j9fd3s was nice enough to provide some specs on the 13B-RE and 20B AFM, courtesy of the Cosmo training manual (in Japanese).



Y axis appears to be the ratio of the sensor voltage (Vs) to constant (reference?) voltage (Vc). X axis is volume airflow in cubic meters per hour.

The graph is deceptive because it has a straight trendline but the signal is more logarithmic/exponential, just like basically every other type of airflow sensor. This means that the sensor has the highest resolution at lower airflow levels. As airflow increases (voltage decreases), the sensor loses resolution. Here is an example relationship between volume AFM opening and voltage:



Here is a section on inspecting the S5 T2 AFM from the fuel & emissions control section of the s5 service manual:



Using our handy mass and volume rate flow calculator program (available as freeware here http://www.katmarsoftware.com/uconflow.htm), we know the following

13B-RE Cosmo AFM:

max volume airflow according to Mazda: 750 m^3/h or 441 cfm
max mass airflow @ 20C/68F & 1 atm barometric : 33.2 lb/min or 251 g/sec



20B-REW Cosmo AFM:

max volume airflow according to Mazda: 1000 m^3/h or 589 cfm
max mass airflow @ 20C/68F & 1 atm barometric: 44.2 lb/min or 335 g/sec



I have not seen actual published specs on any of the Rx-7 airflow meters, but it is highly likely that the Cosmo 20B AFM is an upgrade over an s5 T2 AFM in the sense that it was designed to measure more airflow.

 

S4 FC engines use the flap type and almost all of them use the 0-5v voltage scale:



The s5 and Cosmo use the sliding type:

 

that calculator is f-ing cool!

ive never held a 13B cosmo AFM, but the 20B AFM is physically larger than the S5 t2.

 

the 13B-RE cosmo AFM is physically the same size as the 20B AFM. They are both larger than S5.

 

Very nice. Based on various Rtek logs I've seen, the S4 TII AFM seems to top-out at about the same airflow as the 13B-RE: about 33.7 lb/min.

 

Just found specs on the s4 airflow meters:

s4 non turbo is 600 m^3/h
s4 turbo is 720 m^3/h

so the 13B Cosmo AFM (750 m^3/h is slightly higher flowing and the 20B AFM (1000 m^3/h) is way higher flowing.

 

oh sweet!

something something complete

 

Yes! Excellent info!

Now just have to figure out how to make the 20b afm work on the s4 ecu (not wiring, that's the easy part). My hope is that running 720 primaries and telling the ecu they are 550's should be enough.

While the s5 and 20b afm are both cone style they are designed differently. The s5 afm had a divergent path style bypass that goes around a cone that seals completely against the outer wall. By contrast, the 20b (maybe cosmo to?) Afm has no bypass tube. Instead the cone is smaller than the inner bore of the unit and never completely seals.

My guess is that the 20b afm bypasses a significantly larger volume of air (since a 20b needs more air to idle) which shifts its metering range up, and consequently makes the air fuel ratio too lean too idle on the s4 t2 ecu (w/ stock primaries).

 

I have been running a RE AFM for a while now. I never tested or dynoed the difference. But i'll only use RE AFM's on all my TII's now... The car ran so much smoother and seemed to pull harder on the top end.

 

I just got a NF01 (20b afm) from japan2la and hooked it up, so far, just driving it around it is so much smoother. I dont know why, but it is. I didnt have to change any idle setting when I switched to the NF01.
I ran it at the exact setting I had when I had the n350 and it ran fine, until at 8psi it ran horribly rich, which was the opposite when I had the n350 which ran very lean.

The NF01 is definitely much bigger then the n350.
On the Rtek, it seems max airflow is 36.7lbs, but I will see when I get back from work. The n350 maxes out at 23.3lbs

 

Isn't the s5 t2 afm the n370....n350 is non turbo.....might explain your lean spot and low airflow readings

 

N304 is the 84-85 GSL-SE, its a flapper, its small, like a 90-93 miata.
the 90-93 miata is a B61P i think

the 94+ miatas use a hotwire, so its totally different

N326 is the 86-88 NON turbo FC
N318 is the TURBO FC. both of these are bigger then the Gsl-se, both are the same physical size

N350 is the 89-91 non turbo. these are the first sliding vane AFM's, they were almost all replaced under warranty, the updated ones have a white dot on em.
N370 is the 89-91 turbo

N390 is the JC cosmo 13B
NF01 is the JC cosmo 20B

there are a bunch of KL/KJ meters that are from the V6 mazdas. these are physically cosmo sized.

 

Can the N370, 390 or NF01 be used on a S4?

 

The N370 will work with the S4 T2, I have idled and driven around with it, but I didn't push it , try to tune it or log airflow readings for comparison with the N318.

The N390 will not work with the S4 T2......errr.....it will not idle with the s4 T2.......on stock injectors. I installed it when the car was cold (1200rpm warmup idle) and drove around the block until the car warmed up, thats when I discovered that it won't idle, runs too lean (at idle) because it bypasses more air than the N318 (and N370 apparently). In theory larger primary injectors could fix this problem, but you would need some way to tune them outside of idle.

NF01 would be even worse than the N390 in terms of trying to work with an S4.

 

Well with my FD twin'd FC about to hit the knife, i might as well jump on this band wagon.
i'll be ordering the 20B AFM NF01.

as for the idle air port, well it's too big to idle huh? i'mma take a heat gun and melt it smaller until it does

can somebody beat me to it? ****, now i need an s5 afm pigtail.

 

This sounds like an awful idea...

You buy an afm because it is larger and flows more than stock, which unfortunately makes it impossible to idle. So your solution is to modify it so that it flows less to the point were it will idle, and thus flow roughly the same cfm as a stock afm.........?