Has anyone adjusted their map for ethanol?

Cliff generously gave me his Breva map to use on my V7C, and I am experimenting with it.  Cliff's map was designed for pure gasoline, but in the northeast USA, all fuel is either 10% or 15% ethanol (E-10 or E-15). 

References give these stoichiometric fuel-air ratios;
Gasoline at  1:14.7
E-10 at       1:14.1

My calculations suggest that a gasoline map should be enriched 4% for use with E-10.  Does this sound right?  Anyone have direct experience with this?

Joe

Hi Joe - in the UK our ethanol content is currently less than in US fuel and I believe that of Sweden (Raz?). The UK will follow in due course no doubt. A Qinetiq reporthttp://www.dft.gov.uk/publications/qinetiq-10-02471/ mentions "enleanment" (section 3.1) and at E10 your 4% may be prudent however even more fuel may be required to counteract the high oxygen content in ethanol - 35% of ethanol is oxygen - as mentioned in this report. Interesting as the lambda sensor suggests richer settings on my 1100 sporti than I imagined necessary. Not a chemist but worth considering.

Gavin

Gavin, it may never happen in the UK.  A major reason we have it in the US is that corn is a big crop, and farmers have political friends. 

I have just started doing road tests, and will probably base map enrichment on spark plug condition. 

If you have a better way to measure, let me know!

Have you considered the ecu tie in with a lambda sensor? It takes some of the guess work out of the process and enables pretty serious logging. My bike seems to like more fuel than the ideal stoichiometry value suggests although this may be a calibration issue with the sensor - to be checked this year with a sensor from my friendly local http://www.hmdtuning.co.uk/motorcycletuning.htm dyno centre. As I am unclear what motorcycle / ecu combination you are running it may be worth checking this out with Cliff. Plug chops are ok but for continuous monitoring you can't beat a wide band sensor - a "dyno on the road" basically.
I, and other MyECU users operate the Innovate LC1 http://www.innovatemotorsports.com/products/lc1.php system - not particularly cheap, but converts my 1100 sport FI into a closed loop system - if I want.

Worth checking out the threads on here about this facility and user experiences.

Our petrol is presently at E5 - I think. I run an Isuzu diesel p/up day to day and so haven't looked at the petrol pumps this year yet - still some snow on the roads around here. Hopefully the doomsters warning of E10 or E15 will be proved wrong but the EU trend is to move towards less CO2 fuel by-product so it's inevitable I guess.
Gavin

I've got a 2009 V7C, which has a single lambda sensor for both cylinders.  I haven't run it closed loop yet, as I want to get the map as correct as possible first.

The real bug is that one cylinder is running much richer than the other.  Does MyECU allow for separate maps for each cylinder?

In that case I'd check the injectors before I'd start fiddling with the map. I had this on my V11, when one injector was obviously worn and so let through about 15% more than the other one. First I thought the other one needed cleaning and tried this and that without any good results. Only changing them could help. This was after 120.000 km, though.
Of course MyEcu allows you to address one of the injectors with different (offset) values, but I can't say whether the offset goes on the right or on the left side (from the rider's view).

Interesting thread, btw :)

Hubert

Hubert beat me to it! First off why the "much richer" issue? Check for air leaks around inlets, flush the injectors with cleaner and go through the set up procedure to ensure valves, balance and bypass are all tickety boo.

Check out the MyECU website map section - the map can be offset + or - 7% for 2nd cylinder.

for what it's worth I did a LOT of dry-running (you could just as well call it masturb**ing) with ethanol (from E-10 up to E-85) and I don't really think you need to do anything for E-10. But 4% does sound plausible.

The beauty of a wide band sensor is it will tell the needed adjustment regardless of how much ethanol is in there. And you don't need to tell the sensor about the ethanol, it will tell lambda anyway. Regard any fuel as stoich 14.7, set target at 13.2 and you'll be more than fine.

