Is the MAF the limit?

Something is wrong... maybe a slight boost leak? Stock housing 17 psi (1.2bar) should be MAX 1080 or so. Check another TDI....

That might also be why your turbo is making funky noises. A boost leak is dangerous, you shift the map to the right, same PR ratio (up down on the comp map) but to the right, makes it overspeed etc.

I like the map based limiter! So far so good, similar results to maf. If I pop off a hose, I won't get smoke, lol....
Similar fueling... I set 2700 to 52 mg/s and I was hitting 51mg/s with 2650mbar. Smoky with 11mm and the titan 764. Idle IQ at stock setting hammered to 4 @ 1.56v (to match the pump map exactly). To get rid of the smoke, adapted to 32688, a reading of 6.6. So my fueling was cut down 2.6 to be smokeless at WOT.

51-2.6=48.4 which is about what I was smokeless when the MAF columns were limiting me. To find out what my injectors+pump really does, I take my maf readings of 1275 around the midrange when boost is 2650mbar. Divide by 16.5 afr, and I get 77mg/s... which my tune says 48.4. 77/48.4= 1.59x!
Makes sense, stock pump, stock injectors can do 51 max. The guy who did the pump voltage map did the last column at 65, but on .230 injectors. So it balances out.
I think I might want to do what bhodgkiss posted, his pump map goes to 80... maybe that's better calibrated for my setup.

BTW, why do so many tune to only 4500-5000 rpm? Our ALH 11 and 10mm rotaries can make flat hp until 5500 easily... scale back the fueling, ex 51mg at 4000 gives same hp as 5500 with 37mg/s!

I have no boost leaks.. I'm 110% sure of that.. My 2260 has been failing from the day I installed it.. I got it used with no history, turns out is was some pre-production prototype 2260VK that probably got the snot run out of it.. It leaked oil into the exhaust from day 1.. That might be why it was so cheap haha..

Just where am I going to find another TDI with a heavily ported head, race cam, 2260vk, and a stock MAF housing?

Don't forget, 17psi through an engine with heavy airflow mods will have more total flow than a stock engine at the same pressure..

At idle what do you get w stock, 500?
Idle is where you aren't restricted AT ALL. The air flows in like a baby.. no backpressure to leave in some exhaust.
You wanna fill a container that holds 500 at 1 bar... with 2.2 bar.. it will hold 2.2 times of 500= 1100. There's no way to push more air in, it's pressure, you are squeezing the air 2.2x into that same space.
At higher rpm, it will be the same OR less if restrictions (like backpressure, not enough time at high rpm- etc).



Volumetric efficiency is always highest at idle on a diesel, no throttle... slow speed, low backpressure. Porting/cams/better breathing exhaust+turbine make that VE stay the same up higher.

I'm still confused how you can squeeze more air in than the pressure says... No, porting and valves don't do that(except variable valve timing or overlap), though in the corvette z06 they use special runners to force air in, but thats not under boost. I doubt our tuned cams run overlap, then the exhaust would push back into the intake cause EMP is usually above IMP especially when spooling up.

If you have a compressor set to 145 psi (10bar) , and fill another tank- taking however long you want- wide open... that container will have 10x of it's volume as air. Can't fit in more than 10x, unless you cool it down. Doesn't matter whether you fill it faster or slower, it still can only max out at 10x.

I haven't run a stock housing since I had the 17/22.. I wouldn't even try it with the 2260 since the MAF itself will be a significant restriction if it's the later revision with the mesh.. It has enough vacuum in the intake path to max out one of those filter-minder things with a stock MAF in place.. It even did that with the 17/22..

If porting the head and changing cams doesn't allow more air in, there would be ZERO power gains to be had with head porting and cam swaps.. The gains are significant in higher powered ALH's since the stock intake manifold, intake runners in the head, and valve opening events themselves are very restrictive for anything but sitting there idling..

One other thing I noticed which seems to hold true on most any ALH regarding MAF readings.. 900rpm no load look at the maf reading.. now bring rpm up to 1800 ensuring it makes no boost (unhook the actuator if you have to), and look at the MAF reading.. rpm is double, but maf readings are VERY close..

Yes, because the ecu is reporting MAF reading per stroke, per revolution. It's not a total flow. The ecu takes maf volts 0-5 (0-460kg/h) and divides by rpm to get mg/s.
So it's how much is going in per intake cycle.

