench would measure the airflow by virtue of the pressure differential over a sharp edge orifice. Could I measure the airflow with the 4 maps, the baro and the manifold map based on teh pressure differential say over the maf sensor ?

[Laminar]

Do you know the correct transfer curve for one MAF on its own? If so, just multiply the flow value of each voltage by two, it will reasonably approximate your flow. The closer the two MAF voltages are to each other, the better the accuracy, and the error decreases as voltage increases.



I did some playing around in Excel to confirm, though this board won't let me attach the sheet.

If you don't know the curves for your MAFs, you have a lot more work cut out for you. I've found lots of people wanting to build flowbenches to test their MAFs, but very little info on how to do it reliably.

[Yves]

I have 2 LS1 voltage mafs (90 mm).

Actually how I intend to measure them is based off this, which is the setup of a basic flow bench :



I'm thinking that the pressure at point B is the baro reading. The pressures at C and D could be done by inserting a sharp edge orifice of a specific size in the inlet tract after the maf and with 2 connections for the extra map sensors. That way the airflow could be calculated I would think.

[Laminar]

Are you doing this to figure out the curve of the MAF or do you plan to run like this?

[Yves]

Just to figure out the maf curve

[Laminar]

And no one has published curves for this MAF?

[Yves]

And no one has published curves for this MAF?

the way I understood it, the curves are very dependant on the intake system (tubes, length,..) they're used in.

[Laminar]

Okay, I see what you're saying. So you could temporarily run an orifice plate in each intake tract and compare the calculated flowrate from the pressure drop to the voltage signal from the MAFs to get an idea of the transfer curve of the MAFs as they sit in your system.

I just wonder if the assumptions and inaccuracies inherent to the pressure drop calculation and subsequent removal of the orifices from the tract would be more or less error than just rolling with the given MAF curve to begin with. Either way you'll probably need to tweak the curve once you start tuning.

If you do get a MAF curve flowbench set up I'd be very interested in seeing the results and can help double check orifice calculations, my company has some proprietary calculation software we use for orifices/venturis/annubars.

[Yves]

Okay, I see what you're saying. So you could temporarily run an orifice plate in each intake tract and compare the calculated flowrate from the pressure drop to the voltage signal from the MAFs to get an idea of the transfer curve of the MAFs as they sit in your system.

I just wonder if the assumptions and inaccuracies inherent to the pressure drop calculation and subsequent removal of the orifices from the tract would be more or less error than just rolling with the given MAF curve to begin with. Either way you'll probably need to tweak the curve once you start tuning.

If you do get a MAF curve flowbench set up I'd be very interested in seeing the results and can help double check orifice calculations, my company has some proprietary calculation software we use for orifices/venturis/annubars.

That would be great.

I see your point regarding the orifice. It's in my opinion however the flow through the maf that counts.

The alternative would be to attach a vacuum cleaner or to try and see what it's doing on the scatterplots.

[subwoofer]

No need for messing around with vacuum cleaners and orifices, tuning the MAF curve while on the engine is actually easy - provided you get it started in the first place. But the leaway there is actually larger than you think, with MAFs you don't have the lean/rich hunting idle problem you have with a poorly tuned VE table.

http://www.msextra.com/forums/viewtopic ... 34&t=49504

It appears the illustration links have been changed somehow, I may go back in and edit one day.

[Yves]

Sub,

That seems actually quite good.

What that about the resonance ?

[subwoofer]

The resonnance problem is for all intents and purposes gone with the MS3 1.3 release.

I just installed a factory fresh MAF today, the old MAF curve I had was compensating for a slowly dying sensor. I put the new one in, the engine coughed a bit as it was getting way too much fuel. By the time it had warmed up I had decent idle with EGO running. I then guesstimated a midrange, left EGO at +-25% authority and went for a 10 minute drive. I am now going out for another short drive, 15 minutes-ish, that should put me into the +-3% fuel territory. Try that running SD...

[subwoofer]

This is the result from my most recent drive, after 2 iterations on the MAF curve. Above 3.5V the curve was a just a "think of a number" type curve, but it turned out it wasn't too shabby after all. When too rich, MAFcorr accuracy goes out the window as the AFR readout saturates. After correcting it now, I expect +-/2% fuel accuracy up to around 4V MAF voltage, that equates to WOT up to about 4500 RPM for this engine.

[Yves]

Thank.

Would you be willing to guide me through the process when I get there ?

[subwoofer]

No problem, will be as helpful as I can.

[bradyzq]

Hi,

I noticed a couple of things that seem to have slipped under the radar.

First, LS1 MAFs are frequency, not voltage, so the circuit to average the outputs will be very different.

Second, the LS1 MAF transfer curve should be available as a pull down option in TS, or so I've read. The housing is 75mm stock, so if yours are in 90mm housings, you'll have to scale the transfer curve.

Using one MAF and halving the displacement should get you running, with fuel trims doing the rest of the work to dial in the curve.

How much power are you expecting to make? Twin 75mm MAFs can meter a LOT of air!! And twin 90's, wow, 40+% more than "a LOT."

[subwoofer]

I don't know the LS1 MAFs, but I think I would recommend voltage output units, I don't know the state of the frequency code.
90mm MAFs can support ~900bhp or so - each, are you really building a 1200+ hp engine? You give up idle and cruise resolution for a top end you may never need or reach.

[bradyzq]

This exists, aimed at the 300ZX TT crowd, but it uses Bosch connectors, so I imagine it would work with any Bosch MAFs, though stock 300ZXTT MAFs seem to be 30$ on ebay.

http://www.selindesign.net/products/2nd ... translator

[Yves]

I think the ls1 voltage mafs were used untill 2000 or 2002.

This is the type I'm talking about :

http://www.ebay.com/itm/3-5-LS1-MAF-for ... f8&vxp=mtr

However, now you have me wondering if the curve in TS is for a voltage or frequency maf

[Yves]

I don't know the LS1 MAFs, but I think I would recommend voltage output units, I don't know the state of the frequency code.
90mm MAFs can support ~900bhp or so - each, are you really building a 1200+ hp engine? You give up idle and cruise resolution for a top end you may never need or reach.

I don't think the conclusion can be drawn based on hp. The maximum flow is 512 gr/sec I believe. I found that to equate to around 700-800 cfm. The engine needs at least 800 cfm at maximum hp. The additional feature is I use 2 airboxes because of the crossram ITB manifold.

Of course it depends on the airspeed you want in the inlet. If you allow a lot of airspeed and even some vacuum, you might squeeze by with just one. A calculation I have made has shown me that a 90mm tube per bank is just enough in order to reduce the airspeed and the vacuum.