The richness of one cylinder existed with the orig. ECU and orig. pistons (got a big bore kit now).  It may well be a faulty injector, but they are usually reliable items, and I hate replacing expensive parts just to "hunt" for the problem.  Changing the cyl offset would be a short term fix, and if all is well I could leave it at that. 

I've adjusted the valves, balanced the throttle bodies, and chased some earlier intake air leaks that existed when I put in the big bore kit. It's unlikely the cause here, but I am open to ideas as to the problem.

Your suggestions are right; in that I've got to fix this before I start changing the map to correct for ethanol.

In part, I am loathe to replace the injectors because I've never done it before.  Changing carburetor jets was so much easier (sigh).

There are companies that will service injectors and check their flow rates etc. I think their pricing will be a lot less than buying a new injector.
I'd start with a good injector cleaner first.

YaBB Administrator wrote on 04/10/12 at 06:56:29:

There are companies that will service injectors and check their flow rates etc. I think their pricing will be a lot less than buying a new injector. I'd start with a good injector cleaner first.

Yes, I found a company that'll clean, test, and flow map the injectors for $50 US for the pair.  Right now I'm swapping the right + left injectors, and then a plug chop, to confirm that the injectors are a problem.

I've got a 09 V7C.  How do the injectors attach to the fuel line?  The electrical connector was fixed with a metal clip, but the connection to the fuel line isn't clear.  Anyone have experience with this? 

E-10 fuel has a stoichiometric level that's 4% richer than pure gasoline.  Since the oxygen sensor is what monitors the air-fuel mix, I'll adjust that.  A 4% reduction in the seven O2 sensor target voltages should make it kosher for E-10. 

Voltage changes aren't one-to-one proportional to air fuel ratio changes, but it's the best I can shoot for.

Does this sound reasonable? 

sign216 wrote on 04/28/12 at 02:00:45:

E-10 fuel has a stoichiometric level that's 4% richer than pure gasoline.  Since the oxygen sensor is what monitors the air-fuel mix, I'll adjust that.  A 4% reduction in the seven O2 sensor target voltages should make it kosher for E-10.  Voltage changes aren't one-to-one proportional to air fuel ratio changes, but it's the best I can shoot for. Does this sound reasonable?

If you run open loop, yes. But if running closed loop, you should not alter the targets: If you run closed loop you can pour whatever mix of fuel the engine will accept, and it will adjust automatically. The end result may very well be that it gets 4% more fuel and if you run in AutoTune this might end up in your map too. This is because the sensor does not actually measure A/F but Lambda and your targets are actually Lambda targets.

That was the short answer. The slightly longer answer is that even the Lambda target may actually change a little but this is for other reasons and for E10 (as opposed to E85) I really do not think it will be significant.

This is just off the top of my head and I am a bit rusty  :)

Raz

If you want it as good as possible you must reprogramm your controler. As the WB probes usually give Lambda the shown AFR is the product of Lambda * AFR_stoich. If you don't correct the AFR_stoich value in the lookup tables you will always get slightly shifted results.
I wouldn't care about that. I guess the position of the probe, resonances in the exhaust, losses during valve overlap etc. will have a bigger effect as this AFR_stoich value.

Hubert

Luhbo wrote on 04/28/12 at 05:49:16:

If you want it as good as possible you must reprogramm your controler. As the WB probes usually give Lambda the shown AFR is the product of Lambda * AFR_stoich. If you don't correct the AFR_stoich value in the lookup tables you will always get slightly shifted results.

I believe this is not correct, at least not for most controllers. All sensors (and this is without exception) are actually measuring lambda, nothing else. Everything is lambda inside the black box: The targets, the output, the correction.

But for cosmetical reasons, there is a programmable factor so you can get the output, and set the targets, as A/FR. This defaults to 14.7 (traditional gasoline), ie. Lambda 1.0 == AFR 14.7. But again, this, is purely cosmetical.

