Baro and MAT ( IAT ) correction.

A general forum and a place for initial or prospective users. See Manuals/Documentation
Click here to enter
Contact a Forum Administrator
If unsure where to post, post in this sub-forum.

Moderators: jsmcortina, muythaibxr

jsmcortina
Site Admin
Posts: 39585
Joined: Mon May 03, 2004 1:34 am
Location: Birmingham, UK
Contact:

Re: Baro and MAT correction.

Post by jsmcortina »

Jean makes an excellent point. The reason the existing air density table is not working correctly is because we are using the wrong temperature.

One thing we can all agree on, is that the air that enters the cylinder will almost always be warmer than the ambient air as it has passed through a warm inlet manifold and a hot cylinder head and valves.

The heating effect is likely to be related to rpm and engine (coolant) temperature.

Something proposed a while back on MS1, but was difficult to implement on that platform was to calculate the MAT based on MAT and CLT using an rpm factor lookup table: (made up numbers!!)
e.g.
rpm %age MAT
500 60
1000 65
6000 90

Then we could do MATadj = (MATsensor * MAT%) + (CLTsensor * (100-MAT%))

James
I can repair or upgrade Megasquirts in UK. http://www.jamesmurrayengineering.co.uk

My Success story: http://www.msextra.com/forums/viewtopic ... 04&t=34277
MSEXTRA documentation at: http://www.msextra.com/doc/index.html
New users, please read the "Forum Help Page".
Philip Lochner
Super MS/Extra'er
Posts: 1004
Joined: Thu Feb 16, 2006 6:18 am
Location: George, South Africa

Re: Baro and MAT correction.

Post by Philip Lochner »

arlo#1 wrote: I was wondering if it would be possible to have the mat correction dissabled till the coolent temp was hi enough so they arent fighting each other.
I recon having the MAT sensor downstream of the throttle is preferred but some on the forum recon heat soak of the sensor and thus false readings is a problem. I can see that being the case with an enclosed sensor which will take some time to adapt to the air temp but with an open element sensor I struggle to see that it should be a problem - once air flows over it. I certianly believe that the sensor should be fully exposed to air flow at the maximum velocity possible.

WUE and MAT don't "fight" each other. They compliment each other. Besides, when coolant is cold WUE is by far a bigger enrichment factor than MAT, so there is in my view no need to disable MAT correction while WUE is on.
Kind regards
Philip
'74 Jensen Interceptor 440ci (EFI'ed with MS2 and wasted spark + GM 4L60e GPIO controlled - both on Extra FW)
Philip Lochner
Super MS/Extra'er
Posts: 1004
Joined: Thu Feb 16, 2006 6:18 am
Location: George, South Africa

Re: Baro and MAT correction.

Post by Philip Lochner »

jsmcortina wrote:Jean makes an excellent point. The reason the existing air density table is not working correctly is because we are using the wrong temperature.
Absolutely!!

The question then: "What air temp do we really need to know?"

I recon MAP should also play a role in this modelling equation. Let say baro is 100kpa. If air hits the throttle body at 100kpa (ignoring passage and air filter losses) but after the throttle MAP is 50kpa, the air should have had a temp drop. I wish someone can tell me how much!! As it now passes through the inlet manifold on its way to the cylinder it warms up again. But if that same air passes over the throttle at WOT, that pressure drop (temp drop) is minimal.

By the time it hits the cylinder at BTDC it will surely be much hotter. Is this the temp we need to know?
By the time its been compressed just before ignition it will be even hotter. Is THIS the temp we need to know?

BUT, stepping back a little ....

I, me, Philip, don't care any little bit what the inlet air temp is, or what air temp we need to measure, or how or where. All I care about is that the AFRs should be what I want them to be for all values of MAT. The current non-linear MAT correction table is IMHO perfect!!

Using Jean's analogy, I dont care what the volume of the box is, as long as the colour is right :D
Kind regards
Philip
'74 Jensen Interceptor 440ci (EFI'ed with MS2 and wasted spark + GM 4L60e GPIO controlled - both on Extra FW)
jsmcortina
Site Admin
Posts: 39585
Joined: Mon May 03, 2004 1:34 am
Location: Birmingham, UK
Contact:

Re: Baro and MAT correction.

Post by jsmcortina »

This is somewhat an "empirical" vs. "model based" tuning discussion.

