Results 1 to 14 of 14

Thread: Mitsubishi Karman Vortex Air Flow Meters: How they really work

  1. #1
    Kenneth's Avatar

    Offline
     
    Name
    Kenneth
    Join Date
    Nov 2004
    Last Online
    12-01-2024
    Membership ID
    NZ002
    Posts
    6,968
    Country
    New Zealand
    Location
    Cambridge
    Car
    Kia Sorento :P
    My Garage
    Visit
     

    Mitsubishi Karman Vortex Air Flow Meters: How they really work

    As many know, I have long advocated using the standard airbox setup. I harp on and on about how pod filters don't work etc.

    The following technical explanation gives some technical basis for these beliefs . It may be a bit technical so if there is some point you wish clarified or explained further, say so.

    Some of the material is easier to understand with a diagram, in which case please refer to the picture below. (Unknown origin, found on the net)

    The Mitsubishi air flow meter is what is know as a Karman Vortex Air Flow Meter. (KVAFM)

    The principle of its operation is fairly straight forward and goes like this

    When something is placed in the path of a fluid, vortices are created in that fluid which vary in frequency proportionally with the speed of the fluid.

    Mumbo Jumbo? The simplest demonstration of this principle is the wake of a boat. As the boat moves through the water a bow wave is created in the water. This bow wave moves faster the faster you go. (More accurately, the faster you go the more bow wave ripples you create per second)

    Air is the same, you just cant see the evidence unless there is some form of colour added to the air stream.

    These "bow waves" in the air cause fluctuations in pressure, just as a car has a low-pressure zone behind it when moving, so do the vortices.

    In the KVAFM, there is a obstacle placed in the path of the moving air. This obstacle is of a certain shape so that the pressure fluctuations are directed at an area that carries that pressure to a mirror surface. (See pressure directing hole in the diagram)

    The fluctuations cause the mirror to oscillate at a frequency proportional to that of the vortex frequency. That mirror reflects or cuts light to a photo transistor which in turn switches on and off a voltage as the light hits it.

    The result is a voltage frequency output which is (all being well) proportional to the frequency of the vortices in the air.

    As the frequency of the vortices increases with the speed of the air, this can be used to calculate how much air is moving through the sensor.

    There are 3 parameters that can be used in the calculation of air flow through a pipe, Diameter of the pipe, the speed of the air, the volume of the air.

    with any 2 of those parameters you can calculate the other.

    The KVAFM provides the speed of the air, the ECU knows the diameter of the tube the air moves through and as a result it is able to calculate the volume of air moving through the KVAFM. (Air Density is then taken care of with the Intake Air Temp sensor (IAT) )

    So far so good, to summarise:
    The KVAFM uses the disruptions in the air flow to measure how fast air flows past the sensor, this is used to calculate the volume of air flow, the IAT is used to calculate the air density. Providing this information is accurate, the ECU can calculate EXACTLY how long to open the injectors per RPM to attain an Air Fuel Ratio (AFR). The desired AFR is found using a Load/RPM lookup.

    To make things quicker, there is a table in the ECU that tells it the volume of air in relation to the frequency of the KVAFM output. The reason for this is because the KVAFM size is fixed and therefore for a given air speed the volume will be the same. Instead of calculating every time (costing valuable processor time) it can look up a table.


    As you can (hopefully) see, the whole AFR calculation relies on the frequency generated by the KVAFM being correct. This is originally calibrated and tested using the standard intake system. This intake system consisting of the airbox, air filter and KVAFM is a known environment which is used to give accurate data to the ECU.

    If you make a change to that system, by adding a pod filter, a bend in some pipe work before the MAF (Haven't you wondered why the MAF is bolted to the side of a box? It isn't the best shape for air movement) then you run the risk of upsetting the way the air hits the obstruction in the KVAFM.

    Fortunately one assumes Mitsubishi made the airbox setup to provide a stable air environment for the KVAFM, thus when you do make these modifications the most likely result is that you INCREASE The rate of pressure fluctuation past the sensor.

    So, you add some turbulence to the air going through the sensor what happens? Since the ECU relies so heavily on the sensor for accurate AFR calculations, any increase in turbulence is likely (unless you are unlucky in which case your engine blows up) to cause your AFRs to be richer than they should be.

    In practice the results have shown this to be the case also, those who add pod filters put bends in their pipework before the KVAFM (usually trying to achieve a cold air intake) have fueling issues.

    So, unless you run MAP (manifold absolute pressure) based engine management, you had best leave the airbox alone, buy an ARC box or fabricate one that behaves in the same way as the standard airbox.

    Addition:
    K&N do sell a pod filter which is especially designed to work with the Mitsubishi KVAFM. I didn't add this information originally because I don't like to endorse manufacturers claims without any proof of them working. However Mitsiman (RPW?) has endorsed this product and on his word I post a picture of it. Perhaps Mitsiman would would to post a product code or some such to help out those looking for one.

    I have uploaded a picture of the unit, as you can see it is distinctive in shape and you should know at a glance if you have the real deal!
    Attached Images Attached Images
    Last edited by Kenneth; 24-09-2008 at 10:41 PM.

