rotarygod

Just wondering if anyone has studied the schematic of the RX-8 maf and it's voltage outputs. I am curious to see if there is another one from a different car that may be compatible. We need to find one that is more consistent across the board. The RX-8 maf is a really poor piece of work. They have inconsistent outputs and tuning is a pain because of this. Anyone studied this? Does anyone have the actual schematic of it? I'll look it over to find one.

Magic8

I thought Maurice was looking into a MAF replacement. Maybe RX-Friend and their posse can chime in.

zoom44

i am looking into it

WHealy

The MAF is the only part I have gotten without ordering thus far. It seems it's used on several Mazda's. Hence they had it in stock. Not exactly the question RotaryGod is asking, but wanted to add it.

IKnowNot'ing

RG,
I'd like to help. However I don't understand why on earth this MAFS should cause a problem unless :
- the intake system upstream of the sensor has been modified
- the engine runs in the backflow region (provided it's not a two way sensor that can handle it), in this case, the MAFS signal is NOT USED AS AN INPUT BY THE ECU, vut the ECU use default values contained in a table, whatever the MAFS reads!!

Can you please tell me :
- what kind of incensistency you measure ?
- where the engine operates (RPM, load) when this occurs?

IKN

dcfc3s

Dunno if this helps any, but the Greddy E-manage can change hotwire MAF's no sweat. Since it works with the RX-8's AFM just fine, you can set the dials on the E-manage to the RX-8's MAF, then select the new airflow meter type in the configuration page. The E-manage handles all the signal work for the new MAF, as long as it's wired up right.

Nissan has some big, quality hotwire MAF's - the one from the 300zx twin turbo (80mm diameter) and the early Infiniti Q45 (90mm) come to mind.

What are you trying to accomplish?

Dale

rotarygod

The reason I am curious is because there is alot of error between the maf's of different RX-8's. Some read a higher voltage at a lower or higher rpm than others. Surely, that's got to do something with the tuning. SSR has reported that the maf on a certain turbo RX-8 is maxing output voltage at about 50% throttle and is unable to be tuned above this which is what is holding up development. I want a better maf that doesn't have tuning issues for forced induction use. GM, Ford, Nissan, etc. don't have these problems with theirs so it just makes more sense to use one of theirs if possible.

I do have my own little plans for the future.

robertdot

Well, Ford seems like the natural choice if you are replacing it... keep it in the family.

That is, unless the same MAF is used on other Ford cars, which may or may not be possible depending on how much Ford wants to cost cut, and how much Mazda has to actually listen to Ford. I'm not familiar with the internal politics of Ford, but, anyway, maybe the designs of the MAFs are similar enough for an easy conversion.

shelleys_man_06

I agree with robertdot. What about using a Mustang MAF? JWT has a tech article about installing a Mustang Cobra MAF to replace the Z32's. According to the PDF file, the MAF will only work with the JWT-tuned PCM. The issue I see is how to program the RX-8's PCM to accept the Cobra's MAF. Is that possible, if not now, but in the future? What kind of effect would this have on the RX-8?

IKnowNot'ing

Was the signal measured with a voltmeter from the MAFS signal wire or using a CAN diagnostic tool?
If it's measured from the MAFS directly, the acoustic tuning of the Renesis and the specific design characteristics of the Wankel (no valve) might create back flow in many more areas than for a I-4 piston engine. The OEM PCM would then ignore these measurement and rely on other means to schedule fuel and spark (inferred MAF tables, linear lambda sensor...).

Is the MAFS placed upstream or dowstream of the turbo? Downstream would be a bad idea...



Maxing at 50% TPS under boost is normal! 50% TPS is more than 50% open area.

If the problem is back flow, you can try all the MAFS in the world without improvement, unless you can find a compatible bi-directional MAFS (these were still future technology when I was at Ford in 2000).

SWSDuvall

Hi all,

I've been reading about how the tolerance of the MAF on the 8 is wide, but the 8's ECU compensates. It must create a relative zero at intialization to compensate for each vehicles relative "zero". I want to build that adjustment into my Stage 1.1.