Philip is right that what matters at the end of the day is an engine that has smooth operation with the desired AFRs over all operating conditions.

How we get there is open to debate.

I would argue that if we get the engine model closer to reality and make it easier for newbies and experienced tuners alike to do that, then surely we have a better system.

The simpler we make it to get it right and the closer the default settings are to the average engine the less install headaches there will be.

If, for example, everyone tried out Philip's airden correction numbers and found them to be a big improvement, I'll put those numbers in the code as a default.
Or, if I implement the %age correction and there is a number that works well for the majority of users, likewise, that will be the default.

James
I can repair or upgrade Megasquirts in UK. http://www.jamesmurrayengineering.co.uk

My Success story: http://www.msextra.com/forums/viewtopic ... 04&t=34277
MSEXTRA documentation at: http://www.msextra.com/doc/index.html
New users, please read the "Forum Help Page".
arlo#1
Helpful MS/Extra'er
Posts: 141
Joined: Mon Aug 21, 2006 6:01 pm
Location: Nanaimo B.C. Canada

Re: Baro and MAT correction.

Post by arlo#1 »

In my opinion the temp we need to mesure is right befor the throtle blade because as you go closer to the intake valve the teperature is already somewhat known by the temp of the coolent. And by not being under vacume makes the code alot more simple. My stock srt4 iat location was in the air box which is along way fom the engine at all concidering it is in the front of the fender and the pipeing goes through the turbo then the intercooler then the itake ect. And in stock trim it runs pretty damb good. Maybe as a user friendly way of getting around this we could have a 0-100% control slider for the mat correction.
Its finaly on the road! Self Custom Made 440 Road Runner efi. http://www.youtube.com/watch?v=974YsInWxhQ ms2extra 2.1.01 code
Philip Lochner
Super MS/Extra'er
Posts: 1004
Joined: Thu Feb 16, 2006 6:18 am
Location: George, South Africa

Re: Baro and MAT correction.

Post by Philip Lochner »

Philip Lochner wrote:Let say baro is 100kpa. If air hits the throttle body at 100kpa (ignoring passage and air filter losses) but after the throttle MAP is 50kpa, the air should have had a temp drop. I wish someone can tell me how much!!
I asked a friend of mine who's sums I trust. According to his sums the air temp immediately after the TB will be -8degC (minus 8 deg C) !! (that's assuming no heat is gained over the TB and no work is done on the air while passing over the TB)

This means that at full throttle it will be 50degC and at partial throttle it will be around -8degC. Of course on its way to the cylinder it heats up again.
Kind regards
Philip
'74 Jensen Interceptor 440ci (EFI'ed with MS2 and wasted spark + GM 4L60e GPIO controlled - both on Extra FW)
Philip Lochner
Super MS/Extra'er
Posts: 1004
Joined: Thu Feb 16, 2006 6:18 am
Location: George, South Africa

Re: Baro and MAT correction.

Post by Philip Lochner »

More info (as received from my mechanical engineer friend):

Formula for getting to these figures attached. Gamma = 7/5 for normal air.

Look at what happens at the cooling effect as one rises above sea level. I live at 85kpa which is what prompted me to draw up the table for .85 bar.

At sea level having 50degC before the TB and driving at 50kpa one would have (in theory) -8degC after the TB.

But at 85 kpa driving with the same conditions, 50degC before the TB at 50kpa one would have 4.5degC after the TB - air almost 13 degC warmer. I wonder if this does not also contribute in a big way to the AFR deviations I've seen with barometric changes....

I must say that I find the figures hard to believe more so around 30 and 40 kpa ....
Kind regards
Philip
'74 Jensen Interceptor 440ci (EFI'ed with MS2 and wasted spark + GM 4L60e GPIO controlled - both on Extra FW)
arlo#1
Helpful MS/Extra'er
Posts: 141
Joined: Mon Aug 21, 2006 6:01 pm
Location: Nanaimo B.C. Canada

Re: Baro and MAT correction.

