Practical Reflash Tuning: Step 5: Optimising the Tune
Watch This Course
$229 USD $114.50 USD
-OR-
Or
8 easy payments of only $14.31
Instant access. Easy checkout. No fees.
Step 5: Optimising the Tune
48.44
| 00:00 | - The next step in our re-flashing purpose is to begin optimising our fuel and ignition timing and the wide open throttle conditions using ramp runs on our dyno to replicate real world acceleration out on the road or the race track. |
| 00:14 | Now at this point we've made some modifications to our base ROM file, however we haven't actually flashed this into our PCM yet so for all intents and purposes right now our car is absolutely factory. |
| 00:28 | Now if we're in a situation where we can run the engine on the dyno as delivered, it's always best to do so. |
| 00:35 | This gives us a baseline that we can then compare our modifications against, so when we're altering our fuel or ignition timing, we're going to be able to compare against that baseline figure and see if we've made improvements or if we've gone the wrong way and we're losing power. |
| 00:51 | So let's get our car running on the dyno now and we're going to do a full power run in fourth gear from around 1500 RPM out to about 6300 RPM. |
| 01:02 | Right, we've got our engine running on the dyno, we're ready to go, so let's see what our stock calibration gives us. |
| 01:33 | Okay so there's our baseline run complete, and that's readjusted to 237 kilowatts at the back wheels or 318 horsepower. |
| 01:42 | You can see on our dyno screen here we have our power in kilowatts, which is the green line on the bottom of our screen, the red line above shows us our air fuel ratio being measured in units of lambda. |
| 01:54 | So now we've got our baseline file, we can actually flash our modified calibration into the PCM, and we can start making some tuning changes. |
| 02:04 | Let's have a look first though at our scanner and we'll see what exactly we've got going on, what we're interested in looking at, what we're going to be paying attention to in our scanner. |
| 02:14 | Now on the left hand side of our scanner screen you can see we have our individual channel, so these are individual channels that we are logging and we can see they're instantaneous values, how well the engine is running. |
| 02:29 | Now in this particular instance, the instantaneous values aren't necessarily that useful to us. |
| 02:35 | It's much easier, much more useful to view how these parameters that we're logging and changing in relation to time, you can see it in the middle of our screen here, we have our chart versus time, so this is graphing the parameters that we are logging versus time. |
| 02:52 | And a couple that we're interested in, obviously we have our RPM at the top here, and you can see that we have performed our ramp run here between 1500 RPM and 6300 RPM. |
| 03:05 | Anytime we have performed a run on the dyno, be it with a stock calibration or with a modified calibration, one of the first things we always want to look for is our knock retard. |
| 03:17 | Our knock retard parameter is going to tell us whether or not the PCM was removing ignition timing and response to detonation occurring. |
| 03:25 | And we can see our knock retard parameter is displayed here in red with the name KR, and on the same chart we also have our ignition timing being displayed in white, so this is the actual ignition advance being delivered to the engine. |
| 03:41 | You can see here that we don't have any real knock retard occurring, which means that our pull on the dyno there was performed with no detonation occurring. |
| 03:50 | Moving down you can see we have our measured lambda versus, our commanded lambda, and this is what we're going to be using for deciding how well the PCM is managing to track our target air fuel ratios. |
| 04:04 | And you can see for the most part that our measured air fuel ratio is tracking very closely to our commanded, we do have a spot here where the air fuel ratio was a little rich, we see a measured lambda of 0.847 or eight five, versus a target of .87, so around about 2-3% rich in this area. |
| 04:27 | Now remember at this point in time we've got a stock calibration here, we haven't made any changes in our car, it's 100% stock, so this is the sort of variation we're seeing even in a 100% stock calibration. |
| 04:40 | We're now going to flash our modified calibration into the PCM, and we can do another run. |
| 04:48 | Now really the only changes we're going to make here are for our first run, remember we've modified our power enrichment targets, we've removed some torque limits but those probably wouldn't have been having much effect for this ramp run here anyway. |
| 05:02 | So at the moment we haven't changed our ignition timing, we're probably not going to expect to see much difference in our measured power for our first change. |
| 05:11 | So let's shut our engine down now. |
| 05:15 | We'll disconnect from our scanner, and we can go back to our VCM editor. |
| 05:20 | So to flash the file into our PCM, we need to click on the write vehicle icon up on the top tool bar. |
| 05:29 | I've got some options here that we can choose depending on what we're actually trying to do. |
