map sensor theoretical curve [Archive]

1975BMW2002

08-04-2004, 09:08 AM

Hey All,

I'm trying to extablish a base fuel curve using a map sensor on our engine. I have a few questions...

first-Should the map sensor go closer to the runner end of the plenum, or the restrictor end? We have a 1-4 splitter that is a few inches long, and have run a TPS on it. now i want to switch to a MAP sensor, and am not absolutely sure where to put it.

Second-I talked to TExas A&M at competition this past year, and they said that they started off with a theoretical fuel map using a map sensor. I have a few thoughts on how to do this. coudl anyone tell me if I am on the right track or totally off base on this? I theorize that all you need to do is calculate how much air your engine could suck in at a specific pressure-say atmospheric since we know what the density of air at atmospheric pressure is- and engine speed. Then put all those values into a spreadsheet for the full throttle values, then scale them down accordingly as the pressure drops. once it hits a half a bar, inject half the fuel, and so on... then at the end of compiling the spreadsheet multiply all of the values by the volumetric efficiency of the engine. It seems so easy in theory. What am I missing. Am I only missing that the fine tuning takes days longer, and idling and cold start up are a whole different can of worms?

Thank you for any advice/comments.

Bill

1975BMW2002

08-04-2004, 09:08 AM

Dan Deussen @ Weber Motor

08-04-2004, 12:29 PM

I don't see the reason to run a MAP based system on a N/A engine. A TPS based system on these restricted engines bas proven to be quite reliable and accurate. Throttle response of a TPS based system is also much better.

If you do decide to run a MAP sensor for primary load sensing, I would experiment with the sensor location. Finding a location at which you can get a nice steady reading can be a hard task on these typically small intake manifolds. There are a lot of pressure waves going through the manifold which can affect your MAP readings by quite a bit.

Charlie

08-04-2004, 04:48 PM

I would be interested in a discussion of TPS based mapping versus MAP based.

In 2003 we had to start over (and really, the stuff we'd done in the past wasn't that great, at least under my watch) because we had a new MoTec ECU. There was quite a bit of discussion about TPS based mapping versus manifold pressure. I was in favor of TPS because I thought it would be easier to simply load the engine and run different throttle angles. You can likely achieve all your data points easier with a TPS system than MAP (you just can't get to all the MAP points, some interpolation is always required).

However our cheif engineer that year did basically all the mapping (I was captain and got to do cost report instead, ouch). He started out with a TPS map but was not satisified at all with the progress. I'll have to ask him why, but I think the delay because of the large plenum was a major factor. He ended up using a MAP setup and got it done quickly and honestly did a fantastic job for the timeframe and the crappy equipment we have. He did spend some time creating a spreadsheet like you are talking about Bill, and although I don't remember how well it correlated it is definitely a good starting point.

As far as MAP sensor placement, I would expect that most places are fine, as long as your calibration is done there and it doesn't change. We've always run it around mid-way between the runners and restrictor on the plenum itself. A good steady reading has never been an issue, even when logging at 50 hz it looks fine. If you want to dampen the signal put it on a vacuum line with the nipple drilled with a small orifice. But I don't think its necessary.

Our throttle response is great. I couldn't be happier with it.

I'd still like to know the benefits of a TPS setup. I know some race teams that use that route, but honestly seems like a MAP setup is always going to be more representative of airflow.

BryanH

08-05-2004, 06:20 AM

Charlie, Speed/density (map sensor load sensing) based mapping actually uses manifold vac. to estimate cylinder filling. If the engine has less than 225deg duration cam timing, uses a plenum < 1.5 x total capacity and single throttle body it is quite accurate. BUT if volumetric eff. changes in any way,usually a change in ex. backpressure, part throttle A/F ratios will vary dramatically. (Full load very little change).
As long as you are aware this is not a problem.

A far more dangerous problem occours when you use speed/density on an engine that doesn't pull very good manifold vacuum combined with individual throttle plates. Full load A/F ratio becomes overly sensitive to barometric pressure, eg -5% baro = -20% fuel!!!

TPS based mapping is called Alpha N. The throttle plate angle is a very good measure of mass air flow, combined with MAT and barometric pressure compensation. I find mapping an Alpha N system quicker and far more accurate. But it can initially appear impossible until you understand how the maps should look. First start up is more difficult than speed/density.

