The DSM has a Karmann Vortex "Mass Air Flow" meter (MAF). The principle behind von Karmann shedding is a complex relationship between fluid cohesiveness, surface shape, cross sectional area, fluid velocity, etc. etc. The important thing to realize is: if air is moving fast, it will want to tend toward being turbulent at a somewhat linear rate based on how much air is moving through an infintely thin cross-section and how fast the air is moving. Once the air is turbulent vorticies appear in the fluid stream and it is these vorticies that the MAF measures. It counts the number of vorticies and returns the value in a cycles per second measurement (Hz = 1/s). Using a table of known temperatures and pressures vs. the flow rate established as a frequency domain signal, the engine computer can make fuel trims, timing trims, etc. This frequency signal is what these people are qouting.

Without knowing the tables, it is rather meaningless. Other than to say, "Holy shit, I have a lot of air moving! Its at 1500 Hz and the last run it was at 1200 Hz!" But you cannot say, "I have 276 cfm of air moving through my throttle body this run where I had 220 cfm last time."

 

So would the "corrected" lb/min figure be a good way to compare flow rates of different turbos/cars?

 

The basic MAS info is in the Tech Info Manual in Chapter 14, "Fuel System," pages 14, 15, 16, 17 and 18.

The MAS takes three different readings, baro. press., temp. and Karmen Hz. It sends these three signals to the ECU, then it figures out the actual air mass or lbs/min and injects the right amount of fuel for the given air mass and rpm.
Karmen Hz simply shows air flow. It is one of the three things needed to figure out the lbs/min. The temp and altitude/pressure will figure into the actual lbs/min of air moved.
So when you "plug" in a S-AFC, you are only molesting the Karmen Hz number.

So IMHO, if I'm "here" and your "there" and we are at different altitudes/pressures and temps then we will see different lbs/min readings. Whereas the Karmen Hz reading wouldn't show these "ambient" differences; only the air flow sensed moving through the Karmen Hz sensor.

Now you also have a secondary problem, MODS!, and perhaps state of tune. IMHO, you would have to swap one turbo for another on the same car and do a comparative run at the exact same temp., pressure and even load for lbs/min to be a decisive measure. Whereas Karmen Hz is a "pure" airflow reading.
Then you have at least three different MAS's. (1g & 2g and hacked mas's.) ie: you couldn't compare my Hz readings with anothers as I have a 2g mas on my 1g, now what about a hacked 1g mas where alot of air is unmetered.

If I were you, I would only pay attention to your Hz readings and look for patterns, abnomalies, etc. Let your timeslip or a dyno run tell you what you are really looking for, IMHO.

BUT, when choosing a turbo for purchase. That is when you wanna look at lbs/min a compressor wheel is capable of moving, NOT cfm.


[Edited by Van on 12-18-2001 at 12:40 AM]

 

Thanks guys.

Jon

 

Some more help?....I'm still confused....

dsm4eva and Van

As opposed to responding via Jon, he signed me up for my own account....so we'll see how this goes.

I like the discussion, but I am still confused regarding the Hz measurement being a "true" or "universal" measurement. I am not saying this to defend Jon or to become confrontational, this is simply something that I want to understand and I am still struggling with the explanations.

For the very fact that the ECU needs the pressure and temp measurements coupled with the Hz reading to determing mass flow makes me believe that this measurement (MAF in Hz) is not independent of ambient conditions. I guess I am thinking more along the lines of the behavior of the air and its transition to turbulent flow and what not. I would argue that its turbulent behavior will differ with temperature and pressure and therfore the number of counts via the MAF would change based on these condtions.

I am also thinking about the change in the air's viscosity with temp and pressure and how this would change the air flow based on friction losses through the intake and exhaust pipes. Anyway, I feel like I'm going round and round...so let me know what you think.

On the conversation about mods....I totally agree with your thoughts. The entire "plumbing" system for intake and exhaust could have a huge affect on the flow resistance imposed on a turbo and it is not fair to compare turbos unless all these variables are the same.

 

I think WRX is over complicating things. Hes, HZ is corrected for weather conditions (baro and temp). That is why you shouldn't look at your HZ readings ONCE and assume that is indicative of the turbos capacity. You observe the number over a given amount of time and see the variance and then quote the range. Unless you are seeing some sort of extreme weather that no one else in the world sees, then it will not make a difference.