| 05:34 | Here we can write to our engine control module or PCM, this is what we want to do. |
| 05:40 | We also have the option here of writing to our transmission control module, since in this particular vehicle we're running a 6-speed automatic. |
| 05:48 | In this case I haven't made any changes to the automatic transmission tuning, so I'm not going to write to that controller, this simply speeds up the process because we're not needlessly writing to a controller when we don't need to. |
| 06:01 | Pressing the write button will flash our new calibration into the PCM, this takes around about 40 seconds, and then we can get the engine back up and running. |
| 06:10 | All right we've got our engine back up and running now and our scanners operating. |
| 06:14 | And let's have a look, first of all we can see that our long term fuel trims are now sitting at zero, and remember that's because we have disabled our long term fuel trims here. |
| 06:24 | We've still got our short term fuel trims active, and you can see right now just after we've restarted that these short term fuel trims are actually quite negative, they're sitting at around about minus seven or 7-8%. |
| 06:35 | And now this is another reminder here that when we're doing any tuning changes, when we're using our scanner, we always want to let the engine operate, get up to a normal operating temperature, we don't want to be doing any scanning while the engine is warming up, or also after a hot restart when everything's heat soaked. |
| 06:55 | Now that brings me to another point that's worth discussing here as well when we're performing any runs on our dyno and we're trying to make tuning changes. |
| 07:04 | We always want to make sure that our operating conditions at the start of the run are as consistent as possible. |
| 07:10 | And in particular, if we look here in our scanner, we want to take note of our engine coolant temperature as well as our intake air temperature. |
| 07:18 | And we want to make sure that those are always as consistent as possible from one run to another, and that's going to ensure that if we see an increase in power, or a decrease in power, that that's as a result of the tuning changes we've just made rather than differing conditions that the engines operating under. |
| 07:37 | All right, so let's perform our first run with our new calibration, and we'll see what everything looks like in terms of our fuel and our ignition timing. |
| 08:10 | All right that's our next run complete there, so this is the first run we've made with our new calibration. |
| 08:16 | Now our power is essentially identical to our first run and that's what we would expect. |
| 08:22 | What we can see though is our lambda plot versus our last run with our base or factory calibration. |
| 08:29 | You can see that we're a little bit leaner in the beginning of the run, and right at the top of the run we're around about the same, so we're essentially leaned out slightly at the bottom end of our power enrichment table. |
| 08:43 | Let's have a look though at our scanner and we'll have a look at what we need to be considering when we're performing these full throttle runs on the dyno. |
| 08:52 | So the very first thing we're still going to have a look at is our knock retard, we want to again make sure that our engine hasn't been detonating during our run and you can see that our entire log file there and our entire scan file is clean, we don't have any knock retard occurring. |
| 09:11 | Now this is always the first thing we want to look at. |
| 09:13 | Now provided that we aren't suffering from any knock and we don't have any immediate work to do in order to correct that, it's easiest when we're beginning to learn how to tune if we can separate the tuning task into two separate jobs. |
| 09:29 | I like to look at the fuel first and then follow that up with the ignition timing. |
| 09:34 | Just means that we've only got one thing to concentrate on. |
| 09:37 | Once you're a little bit more competent, you're used to both the tuning software and operating the dyno, we tend to make changes to both our fuel delivery and our ignition timing simultaneously, but for now considering we don't have any knock occurring, we can concentrate on our fueling. |
| 09:52 | I'll point out right now as well, that while I'm presenting a worked example here, normally when we're tuning I always recommend that you use audio knock detection equipment, just to listen audibly for detonation occurring during a ramp run. |
| 10:09 | Now while the GM PCM does have a very good knock control strategy, so we can usually trust the scanner, if the scanner is showing no knock, then generally we can be relatively safe in saying that the engine wasn't suffering from knock. |
| 10:24 | It's always good to be able to confirm and validate that the PCM is in fact correctly identifying knock. |
| 10:32 | So when we're tuning I always recommend using knock detection equipment. |
| 10:36 | At the same time, also when we're doing a full throttle ramp run on the dyno, well right now our car is 100% stock standard, so we can obviously expect to run the car right through the rev range relatively safely, particularly once we've started making some more dramatic modifications. |