Eg.600cc 4cyl at a set rpm point say 6k the tps load points can be used to smoothly bring lamda from 1 at 5%, to .93 at 40%, .9 at 60% and .88 above that.

With speed/density this is not possible as vac has dropped to atmospheric at around 40% tps and you have only one load point to adjust. The lower you go in the rpm range the smaller this tps % becomes for map to go atmo.

This effect is even more important for the timing map, Torque output can be controlled by throttle position, and idle speed can easily be locked to a desired range. Virtually every atmo race car in the world uses Alpha N load sensing. There is lots more but post is already too long.

"The toes you step on today are attached to the legs holding up the ass you may have to kiss tomorrow"

Charlie

08-05-2004, 07:49 AM

Of course I know that MAP estimates airflow, not sure what your point is? Does a TPS system not estimate airflow, but somehow just know?

Thanks for the information. I'm not sure I agree that with 40% throttle you see atmopsheric. That is highly dependent on your throttle. But thanks for continuing the discussion, I will be looking into this a bit more, any more comments you want to throw in please do.

Just curious, since you mentioned multiple throttle plates obviously this is not FSAE. What is your experience with other engines? I do know that most of the race engines I've been told are TPS based are in fact multiple throttle body engines.

BryanH

08-05-2004, 09:05 AM

Charlie, throttle plate angle appears to be a very acurate measure of mass air flow. (If the engine has a pressure fed air box baro needs to measured in air box)
In fact I have twice had modified throttle bodies (lumpy polishing!) give very weird varying lambda at part throttle which must be down to eneven flow rates/turbulance. Stock throttle bodies always come up with very linear maps.

The 40% was a very hypothetical #

maybe 80% of the atmo cars I do are single throttle body (500+hp Ls1 last month) Road cars are usually speed/density so as to use isc if they are near stock.
Turbo road cars use speed/density
Nissan GTR twin turbo must be done AlphaN with map conpensation
Big overlap cam supercharged engines use AlphaN with map compensation!

F-SAE engine doesn't know its a F-SAE engine.

btw RMIT's wr450 used Haltech E6X in AlphaN mode
std. trigger.

Your post on engine design should be stuck up on the wall of every f-sae team in the world.

Winners don't build excuses into their cars

Charlie

08-05-2004, 10:45 AM

See I guess my big question is how can a throttle give you a good representation of mass flow? I understand you are talking from experience so I'm not saying it isn't true. But if pressure was equal on each side of the throttle it could not represent mass flow at all. If you have an orifice, you can accurately determine mass flow only if you know the pressure on each side. And I suppose that with RPM that is estimated, but also changes with VE.

What post are you talking about? As you can see I have so many I have never really accomplished anything in the real world. http://fsae.com/groupee_common/emoticons/icon_wink.gif Thanks though.

murpia

08-06-2004, 02:19 AM

For me, the main difference between 'speed density' (MAP based) and 'Alpha-N' (TPS based) fuel strategy is the ease of tuning when making physical changes to your engine, and the error tolerance.

For example, say you decide to swap your butterfly throttle for a barrel, to gain performance. That would require a complete re-map from scratch with Alpha-N, while with Speed Density you might need to make very few (if any) changes.

Once properly mapped with A1 hardware there will be little if any performance difference, but in control system terms the Speed Density system will be more 'robust' to things like throttle cable stretch, idle stop adjustment, noisy TPS signal, etc.

Ian

BryanH

08-06-2004, 05:54 AM

Charlie 906 posts! time to get a real job.

post was; posted August 03, 2004 06:36 PM
Typically new teamers.........

The question is why wouldn't it?
The VE won't change,(unless rod knocks plenum off etc.) mass air temp is compensated for, and baro pressure is realtime compensated for. btw baro compensation is linear- 500 Mbar is -50% fuel compared to 1000 Mbar.
After mapping near 1000 engnes imho Murphia has got it ass a peak, Alpha N is insensitive to setup changes unless the VE actually changes and any change in VE can be seen in the lamda graph on the dyno sheet, leaner means VE has improved.
If the throttle body size/style is changed the only change to the fuel/rpm maps needed is to the slope of the inj time between 0% tps and tps % @ max VE on each rpm map.