Now, you can set that factor to whatever you want and it does not actually affect operation at all. If you actually run a fuel with a stoich of 16.92 or 3.14 but leave the setting as 14.7 you can just set your targets like you are used to: eg. 12.8-13.2. Even people that know all about this do this on purpose because they are used to these figures. What really happens is that your REAL targets is Lambda 0.8-0.9 and the sensor will read Lambda so everything will turn out OK.

Raz

Raz and Luhbo, my equipment is different.  I have an 09 V7 with a factory narrow band sensor.  My MyECU module is a later one, that shows the O2 sensor output as a voltage from 0 to 1 (although 5 is listed as a possible maximum).

I made a mistake in my previous posting.  I think reducing the O2 target voltage makes the air fuel ratio (AFR) leaner.  I did two runs, logging both, with two different O2 volt adjustments (-4, +4).   It seemed to run better with the +4 map, but I'm not an unbiased observer.

I'm having a little trouble reading the data logs.  I'll put that in another post, it will be more "searchable" for the next guy starting out.

The factory narrow band sensors are a small step up from the earlier narrow band sensors. There is some gradual curve.
You should stick with the voltage targets in the original map that I use on my Breva 750

Did you change the targets from +ve to -ve ? You can't do that. The sign is to reflect they slope of the curve from the sensor. +ve is for the richer is a higher voltage.
Using the correct sign will give closed loop operation. Using the wrong sign will just peg the compensation at + or - 25%

YaBB Administrator wrote on 04/28/12 at 09:38:38:

The factory narrow band sensors are a small step up from the earlier narrow band sensors. There is some gradual curve. You should stick with the voltage targets in the original map that I use on my Breva 750 Did you change the targets from +ve to -ve ? You can't do that. The sign is to reflect they slope of the curve from the sensor. +ve is for the richer is a higher voltage. Using the correct sign will give closed loop operation. Using the wrong sign will just peg the compensation at + or - 25%

I didn't change any signs, - or +.  When I wrote -4, I meant a 0.04 volt decrease of the target voltage.  i.e. 0.81 instead of 0.85.

E10 fuel operates best with a 4% richer mixture.  Yea, you can ignore it and be happy with slightly diminished performance.  But being happy with "slightly diminished performance" isn't what MyECU is all about, is it?

MyECU allows a rider to account for the variations of his conditions, and I'd like to take advantage of that.

sign216 wrote on 04/29/12 at 08:24:39:

E10 fuel operates best with a 4% richer mixture.  Yea, you can ignore it and be happy with slightly diminished performance.  But being happy with "slightly diminished performance" isn't what MyECU is all about, is it?

I'm all for it. But your sensor (even if it's a narrowband) reads lambda. And your targets are lambda (no matter in what form they are represented). So if you go from pure gasoline to E10, your targets (if correct in the first place) will need no change: Your closed loop system will end up with a 4% richer mixture if that is what's needed. This is the whole point of running closed loop.

Raz

raz wrote on 04/29/12 at 08:43:53:

sign216 wrote on 04/29/12 at 08:24:39: E10 fuel operates best with a 4% richer mixture.  Yea, you can ignore it and be happy with slightly diminished performance.  But being happy with "slightly diminished performance" isn't what MyECU is all about, is it? I'm all for it. But your sensor (even if it's a narrowband) reads lambda. And your targets are lambda (no matter in what form they are represented). So if you go from pure gasoline to E10, your targets (if correct in the first place) will need no change: Your closed loop system will end up with a 4% richer mixture if that is what's needed. This is the whole point of running closed loop. Raz

The lambda target is set up for pure gasoline.  E10 has a different stoichiometry.  Here are the different AFRs;

Gasoline at  1:14.7
E-10 at       1:14.1

The O2 sensor will give you an adjustment, but it's trying to get you to 1:14.7, which isn't what E10 needs.