Plus, taking lbs/min from the logger and modifying it to remove the correction from the AFC is impossible.

 

Plus, taking lbs/min from the logger and modifying it to remove the correction from the AFC is impossible.

Really? Why is that? I've been hoping for somebody to comment on that equation I drew up either pro or con because it was just an educated guess.....you're the first one, so let me know why it cant be done...can it even be done in a general comparison way, such as we are doing with Hz?

I assume its becuase the ecu is taking the hz number and putting that into an equation before it spits out the lb/min...and without knowing those equations, its impossible to compensate?


ps, if it is impossible, I'll definately go in the level 1 review and correct myself so im not misleading anyone.

[Edited by 99mmcgsx on 12-18-2001 at 05:48 PM]

 

I am no expert in this area but let me throw in my 2 cents. Per what Van stated, there are three variables (baro pressure, temp and karmen) that determine lb/min. If baro pressure and temp are constants, then using formula to do the correction makes sense. However, in real life, I dont think they are. And I dont think airflow correction is accurate also.

Lets take for example using 660cc injectors w/ AFC for correction. Since 660cc is approximately 50% larger than 450cc injectors, does it mean we should use almost 50% airflow correction on the AFC? Most definitely not, the most I have heard of is 30% correction.

 

I never knew that th 2G spit out actual lbs/min readings on the logger. If it is at all acurate then you can get a great idea of flow rates of a given turbo/parts combination a a given boost level. ( of course when a figure is quoted you would have to supply an air temp and altitude so people would be on the same page) You could also cross reference your 1/4 ets to your flow rate and figure out a rough volumetric efficiency for horsepower on your particular vehicle. If enough people did this we could compare and figure out how similar each motor runs car to car, or if not at all. I like what you are on to Jon. This is very interesting. Even if not as a whole or for comparison basis i think it could help with power tuning and parts choices for your own vehicle if it could be proven at all consistant.
And, theoretically, you could take a 1G Hz signal, and the baro, and temp, and use an equation to figure the lbs./min also. Is any one on the same page? I find this exciting. What i cant do is figure out the equation for the 1G.
ALSO. If this is at all accurate, you could run your given turbo ( we'll use Jon as the example) at a random boost pressure, and see if the lbs./min rating is at all close to the compressor map of a given turbo. So, in Jons case, in which ( correct me if I am at all wrong) flows a maximum of 36 lbs/min. You could run it at say 20 psi and see where this falls. Or possibly how close you can get an actual 36 lbs/min. Do you follow? Am i rambling?

 

This post might not make sense at first, but I know why you guys are talking about this. Check any of the eighty million turbo threads on the board if you don't know what I'm talking about.
I think you guys are completely over analyzing things. A turbo is a turbo, yes some are better than others. Pick one that seems to suit your et goals and pricerange. Talk to people who have the turbos you're interested in. Use the search button, read the times page etc..
Then worry about the rest of your setup. Having a nice turbo is great, but it won't do shit if your FMIC has too much pressure drop, you have a unported o2 housing, etc.. Get rid of any bottlenecks in the turbo system you can find.

Tune until you can get rid of knock. Go to the track and put in C16. Up the boost. Tune, tune, and tune some more. Electronic gizmos and gadgets are great, but remember it's an engine controlled by a computer, not a computer.

 

Forgive me if my terminology sounds incorrect..........
What I seem to gather is that the Hz is soley dependant an actual VOLUME moving through the meter. Thats how i've always grasped it. And while volume is significant, knowing how DENSE it is is dueley important. Therefore the meter determines density based on altitude and actual tempature. Both of these affecting density.
Soooo...... (This is entirely speculation) would you tend to believe that no matter what the altitude, air VOLUME would remain consistant. Its DENSITY would change, due to altitude and tempature, but that wouldnt be reflected in the Hz signal. If there is another factor here, or anything im overlooking please respond. The horsepower loss from elavation and tempature is due to lack of DENSITY, not loss of volume.
For comparison purposes, to people would have to share equally configured meters to be able to compare results. For 2G's this commonly won't be a problem.
Now, i'm not certain about the meters themselves being calibrated differently from meter to meter. How off could they be? They have to be within some sort of tolerance. They all have to be run by the same computer.
I'm not so sure how to determine if the air flow reading from the logger can be corrected to determine actual air flow, but simply adding or subtracting a percentage ( altough i realize your equation was more in depth than that) doesnt seem to be the most effective route, though I certainly dont have a better solution.
AN IDEA............. Say you make a pull and at 6500 rpm it shows a flow rate of 30 lbs./min of airflow. Then you richen up 6500 on the AFC by 10 ( assuming you were negative in the first place). This will put you closer to the true value, correct? You might be able to see a ratio reflected somehow. (of course the car wont run so well) Do you follow? If you put all your settings at zero ( with over sized injectors that NEED to be leaned out ) and made a pull, although it will run poorly it should show ACTUAL air flow.