| 10:55 | We may find that the air fuel ratio is either excessively rich or excessively lean in some areas, or alternatively we may find that at some point in the rev range the engine suffers from knock. |
| 11:06 | Now it's always safest in these situations to simply abort the run, come back to idle, look at our scan data, decide what we need to do, and then make the appropriate changes. |
| 11:17 | It's always safest to do this rather than try and pull through to the end of the run and potentially risk damaging the engine. |
| 11:24 | We can always do another run on the dyno, and we can do that very easily. |
| 11:28 | Now let's have a look at our commanded air fuel ratio, and again because we're running what is essentially a stock standard vehicle here, we can expect the measured air fuel ratio to very closely match our commanded, and for the most part here you can see that that is the case if we click through our file, we can see that for the most part here we are generally within 1% or better, some places we're absolutely bang on our target, which is what we'd expect. |
| 12:00 | However, we can see that there are some areas where we're a little bit rich, particularly right here up in the top end, we're still within 1%, you can see we do have a little area here we're a touch rich, and this area here at 3500 RPM, you can see that our measured lambda are sitting at 0.86, whereas our commanded is 0.886, or 0.89 almost, so we're a little rich in this area, and at the same time you can see that right at the start of the run, we were a little bit too lean at about 2-3% lean, we're sitting at around about 0.91, 0.92, versus our target of 0.89. |
| 12:47 | So what does this mean and how can we fix it? What this simply means is that at any point where our air fuel ratio doesn't match our target, it means that the PCM thinks that either more or less air is entering the engine than what actually is. |
| 13:03 | So what this means is that the MAF scaling as well as the virtual volumetric efficiency system that the GM PCM uses may not quite be accurate in these areas. |
| 13:17 | And depending on how fussy we want to be and how close we want to be to our target depends on what we're going to do about this. |
| 13:25 | First of all, we need to consider what we're actually aiming for, and while we've got a target here, our commanded air fuel ratio or commanded lambda, what we need to understand is that the measured lambda in the exhaust is always going to be fluctuating a little bit, we're never going to be achieving one single fixed value. |
| 13:44 | So when I say, for example, that I'm targeting perhaps a lambda value of 0.90, what I'm generally really meaning is that I'm looking for a range around that area that I'm happy with, so if I'm looking for 0.90, I might be happy with a measured lambda in the range of 0.89 to perhaps 0.91. |
| 14:06 | So generally, something with around about 1% or better of my target, I'm going to be pretty happy with. |
| 14:13 | We also need to consider that run to run we will probably see some variation in our lambda, we're never going to see an exact back to back replica of our lambda plot every time we run the car on the dyno. |
| 14:25 | So if we're trying to achieve an absolute error of zero, it's often relatively futile, we're probably not going to achieve that. |
| 14:34 | If we are going to accept an error in our measured lambda, generally I'd like to see us being just a little bit richer than our commanded lambda, as opposed to a little bit lean. |
| 14:46 | Now how can we go about fixing these errors? Well in our next worked example, we're going to look at the technically correct approach which is to go through and rescale both our mass air flow sensor as well as our virtual volumetric efficiency tables. |
| 15:05 | I won't get into detail right here, but the GM PCM does use both a speed density and a MAF system for measuring load, and it blends between those two systems depending on the operating conditions. |
| 15:22 | So in the next worked example, we'll look at how to correctly address our scaling and calibrating both of those. |
| 15:28 | Since the PCM is primarily using the input from the mass air flow sensor, and under steady state conditions when we don't have large transience going on, and particularly our higher RPM, in a situation like this when we only have a couple small errors that we may want to correct, we can get away with adjusting our mass air flow sensors scaling to account for those. |
| 15:52 | So let's look at how we can do that now. |
| 15:55 | So how can we go about correcting these errors and our MAF scaling? Well in order to help us, what I've done is set up a histogram which we'll have a look at now, let's just drag over our graphs, and you can see we have a histogram set up here which is EQ error MAF. |
| 16:11 | So this is the error between our commanded lambda and our measured lambda, and this is a MAF channel available inside the VCM scanner software. |
| 16:22 | And what we've got here on our axis is we've got the mass air flow frequency. |
| 16:28 | So this is the frequency output from the mass air flow sensor, and what essentially we have inside our VCM editor is a table that converts mass air flow frequency into air flow in grammes per second, so this is the primary load input that the PCM is using. |
| 16:46 | So any time we have an error between our commanded and our measured lambda, this simply means that there's an inaccuracy in our MAF scaling. |
| 16:56 | So let's go back and have a look at the areas we're going to want to address here, and we'll go back to our chart logger here, and we're going to make two changes. |
| 17:07 | We're going to have a look at this area here at 3500 RPM, and you'll remember we're about 2-3% rich. |
| 17:16 | And what we want to do is have a look at the mass air flow sensor frequency at this particular point in time. |
| 17:23 | Now we have our mass air flow frequency logged here in our channels list, and if we cycle through we can see that we start the area where our air fuel ratio starts to move a little rich, so we're gonna go about 8600 hertz, and goes through to around about 8800 hertz, or 9000 hertz. |
| 17:47 | So we're looking at an area between 8600 and 9000 hertz. |
| 17:52 | Let's pull our histogram across again and we'll have a look at that particular area in our histogram data. |
| 17:59 | And when we get through to that area, you can see that that same area is being demonstrated here, and in fact all the way from about 7800 through to about 9000, we have an error there, and any time our error is negative this means that our measured air fuel ratio was a little bit richer than our target. |
| 18:23 | Now there's a few ways that we can deal with this data. |
| 18:26 | A lot of tuners will use the paste special function that we will be looking at in another worked example. |
| 18:33 | Here, because we're only making small changes, I'm going to simply manually make changes to our mass air flow sensor scaling, so let's look at what we want to do. |
| 18:45 | Essentially from about 7650 through to around about 8900, we're going to make a change of minus 1%. |
| 18:54 | What I'm going to do here is look at our data, I'm looking for a trend, the general amount of error that we're logging. |
| 19:01 | So in this case from 7500 through to about 8250 hertz, generally we're moving around about 1%. |
| 19:11 | So let's go to our VCM editor, so we're looking at 7500 to 8250, and what we want to do is go to our air flow, and our air flow versus frequency calibration. |
| 19:22 | Let's move through until we get to 7500, and we're going to highlight 7500 through to 8250. |
| 19:29 | Now what we want to do is remove 1% here, and we can do that by multiplying our value here by 0.99. |
| 19:38 | So we enter 0.99, press the multiply symbol, and that will simply remove 1% from those values. |
| 19:44 | Let's go back to our scanner, so we've adjusted up to 8250 hertz, now from 8400 hertz up to 8700 hertz, we're actually removing slightly more, we're around about 1 1/2, perhaps 2%. |
| 20:00 | So let's make a slightly larger change there, so we're going to go up to 8700 hertz, and we'll highlight those cells, and this time we're going to remove 2% by multiplying by 0.98. |
| 20:16 | We'll jump back to our scanner, and we're going to make a further change, at 8850 and 9000 hertz we're going to remove 1%. |
| 20:28 | So this is an easy way of manually making changes to our mass air flow sensor scaling. |
| 20:36 | And important aspect when we're adjusting our MAF scaling is to make sure we maintain a smooth and consistent shape to our mass air flow sensor calibration curve, if we start getting really erratic, change is needed in our mass air flow sensor scaling, then that indicates that we've probably got something else that's an issue that we need to look at. |
| 20:58 | So we've adjusted our areas there, between 7500 and 9000 hertz. |
| 21:06 | I'm not going to go through and make every correction, I just wanted to demonstrate the procedure, let's have another look at our scan data. |
| 21:15 | If we look at our chart logger, you can see that we also hit this area at lower RPM where we were a little bit lean. |
| 21:23 | And that area there started at around about 5800, 5900 hertz, you can see we've got 5900 hertz showing at 1260 RPM, and we were lean through to about 2300, 2250 RPM, and at that point we were 7500 hertz. |
| 21:46 | So we're going to make one change to try and correct that as well, let's move our histogram back over and we'll have a look at our data there. |
| 21:54 | And we'll just move across so we can see everything we're interested in. |
| 21:59 | So you can see from 5850 through to right about 6900, we're around about, I'm going to make a change of 2% here. |
| 22:10 | We do have a little outlier here of 4%, but what I'm going to do is simply make a change from 5850 through to 6750, we'll add 2%, and we can do that by multiplying by 1.02, now let's go back to our scanner, our histogram, and we can see that at 6900 hertz we were still a little lean, so we had an error of 1.73, and then at 7072 we're about 1% lean. |
| 22:44 | So let's make those changes now as well. |
| 22:47 | So that's 6900 hertz, let's add 1 1/2%, and then for our next two cells we're going to add 1%. |
| 22:59 | Now if we have a look at this as a 2D graph, it's hard to get exact detail in this graph, but you can see we've still got a relatively smooth shape, so this is quite important, we want to make sure that we don't end up with any really large irregularities in the MAF scaling, otherwise that's going to introduce driveability problems. |
| 23:20 | Let's shut that down, we've made our first change there to correct those errors that we've seen. |
| 23:26 | What I'm going to do now is save my calibration, we'll just disconnect our scanner, and we can flash that change into the PCM and see if that's corrected our error between our commanded and our measured lambda. |
| 23:42 | All right we're got our engine running again, we've got our scanner operating, so let's do another pull on the dyno and check our results. |
| 24:10 | All right so that's our next run complete there, and again because we're already making very small changes here to our fueling, we're not expecting any really dramatic changes to our power, and we've essentially overlayed with that last run. |
| 24:23 | What we can see though is we've corrected, remember we were a little bit lean on our last run at low RPM, we've now corrected that. |
| 24:31 | We were also a little bit rich through the mid range, and we've also corrected that. |
| 24:36 | Let's have a look at our scanner and we'll see how well those changes have worked out. |
| 24:41 | Now you will remember on our last run I had pointed out we were a little bit rich in the top end of our run, we didn't make any changes there, and you can see that we still are around about 1% rich in the top end of our run. |
| 24:56 | But we can address that exactly the same way as we just have, however the area that we were looking at, remember that we were looking at around about 3500RPM and down, so anywhere from around about the start of our run, about 6000 hertz through to about 9000 hertz on our mass air flow sensor output, and you can see now the couple of small changes that we've made to our mass air flow sensor scaling. |
| 25:25 | Our error is almost zero now, we're right on our target which is really good, this is what we want, and that's just how simple it is to use the histogram to help us optimise the mass air flow sensor scaling. |
| 25:40 | So in this case, I would remove a further 1% from 9600 hertz through to 9900 hertz, let's do that now. |
| 25:49 | We'll just pull up our editor again, and we're going to cycle through from 9600 hertz to 9900 hertz, so we want to remove 1%, so we multiply by 0.99, and then if we go back to our scanner, from 90, from 10,000 hertz and above we're going to remove in this case 1 1/2%. |
| 26:15 | You can see we do have a slightly larger error here, but again I'm always going to want to be slightly richer rather than slightly leaner, so we're going to try 1 1/2% and see how close that gets us. |
| 26:27 | Now while our run is only taking us out to 10,500 hertz, I'm going to extrapolate that change through to the end of my MAF sensor scaling. |
| 26:37 | So let's do that now. |
| 26:40 | So we're going to highlight everything from 10,000 hertz and above, and we're going to make a change of 0.985, which will remove 1 1/2%. |
| 26:50 | So that small change should now correct the slight rich area we're seeing at a higher RPM, and this also shows how powerful the histogram is in quickly dialling in our MAF scaling. |
| 27:05 | Okay so we've corrected that high RPM error, but now we're also going to move on and we're going to start looking at our ignition timing. |
| 27:13 | And what we want to do is start, when we're optimising our ignition timing, I like to start as long as long as we have no knock occurring, by making across the board change to our timing right through our entire run. |
| 27:26 | However, our next problem is where exactly was the PCM operating during that ramp run? So we need to know where exactly in the spark advance map the PCM was accessing. |
| 27:41 | Now let's jump across to our spark map, and if we have a look at our high octane spark advance table, you can see that we're using a load excess of spark airmass, and this is airmass in grammes per cylinder. |
| 27:56 | And obviously the x-axis is relatively straightforward, just engine RPM. |
| 28:01 | So let's have a look again at our scanner, and you can see on the bottom of our charts here, I'm logging the parameter cylinder airmass, so this is helpful in showing exactly where the PCM was accessing, so we can see at the very start of our run, we're sitting at around about 0.64 grammes per cylinder, at the peak at about 5000 RPM we're sitting at 0.82, and then right at the very top of the run we actually drop off a little bit again to 0.75. |
| 28:33 | So this will let us know whereabouts in our ignition tables we're accessing. |
| 28:38 | Again, we can use our histograms as well to help us with that, so let's have a look at how we can do that now. |
| 28:44 | By default, we have our histogram set up here for spark advance. |
| 28:49 | So this histogram is filled out with the current ignition timing that's being delivered as we move through the cells in that histogram. |
| 28:59 | Just updates the cells with the average of whatever the ignition timing is for that particular combination of load and RPM. |
| 29:08 | And on the face of it that's quite helpful until we actually look at our axis. |
| 29:14 | Of course, we have engine speed on our x-axis, which is great, that's what we need. |
| 29:18 | However, if we look at our load axis, our y-axis, now we have manifold absolute pressure, and that's not particularly helpful because remember our spark advance maps are in units, use load units of grammes per cylinder. |
| 29:32 | So in this case we're sitting constantly at around about 90 to 95 KPA through our wide open throttle ramp run on our dyno. |
| 29:42 | So what we can do though is correct this, we can change the load axis so that this histogram is useful. |
| 29:47 | If we right click and go to graphs layout, and we have all of our graphs available here, we're going to make a change here to our spark advance histogram, and you can see first of all the parameter that's being logged into our histogram is spark advance, which is what we want. |
| 30:05 | We have our settings here for both our column and our row axis, and this is where we need to make a change. |
| 30:13 | Now we want to leave our engine speed as our column axis, but we want to change our manifold absolute pressure row axis to cylinder airmass. |
| 30:23 | And we can do that simply by left clicking, and there's a couple of ways we can find our new parameter that we want to log here. |
| 30:30 | We can start by simply typing the parameter that we're looking for, and in this case I've typed in cylinder, everything with cylinder in the title is displayed, and the particular parameter that we want, cylinder airmass, is showing here. |
| 30:45 | So we can select that by double clicking on it, and cylinder airmass is added. |
| 30:51 | The other way we can do this is to simply navigate through the structure, the menu structure, and find exactly what we're looking for, and we've got the same parameter. |
| 31:03 | Okay, so that's part of the job done, however the values for the break points on this axis are still manifold absolute pressure values, which aren't any use. |
| 31:12 | We can correct that and grab the correct break point values from our VCM editor, so let's pop back to our editor. |
| 31:20 | We're on our spark airmass, we're on our spark advance table with our spark airmass load axis. |
| 31:27 | If I right click here on our table, we go down to row axis and I use the copy labels option. |
| 31:36 | If we now go back to our scanner, we can highlight the current axis values, we can delete those and if I press Control + V to paste, you can see now that has filled our axis break points for the same ones we used on our spark advance tables, so this is really useful. |
| 31:57 | Now we've actually got a histogram that's going to guide us to where the PCM is accessing in our spark advance table. |
| 32:05 | And you can see that at the beginning of the run here, we're sitting at around about 0.64 grammes per cylinder, as we move through the rev range you can see that the mass air flow grammes per cylinder increases, and it peaks down to around about .82, as we already found and then it drops off a little bit. |
| 32:24 | So we now know exactly where we're operating in our spark table, and this is really powerful. |
| 32:29 | So now that we know where the PCM was operating in our spark advance map, let's go ahead and make an across the board change. |
| 32:37 | And what I'm going to do is advance the timing by two degrees in that entire area that the PCM was accessing under wide open throttle. |
| 32:45 | So what we want to do there is make that change anywhere above .64 grammes per cylinder, and I'm going to make the change from 1400 RPM and above, so let's go back to our editor, and we're going to do exactly that. |
| 33:01 | So we're going to highlight, so I'm actually going to come down to 0.60 grammes per cylinder, we're gonna highlight everything that we're going to run through, and I'm going to just advance that timing by pressing the plus key twice. |
| 33:16 | Now when we are making changes to our ignition timing like this, it does pay to make sure we keep a smooth shape to our curve, to our timing curve. |
| 33:27 | We don't want to end up with a really erratic shape to our timing curve, because that is going to end up resulting in the possibility of erratic engine operation as we move through these areas. |
| 33:40 | So what I'm going to do now is just hand blend the changes that I've made here, just so that we are seeing a relatively consistent shape to this curve, so here I've just advanced the timing in this lighter load area, because we were finding the timing dropped away and then advanced again, and that's not really likely that the engine's going to want that. |
| 34:06 | Likewise, here at lower RPM I'm also going to advance our timing up just again to smooth the changes that I've made. |
| 34:17 | Let's see what our single course change there has resulted in, so we'll shut our engine down and we'll flash that modification into the PCM, and we'll see the result of that change. |
| 34:31 | All right, we've got our engine back up and running with our advance timing map, so let's do another run on the dyno and check what that additional two degrees has resulted in. |
| 35:00 | Okay so let's start by having a look at the results on our dyno, and we've seen essentially no difference in our power here. |
| 35:08 | We've got our current green line has overlayed almost exactly over the top of our last one. |
| 35:13 | We've got a little blip in our power right up at the top, and we've potentially gained a little bit of power throughout mid range, from right about 2800 RPM through to perhaps around about 5000 RPM. |
| 35:29 | Let's have a look at our scan data though, and straight away we can see that we are now with our advance timing suffering from knock, and that's being indicated here by our red knock retard. |
| 35:43 | Now there's a couple of things that I wanna point out here. |
| 35:46 | First of all, you can see that when knock occurs we get a sharp increase in our knock retard value, so at this particular point at 5000 RPM. |
| 35:55 | You can see we have a knock retard of 2.4 degrees. |
| 35:58 | Now this also demonstrates the decay aspect which we talked about, when we're setting up our base ROM. |
| 36:04 | You can see that our knock retard is gradually removed. |
| 36:08 | Now in this particular instance, our timing in this area is obviously too advanced, because as the timing is advanced back as our knock retard is removed, we can see another spike here of knock retard, and a third one, so we actually have a little bit of work to do in this area. |
| 36:27 | What this means is that our calibration was probably already pretty close to optimal in this particular area. |
| 36:33 | What we can see though is that at higher RPM, from around about 6000 RPM and up, at this point we don't have any knock retard active, and you can see that everything's clean in our scanner, and we also have shown a very small improvement in power in the top end, so that indicates that our engine wanted that additional advance there. |
| 36:57 | Before we continue with our ignition timing, let's just have a look at our air fuel ratio again, and remember we did make one more change to our MAF scaling at higher RPM where we were just a touch rich, and you can see that that's corrected that, that one change that we've made there at high RPM, our higher frequency and our MAF scaling. |
| 37:17 | That's just an indication of how easy it is to correct any errors we have in our air fuel ratio. |
| 37:24 | Now let's look at what we're going to do with our ignition timing. |
| 37:27 | Let's jump back first of all, and look at our dyno plot, and as we discussed we haven't seen any improvement from that additional timing through the lower RPM area of our run, we haven't really see any change up to about 2800 RPM. |
| 37:43 | So any time we've advanced the timing but we haven't seen any improvement in torque or power, this means that this hasn't been beneficial to the engine, we were already at MBT timing, or perhaps even beyond MBT timing. |
| 37:58 | So if that hasn't been beneficial, we're simply going to remove that timing. |
| 38:03 | So let's go back to our editor now, we'll go back to our high octane spark map, and remember we made that change from 0.60 grammes per cylinder, so what I'm going to do is take that two degrees back out, up to 2600 RPM, and I'm also just to keep everything smooth, I'm going to take one degree out at 2800 RPM. |
| 38:25 | This is about the first point we started seeing a really small change in our power, albeit it is incredibly small, so we're really starting to split hairs, and I'm also just changing the areas down at lower load, lower RPM, which I'd hand blended previously. |
| 38:43 | So now the lower area of our timing map is back to the factory calibration. |
| 38:48 | Let's jump back to our scanner. |
| 38:50 | So this is the area we really want to pay some attention to though, and you can see that our first knock, and we just did it at about 5000 RPM, and generally when I see this I want to remove timing obviously in that area, but generally I'm also going to remove some timing slightly before that as well. |
| 39:07 | So we actually see some very minor activity down here from around about 4500 RPM. |
| 39:15 | So what I'm going to do is between four and a half and 6000 RPM, I'm going to remove the timing that I've added back in. |
| 39:22 | So we don't have a break point at 4500 RPM, so what I'm going to do is remove that timing from 4400 RPM through to 5600 RPM. |
| 39:36 | So I'm going to remove two degrees in that particular area. |
| 39:40 | Again, just for the sake of keeping our ignition table smooth, I'm also going to blend the change that I've just made there into the surrounding areas. |
| 39:51 | Now what I'm going to do is at 6000 RPM, I'm going to advance our timing by one degree, and we're going to try advancing our timing by two degrees from 6400 RPM and above. |
| 40:07 | So we're going to advance our timing in that area above where we had our knock occurring, but where we also saw an improvement in power on our dyno. |
| 40:16 | So let's flash that change into our PCM and see what our results are. |
| 40:21 | All right we're back up and running, so let's perform our next run. |