Alpa N is used on competition cars because the laptimes say its better.
To drive I'd liken it to the difference between condom and no condom.........
Bryan H

Charlie

08-06-2004, 07:15 AM

I guess what I take from this is that either method can bring the same result. However a proper MAP setup must include throttle compensation, and proper AlphaN setup must include MAP compensation.

Except throttle compensation on a MAP is a tricky thing to do, but MAP compensation on a throttle setup is easy.

1975BMW2002

08-10-2004, 08:53 AM

Wow,

I think I am more confused now than before. The reason that i wanted to go with the map sensor was that i wanted to build a knowledge base that woudl be applicable to a future team that may be running forced induction. Next year's team is still deciding on it. I was not aware that TPS was superior in some respects, just different.

I guess my confusion comes from the fact that I talked to an egine guru for a few minutes. in our short conversation he said to calibrate the engine for .5 pounds per horsepower-hour. once i did the math with my assumed hp of 78, an assumed VE of .90, and 12 pound injectors, it meant having the injectors open for a lot longer than I know they should be open.

I figured I should base the map on a 13.0 to 1 air-fuel ratio, and try the map like that.

Is there any wisdom on starting points for a MAP based curve?

Bill

Chris Boyden

08-10-2004, 09:48 AM

What ECU are you running? Some ECU's build an initial map (tuning wizard) to get you started. But that's all they do...

The pulse width's vary depending on several ariables...namely fuel pressure and injector size. So it's hard to say what the exact pulse width should be...if you can calculate it, more power to you.

We did the initial tuning by ear (no load) until it sounded good and didn't smell too rich. A wide band o2 sensor can help alot here. You can go through all of the throttle positions and adjust the pulse width until you get the A/F ratio you want...13:1 is good. But, developing a good feel for the sound and response also works well. Also 13:1 is a good place to start, but you want to set the A/F ratio to whatever the engine wants to make make power, which may change at different rpms.

We placed the map sensor at the top of our triangle shaped plenum right next to the plenum restrictor flange. We didn't seem to have pulsation problems with the sensor. We also used TPS compensation, which helps the throttle response alot! It wasn't that hard to figure out....

The general shape of the map follows the torque curve of the engine.

When we went to the dyno, we followed the same procedure that you'd follow for a TPS system.
We started at a low rpm and set the load, adjusted the pulse width until we got a good A/F or max power at that point. We'd then increase the load and open the throttle a little to try and maintain the same rpm. Then adjust the pulse width....over and over ad nauseum. Actually, it's alot of fun.
Dyno setups vary, so there are several approaches to take.

The whole process isn't terribly difficult. remember if it sounds good...it's probably not too far off. We pulled 60 hp on the first dyno run, just tuning by ear. From there, we got 20hp more (at a lower elevation than Albuquerque) Just make sure you change one thing at a time, and be systematic, save changes as you go into new files, and keep notes. The two engines that I setup both run map w/ tps compensation. I've had no problems. Both have good throttle response.

As far as your calculations for pulse width, you can also raise the fuel pressure if you're careful to not lower the fuel flow too much. That'll decrease the duty cycle to an acceptable range.

Good luck

Moody

08-17-2004, 01:26 PM

Awesome discussion, guys - funny how questions are answered even before you ask them around here ...

Chris, it seems that you've had some success with throttle compensation; what is your peak throttle percentage? Our MAP signal is erratic to say the least, and putting in enough TPS to calm it down has made me second-guess our decision to use a MAP-based load system.

Thanks!

- matt

Chris Boyden

08-17-2004, 04:51 PM

Moody,

We're running about 150% adjustment.
and 100% for the tps%/s. and I have it kick on
at about 30% tps.

If the map is erratic, go with tps. But we've never had problems with instability with the map sensor....so never felt like we had to go there.

The tps tuning was done off the dyno. Maybe somebody else out there has figured out a good method for map/tps compensation on the dyno.

Moody

08-19-2004, 08:41 AM

Has anyone else had erratic MAP signals? I haven't considered the idea of noise, just assumed that the strange flow inside the plenum made for non-stable pressures. Maybe we oughta check the noise factor a little more closely. =-/

(?)

Trans Am

08-31-2004, 08:14 AM

I'm loving this Post...

For the last few years we have been using speed/density for our engine management. so far it has proven to be a much more robust control stratgy over Alpha-N. The TPS that we run only compensates for transient fueling as their tends to be a short delay in plenum pressure with a change in demand. I don't know about cranking being more complex, I do know that speed/density control offers a much more accurate (initial) calculation of the true air entering each cylinder, meaning, less compensation is needed to correct the A/F ratio.

I'd be happy to keep this debate going, this is a good information thread and provides a look into two unique control stratigies.

BryanH

09-01-2004, 06:10 AM

Holy Crap! you had better get on the phone to Ferrari F1, Ilmor, Renault F1,Cosworth etc. (and RMIT) and tell them they are doing it wrong.
Bryan Hester

Igor

09-01-2004, 06:56 AM

Well, that's not really fair as I don't think formula one cars have an intake plenum holding a relatively predictable low pressure. If you only have some runners sticking into open air it is going to be impossible to do a MAP control there. (unless you do some really slick signal processing)

Igor

BryanH

09-01-2004, 07:58 AM

Igor, if Map was better F1 would be using it!
Instead of trying to justify mapping directions you have taken, would it not be prudent to research methods tried and proven in F1 since the sixties. Lucas shuttle mechanical injection that was used on the Cosworth 3litre F1 engine was an Alpha N system. As was the Bosch mechanical system used by Porsche in the 70's. They won a few races........
Lucas did a Map compensated version for the Triumph 2.5PI and IMO was it was a thirsty dog.

Please name one winning N.A. race car in the entire world using map. (if it has tyres outside the fenders or a gun rack it doesn't count) Prize will be 1 RMIT demon tweak

This is not about the Hockey Gold medal is it?

Mark Peugeot

09-01-2004, 08:13 AM

<BLOCKQUOTE class="ip-ubbcode-quote">quote:<HR>Originally posted by Halfast:
Igor, if Map was better F1 would be using it!
Instead of trying to justify mapping directions you have taken, would it not be prudent to research methods tried and proven in F1 since the sixties. Lucas shuttle mechanical injection that was used on the Cosworth 3litre F1 engine was an Alpha N system. As was the Bosch mechanical system used by Porsche in the 70's. They won a few races........
Lucas did a Map compensated version for the Triumph 2.5PI and IMO was it was a thirsty dog.

Please name one winning N.A. race car in the entire world using map. (if it has tyres outside the fenders or a gun rack it doesn't count) Prize will be 1 RMIT demon tweak

This is not about the Hockey Gold medal is it? <HR></BLOCKQUOTE>

There are much to your dismay cars that have won using speed density. I can name several drag race cars for starters and a number of road race cars that have done it, including some endurance cars running peripheral port rotary engines.

If you can make a P-Port rotary successfully win road races with a MAP based system, just about anything can.

Am I saying that I like speed density better than Alpha-N? Nope, I think it does not really matter... either the motor is tuned right or it ain't. How the fuel and air gets in the cylinder is unimportant. (yes, I understand that it would be best to have a homogeneous mixture!) Having the right amount at the right time is what is important. Speed density or alpha-N, it does not matter. Speed Density with the right throttle pump values will work just fine if properly programmed.

Mark

Igor

09-01-2004, 09:01 AM

Well, I don't claim MAP is better at all for NA applications. In fact I think it's almost impossible to do MAP on an intake system that only consists of a couple of runners like on race engines.
I'm still not convinced though alpha-n is the best method for an intake with a big plenum behind a restrictor. But I am not blessed with any practical experience either....

Not mad about the hockey at all, somebody has to win. Besides, we have the 100m freestyle in swimming :-)

Igor

Trans Am

09-01-2004, 10:32 AM

Alpha-n is easier to tune, setup, and more wildly used in production vehicles(and therefore proven); these reasons play a big roll in why the majority of teams run alpha-n. Speed/density provides a better air flow calculation to begin with and offers more precise control of fueling at the disadvantage of being super complex to tune and setup. So am I saying one is better than the other? No. It all comes down to how well you tune your engine.

I prefer speed/density because I have the most experience tuning it.

BryanH

09-02-2004, 05:52 AM

<BLOCKQUOTE class="ip-ubbcode-quote">quote:<HR>Originally posted by Mark Peugot

There are much to your dismay cars that have won using speed density. I can name several drag race cars for starters and a number of road race cars that have done it, including some endurance cars running peripheral port rotary engines.

Mark <HR></BLOCKQUOTE>
No dismay in OZ, the home of aftermarket engine management buddy! Drag cars and 20 year old race cars are simply not relevent to this forum.
Map tuning a 13B P.P. is a bit like getting a women to tie you to a barbwire fence and whip the crap out of you before making love, a few people enjoy it but most of us can't see the point.

Alpha N (TPS load control, Trans Am) has several advantages over map based which is why every current F1 car,F3 car,OZ V8 Supercar, British Supertourer,German DTM car uses it and I want to prod you guys into finding out why.
Bryan Hester

Mark Peugeot

09-02-2004, 08:13 AM

<BLOCKQUOTE class="ip-ubbcode-quote">quote:<HR>Originally posted by Halfast:No dismay in OZ, the home of aftermarket engine management buddy! Drag cars and 20 year old race cars are simply not relevent to this forum.
Map tuning a 13B P.P. is a bit like getting a women to tie you to a barbwire fence and whip the crap out of you before making love, a few people enjoy it but most of us can't see the point.

Alpha N (TPS load control, Trans Am) has several advantages over map based which is why every current F1 car,F3 car,OZ V8 Supercar, British Supertourer,German DTM car uses it and I want to prod you guys into finding out why.
Bryan Hester <HR></BLOCKQUOTE>

Using MAP based tuning on the P-Port was actually rather trivial. I disagree that 20 year old race cars are not relevant to the discussion. There is a lot that can be learned from some of those older race cars.

I've personally learned a lot from some of these "old" race cars. Some of which have technology that is well ahead of what I have seen on many FSAE cars.

I'm not trying to say that MAP is the answer for the FSAE cars, as a matter of fact I think that with the restrictor I'd want to stay away from the MAP sensor since it will go to full vac long before we see full throttle at speed.

Does this mean we can't use speed density? No, it means that we have to pay careful attention to things like throttle pumps and IDLE & WOT maps that we normally would not do in these cases.

If you want to know why everyone is using Alpha-N that's pretty simple in my mind. It's easy to do, it doesn't suffer from some of the inherent problems that occur with speed density cars and with feedback from a variety of sensors and extensive dyno programs there is no reason to do it any other way. Accurate fueling is easy to accomplish when you are monitoring the exhaust stream as carefully as the average F1 team.

Mark

Trans Am

09-02-2004, 09:11 PM

Mark makes a good point that 20 year old race cars are still very relevant and should not be overlooked on this forum. Engine power is still made the same way, it's just managed differently. Suspensions of yesturday define the basics of every car today. Take a step back, Halfast, and maybe you'll learn a thing or too.

Anyhow, speaking about vacuum, it's the same way with Alpha-n, you reach full vac long before full throttle (unless you have a really tiny throttle). Each stratgy has it's own trade offs and it all comes down to tuning.

Moody

09-06-2004, 12:24 PM

looks like this is turning into a somewhat circular discussion, so allow me to branch off with some ignorance:

What exactly is meant by throttle pump?

I'm hearing lots of general talk here; while I understand that past experience is all well and good, and 20 yr old road cars can give some insight into possible solutions to the FSAE powerplant problem, what are some of your specific experiences with a restricted, 600cc motor? Obviously, we're all hitting full vacuum before full throttle; doesn't that effectively nullify any throttle beyond that point? (other than richening the mix)?

Can we all agree that start-up teams in this competition should probably start with alpha-N (TPS) load calcs because of resource and time constraints?

- matt

Charlie

09-06-2004, 12:34 PM

<BLOCKQUOTE class="ip-ubbcode-quote">quote:<HR>Originally posted by Moody:
What exactly is meant by throttle pump?
<HR></BLOCKQUOTE>

Extra burst of fuel during rapid throttle opening

<BLOCKQUOTE class="ip-ubbcode-quote">quote:<HR>
Obviously, we're all hitting full vacuum before full throttle; doesn't that effectively nullify any throttle beyond that point? (other than richening the mix)? <HR></BLOCKQUOTE>

Yes it means that after a certain throttle position, extra throttle does not increase airflow. And if you engine is mapped correctly it should not increase fuel either! Personally I think it is good to have 'full throttle' a bit before 100% to ensure that the drive is always at full throttle if he wants to be (if he's only pressing 95% throttle, he probably wants 100%). Plus it gives you a bit of extra assurance that your throttle is not too small. But if you are hitting full airflow at 70% throttle, your throttle is way too big.

<BLOCKQUOTE class="ip-ubbcode-quote">quote:<HR>
Can we all agree that start-up teams in this competition should probably start with alpha-N (TPS) load calcs because of resource and time constraints?

- matt <HR></BLOCKQUOTE>

No. Like I stated earlier our guys attempted TPS based mapping as beginners and found MAP based to be much simpler, and had very satisfactory results. Which one is superior is (obviously http://fsae.com/groupee_common/emoticons/icon_wink.gif ) debatable... http://fsae.com/groupee_common/emoticons/icon_razz.gif

John Bucknell

09-06-2004, 04:53 PM

As somebody who does this control system stuff professionally, I gotta say there are all sorts of control systems out there - each with their specific advantages:

Alpha-N is about the only system practical on individual runner (or individually throttled) engines. This is due to the fact you can't get a decent density signal with a highly tuned intake manifold for ultimate throttle response. It isn't because it is superior for control, in fact look at your F4i or even F1 engines that have airbox pressure input to compensate for density changes. So you have to tune on the dyno, ie you can't make a theoretical model that works and dial it on the computer.

Speed-density is the primary solution for production cars out there, and every boosted car except those that are individually throttled (like Skyline). This setup eliminates the fact that your orifice stops throttling at 10% of travel at low engine speeds, so fine fuel control is easy. However, it is slower than Alpha-N and needs some acceleration compensation. One extra bonus is the altitude compensation is built in.

MAF systems are the least easy to develop for race cars, even though they should be the easiest. This is due to drasticly worse transient fuelling compensation due to measurement lag, so all sorts of compensation are necessary - even up to adding a MAP sensor. On the plus side they're good for engines that run pretty steady-state.

I'm open to the idea of Q&A on control systems....

Trans Am

09-06-2004, 06:26 PM

Hey John,

I've been searching for some good books, or papers, on control stratigies, can you recommend any?

John Bucknell

09-06-2004, 09:22 PM

I'd suggest the Bosch handbooks as a starting point Bosch Site (http://www.bosch.de/start/content/language2/html/734_2812.htm) . I found this link just now on control systems referring to the Bosch stuff-> MoBIES (http://vehicle.me.berkeley.edu/mobies/powertrain/)

Also, I subscribed to the DIY EFI mailling list for the longest time, which can be found here -> DIY EFI (http://www.diy-efi.org/diy_efi.html) out of which the EFI332 project emerged which has descriptions of algorithms.

I guess ultimately I've been trained on the job, and there wasn't any assigned reading giving a good overview. I've had to distill out opinions based on experience and where the industry is. Which is why I suppose I've offered to answer questions, as I don't think there is any good literature. Motec's literature is okay, as is some of the other aftermarket control systems - but pretty basic, with no objective comparisons. I'll do some more research and get back to you guys...

murpia

09-07-2004, 02:56 AM

<BLOCKQUOTE class="ip-ubbcode-quote">quote:<HR>Originally posted by John Bucknell:
Also, I subscribed to the DIY EFI mailling list for the longest time, which can be found here -> http://www.diy-efi.org/diy_efi.html out of which the EFI332 project emerged which has descriptions of algorithms.
<HR></BLOCKQUOTE>

DIY-EFI evolved into MegaSquirt, which I know has been implemented by at least 1 FSAE team. The http://www.megasquirt.info/index.html site has a lot of useful info, much of which can be generalised to any speed-density system with (optionally) TPS compensation and closed-loop lambda.

Ian