Powercommander LCD unit

[JuhaV]

Hi,

 

After having sold my Sport Corsa 1100i, I have not been writing to this forum lately, but remembering that here are people that have considered lamda sensors, ECUs and especially Cliff's remarkable My16M and My15M units in depth, I decided to ask your opinion regarding the following.

 

Please see the threat here.

 

Sorry, that is a Honda RC51 forum but I am driving one of those nowadays

 

The key of the question that I have tried to present on that forum is whether it would be possible to "misuse" the pressure sensor input of the Powercommander LCD unit to trim the fuel injection in real time based on wideband lambda sensor output ?

 

Any comments regarding this would be welcome.

 

br, JuhaV

 

PS. I had a My16M in my Sport Corsa together with wideband lambda and closed loop feedback to trim AFR. It worked out very well.

[raz]

Welcome back Juha! I read all about your MyECU history before buying one myself.

 

I think, if nothing else, it is a cool bodge and good out-of-the-box thinking

 

I can't see any reason it wouldn't work. But that doesn't necessarily mean it will work out extremely well. If you try it with good results, please come back here and tell!

 

I guess the ECU won't damp the pressure signal so if you can set that in the sensor controller (with LC-1 you can) it may mute some noise. Also, with the LC-1 you can set what voltage you want in an error-condition or warm-up. You would want to set that at like 0% fuel trim.

 

Good luck!

[Guest Phil_P]

Maybe I misunderstand, but the Power Commander doesn't accept sensor inputs as such, it just alters the injection and ignition settings as they are sent from the ECU to the injectors and coils. I use the LCD display to upload maps, but as far as I can see, you can't directly change any settings with it. I was contemplating using a 'Palm' device as an onboard 'tweaker'.

[JuhaV]

Phil,

 

You are correct, the Powercommander will not accept any sensor inputs, however, the LCD unit equipped with the HUB will do so.

 

In principle,what is needed would be

 

1) Disconnect original lambda from the ECU (Honda RC51 nor MG Sport 1100i does not have one so we can skip this)

2) Plug LCD unit to Powercommander

3) Plug HUB to LCD to provide "pressure sensor" voltage input to be used now as our lambda voltage input

4) Install wideband lambda sensor and connect the 0-5 V output to "pressure sensor" input via HUB to LCD

5) Program wideband lambda sensor controller to provide suitable AFR to voltage conversion, for example, 0 V = AFR 12 and 5 V = AFR 14

6) Program LCD to trim fuel injection based on the "pressure sensor" input voltage. According to Powercommander manuals, the HUB will provide 0.25 V resolution for the 0-5 V signal, so between AFR 12 and AFR 14, according to the example above, we could have 20 trim values.

7) Go and ride.

 

Raz, thanks for the welcome ! I will need to do some testing with the wideband lambda unit that I have (Australian 2CO) to see how the 0-5 V output behaves during warmup and so. And how easy it will be in practise to reprogram the AFR to voltage conversion in it. Then, if that part goes well, I perhaps need to go and buy the LCD + HUB.

 

Powercommander does not specify how many times per second the pressure sensor input is read, but I think that it is likely that it will be fast enough, perhaps once per each injection cycle. Even something like 10 Hz would be ok because the response rate of the lambda sensor is not much faster.

 

The trim values keyed in the LCD can be limited to max. plus minus 15-20 % from the base map values, so the base map needs to be good enough that is easily achievable.

 

Phil, according to the LCD unit manual the unit provides only the LOW, MID, HIGH fuel trims (basically the same as with the buttons in the Powercommander unit) so as you said, an additional Palm or laptop is needed to really edit the trim tables.

 

As you Phil have the LCD unit, perhaps you could comment whether the LCD display is bright enough to be visible in direct sunlight ?

If so, and weatherproof enough, I could consider removing the original speed and rpm display unit from my bike and replacing it with this cool LCD display instead

 

br, JuhaV

[Guest Phil_P]

Although Dynojet describe the LCD as being lightly weatherproof, to be quite honest, I wouldn't trust it much unless I had done some serious extra sealing with some RTV. The visibility is quite good with a fair contrast screen, but it is pretty (read VERY) small. It is in fact too small to be able to access any of the functions with gloved hands.

 

I really like the idea of making a closed loop system using the pressure sensor for lambda feedback, although I think at the moment I need another trip to the dynamometer to remap my system since I fitted the Creedon chip.

 

I have incidentally taken the temp sensor from the left rocker cover, made up a solid copper mount for it, with conductive paste between the sensor element and the copper. The whole block of copper is now wedged firmly between the cylinder fins. At a standstill I certainly get a good resistance profile, but I've yet to try plotting it while on the move with a cooling air flow.

 

Will happily try to take some pictures of the LCD for you.

[JuhaV]

Will happily try to take some pictures of the LCD for you.

 

Phil, pictures would be nice, especially when the unit is on the bike to get better idea of the size/visibility.

 

My bike is most of the time for track duty only so the weatherproofness is not a big issue. It also normally stays inside during nights and for the time it being outside, the LCD unit needs to come out so it will not be stolen.

 

The gist here is that with the more and more restrictive dB-limits at the tracks, I need to use dB-killers at some tracks and no dB-killers at some other tracks. It would be so nice if the bike would "automatically" compensate in the fuel injection for these changes to always provide optimal AFR.

