Idea for ignition tuning for best torque

[cygnus x-1]

Over the last several weeks I've been trying to get my ignition timing tuned for best torque. I'm too cheap to get it done on a dyno so I've been using a straight level back road to make repeated pulls with different timing settings. Megalogviewer can then tell me what the HP and torque were with each pull and I can compare settings to see what makes the most torque. This process works but it's tedious and you have to do at least a couple runs with each set of settings (timing map) because there is some variability from run to run, even with the same settings. You also can't compare runs from different days or sometimes even different times during the same day because of changes in air temperature and such.

So it occurred to me that it would be really useful if there was a way to add a step function to the ignition advance, on command. The idea being that a sudden change in timing advance would show up in the datalogs as a sudden change in torque. This would eliminate the variability you get from pull to pull. This would also make part throttle tuning much easier. You could cruise along and toggle the step on and off and get instant feedback on what the timing change did. I also do all my tuning by myself and I use a tablet PC (no hard keys) for data logging, so I can't easily push buttons and drive at the same time. The tuning window in Megatune is too difficult to manage on a tablet PC so it's pretty much useless to me.

Something like this already exists in the "Other Fuel Settings" in Megatune to cyclically add extra fuel. It would be cool if the same thing could be done for ignition timing.

Thoughts?


C|

[242ATL]

what version of MS are you using? The extra code allows for table switching. I imagine you could easily hook up a switch to activate the table switching on the fly.

I like the idea. Street tuning for maximum torque is difficult, even on the same stretch of road. Unfortunately, I never really got the hang of the torque and hp figures coming form MLV... most likely something incorrect in my settings. I've been meaning to get back out there and try logging RPM/second during a tuning run, but haven't had time yet.

Mike

[cygnus x-1]

I'm using the regular MSII. Table switching could work but is really more than I would need. Simply adding a couple degrees of timing while a particular key is pressed would do it.
Tuning for max torque is certainly difficult but it's where all the real gains are to be had. We have MLV VE analysis for fuel tuning, but there is nothing for ignition timing.

The HP and TQ graphs are pretty straightforward. You just need to turn them on in the optional fields and fill in the parameters. The numbers won't be comparable to dyno numbers but they are useful for relative comparisons. Incorrect settings will cause funny graphs as will an unstable tach signal.

C|

[cygnus x-1]

Any other comments? Bueller? Bueller?

C|

[newtyres1]

You have a great idea there IMO. I used to have an old motorbike with the advance/retard lever on the handlebars, it was very easy to find the sweet spot. A 3 pole switch that could leave a port pin (if there are any spare) either open, grounded or tied to 5V could add or subtract a few degrees while driving along. Direct A/B comparisons while driving along would make it a lot easier.

[vinister]

I always wished I could have fuel and ignition trim knobs with ms. Is there any way to implement that?

[pb3]

A 3 pole switch that could leave a port pin (if there are any spare) either open, grounded or tied to 5V could add or subtract a few degrees while driving along.

You would need two inputs for that, then you could have four positions. The ECM will see and open and 5V input as the same thing with a switch to ground type input, that is because the ECU will have a pull-up resister. So the normal way of wiring the switch input would be open in one position and shorted to +5V in the other, with no in between.

[newtyres1]

Yes, of course, you are right, there are only two states for a switch. I like the idea of a trim as vinister suggested, if there was a spare ADC channel a pot could be wired up to it to give a variable range of +/- ignition timing.

Ian.

[vinister]

I've seen many race car dashboards and steering wheels with fuel and timing trim knobs, I really think this feature should be high on the priority list. It would make it so easy to adjust on-the-fly at your heart's content.

[dirtytorque]

Problem is though by the time you have heard the det and then pressed the button the engine may have moved on to another point and so you end up retarding the wrong point robbing you of power unecessarily?!?
I have just ordered a TPIC8101 microchip so I can get MS to log knock and retard the timing .

There is also Knocksense that works with MS so I understand.
I am in the same boat.
I have dyno'd my car and I am about 20 ft lb of torque down.
I want my torque back.

[newtyres1]

Problem is though by the time you have heard the det and then pressed the button the engine may have moved on to another point and so you end up retarding the wrong point robbing you of power unecessarily?!?

I was thinking more along the lines of this scenario - you have a motor that you need to tune the ignition table from scratch. At certain load points you can hold the engine at that load point and vary the ignition trim with a pot to see what effect advancing or retarding the ignition has, in real time. Then you can adjust the ignition table for that load point with MT and move on to another one. Another scenario - the engine just doesn't feel quite right at a certain load point while driving along to some destination, you can turn the trim up or down and confirm if it is an ignition timing issue or not. Or this scenario - you get knock at a certain rpm/load, wind the trim down until you don't get the knock at that rpm/load bin, adjust the ignition table in MT and go from there. Even another one - you are tuning for max economy on the highway, you pull a bit of fuel out with the fuel trim, then adjust the timing trim to suit, seeing how small you can get get the injector pulsewidth (in MT) and still hold a steady speed and still have the engine run smooth.

I think it would be extremely useful.

Ian.

[vinister]

I was thinking more along the lines of this scenario - you have a motor that you need to tune the ignition table from scratch. At certain load points you can hold the engine at that load point and vary the ignition trim with a pot to see what effect advancing or retarding the ignition has, in real time. Then you can adjust the ignition table for that load point with MT and move on to another one. Another scenario - the engine just doesn't feel quite right at a certain load point while driving along to some destination, you can turn the trim up or down and confirm if it is an ignition timing issue or not. Or this scenario - you get knock at a certain rpm/load, wind the trim down until you don't get the knock at that rpm/load bin, adjust the ignition table in MT and go from there. Even another one - you are tuning for max economy on the highway, you pull a bit of fuel out with the fuel trim, then adjust the timing trim to suit, seeing how small you can get get the injector pulsewidth (in MT) and still hold a steady speed and still have the engine run smooth.
I think it would be extremely useful.
Ian.

Yes! Ian that's spot on, I use the trim angle setting to do that but it is tedious and dangerous to be fiddling with the computer while driving, constantly pulling over to enter settings... a knob would completely eliminate that. So what would we need, and is it possible with MS2? I suppose the requirement is 2 seperate ADC pins and some snippets of code. We can probably look at the knock system code for a good example of how its done. But, do we have the pins available? Or at least one so we can just get the ignition done, worry about the fuel later.

As you mentioned it would be extremely useful for economy tuning. Just pick a road, hold a steady speed, and reduce fuel/advance timing to find the sweet spot. I imagine we could all save some fuel money ;)

[newtyres1]

There appears to be two spare ADC channells connected to jumper pads JS4 and JS5 on the V3.0 PCB. For hardware I would think a 5K linear pot with a 5V Vref feed and a ground would do, with the wiper supplying the analogue signal. Center position (2.5V) = no change, each 0.5V above or below that adds or subtracts one degree from the whole ignition table. That would give +/- 5 degrees.

I wonder if this has been discussed before, I have not seen this idea suggested before, but its quite possible it has been brought up.

Ian.

[cygnus x-1]

Hmm, knobs. Originally I was thinking more of a single button that would just add or subtract a certain amount of timing, that would be configurable from Megatune. I would probably set it for 2 degrees or so. But an ADC channel with a knob would be more flexible because you could either go up or down without having to reconfigure in the software. You could either use a knob for trimming or some switches with fixed resistors. I would rather have momentary switches because I could instantly change timing by a fixed amount without having to turn a knob and guess how much it's changing. The *instant* part is important because if you're doing this while accelerating you need to see the effect the timing change had right at that load area. A knob would just be too slow. But the idea is still good and a knob could also be useful as well.

Now we just need to figure out how to do this. Are those extra ADC channels brought out on the relay board? I don't remember there being any spare channels left, but if there are that would be very cool if there was at least one. And then there's the code changes and recompiling.

I'm pretty sure this idea has been brought up for fuel trimming, but I've never seen it for timing trimming.

C|

[vinister]

The whole project seems very straight forward, I just don't have any time right now to devote to it.

The spare channels are on the board, they are not typically wired into any pins on the DB-37, but that's as easy as a couple of jumpers.

I think the code is very basic to write, especially since good examples already exist with the knock-retard code segments. I would love to make this my project, but i just started my 2nd year in school, and a new job, and I'm a single father, so there is just no way I can fit it in.

We just need to get someone interested that knows how to write the code. Other than that, we pretty much have it nailed.

[cygnus x-1]

Heh! Well, as it turns out my degree was computer science, so I should be able to figure this out without too much difficulty. But oddly enough though I never ended up working as a programmer. Funny how that works. Anyway I'm trying to get to where I can successfully compile the source into a working bin file. Unfortunately the free version of CodeWarrior doesn't work because it has a code size limit and the Megasquirt code exceeds that limit. So now I'm trying to get the gcc version to work.

C|

[newtyres1]

Cygnus, having any joy with the gcc version?

Ian.

[cygnus x-1]

Cygnus, having any joy with the gcc version?

Ian.

Not yet. Depending on the code optimization settings I either get an internal compiler error, or an error about text regions being full.

C|

[cygnus x-1]

After some more digging it appears that the current (2.886) B&G code is just not gcc friendly. And the the free version of Codewarrior has a 64k limit that the MS code goes beyond, so it won't work either. The MS2/Extra code however IS gcc friendly, so I took a look at that to see what the difference is. Well it's significant. Gcc being a more generic compiler does not optimize as well as Codewarrior does (not surprising since Codewarrior is tailored to Motorola chips). The MS2/Extra code had to be reorganized/reworked to make it compatible with gcc. So what this all means is that the only people that could add this feature would be either someone with an unlimited version of Codewarrior, someone willing to rework the B&G code to work with gcc, or B&G themselves. Unfortunately I am none of those.

The last option is to try it on the MS2/Extra code, but my free time for the near future is about to evaporate. And of course it's also out of scope for this forum so any discussion pertaining to MS2/Extra would have to the other forum.

C|

[newtyres1]

Well, we can't say you haven't tried. Thanks for looking in to it. There may be a small chance that Al Grippo might do it. Al, if you read this, is there any chance of doing an experimental version for us to try, on the understanding that there will be no support and if we don't like it we just go straight back to regular code? i.e. we just want to try the idea to see if it is a good way to go or not.

Ian.