Post by arlo#1 »

This is some great reading! Thanks for the chart Philip. So the question is how is the code currently writen to correct for iat and how are we going to have it in the future? I can mount my iat in the acutal intake it self but it is a big chunk of pollished aluminum and it would probably heat soak easily and I would have alot of work to pull it off to drill and tap it. Would it be possible to have two options for the placement like in general settings Iat placement before throtle plate enter 1 and after throtle placement enter 2?
Its finaly on the road! Self Custom Made 440 Road Runner efi. http://www.youtube.com/watch?v=974YsInWxhQ ms2extra 2.1.01 code
vinister
Experienced MS/Extra'er
Posts: 159
Joined: Sat Jul 19, 2008 11:04 pm

Re: Baro and MAT correction.

Post by vinister »

I also recently realized that the MAT correction has issues. I would never have the same AFRS in the morning/afternoon, even after turning off all the corrections and running straight from the VE table. I decided to put all zeros in the mat correction, leave my VE table alone, and simply tune out the AFR deviations via the MAT correction table.

After a couple weeks of tuning, I am now reaching my target AFR's almost exactly no matter what the MAT is, and this is with NO ego correction. I too believe that the car should hit its targets without relying on EGO correction.

My MAT table may be technically 'wrong' from a physics standpoint, but I think the point here is to reach target AFR's no matter what method is used.

I think my next step is constant baro because I also notice the AFR change with altitude, and once per start is def. not enough.

I have my MAT sensor mounted in the end of the filter element, there is a plastic-chrome cap (one of the cheapo K&N knockoffs), and I threaded it directly into the plastic cap. So all the air entering the filter is going straight past the sensor, and the sensor is insulated from any metal by the filter element itself. There is about 14" of manifold distance between the sensor and the valves.

I think the code is great as-is. The code reflects the physics calculation, and the correction table is used to correct for this 'air-heating effect' we are getting from the manifold runners etc. Everyone has a different MAT sensor mounting position, different engine layout, so the correction table allows us all to have proper MAT correction while leaving the proper calculation in the code.
Philip Lochner
Super MS/Extra'er
Posts: 1004
Joined: Thu Feb 16, 2006 6:18 am
Location: George, South Africa

Re: Baro and MAT correction.

Post by Philip Lochner »

Philip Lochner wrote:
arlo#1 wrote: I was wondering if it would be possible to have the mat correction dissabled till the coolent temp was hi enough so they arent fighting each other.
WUE and MAT don't "fight" each other.
I've been thinking about this....

Its true that they don't "fight each other - but only after the MAT correction table has been tuned. In practise though, as the engine warms up the engine bay warms up and so IAT will warm up. So with low CLT you get low MAT's. So now you end up tuning WUE with an untuned MAT table and once the engine is warm, lets say you get to drive in very cold weather so now you get cold MATs with a warm engine, you tune MAT to correct and next time you start the engine, WUE will be overly rich.

So I can see that they could end up "fighting" each other.

I just concluded that in an ideal world one would want to tune the MAT correction table even before you do a final tune on the WUE table.
Kind regards
Philip
'74 Jensen Interceptor 440ci (EFI'ed with MS2 and wasted spark + GM 4L60e GPIO controlled - both on Extra FW)
Kurt A
Experienced MS/Extra'er
Posts: 311
Joined: Sun Jun 06, 2004 11:26 pm
Location: Quesnel B.C. Canada

Re: Baro and MAT correction.

Post by Kurt A »

This is an awesome thread, it also explains why that when I park the car in perfect tune during the fall it is different when start back up in the spring. I understand that the gas law is accurate but in the real world tweaking of the table is required. Is it possible to set up a tunable table much like Phil Tobin has done in Tuner Studio for the WUE?
Kurt
Rotax 912 turbo'd under construction MS2 Extra
2005 Mustang GT, turbo'd, aluminum heads, MS2 Extra
1928 Ford Coupe, 545BBF,671 blower, Squirted by MS2 extra 3.1.0, V3.0 Board
jsmcortina
Site Admin
Posts: 39585
Joined: Mon May 03, 2004 1:34 am
Location: Birmingham, UK
Contact:

Re: Baro and MAT correction.

Post by jsmcortina »

I implemented the scaling factor in the code that will be the next beta.

James
I can repair or upgrade Megasquirts in UK. http://www.jamesmurrayengineering.co.uk

My Success story: http://www.msextra.com/forums/viewtopic ... 04&t=34277
MSEXTRA documentation at: http://www.msextra.com/doc/index.html
New users, please read the "Forum Help Page".
arlo#1
Helpful MS/Extra'er
Posts: 141
Joined: Mon Aug 21, 2006 6:01 pm
Location: Nanaimo B.C. Canada

Re: Baro and MAT correction.

Post by arlo#1 »

Sweet thanks
Its finaly on the road! Self Custom Made 440 Road Runner efi. http://www.youtube.com/watch?v=974YsInWxhQ ms2extra 2.1.01 code
mike_robert
Helpful MS/Extra'er
Posts: 139
Joined: Sun Feb 10, 2008 8:26 am

Re: Baro and MAT correction.

Post by mike_robert »

I have the same misgivings as Jean.

We have a recurring observed problem.

1)The physics formulas work very well, accurately and reproducibly in many other non-MS applications.

2) The OEMs have made these same formulas and sensors work just fine - enough to satisfy the emissions people in all countries that I know of.

3) The math in the code looks good. There have been several very intelligent people generating multiple versions of this code for 2 different processors in assembler and 2 different, very disparate C compilers on many different versions of the code. Elementary code/logic flaw would therefore seem very unlikely.

4) Given 1,2 and 3, Solve for our problem "x" - but what is "x"?
Black99rt
Experienced MS/Extra'er
Posts: 364
Joined: Tue Mar 14, 2006 10:38 am

Re: Baro and MAT correction.

Post by Black99rt »

I'm very interested to see what works as a good scaling factor for this. I think it is a step in the right (albiet physically wrong) direction. It would be interesting to have some OEM speed density input on this, although I suspect their CL hides sins and their open loop is usually horrid rich regardless.
1999 Dakota 5.9L R/T-
MSIIe 3.1.2 Batch Fuel/MSD single coil dizzy/Stepper idle
LC-1
Philip Lochner
Super MS/Extra'er
Posts: 1004
Joined: Thu Feb 16, 2006 6:18 am
Location: George, South Africa

Re: Baro and MAT correction.

Post by Philip Lochner »

Here is another interesting table (confirmed by my mechanical engineer friend). Hope I'm not duplicating this from MS documentation manual somewhere

If I read this right, then if I can reduce the IAT from 80degC( 176F) to 20degC (68F) I should get 17% more air (ie 17% more oxygen) into the combustion chamber. Compliment that with 17% more fuel and that should translate into a rather tangible increase in power not so?

BTW: those % figures are very close to the figures needed to "zero" the effect of the MAT correction table - almost as if to confirm the built-in table should be correct.

Could it be that the internal table does not work as well as expected because it does not take the temp drop over the TB into account?
Kind regards
Philip
'74 Jensen Interceptor 440ci (EFI'ed with MS2 and wasted spark + GM 4L60e GPIO controlled - both on Extra FW)
jsmcortina
Site Admin
Posts: 39585
Joined: Mon May 03, 2004 1:34 am
Location: Birmingham, UK
Contact:

Re: Baro and MAT correction.

Post by jsmcortina »

If the temp sensor is supposed to be installed in the manifold, wouldn't that temp drop already be allowed for?

On an earlier point you made, the temperature we want in the PV=nRT equation is, I believe, the temperature of the air as it passes the intake valve. This surely is the limiting factor, we are only really interested in the air/fuel mix that makes it into the cylinder (but what about valve overlap?!) So the coolant temp and intake temp drop do have an impact.

To quantify the TB temp drop, how about installing a temp sensor either side?

(I've just the other day received the copy of the Heywood authoritative text on internal combustion engines, but have not had the time to see what he says on the matter.)

James

Ref: "Internal Combustion Engine Fundamentals", John B. Heywood. ISBN 0-07-100499-8
I can repair or upgrade Megasquirts in UK. http://www.jamesmurrayengineering.co.uk

My Success story: http://www.msextra.com/forums/viewtopic ... 04&t=34277
MSEXTRA documentation at: http://www.msextra.com/doc/index.html
New users, please read the "Forum Help Page".
Keithg
Super MS/Extra'er
Posts: 2413
Joined: Sun Mar 06, 2005 9:15 am
Location: Chicago, IL, USA
Contact:

Re: Baro and MAT correction.

Post by Keithg »

This is an interesting discussion. A couple things get confusing for me. One is the comment of altitude changes affecting the spark and/or fuel. I thought that the addition of a load calculation (map/baro) would allow you to use 'the appropriate part' of the VE map for the fuelling calculations. I want to keep spark out of this and completely separate as it confuses things and I think we are talking totally about fuel calculations.

I have trimmed my WUE table so that my car idles at or near 14.7 at idle during warmup. It is very rich at certain parts of the map as I drive, though. This needs to have a RPM trim to it or, maybe WUE is only useful at idle, anyway. I do not know.

Another thing that I mentioned to Ken is that I think that *all* fuel modifications should be scaled based on injector size. If you have huge injectors relative to displacement, a percent based on current PW can be huge and should be scaled to the injectors used (or req fuel, take your pick)

My goal has always been to tune my complete VE at a certain MAT. Right now, I cannot do that and only the low end of my map is tuned at a proper temp. I say proper as my intercooler is in my engine bay and at low speeds suffers heat soak, so I try to tune below 70F MAT. Heat soak is a terrible problem and seems to be worse with MS2, somehow. I have a very aggressive MAT correction and it still seems to be too lean on a hot start. I do not have an answer, just adding fodder. One think I have noted is that I think I could use a few more bins with the MAT correction curve.

KeithG
pigga
Master MS/Extra'er
Posts: 618
Joined: Sat Feb 09, 2008 12:44 pm

Re: Baro and MAT correction.

Post by pigga »

Hi all,
1st of all: A very interesting thread! I got here because we discussed exactly the same issue in the german MS board ans someone linked to this topic. I suppose as well that there is some Issue with the MAT-Correction. However I dunno WHAT.
The fact: I used to drive MS1 and encountered the effect hat the AFR varies. So...if the engine idles for a while, the AFR gets more and more rich. If I correct this (e.g. by using MLV) and I try to start over after a while (if the engine has cooled down), the VE at idle is undriveable lean.
In the meantime my MS2 offers the great feature of an addition MAT-correction als discussed before.
In my case I found out that I have to decrease the fuel amount with rising Manifold Air temperature. Actually 4%/°C seems to work for me quite well.
Philip Lochner
Super MS/Extra'er
Posts: 1004
Joined: Thu Feb 16, 2006 6:18 am
Location: George, South Africa

Re: Baro and MAT correction.

Post by Philip Lochner »

jsmcortina wrote:If the temp sensor is supposed to be installed in the manifold, wouldn't that temp drop already be allowed for?
All my logic tells me that it should be so. I remember seeing so many debates, as to whether the MAT sensor should be before or after the TB - and those are just the ones I did see - let alone the ones I did not see. Are we perhaps now arriving, with good reason, at the conclusion that the MAT sensor SHOULD be after the TB (thereby taking into account not only the temp before the TB but also the pressure drop over the TB and its resulting cooling/density effect) ?

Now a new dilemma. I'm driving along at WOT with 100kpa (ie at sea level) and 50degC (122F) MAT (before and after the TB) and suddenly I lift my foot and MAP goes to 40kpa. IF the table I presented above is anywhere close, the inlet air temp after the TB at this new condition should very suddenly move to -74degC (-101F) - too fast for even an open element sensor to register besides being out of range most likely. This is distorting my brain... I MUST be missing something here...
jsmcortina wrote:To quantify the TB temp drop, how about installing a temp sensor either side?
I'll do this and I'm hoping others will too so we can get several before and after (the TB) logs to see what we get. I'll connect my CLT input to the other MAT sensor thereby logging both in one log. Will do this when the engine is hot and will set WUE to 100% during the run.

But now we get to another issue which I've had trouble getting to grips with. Several installers reported "heat soak" from installing the sensors in the inlet manifold meaning that heat is transferred from the inlet manifold via the sensor body, via the electrical wires running inside the sensor up to the heat resistant material itself, thereby transferring heat to the actual heat sensitive sensor, thereby causing false MAT readings. Can heat travelling up those thin wires (pic attached) holding the heat sensitive element, add heat to the element?

I can see heat soak being a problem if enclosed sensors are used such as a CLT sensor but as long as air flows over and open element sensor, I find that very hard to accept. I now wonder if what they describe as "heat soak" is not just a symptom of inappropriate MAT correction?
Kind regards
Philip
'74 Jensen Interceptor 440ci (EFI'ed with MS2 and wasted spark + GM 4L60e GPIO controlled - both on Extra FW)
Post Reply