  2. #2
    MPBVr4's Avatar

    Offline
     
    Name
    Malc
    Join Date
    Jun 2004
    Last Online
    29-11-2023
    Membership ID
    218
    Posts
    2,586
    Country
    United Kingdom
    Location
    Taunton
    Car
    Audi S3 3000GT
    My Garage
    Visit
     
    Interesting write-up. I've never looked at the air flow meter in the VR4. Is there a flow straightner at the inlet (possibly a "honeycombe" section) if so this should give a sufficiently developed flow profile for the vortex meter irrespective of the approaching pipe configuration. Having said that I agree the standard air box should give more reliable measurements.

  3. #3
    Kieran's Avatar

    Offline
     
    Name
    K
    Join Date
    Aug 2003
    Last Online
    04-12-2011
    Membership ID
    10
    Posts
    21,149
    Country
    United Kingdom
    Location
    The Midlands
    Car
    Das LuftwaftenW
    My Garage
    Visit
     
    Quote Originally Posted by MPBVr4
    Is there a flow straightner at the inlet (possibly a "honeycombe" section)
    There is indeed - the 'mouth' of the MAF contains it.

  4. #4
    Kenneth's Avatar

    Offline
     
    Name
    Kenneth
    Join Date
    Nov 2004
    Last Online
    12-01-2024
    Membership ID
    NZ002
    Posts
    6,968
    Country
    New Zealand
    Location
    Cambridge
    Car
    Kia Sorento :P
    My Garage
    Visit
     
    The air straightener section will alter the air shape to a degree, but don't forget it also has to let the air past so it is somewhat of a balancing act between flowing ok and massaging the air flow.

    If there is sufficient turbulence in the air flow, this section of the AFM will not be able to correct it.

    A real world example is what happened to Carsten (Gly) who had massive fueling issues and fuel cut at 10psi after fitting hard pipes and doing a custom air box.
    He had a bend right in front of the AFM, putting a straight pipe between the air box and the AFM reduced this problem significantly, though it did not eliminate it.

    The AFM measures the speed of the air, but it only measures the speed through a portion of the AFM assembly. It then assumes that the air flow is uniform and calculates the volume through the whole AFM. If the air distribution is not uniform due to turbulence, then this measurement will be incorrect.

    The key issue is that its the speed of the air going past the sensor that is being measured.

  5. #5
    Gly's Avatar

    Offline
     
    Name
    Carsten
    Join Date
    Jan 2005
    Last Online
    22-10-2023
    Membership ID
    NZ019
    Posts
    7,283
    Country
    New Zealand
    Location
    Auckland
    Car
    Nissan Navara
    My Garage
    Visit
     
    actually i had the issue with the pod directly on the afm....


    after i moved the pod into the box and had a straight section befor the afm its better, boost cut at 15psi,
    but still have over fueling issues.

  6. #6
    __ET__'s Avatar

    Offline
     
    Join Date
    Oct 2007
    Last Online
    17-01-2010
    Posts
    33
    Country
    Australia
    Car
    Legnum VR4 Type
     
    Would a high flow pannel filter be ok to use? Or is there no point as it chages the airflow?

  7. #7

    Offline
     
    Name
    Brad
    Join Date
    Jan 2004
    Last Online
    20-10-2011
    Posts
    22,175
    Country
    New Zealand
    Location
    Karaka
    Car
    F/lift 5MT VR-4
     
    yep, Panel Filters are fine because they don't change the shape of the air

  8. #8
    MPBVr4's Avatar

    Offline
     
    Name
    Malc
    Join Date
    Jun 2004
    Last Online
    29-11-2023
    Membership ID
    218
    Posts
    2,586
    Country
    United Kingdom
    Location
    Taunton
    Car
    Audi S3 3000GT
    My Garage
    Visit
     
    Out of interest, I wondered if anybody has tried a proper "Mass Flowmeter" as used on most cars these days. These measure the actual mass of air entering the engine using a "hot wire" sensor. By using one, additional computation using temperature and pressure isn't required and should give significantly better results than a volumetric flowmeter (vortex meter). There are units available with frequency outputs but would probably need scaling to obtain decent results.

    My concern with the standard setup, if you're being really fussy, is that the pressure at the measurement point isn't measured. If you use a different filter or your standard filter is blocked up a pressure drop will occur making the air at the measurement point less dense than air is at atmospheric pressure which the ECU assumes it is. Thus the mass flow calculation is flawed

    zzzzzzzzzzzzzzzzzzz........................