 

This post might not make sense at first, but I know why you guys are talking about this. Check any of the eighty million turbo threads on the board if you don't know what I'm talking about.
I think you guys are completely over analyzing things. A turbo is a turbo, yes some are better than others. Pick one that seems to suit your et goals and pricerange. Talk to people who have the turbos you're interested in. Use the search button, read the times page etc..
Then worry about the rest of your setup. Having a nice turbo is great, but it won't do shit if your FMIC has too much pressure drop, you have a unported o2 housing, etc.. Get rid of any bottlenecks in the turbo system you can find.

Tune until you can get rid of knock. Go to the track and put in C16. Up the boost. Tune, tune, and tune some more. Electronic gizmos and gadgets are great, but remember it's an engine controlled by a computer, not a computer.

Nick I understand all your points here and basically I agree with you. Definately if you take a step back, the quarter mile times and dyno results are what matters.....no matter what gadgets or equations you are using. I totally agree with you there.

I guess thats not really what I'm getting at here. I, like quite a few people here, are pretty sick of waste-of-time threads about why one turbo is better than another. The reason they are a waste of time is because people have no way to really say whether the turbo is good. Yes the quarter mile is best for overall car performance but we all know that number is based on a ton of factors....the turbo is just one. So what happens is you end up basing your opinion of a turbo on how many adjectives somebody uses...."it pulls really really really hard!" OR, you base your opinion on their quarter mile results....which we all know doesn't depend on the turbo alone....just think of how many turbos have gone 12.5 in the quarter....many different kinds. So what happens is when somebody gets a new turbo (like me) and tries to write a review of it based on turbo-specific data instead of quarter mile times or adjectives, you end up with nothing really to go on other than Hz....which I dont think is that accurate, and I for one can't log it. I just think it sucks that when you buy a turbo, the shop website quotes CFM and airflow and all that technical stuff, but once it is on the car nobody really verifies that....the technical aspects are over basically and it becomes a war of words, not data. It just makes it really hard for people to choose a turbo IMO. Case in point...the level 1 review Ive been writing. Everybody wants numbers, not opinions...and I want to be able to give them actual data. The problem is, I think everybody is waiting just for a quarter mile time....which again, is not really the best way to judge one turbo vs. another. I want to be able to show some other data that is more useful....thats why I was hoping to get some actual airflow data and airflow curves from the turbo. (and yes, I understand that even these numbers are dependent on other parts on the car, but its still better than quarter mile time for comparing turbos)

The more hard data that common dsm people can get from their cars the better, because it will push the manufacturers into designing better products IMHO, and it will help other dsm people have an easier time choosing a turbo.

Jon

 

Though it would require a dedicated investment for no other purpose than monitoring, I would like to see air temps at given boost pressures. If more than one turbo could be tested it could give some interesting data. I doubt this will ever happen.
Of course you would place the probe before the intercooler.

 

Well, simply put, as Van and DSM4eva states, the Hz read by the MAF is related to velocity of the airflow through the MAF. And with a fixed MAF x-section, that means it's related to the volumetric flow rate. That coupled with the pressure and temperature will allow you to calculate a mass flow rate. Hence if the ambient conditions are similar, then a Hz comparison ON THE SAME MAF will allow a simple determination as to WHEN more air was flowing. How much more air would be a tad tedious to calculate.

 

When Garrett develops compressor maps for their turbos they use lbs/min instead of some measurement of volume. WHY?

MAF and MAS are used to calculate what?

MASS!

Its all about air mass.

When you measure the mass of something, you can know exactly how many particals of that substance you have.

If you hold your air/fuel ratio constant, lbs/min has a direct relationship to HP.

The karmen readings on your AFC are nearly useless without knowing the density of the air.

Forget Hz

Use the lbs/min reading. It will tell you everything you need to know without all the complicted calculations.