 

I was thinking more of the possible issue that if the lambda voltage signal fluctuates in some occasions wildly and if the LCD unit tries to trim the fuel injection without any damping these fast changes, the most straightforward solution would be to arrange an electronic (active or passive) low-pass filter to the 0-5 V signal line between the lambda unit and the HUB/LCD. There it is possible to select a response time for the filter to average the signal mildly to make it more stable. This is basic electronics and can be done at a cost of a few euros.

 

This is becoming interesting, looks like it needs to tested in practise

 

br, JuhaV

[dlaing]

5) Program wideband lambda sensor controller to provide suitable AFR to voltage conversion, for example, 0 V = AFR 12 and 5 V = AFR 14

As an alternative idea, you should be able to ignore the DynoJet equipment and set the lambda sensor controller to provide a suitable AFR to Voltage conversion, and then pipe it in place of the engine temperature sensor output.

It probably would not work well until somewhat warmed up, so you would need a toggle switch to change from temperature sensor to O2 sensor.

The temperature sensor is about 4.9V at -40C, once the engine is somewhat warmed up, ie., 40C it reads about 2.5V and then when it fully warms up, 500-600mV at 100C. I'd have to check but I think at 125C it gets pretty close to zero, but I don't think it ever reaches.

Once at about 40C you could probably switch to closed loop.

If the voltage output is 2.5 or higher it should enrichen, and if it approaches 500mV it should lean it out.

You might have to play with the settings a bit to get the right closed loop feed back.

But the beauty of it is that it might even work just by using 0 V = AFR 12 and 5 V = AFR 14

But I would be careful as you might get false readings, and dumping more or less fuel sometimes results in the opposite lambda reading of what you expect.

To be more conservative, you could set the range to maybe 1 V = AFR 12 and 3 V = AFR 14

If the AFR tends to be too high, try 1.5 V = AFR 12 and 3.5V = AFR 14

If tthe AFR tends to be too low, try 0.5 V = AFR 12 and 2.5V = AFR 14

If sometimes it is too high and sometimes too low, you might need to expand the range to something like 0.5 V = AFR 12 and 4 V = AFR 14

But watch or better yet, log the lambda meter!

I think that as long as you are somewhat warmed up, lambda feedback will be better than temperature feedback.

But if you already have all the dynojet gear, you might as well do it that way, assuming it is doable. The advantage is that you could change the fueling map, while if you do it the way I suggest it, it is more of kluge, unless you have software to re-map the ECU.

If anyone is interested I can take some screen shots of the Tuneboy engine temperature compensation map that indicates how the ECU handles the change in voltage.

One thing I would be concerned about is the durability of the sensor in wet conditions.

[JuhaV]

dlaing,

 

That is a very interesting idea ! To substitute the temp sensor output with the lambda generated voltage.

Yes, that can work just fine if the ECU will use the temp sensor output in some substantially linear manner to rich or lean the mixture.

In an aircooled Guzzi with just one temp sensor this is most likely the case. The lambda sensor controller could be programmed to "match" that effect. If the mixture is too lean, the temp sensor input to ECU would be fooled to indicate that weather is colder and vice versa.

 

However, I will try to apply this to a HRC Honda, that has quite a bit more complicated FI system and several temp sensors in various places. I have no idea whether any of these would affect the mixture in a linear manner. I do know that in a case when the cooling water temp is rising above certain level, ECU enrichens the mixture to help to cool down the engine. Forward and rear cylinders have different fuel and ignition maps etc. Thus the mixture that ECU chooses to inject depends on several sensors. Might be that one of the sensor inputs could be used for that but that would need more studies.

 

As I already have the Powercommander USB and wideband lambda, I am only trying to find a proper reason to buy the new gadget, the very technical and nice looking LCD unit

 

br, JuhaV

 

PS. Tuneboy info and pics would be nice. We might get some new ideas ...

[dlaing]

Here is the engine temperature compensation map

The hotter it gets, the leaner it gets, and I presume the leaner it gets, the hotter it gets.

I don't think it is a well designed curve, but I can't prove it just yet, but once I get a WBO2 sensor I will be modifying ALL the maps.

The barometric curve is interesting

Note how they avoided leaning at too high an altitude. I think they did this to avoid overly leaning if the sensor failed.

[raz]

FWIW, the pressure sensor outputs a linear voltage (it's three-wired with its' own +5 and ground) between 0 and 5 volts.

   Volt	mmHg	mBar
  0.250	127.5	170
  4.750	787.5	1050

[JuhaV]

Using the barometric ECU input for the lambda feedback might work just fine !!!

 

Pressure indicates air density in the airbox. Increasing air density requires more fuel and vice versa.

This relationship should be linear, as with increasing pressure the amount oxygen available for combustion increases at the same rate.

 

So, programming the lambda controller to match the barometric pressure sensor voltage would to the trick most likely.

Because lambda will indicate whether there is enough or excess oxygen for combustion, one might even completely dump the pressure sensor.

 

Hmm, Honda has also a MAP sensor (Manifold Air Pressure) ...

 

Seems that with clever programming of the wideband lambda controller one could try quite many kind of approaches to "misuse" any of the ECU inputs that will affect fule injection.

 

Hope that I had more free time to experiment with these things.

 

br, JuhaV