  9. #9
    Nutter_John's Avatar

    Offline
     
    Name
    John
    Join Date
    Sep 2005
    Last Online
    Membership ID
    269
    Posts
    11,656
    Country
    United Kingdom
    Location
    Luton
    Car
    FL Galant VR4
    My Garage
    Visit
     
    thanks Prof Malc , just what I was thinking

    but a very good thread this is


    Kanji Automotive Solutions - Looking after your Pride and Joy

    Servicing - Upgrades - Tuning - Pre Used Parts - Undersealing - Advice - Consultancy
    PM or Call me for any work requests or to discuss your requirements
    Ebay Shop -> http://stores.ebay.co.uk/sun-wizzard...=p4634.c0.m322


    To contact the CVR4 Staff please see HOW TO: Contact Committee or Moderators

  10. #10
    Kenneth's Avatar

    Offline
     
    Name
    Kenneth
    Join Date
    Nov 2004
    Last Online
    12-01-2024
    Membership ID
    NZ002
    Posts
    6,968
    Country
    New Zealand
    Location
    Cambridge
    Car
    Kia Sorento :P
    My Garage
    Visit
     
    Quote Originally Posted by MPBVr4
    Out of interest, I wondered if anybody has tried a proper "Mass Flowmeter" as used on most cars these days. These measure the actual mass of air entering the engine using a "hot wire" sensor. By using one, additional computation using temperature and pressure isn't required and should give significantly better results than a volumetric flowmeter (vortex meter). There are units available with frequency outputs but would probably need scaling to obtain decent results.
    I am looking at trying this. The holden AFM is done this way and has a frequency output also, however the frequency range is much bigger (up to ~12000Hz) so there would have to be some calibrating circuit. Like many things on my plate at the moment, this will have to wait till I actually get the Galant back on the road though.

    Quote Originally Posted by MPBVr4
    My concern with the standard setup, if you're being really fussy, is that the pressure at the measurement point isn't measured. If you use a different filter or your standard filter is blocked up a pressure drop will occur making the air at the measurement point less dense than air is at atmospheric pressure which the ECU assumes it is. Thus the mass flow calculation is flawed
    I think you are splitting hairs. The pressure difference across the air filter is going to such a small drop in the ocean of potential causes of inaccuracy that it probably isn't worth thinking about.

    But yes, if you want to be really fussy you could say the sensor isn't 100% accurate. You can safely say that about any form of air flow sensor though... While you could calibrate it very accurately, the expense would not be worth it and engines are not THAT sensitive to small inaccuracies.

  11. #11

    Offline
     
    Name
    Christopher
    Join Date
    Apr 2008
    Last Online
    06-03-2019
    Posts
    278
    Country
    Other
    Location
    Curacao
    Car
    Galant VR4
     
    Dudes , Sorry the change somewhat on the topic, but I think my MAF is dead. But I want to know sure, how can I Test it to know for sure that I need to change it?

  12. #12
    swinks's Avatar

    Offline
     
    Name
    Tomasz
    Join Date
    Feb 2008
    Last Online
    21-10-2022
    Posts
    4,578
    Country
    United Kingdom
    Location
    Bourne, Lincs.
    Car
    ex-Galant VR4
     
    Just another question.
    I need a picture of MAF flow profile (cross section). Just to compare fitting of that screw/bolt which is in the bottom section secured by silicone glue.
    After reading that thread I assume it's so called obstacle to create vortex flow. And I remember article in library section saying that undoing that screw makes engine better response.
    Well, I have done few trials, and it's like undoing screw by 8mm makes engine run noticable leaner (according to my LC-1).
    And now... I'm bit screwed because I forgot to mark how deep was that sitting in stock setting.
    Ex: Galant VR4
    Running 268 HP ATW and 443 Nm torque at 0.9 bar
    Now: Lancer Evolution 8 FQ-300
    Running 325 HP ATW and 510 Nm torque at 1.6 bar

  13. #13

    Offline
     
    Name
    ralf
    Join Date
    Nov 2007
    Last Online
    10-11-2013
    Posts
    205
    Country
    Germany
    Car
    evo
     
    Quote Originally Posted by MPBVr4
    Out of interest, I wondered if anybody has tried a proper "Mass Flowmeter" as used on most cars these days. These measure the actual mass of air entering the engine using a "hot wire" sensor. By using one, additional computation using temperature and pressure isn't required and should give significantly better results than a volumetric flowmeter (vortex meter). There are units available with frequency outputs but would probably need scaling to obtain decent results.

    My concern with the standard setup, if you're being really fussy, is that the pressure at the measurement point isn't measured. If you use a different filter or your standard filter is blocked up a pressure drop will occur making the air at the measurement point less dense than air is at atmospheric pressure which the ECU assumes it is. Thus the mass flow calculation is flawed

    zzzzzzzzzzzzzzzzzzz........................
    bring it back from the dead

    I used a GM 3" MAF on my mitsi after the Blow off valve to cure my overfueling issues (my piggy back supported it and translated the signal_)

    You can also use a MAF translator..it translates Mitsi AFM signals to GM 3" Maf signals.. pretty easy to wire up and doest cost a fortune..you can also ajust fueling with the provided control box


    http://www.maftpro.com/

  14. #14
    Turbo_Steve's Avatar

    Offline
     
    Name
    Steve
    Join Date
    Oct 2005
    Last Online
    19-06-2023
    Posts
    7,051
    Country
    United Kingdom
    Car
    Silver
     
    bodge bodge bodge

    You're always going to be fighting for timing!

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •