Air Temperature Sensor - Improving Engine Performance?

[AndrewA]

Ben, please bash the FRK in the FRK thread! Not here!

Dr. Foster, please stop with the spam! This thread is not about the FRK!

[boxercup]

Originally Posted by boxercup

Right you are Bill!

It takes a lot of data to analyze and perfect!

R&D is not riding around a lake!

Without repeatable verifiable empirical data you can not quantify anything!

The Dyno is the only way you can quantify!

Then you can take a ride around the lake with a portable 5 gas analyzer/data logger to proof your work.

Ultimately, only then you can have the data needed!

Sorry Andrew,

I see no spam here!

Just good advice.

Best!

[jduke]

Andrew, I hope you follow this to some kind of finish by taking up Robert and his FRK's return policy and actually comparing what you have done with what the FRK can do.

[Bob Hancock]

Now there's a good idea. Owen and Andrew do the very best temp sensor mod they can using whatever resources they care to throw at the project. Evaluate. Then, get an FRK and evaluate. Good news is if you don't like the FRK, send it back for a full refund and all you've invested is your resistor research. I'd be all ears.

[bill pierce]

Originally Posted by Bob Hancock

Evaluate. Then, get an FRK and evaluate.

+1 to that. I don't see how you could tune it, without at least an exhaust gas analyzer though. We just bought a new one at work that I use on boilers. ($5000).

Seat of the pants is no way to tune an engine. It's just asking for a melted down piston.

[OwenM]

Adding a resistor in series with the temperature sensor can never make the bike run leaner than it would without it, you can only go richer which will not melt pistons or cause engine damage.
If you go too rich you will feel it, but it appears there is just not enough trim factor available with changes in temperature sensor input to give a significantly over rich condition.
Using a conventional pot which gives a visual indication of what setting is in use rather then the multi turn trim pot shown in the pictures above is a better idea until you work out what setting to finally use. Remember that you must allow 30 seconds or more for adjustments to take effect as the temperature sensor input is very slow to react, you can’t just turn the pot and get an instant reaction.

I see no need for a dyno or gas analyzer for set up unless you are only interested in numbers; most riders want a setup that feels the best to them for the way they ride. You are more likely to find that point by experimenting with different settings your self rather than using a predetermined setting with a FRK.

If you have a GS-911 code reader or go to your dealer/mechanic who has the factory diagnostics tool (GT1) you can see exactly what temperature the FRK or your resistor is sending to the ECU. If you have a FRK plug it in and measure what the ECU sees, then plug in your home made adjustable FRK and dial in the same temperature reading, you have then established what resistance is required to do the same job as the FRK.
Some mechanics may offer to do this for you free of charge just out of personal curiosity, but even if they don’t it only take a few minutes so it cant cost much.

Picture of GS-911 data showing input intake air temperature (21 degrees) as seen by the ECU on right.




GS-911 site.
http://www.hexcode.co.za/


Personally I would much rather spend $299US on the GS-911 than $293US on a FRK.

[AndrewA]

Originally Posted by jduke

Andrew, I hope you follow this to some kind of finish by taking up Robert and his FRK's return policy and actually comparing what you have done with what the FRK can do.

The goal of this project is not to prove that I can make a cloned FRK, or to discredit anyone (either those in the pro-FRK camp, or those few vocal anti-FRKers). Dr. Foster insists that his devise is much more involved than a simple resistor dongle. I’m willing to take him at his word. I don’t have a dog in this fight either way. If it works, great. If not, Ill probably just live with my flat spot and move on to something else.

Noting that, there are already accounts on the web from FRK owners that have compared their FRKs to resistor dongles. There wouldn't be much that I could add...

Originally Posted by Bob Hancock

Now there's a good idea. Owen and Andrew do the very best temp sensor mod they can using whatever resources they care to throw at the project. Evaluate. Then, get an FRK and evaluate. Good news is if you don't like the FRK, send it back for a full refund and all you've invested is your resistor research. I'd be all ears.

I had the pleasure of riding X-Triple’s BCR that has the FRK installed, and came away extremely impressed. It was buttery smooth from idle all the way to the rev limiter. Also, for the few seconds that I was behind X-Triple on his BCR, there was no overly rich smelling exhaust smell. It seems perfect. The FRK is a great product that works.

I am more than willing to document one version of how a resistor dongle might be built, and will provide as much seat of the pants feedback as I can - but I am not really prepared (nor committed) enough to do a full comparison between the two. Also, I'm not sure how valid any test between the two would be without any hard dyno figures (which will inevitably be demanded). Again, I'm just trying to make the bike more rideable within a particular rev range, and couldn't begin to address all of the issues that the FRK portends to improve on.

Regardless, I am more than willing to send the dongle to anyone that might want to invest some time in comparing the two.

Originally Posted by bill pierce

+1 to that. I don't see how you could tune it, without at least an exhaust gas analyzer though. We just bought a new one at work that I use on boilers. ($5000).

Seat of the pants is no way to tune an engine. It's just asking for a melted down piston.

I will be employing Tuning 2.1

[AndrewA]

Originally Posted by OwenM

Adding a resistor in series with the temperature sensor can never make the bike run leaner than it would without it, you can only go richer which will not melt pistons or cause engine damage.
If you go too rich you will feel it, but it appears there is just not enough trim factor available with changes in temperature sensor input to give a significantly over rich condition.
Using a conventional pot which gives a visual indication of what setting is in use rather then the multi turn trim pot shown in the pictures above is a better idea until you work out what setting to finally use. Remember that you must allow 30 seconds or more for adjustments to take effect as the temperature sensor input is very slow to react, you can’t just turn the pot and get an instant reaction.

I see no need for a dyno or gas analyzer for set up unless you are only interested in numbers; most riders want a setup that feels the best to them for the way they ride. You are more likely to find that point by experimenting with different settings your self rather than using a predetermined setting with a FRK.

If you have a GS-911 code reader or go to your dealer/mechanic who has the factory diagnostics tool (GT1) you can see exactly what temperature the FRK or your resistor is sending to the ECU. If you have a FRK plug it in and measure what the ECU sees, then plug in your home made adjustable FRK and dial in the same temperature reading, you have then established what resistance is required to do the same job as the FRK.
Some mechanics may offer to do this for you free of charge just out of personal curiosity, but even if they don’t it only take a few minutes so it cant cost much.

Picture of GS-911 data showing input intake air temperature (21 degrees) as seen by the ECU on right.




GS-911 site.
home — Hex code


Personally I would much rather spend $299US on the GS-911 than $293US on a FRK.

I have access to a GS-911, and will initially calibrate down to -25C (I think - Ill verify this later).

I went shopping today: I got a mixed bag of resistors, a 10 ohm pot (single turn, with fancy knob), and a hobby box. Now all I need are my Bosch connectors to arrive,...and some beer.

[boxercup]

Greetings Andrew,

I admire your enthusiasm. I really do!

However, I think that you have been misinformed.

Owen - You mean parallel don't you?

[OwenM]

Originally Posted by boxercup

Greetings Andrew,

I admire your enthusiasm. I really do!

However, I think that you have been misinformed.

Owen - You mean parallel don't you?

Definitely not, I said series and I meant series, using a resistor in parallel with the temperature sensor will give a lower overall resistance and make the ECU think air temperature is high thus causing a lean running condition and possible engine damage, and we would not want that now would we.

Anyone would think you are deliberately attempting to lead people astray there Robert.

[OwenM]

Originally Posted by AndrewA

I have access to a GS-911, and will initially calibrate down to -25C (I think - Ill verify this later).

I went shopping today: I got a mixed bag of resistors, a 10 ohm pot (single turn, with fancy knob), and a hobby box. Now all I need are my Bosch connectors to arrive,...and some beer.

I do hope that’s a typo, you need a least a 10k ohm (10,000) lin pot to get about -5C, a 10 ohm pot will do nothing.
If you want to simulate -25C a 20k or even a 50k ohm pot will probably be required.

Since you have access to a GS-911 can you do some testing and let us know what value resistance is required to get the ECU to see -25C without the temp sensor in circuit.
Or if the temp sensor is in circuit the required resistor value and current ambient temperature (with the engine cold).

It is not likely that many bikes will need so much fuel that simulating -25C will be required, but it’s useful info none the less.

I also wonder how low a temp the ECU can accept before it bottoms out and won’t go any lower.

[boxercup]

Greetings Owen,

Don't do this at home...

I did not see that you are including the IAT sensor and not eliminating it.

Key word "Dongle".

In that case SERIES is the way!

Thanks Mate!

[AndrewA]

Originally Posted by OwenM

I do hope that’s a typo, you need a least a 10k ohm (10,000) lin pot to get about -5C, a 10 ohm pot will do nothing.
If you want to simulate -25C a 20k or even a 50k ohm pot will probably be required.

Since you have access to a GS-911 can you do some testing and let us know what value resistance is required to get the ECU to see -25C without the temp sensor in circuit.
Or if the temp sensor is in circuit the required resistor value and current ambient temperature (with the engine cold).

It is not likely that many bikes will need so much fuel that simulating -25C will be required, but it’s useful info none the less.

I also wonder how low a temp the ECU can accept before it bottoms out and won’t go any lower.

Yes. That's a typo! The 10K pot and a 10K resistor should get me to -25C. If not, Ill just add more/different resistors. The pot is only temporary.

Why is ambient temperature important?

[OwenM]

I was only interested in the air temperature at the time you where taking measurements because the temp sensor is adding its resistance along with the resistor. If I know the temperature I can figure out approximately how much resistance is due to the sensor if you tell me what value resistance you are adding to the series circuit, does that make sense?
What I want to know is what is the total resistance (sensor plus external resistor) required for the ECU to see -25C.
The sensor on my bike reads about 17k at approx. 0 degrees C, I have not tested it to lower temperatures yet and a deep freeze is only about -18C.

[AndrewA]

Originally Posted by OwenM

I was only interested in the air temperature at the time you where taking measurements because the temp sensor is adding its resistance along with the resistor. If I know the temperature I can figure out approximately how much resistance is due to the sensor if you tell me what value resistance you are adding to the series circuit, does that make sense?
What I want to know is what is the total resistance (sensor plus external resistor) required for the ECU to see -25C.
The sensor on my bike reads about 17k at approx. 0 degrees C, I have not tested it to lower temperatures yet and a deep freeze is only about -18C.

That does makes sense.

I've been bench testing the pot that I got, and am fairly confident that I can accurately measure 0 - 10k to the 1/100ths. Beyond 10K, I will add a fixed 10K resistor in series. The pot appears extremely accurate and exhibits consistant, repeatable results.

I will chart resistance all of the way down to -25C. (edit: and beyond).

Of course, I will be converting all of my results to Fahrenheit (because metric hurts the melon, and is just silly).

[shreddr]

Originally Posted by AndrewA

I will chart resistance all of the way down to -25C.

Of course, I will be converting all of my results to Fahrenheit (because metric hurts the melon, and is just silly).

all gas laws relating temperature, pressure and volume, use a temp value in Kelvin, neither Farenheit or Celsius will work for these calculations. If your result isnt linear, that will be why

[roger albert]

I second the idea of letting the truly technical guys explore and trade facts.
The conjecture stuff will be tossed. Thanks to Owen and Andrew etc for keeping it real.
Interesting.

[Boybiker3]

Originally Posted by AndrewA

That does makes sense.

I've been bench testing the pot that I got, and am fairly confident that I can accurately measure 0 - 10k to the 1/100ths. Beyond 10K, I will add a fixed 10K resistor in series. The pot appears extremely accurate and exhibits consistant, repeatable results.

I will chart resistance all of the way down to -25C.

Of course, I will be converting all of my results to Fahrenheit (because metric hurts the melon, and is just silly).

Some suggestions:

Wire the fixed resistor in, but keep the ends accessible so you can jumper them with alligator clips. That way, you have the advantages of both a 0-20Kohm range and a eyeball readable 10kohm pot. Unless your bike is highly modified, I'm betting you won't end up near 20kohms.

Given the challenge of doing A-B comparisons, the trick is to make the switch as quick as you can - even on the run. So I'd mount the pot on your upper triple clamp. If picking up RF is a concern, just short the wires with small silver-mica cap and use the ferrite bead I suggested earlier. Best place for the cap might be the junction of the pot and the fixed resistor.

shreddr - you're right about gas laws, but the ambient temperature trim factors are arbitrary values that BMW kind of thought were about right and they don't follow any "law," Kelvin or otherwise. Likewise, the NTC sensor (and the inverse network inside the ECU) may not be strictly speaking arbitrary, but they are "unlawful."

It may be fun to "track" resistance vis a vis temperature on that GS-gizmo (I'm not sure what it is). But there is no necessary relation between temperature and the butt dyno, because BMW's trim "curve" is an intervening arbitrary variable. Each rider mostly just wants to see what resistance feels right and, if you own a GS-gizmo, you can find out the temperature later on (or just look it up on the Bosch spec sheet... I may have mentioned this before).

[AndrewA]

Originally Posted by shreddr

all gas laws relating temperature, pressure and volume, use a temp value in Kelvin, neither Farenheit or Celsius will work for these calculations. If your result isnt linear, that will be why

I hope that we are not dealing with extreme enough factors were linear deviations relating to temp scales is a huge issue. Besides, I'm not sure that the GS-911 can display in calvin.

[AndrewA]

Originally Posted by Boybiker3

Some suggestions:

Wire the fixed resistor in, but keep the ends accessible so you can jumper them with alligator clips. That way, you have the advantages of both a 0-20Kohm range and a eyeball readable 10kohm pot. Unless your bike is highly modified, I'm betting you won't end up near 20kohms.

Given the challenge of doing A-B comparisons, the trick is to make the switch as quick as you can - even on the run. So I'd mount the pot on your upper triple clamp. If picking up RF is a concern, just short the wires with small silver-mica cap and use the ferrite bead I suggested earlier. Best place for the cap might be the junction of the pot and the fixed resistor.

shreddr - you're right about gas laws, but the ambient temperature trim factors are arbitrary values that BMW kind of thought were about right and they don't follow any "law," Kelvin or otherwise.

Ben, it has been reported that the ECU can take 30 seconds (or longer) to register changes at the Temp Sensor, so having a pot that can be adjustable on the fly wont really help (particularly in the early going).

I have a "fun" route near my house that is about 30 miles. My idea is to run the route, get home, and then collect my wholly subjective impressions. At that time, Ill either add or subtract resistance, and then go on another run. It may take a while. Perhaps a dozen runs, maybe (much) less.

I'd guess (after reading various interweb accounts) that something under 20k ohms should get me to -25C. My pot and a fixed resistor should allow no hassle, quick adjustments from 10kohm to 20kohm. Seeing that my bike runs in stock trim, I should be well below 20k ohms of resistance when I'm done. Or not....