sign216 wrote on 04/29/12 at 09:08:08:

raz wrote on 04/29/12 at 08:43:53: sign216 wrote on 04/29/12 at 08:24:39: E10 fuel operates best with a 4% richer mixture.  Yea, you can ignore it and be happy with slightly diminished performance.  But being happy with "slightly diminished performance" isn't what MyECU is all about, is it? I'm all for it. But your sensor (even if it's a narrowband) reads lambda. And your targets are lambda (no matter in what form they are represented). So if you go from pure gasoline to E10, your targets (if correct in the first place) will need no change: Your closed loop system will end up with a 4% richer mixture if that is what's needed. This is the whole point of running closed loop. Raz The lambda target is set up for pure gasoline.  E10 has a different stoichiometry.  Here are the different AFRs; Gasoline at  1:14.7 E-10 at       1:14.1 The O2 sensor will give you an adjustment, but it's trying to get you to 1:14.7, which isn't what E10 needs.

I'm afraid you totally miss my point.

Gasoline needs an A/F ratio of 14.7 to get lambda 1.
E10 needs an A/F ratio of 14.1 to get lambda 1.

Your closed loop system reads, writes, eats and speaks lambda. That means if you have targets that work fine for gasoline, they will work fine for E10. And they will automatically do the very adjustment you speak about.

Raz

Raz, the lambda ratio that signifies "1" is different for different fuels.  For gasoline it's 14.7 to 1.  For pure ethanol it's 9 to 1.   See here:  http://moodle.student.cnwl.ac.uk/moodledata_shared/cdx%20etextbook/dswmedia/fuelSys/gasoline/fund/stoichiometricratio.html

Since 1999, many modern, sophisticated engines (not Guzzi) designed to run alcohol fuels have sensors that estimate alcohol content and adjust the engine's parameters for ethanol blends.  This is called a closed loop system with adaptive learning.  See this paper from the Nat. Renewable Energy Lab  http://www.nrel.gov/docs/fy03osti/32206.pdf

Ethanol exhaust has different characteristics (see this Swedish paper http://www.best-europe.org/upload/BEST_documents/environment/Exhaust%20emission061129.pdf ), and using an oxygen sensor/ECU set for gasoline will give slightly less power.  Our Guzzis are designed for pure gasoline, and set so that lambda 1 achieves a gasoline ratio.  Ethanol blends require an adjustment to reach gasoline-like performance levels.

sign216 wrote on 04/29/12 at 22:41:34:

Raz, the lambda ratio that signifies "1" is different for different fuels.  For gasoline it's 14.7 to 1.  For pure ethanol it's 9 to 1.

You are still totally missing my point. Let's say you have a target for pure gasoline. Even if you think of it as AFR 13.2, the target is really Lambda 0.9.

Now, run your engine on pure ethanol. Your target is still Lambda 0.9 so without you changing anything, your target is already equivalent of an AFR of 8.1 for this fuel.

If you'd say you want to enrich the base map by 4%, I'd say go ahead. But do not change the targets.

Raz

I apologize.  I did miss your point, and misunderstand what the O2 sensor targets were.  I thought they were actual voltages from the O2 sensor, and not a % of Lambda. 

Some time ago I built a o2 sensor modification that adjusted sensor voltage to enrichen the mix, and I'm accustomed to looking at O2 sensor data in terms of voltage.

I've dropped that mod to go after the broader capabilities of MyECU. 
Thanks for your help.  With the original map I'll log runs, and see if the data points to map adjustments.

sign216 wrote on 04/30/12 at 00:00:09:

I apologize.  I did miss your point, and misunderstand what the O2 sensor targets were.  I thought they were actual voltages from the O2 sensor, and not a % of Lambda.

Oh, but they are the actual voltages. It's just that they represent lambda, not A/F.

Raz

raz wrote on 04/30/12 at 03:03:19:

sign216 wrote on 04/30/12 at 00:00:09: I apologize.  I did miss your point, and misunderstand what the O2 sensor targets were.  I thought they were actual voltages from the O2 sensor, and not a % of Lambda. Oh, but they are the actual voltages. It's just that they represent lambda, not A/F. Raz

We aren't communicating well.