 

Forget Hz

Use the lbs/min reading. It will tell you everything you need to know without all the complicted calculations. [/B]

I want to use lb/min....BUT.....this number is not correct if you have an afc because you are modifying Hz. Hence the calculations.

The thing I take issue with is that it is 'impossible' to do a conversion to get the 'true' lb/min reading after the afc has converted the Hz number. I think it is possible and I haven't heard any real reasons why not. I was talking with WRX and basically here's what we thought:

The lb/min figure given out by the ECU is based on three things as far as I can understand: air temp at the MAS, air pressure, and airflow in Hz. It's been said that you can't convert the 'molested' lb/min figure back to a 'true' reading because the ECU is using these three things to give a lb/min figure...not just Hz, which is the only part I'm using in the equation way at the top of this thread.
Here's the deal though: It doesnt matter if the ECU is factoring in temp and baro pressure, because for all intents and purposes they stay constant during a WOT run. My temps only change by maybe 1-3 degrees during a 2nd and 3rd gear pull, and air pressure will not change at all in the 10-15 seconds I'm usually logging.
SO, with those 2 things (temp and barometric pressure) being basically 'constants', I can see no reason why you can't use the equation I used in the beginning of this thread to re-calculate your ECU's lb/min figure into a 'true' lb/min figure.

Make sense?

 

Some agreement?

gsxalex, I couldn't agree more.
I don't know if anyone has read the original thread by Jon, but this is how this converstion got started between him and I....with doubts about the Hz measurement being used for turbo comparison purposes.

While I don't want to speak for everybody, it appears that we can agree that the Hz measurement is some version of volumetric flow rate and it isn't until coupled with temp and pressure that it gives meaningful information.......mass flow rate.

The turbo manufacturers put together their own compressor curves for their turbos in units of mass/time. In order to determine where on these compressor curves that one's car is going to operate, you need to know what kind of resistance is operating against the turbo. This resistance starts at the air filter and ends at the exhaust tips.......My point to this rambling? There are soooooo many factors that play into a turbo's ability to push more air mass through the motor that judgements & comparisons can't be made unless everything is considered (not just the turbo).

So I agree, the Hz measurement shouldn't be used for comparison purposes. Sorry for beating a dead horse....just wanted to add my 2 cents.

 

I dont think your equation will work because you are over simplifying the way that the AFC works. If you could somehow get the equations that the AFC uses to modify the HZ signal, then it might be possible. Specificaly, if you have a different setting on the AFC at every RPM point, how does the AFC interpolate its change between the RPM points. Untill you understand how the AFC interpolates between the points it will be worthless. You might be able to get a rough estimate at a single point, 5000 rpm for example, but the sort of moving graphs you posted earlier will be impossible.

Another thing to consider is that the sensors used on our cars are not very accurate. Especially after time. Why do you think these cars use so damn many sensors to control the same thing. The ecu just tries to get to a happy medium by using inputs from the various sensors. air flow, baro pres, temp, O2, MAP, TPS. Trying to get a accurate measurement from far less than calibrated sensors is worthless when comparing to other systems.

 

I dont think your equation will work because you are over simplifying the way that the AFC works. If you could somehow get the equations that the AFC uses to modify the HZ signal, then it might be possible. Specificaly, if you have a different setting on the AFC at every RPM point, how does the AFC interpolate its change between the RPM points. Untill you understand how the AFC interpolates between the points it will be worthless.

Good point....I had been assuming that -20% meant they just brought the Hz signal down 20%, but it could be more complicated, you're right. I just don't see why it would be more complicated than that....what else could the afc be doing to the signal?

Also, I assumed that if you had, say, 4000RPM set at -18, and 5000RPM set at -20, then 4500 would be -19. I guess that could be a wrong assumption as well....though it would seem really strange to me if the afc didn't just come up with a smooth curve like that between rpm points.

I suppose both of these questions could be easily answered by someone at APEXi.

Its too bad standalone engine computers cost so much.

jon

 

Settling the equation question

Did a little experiment. While in neutral rev the car till my Karman is at 50hz and see the air flow using a pocketlogger. Then rev the car till the Karman is a 100hz and see air flow using the logger. Since 100hz is twice as much as 50hz, if the difference in air flow as shown by the logger is also double then the equation is correct. If not then it isn't.

Results

50hz= 0.67 lbs/min
100hz= 1.36 lbs/min

In a perfect world the 100hz reading would have been 1.34lbs/min instead but it was 1.36lbs/min. The margin of error is 1.5% which may be accounted by the fact that I can't keep the Karman count exactly at 50hz or 100hz. It fluctuates up and down a 1hz or 2.

Conclusion, the equation is good.

As far as interpolation, the AFC manual explaines that. Suppose 2000rpms is set at -2 and 3000rpms is set at -4, then halfway between these two points (2500rpms) it will interpolate halfway between settings (-2 & -4) which would result in -3 at 2500rpms. This could be taken into account on spreadsheet if you use the right formula.

 

The reason they are a waste of time is because people have no way to really say whether the turbo is good. Yes the quarter mile is best for overall car performance but we all know that number is based on a ton of factors....the turbo is just one.

Jon, I get what you are saying and it does make sense. Factors play a role in your hz reading though too. If you have a more restrictive air filter then there will be a slightly larger vacuum in the intake pipe, which will affect the velocity, and hz reading. Different intake pipes will also play a factor.
Either way hz readings vary from car to car depending on the setup also.

I still think that 1/4 MPH or a real dyno chart are the best ways to show a turbos potential. While different cars will have different mods, there are "groups" of mods that people can fall into.

 

If you have a more restrictive air filter then there will be a slightly larger vacuum in the intake pipe, which will affect the velocity, and hz reading.

Ditto. Before I cleaned my K&N 2-3 weeks ago I was getting 42-43hz at idle. When I cleaned I also reoiled it (like your supposed to) and now I get 32-33hz at idle. Grrrrr, I just added a restriction. Next time I'll just wash it and put it back in without the oil crap.

 

Oil crap? The cotton gauze filters REQUIRE oil to actually do the filtering part Don't use it unoiled...

Also for the AFC settings things... It does draw straight lines between % values. But if you are somewhere between hi an lo throttle it all goes to crap. There is a formula that finds a point in between the settings but it's weighted. Not a straght line at all. Of course at WOT this is not an issue...

The manual says that the air flow signal is converted to an air volume number. That number is corrected by the set factor, and then converted back to an air flow signal... The AFC was designed for several types of air meters, including karmen. But based on what the post above says the hz to mass ratio seems to be linear, though I would like to see larger numbers over a wider spread just for the hell of it. If the ratio is in fact linear at all hz readings then it is simply corrected by the set factor and sent to the ECU.

On top of it all, the manual said the number is converted to a volume number... So if that applies to karmen sensors then the MAS is measuring volume, which backs up what has been said a hundred times about it needing the other factors to determine mass.

The only way to undo what the AFC has corrected is to get our hands on the formula the ECU uses to combine those three factors to get lbs/min. Then you can adjust the hz number and recalculate. Not hard, but who's gonna do that

 

Ok, I'll do measurements at additional Hz readings and and see where we stand. I'll try to get a couple while I'm driving so the engine is under load.

 

additional readings

In neutral

50Hz= 0.67lbs/min
My previous reading was 0.68 so this difference is insignificant

In 4th gear under load
150Hz= 2.38lbs/min
200Hz= 3.06lbs/min

Comparing 50Hz in neutral to 150Hz under load reveals that the Hz to lbs/min comparison is not linear at least when going from no load to load. 150Hz is triple the amount of 50Hz. So taking 0.67lbs/min and muliplying by 3 equals 2.01lbs/min which is a 16% differential from the 2.38lbs/min observed.

Comparing 150Hz to 200Hz both under load, this is more linear. 200Hz is a 33% increase over 150Hz. So if you take 2.38lbs/min (observed at 150Hz) and add 33% you get 3.17lbs/min. Compared to the observed 3.06lbs/min this is a difference of 3.5%

On a side note, it was much more difficult to maintain a steady Hz reading while driving under load than it was when parked in neutral. This may account for this 3.5% difference from 150Hz to 200Hz but not for the 16% difference from no load to load.

I suspect the difference from no load to load is due to the intake air temp which is hotter at idle then when driving. 50Hz at idle was 0.67lbs/min so 150Hz should be 3 times that or 2.01lbs/min and 200Hz should be 4 times that or 2.68lbs/min. The observed air flow was greater while the car was in motion but since the volume was the same and there was no change is baro pressure the intake temp must be the difference.

The equation should be usable as long as air temp is logged as well and there is no significant change in air temp during the run being logged.