I think in some gassers (like 1.8t maf readings) they don't do this because fuel injection isn't direct, so they can just squirt fuel based on total airflow... no need to be exact per cylinder?

that kind of makes sense.. I just wish I could keep communication up over 5300rpm so I can see what's happening in the last 700rpm LOL..

Ok so the map based tune is a bit Laggy... and thats what someone said on ecuconnections too. There's also a confirmation of the 1000 mg/s smokemap limit. And I have a way around it, divide the values in the maf linearization table... so when it is seeing 1300mg/s at some rpm it is actually reporting 1000. I will let y'all know how it goes.

from http://ecuconnections.com/forum/viewtopic.php?f=2&t=13719&p=72874&hilit=maximum+maf
"
Now lets talk about maximum MAF readings. As we know MAF is sensor measure airflow. Inside MAF is controled heated wire. As air enters the maf, it cools that wire, and the resistance of wire is changed. This is way how ECU can calculate airmass. So... as all sensors also MAF has its accuracy. It is calibrated for some values. In 1.9TDI files - 66,74,81,85KW has last valid value -> 1000mg/str. Surprised?

How can it be true if you log for example 1200mg/str? We can just guess. I would say... ECU is calibrated to measure more than 1000, but only readings below 1000 are valid. I would say that sensor is inaccurate over 1000. There is one map I call it "maf linearis" (x axis - temperature, Y axis Maf readings, Z axis compensated MAF readings)... and in that map is maximum 1000mg/str... so if you extend smoke limiter via MAF over 1000 those values over 1000 are not valid, because of "maf linearisation" map."

Seeing the linearization table in winOLS now I know that there is a factor. The column for X is percent of voltage with 5 being max.
The values are MAF total flow readings in kg/h but you have to multiply that number by 0.8 to get the real number.
Example, my map says 974, and it's really 779.
The maf rating to 4.5 v, as the doc says is 480kg/h and thats what you see around 4.5/5= 0.9 reading!

well.. My 2260VK is about to let go.. It's been tossing oil into the exhaust since the day I put it on, but it suddenly got worse.. Ive had it on the car for about 25-30k, and it's never been really tuned to take advantage of the turbo.. So far it's been at 32psi max (these are good for 38-40)... I haven't even gotten it fast enough to catch the red car (VR6 24V) yet LOL..

SO, I'll be messing with a 17/22 on a stock hardware ALH with PP520's for now until I can get another turbo.. Grrrrrr..

Ok, so I made it so that the whole maf calibration table is divided by 1.3. It's not that I am changing the housing, so linearization is not the issue, just that to report to the ecu that number/1.3. So for example when it says 1000, im doing 1300 actual, and so on.

IT WORKS!!!!
Just did some logs and the VE measurement (maf/absBAR) is the same throughout the rpm range. Just have to make sure you know that a maf reading is really 1.3 x bigger! So now I can have EVERYTHING on my smoke map up to 1000 regulated.

This is with my descreened maf which reads 10% less than normal through the range.
The 1.3x values match my other logs with the descreened maf including low boost too...


speed/ MAPmbar,MAF/ REALmaf/ RealMAFcorrected for my descreen
RPM-AbsMAP- -MAF- Maf*1.3--- MAF as screened
2814 2136mb 905mgs 1176.5mgs 1294
3265.5 2652mb 990mgs 1287mgs 1415
3601.5 2712mb 940mgs 1222mgs 1344
3832.5 2604mb 900mgs 1170mg 1287
4074 2544mb 845mgs 1098.5mgs 1209
4368 2484mb 805mgs 1046.5mgs 1152
4588.5 2412mb 775mgs 1007.5mgs 1109

It's also exactly 1.3x ratio for idle and midrange too, but the log is so huge I just picked out WOT.

very nice..

So I'm playing catch up on this thread, but the MAF is mainly used for emissions. De-screening the MAF is just a bad idea and your not going to get much from changing the MAF table.

Here is the MAF delete mod using a diode, you you need your EGR mapped out & blocked off
http://forums.tdiclub.com/showthread.php?t=205671

also you should check out ECU connections http://www.ecuconnections.com/
the have some good threads on understanding the EDC15( ALH) & tuning them.

another cheap mod combo is the Every Mod + Diode MAP mod that you can get some extra fuel & boost out of

MAP Diode Mod:
http://forums.tdiclub.com/showthread.php?t=207336

Every Mod:
http://forums.tdiclub.com/showthread.php?t=12165&

I have a few good website saved on my desktop at home, when I get a chance I will post them up as well

External mods like diodes and resistors don't make any sense if you have the ability to edit the fuel/air tables in the ECU.. The MAF can be mapped out just like the EGR..

Both myself and robnitro are active over on ecuconnections.. There's not much for ALH tuning guides out there.. The EDC15 guide is centered around the PD engine.. The MSA15 guides are a little help, but only a little..

Maf controls fueling. You tell me descreening is bad, but have you looked at the log values (linear 10% decrease that I included in my afr calcs, works perfect)? Keep in mind that without tuning, the maf modifications change fueling.

A diode mod is worse, same with evry. Limiting fuel putely by torque map is not too good. You may have no smoke in winter, but smoke in summer when air is hotter-less density.

In fact, the only other real alternative is map (speed pressure IAT) based limiter.
But it induces lagginess (Ive seen, and also mentioned on ecuconnections). It also cant account for varying VE based on turbo vane EMP. 2nd gear is more smoky as it has vanes closed more than in 4th or 5th.
Thats also noticed by ecuconnections and I.

I think I have mine dialed in fairly well finally, but I need a little more fuel in some areas on the smoke map.. Light throttle is perfect, 80%+ throttle is almost perfect, but there's a weird kind of flat studdery area from about 2500-3500rpm under moderate throttle where there's plenty of air, but it seems like not quite enough fuel.. I have to take some logs and figure out exactly where in the smoke map the problem is..

Are you using the pump voltage map I gave you?

I found that linearization transform: It works great (99.8% accuracies and <2.5% CV's at each calibration point).

The transform is to look at the original factory calibration raw data table:

"x" = airmass (mg/str)................"y" = MAF voltage


drop the values into Excel into column A and B. In column C, make a column of MAF voltage divided by mg/str ('y/x')

Plot a scatterplot of column C (dependent) versus B (independent).

Now fit a 2nd order polynomial curve to the data. Add label & right click to copy the formula.

In Column D, set '= polynomial formula (where the 'x' is pointing to column B)

In column E, set '= column b/column d' --these are your relinearized transforms, giving mass airflow values in mg/str. Just to check, plot them versus original factory calibration table's mg/str and use statistical software analysis for 'Goodness of Fit'. I used JMP software.

I found that the transform keeps linearity well past the curve-over to around 200% of the highest calibration input.

Regarding the max value of 1000mg/str, that sound like the computer is normalizing values to a peak value of 1000mg/str which might be fine as long as 1.) the linearity is okay (statistical goodness of fit acceptable), and 2.) the ECU is using normalized values of mg/str to make relative changes, where other parameters are known to be constant and valid.

IMO, this is why removing the screen or changing things w/o relinearizing (by transform) can make your life so hard...have to work around things and throw in Ad Hoc 'fudge factors' to get things to work. Just my opinion, and I don't mean to be rude so please take it how you wish. I can run an example of transform if someone can post the factory calibration table.

Where do you have that table? It would be interesting to check out.

The 1000 mg/s is a problem when we can do 1300 mg/s with 2.6-2.7 bar absolute. That means there is no regulation possible at all between 1000 and 1300. The post that I quoted was wrong about it though. It is not based on accuracy. Example, 2.7 bar at mid range does not come close to saturating the stock maf. Yet, they clip it down to 1000?!? It's just by design

I found out recently that it is due to the maf correction temperature map, it goes up to "10000" on y axis and z axis.... the top axis is temperature. This is what the smoke map looks up apparently, and since it doesnt have values above 1000, the rest is discarded.

I understand the problem with descreening, but with all of the logging I have done, compared to a friend's untouched maf, the values are pretty much spot on 10% less. That shows me that keeping the screen in front of the element is what keeps it linear. The mesh around the rest, which is at the same level of the element mesh is just something that restricts flow a bit. Without it, a bit less air flows through the element.

Yes, the screen induces local micro-turbulent flow directly in front of the thin film heating element. Turbulent flow is inherently more linear than laminar flow. So the screen actually physically linearizes the airflow; but not totally (is my understanding). It only straightens out the response curve a little. To get a VERY linear and repeatable output, you need to generate a predictive regression equation--the linearization transform in my last post does that part of it. The linearization transform above does a much better job than a simple quadratic fit to non-transformed raw data---I've tried those in the past, and it induces a lot of error (10%-30%) in the upper third of the calibration range.

So, if we take a maf , with the screen, like Bosch C maf... and run a 3 inch housing instead of 2.5 inch stock...
What could we expect for the values?
Area of 3 in is 9 in^2
Area of 2.5 in is 6.25^2

So for the same flow, the airspeed is 1.44x lower (70% of 2 inch).

The maf with screen, gets flow based on differential pressure, due to the restriction of the screen compared to the freeflowing rest of the pipe. Not sure how to calculate the difference of this, perhaps this is what makes the characteristic curve the way it is, more airspeed, more differential pressure?

The C MAF has a element screen with diameter of almost 1.5 inches. That has an area of 2.25 in^2.

So with 2.5in housing, the element is 2.25/6.25= 36% of the area to maf
3 inch housing, the element is 2.25/9=25% of the area to maf
Difference in relative area, 25/36= 70% of 2.5 inch

3 inch housing reading = (2in reading)*0.7*0.7=(2in reading)*0.49

Of course there would be some curve to it, but the 3 inch maf would be reading double the airflow for the same voltage.

robnitro said:

Are you using the pump voltage map I gave you?

Click to expand...

Yeah I'm using that one.. I ended up modifying it down low a little to make it idle a little better..

robnitro said:

So, if we take a maf , with the screen, like Bosch C maf... and run a 3 inch housing instead of 2.5 inch stock...
What could we expect for the values?
Area of 3 in is 9 in^2
Area of 2.5 in is 6.25^2

So for the same flow, the airspeed is 1.44x lower (70% of 2 inch).

The maf with screen, gets flow based on differential pressure, due to the restriction of the screen compared to the freeflowing rest of the pipe. Not sure how to calculate the difference of this, perhaps this is what makes the characteristic curve the way it is, more airspeed, more differential pressure?

The C MAF has a element screen with diameter of almost 1.5 inches. That has an area of 2.25 in^2. 1.77 in^2 (pi * .75^2 = 1.77)

So with 2.5in housing, (4.91 in^2) the element is 2.25/6.25= 36% of the area to maf 1.77/4.91 = 36%
3 inch housing, the element is 2.25/9=25% of the area to maf
Difference in relative area, 25/36= 70% of 2.5 inch

3 inch housing reading (7.1 in^2) = (2in reading)*0.7*0.7=(2in reading)*0.49 What's this referring to? I calculate that 7.1/4.91 ratio between them means the 3" MAF housing will flow 44% more (not accounting for slight difference in dP that comes into play at higher air velocities, AND in a perfect world with turbulent plug flow (non-laminar)--IMO, the 3" MAF housing will be even less linear than the 2.5".

Of course there would be some curve to it, but the 3 inch maf would be reading double the airflow for the same voltage.

Click to expand...

Some calculation comments in blue above.

I think the red part isn't correct. The sensor still depends on a principle of heating a thin film element versus a known reference value. The cool air moving past it means the element has to supply more current to reach a given temperature in a given time. Resistance is fixed at the factory, so as current changes, voltage must too. The screen in front of the element is just there to physically 'trip up' the laminar airflow so that the air directly in front of the MAF thin film sensor is more linear (turbulent flow is more linear "plug flow" than laminar flow). Pressure does not get measured in any way.

TdiJetta99, you mentioned you have a factory calibration table for one of the MAF's. Can you share it? I'll do the transform, compare it to simple linear and quadratic fits, and post results.

nicklockard said:

Some calculation comments in blue above.

I think the red part isn't correct. The sensor still depends on a principle of heating a thin film element versus a known reference value. The cool air moving past it means the element has to supply more current to reach a given temperature in a given time. Resistance is fixed at the factory, so as current changes, voltage must too. The screen in front of the element is just there to physically 'trip up' the laminar airflow so that the air directly in front of the MAF thin film sensor is more linear (turbulent flow is more linear "plug flow" than laminar flow). Pressure does not get measured in any way.

Click to expand...

I may have phrased it wrong. Pressure- resistance to flow is what makes more and more air flow through the somewhat restrictive C screen. If it were less restrictive (no mesh with angle as in C maf screen), it would be easier for the air to flow through. But when you go with a bigger and bigger housing, the air has less tendency to want to flow through that path?

Just thoughts... I'm an electrician (technician) by trade... so I'm thinking like ohm's law... if you have a really strong connection, that will take most if not all the current, compared to a small wire on the same potential. Of course it is nothing like air, and even water is not like air, lol.


To get rid of the 1000 mg/s limit that the smoke map looks at:
Open vagsuite and your file
Click search map content
Just hit ok. You get a big list. (some bug it doesnt find the map even when you enter a value from the map)
Click length to sort by length
Look for maps that have hex 140 length (or decimal 320)
On mine it is called 3D "blablablabla" C01C "blabla"

It has temp on top , left is values of 10000 etc, and map same values as the left. C01C seems to be common on other files, so that might help.
Right click, edit Y axis.
Change the upmost value from 10000 to 15000. Hit save.
Hit save on the map, and hit read. Make sure the 15000 is there.
Now replace that row with 15000.
Hit save, hit read. Do the same for all 3 maps that are similar.

TDIJetta99 said:

Both myself and robnitro are active over on ecuconnections.. There's not much for ALH tuning guides out there.. The EDC15 guide is centered around the PD engine.. The MSA15 guides are a little help, but only a little..

Click to expand...

I'm also over there as well but i'm working on the EDC16. I know there is not much over there for either of these computers....

not sure if you have seen this or not, of it it will help you...
https://sites.google.com/site/vagecumap/diesel-engine-control

bosch maf calibrations from bosch maf spec sheet

From the bosch maf sensor spec sheet I found the values for their different maf's. The one for 2.45 in internal matches my stock map. For them to read more they run bigger housings.
I put the values into a table and how it corresponds to the maf calibration inside vagedcsuite. To get the real maf kg/h reading from the vagedcsuite table, multiply by 0.8 (as I saw in the winols damos).
Check the 2.45 in table and compare to the map. The Y axis is 4 decimal places up to 1.000 (which is 5v). The 2.45 in maf is rated lower but only because of the spec which goes to 0.9x of the 5v (around 4.5v) but it can do more.
For the values in between, I think its best to put it into excel or some graphing app to copy the curve of the original 2.45 map.

So are we now saying to get rid of the 1000mg/str limit we change a single value from 1000 to a greater value which the MAF smoke limiter references? So I don't need to halve all the values on the MAF axis etc now?

Do you know what address I need to change on an 038906018P ecu please?

robnitro said:

From the bosch maf sensor spec sheet I found the values for their different maf's. The one for 2.45 in internal matches my stock map. For them to read more they run bigger housings.
I put the values into a table and how it corresponds to the maf calibration inside vagedcsuite. To get the real maf kg/h reading from the vagedcsuite table, multiply by 0.8 (as I saw in the winols damos).
Check the 2.45 in table and compare to the map. The Y axis is 4 decimal places up to 1.000 (which is 5v). The 2.45 in maf is rated lower but only because of the spec which goes to 0.9x of the 5v (around 4.5v) but it can do more.
For the values in between, I think its best to put it into excel or some graphing app to copy the curve of the original 2.45 map.

Click to expand...

Can you post this as an Excel file or PM me a copy?

bhodgkiss said:

So are we now saying to get rid of the 1000mg/str limit we change a single value from 1000 to a greater value which the MAF smoke limiter references? So I don't need to halve all the values on the MAF axis etc now?

Do you know what address I need to change on an 038906018P ecu please?

Click to expand...

I don't know if you still need the address but it's 04112 for the first coding. 0c112 2nd 3c112 3th

robnitro said:

From the bosch maf sensor spec sheet I found the values for their different maf's. The one for 2.45 in internal matches my stock map. For them to read more they run bigger housings.
...
http://img209.imageshack.us/img209/9954/boschmafcalibrations.jpg

Click to expand...

I've got a 3" housing for mine off of a volvo, but I just did some calculations, and the stock 2.5" MAF housing should be good for a bit past 1500 mg/str, well up to 5k RPM, where it calculates out to 450kg/hr, and the sensor linearization tops out at 480 kg/hr

Assuming the computer can see 1500mg/str. All I've seen out of mine yet was 1275mg/str, and it is in two datacells, so that may be the upper limit even changing that one value from 10000 to 15000

I'm having second thoughts about the 3" MAF being an upgrade. Maybe if you scale the smokemaps to correct for the larger sensor housing, rather than trying to get it to read past that limit.

Why not do it properly and change the MAF, linearization and diagnostic limits. Thats whats holding you back.now.