Using my Canzoomer Stage 1.1 and the Greddy software, I plotted/stored my MAF (Greddy lables them Airflow Meter Input/Output) voltages. I've attached the image. Notice the ignition timing varies as does the voltages. I put the cursor near the max value - RPM ~3700 and Airflow v about 3.7v so you can see some values to the right of the image. The unit IS working and I can feel it, but I want more :D

At idle, I am at 1.25v. As RPM varies, the Greddy's values change. The max value I captured was 3.7v at 7,600 RPM (2x the displayed # on the Greddy).

Here's my question:

Looking at the Ignition Timing adjustment map, I see Maurice has built the table from 0.0 up to 4.0 volts. Since my MAF has a "zero" of 1.25 volts should I adjust my table up by a relative 1.25v? By that I mean should I slide all the rows "up" so that the entries Maurice has under 0.0v would now coincide with the 1.3v line. And the entries at his "max" of 2.3v would now be 3.6v (2.3 + 1.3)?

Does this make sense? In other words, the values that are on the 0.0v line would be slid up and now be at the 1.3v line. All values "under" 1.3v would be set to zero or maybe the same as the 1.3v line all the waydown to 0.0?

The Airflow Adjustment map is based on RPM, so no changes there - I assume...

Any advice appreciated.

Thanks!

Steve
Near Seattle, Wa

zoom44

the previously mentioned voltage variances were observed by looking at readings measured by canzoomer and others using a can interface.


why would maxing output voltage at 50%thrrottle be normal?

Omicron

I'm really glad you brought this up. Rotarygod. I've thought for some time the MAF needs to be upgraded, but I've never had a chance to look into it. BUT, if Zoom44 is on the job, you can be sure we'll get it nailed down. :D

IKnowNot'ing

It does not tell me in which operational conditions this happens.

The OEM MAFS has been dimensioned for a NA Renesis giving at most (assuming 100% VolEff) :

1.3 l x 9000 RPM = 11.7 m³/min = 14 kg air /min, plus a margin.

If you give it say 6 PSI boost at 9000 rpm, you get a pressure ratio of about 1.4 ( (15+6)/15 ), hence the max mass air flow is now :

1.3l x 9000 RPM x P2/P1 = 20 kg air /min.

And, as 50% throttle correspond to more than 50% open throttle cross section area (say 65%), and as the P2/P1 pressure ratio remains equal, you get 65% x 20 kg air /min = 13 kg air /min, which is pretty close to the Max nominal MAF of the OEM sensor.

You can re-do the calc with the boost pressure they actually scheduled, and see how it goes.

NOW, if they blow through the MAFS, no calculation needed. This would be the cause then.

shelleys_man_06

I don't see how you jumped from 11.7 m³/min to 14 kg/min. What did you multiply the volumetric flow rate to get the mass flow rate?

Magic8

Density of air:

Density of air at 20 Celsius = 1.20 kg/m^3

11.7 m^3/min * 1.20 kg/m^3 = 14.0 kg/min

shelleys_man_06

Okay. There is another way to calculate that, but this one is easier :D.

shelleys_man_06

What is the difference between drawing through and blowing through the MAF, in terms of usable readings?

For the RX-8, would it be more fashionable to place the turbo before the MAF, or after?

IKnowNot'ing

> 0 kg/h mass air flow usually correspond with 1V signal.
> 1.25 V signal at idle seems right considering the non linear character of the response. At idle you flow some air.
> Max nominal mass air flow should be reached with engine running close to Max RPM, at WOT. Not at 3600 RPM or 7600 (unless you have a std power).
In these conditions (max RPM / WOT) the MAFS should return a 4.5V signal.

Usual tolerances for MAFS are +/- 3%

SWSDuvall

Thanks for the reply. The 0 airflow value set for 1.0v makes sense. I work with control systems as well is IT. All the 4-20ma sensors work similiar. At "0" flow/value, the device "puts out" 4ma - not zero. A zero ma value on the sensor indicates failure (below 4ma). So, having the MAF put out 1.0v at 0 flow makes sense.

To fine tune my question then, instead of adjusting the table for 0.0 as the starting point, it seems that 1.0v is the "zero". If I adjust the CanZoomer's ignitition advance map "up" slightly - moving the values in the table "up" (the existing zero values get moved up 1.0v), I will be delaying the ignition advance slightly.

Smart move or dumb?

Steve

IKnowNot'ing

I can't answer that. You should ask directly to the CanZoomer people. As it is RX8 specific (I believe), they probably did it correctly to start with. Remember that the Voltages I gave you are 'typical' voltages, not the RX8 specifically.