| 40:46 | All right, so now our last run's given us a, basically the same power, we've got a 239.6 kilowatts at the wheels, you can see that overlayed with our previous run we have picked up again just a very small amount of power in this top area of our rev range. |
| 41:03 | Let's have a look at our scanner, and you can see that we have removed the knock that we were seeing previously through our mid range, through around about 5000 RPM, so retarding the timing in that area has achieved that aim, however you can see now we do have some knock occurring right in that top RPM range, up at 6100 RPM, and remember that's because we've advanced that timing further. |
| 41:31 | So what this means is that we need to go back into our spark advance table and retard the timing up at that high RPM area where we previously advanced it. |
| 41:42 | Now we have seen a very small improvement in the power on the dyno from advancing that timing, but this is where we need to be a little bit sensible and it doesn't really matter how much power our engine's making if the engine is knocking, in that case we need to take care of that knock, make sure the engine isn't knocking by retarding the timing. |
| 42:02 | Make sure that we can get a clean run with no knock occurring. |
| 42:07 | Now let's talk about what we are likely to see or what we can expect to see from our knock control system as well. |
| 42:13 | Obviously in the perfect world, we'd like to make sure our engine is never suffering from any knock at all. |
| 42:18 | However, in the real world we will see some occasional knock activity, and we'll definitely see this even in a completely factory calibration, some of these engine's we find the factory calibration is very aggressive in the timing, and we'll see a lot of knock activity occurring. |
| 42:36 | What I'm trying to avoid is sustained or repeatable knock retard that's occurring run after run in exactly the same regions of the rev range. |
| 42:47 | So this would indicate that we do have a problem, the engine is knocking and we need to address that in our spark advance tables. |
| 42:55 | Particularly once we get our calibration off the dyno and we take our car out onto the road or the race track, it's not uncommon to see an occasional knock event occur, what I'm looking for is to make sure that that knock event isn't repeated or continue in that same load and rev range, if it is, we need to address that in our calibration. |
| 43:17 | So in this case, the last change I'm going to make here is we're going to remove some timing in this high RPM range. |
| 43:24 | Now I'll just show you another use of our histograms that we can, we can use the histograms here to help us with our spark timing in terms of where the engine is knocking, by going to our spark retard histogram. |
| 43:40 | Now in this case, you can see we've got a histogram that looks very much the same as our spark advance, only this time it is showing any areas where retard was being pulled due to the knock retard system. |
| 43:53 | At the moment we're seeing this filled with zeroes, however this is because of the resolution of this particular table, I'm sorry, the precision of this table. |
| 44:02 | If I modify this I can right click and click on graphs layout, go to our spark retard, and now I'm just going to add one decimal place so we can actually see a little bit more precision. |
| 44:14 | Now we can see at higher RPM, we can see that some knock retard was in fact being pulled out. |
| 44:21 | So this helps us again highlight exactly whereabouts in our ignition table we need to modify. |
| 44:27 | So we know we added some timing there, 6000 RPM and above, and I'm simply going to go back to our spark advance table and we'll go back to our 6000 RPM area, and we're going to retard our timing, in this case I'm going to simply take out the two degrees we've put in. |
| 44:49 | We'll close that down, flash our new map into our PCM, and we'll see what our final run gives us. |
| 44:58 | All right we're back up and running, so let's perform another run. |
| 45:21 | All right we've got our last run complete there, our power is essentially exactly the same again, 239.5 kilowatts at the wheels. |
| 45:28 | You can see right in the very top end, we have lost a very little amount of power as a result of retarding that timing up above 6000 RPM, but again, if the engine's knocking, it doesn't really matter how much power the engine's making, it can end up resulting in damage so we need to deal with that knock. |
| 45:47 | Let's have a look at our scanner now, and you can see that this time we have removed our knock event that we had up around about 6000 RPM and above, however we do now have a very small hit of knock retard occurring here at 5200 RPM. |
| 46:04 | Now this is one of the situations where, because we haven't previously had any knock retard occur at that point, and we haven't added any timing in for that particular run, rather than simply going and removing timing right now, I would actually run the engine through again, do another ramp run, and see if that knock was repeated and consistent. |
| 46:26 | If it was, then we'd go and address that in our spark advance table. |
| 46:30 | But there's a pretty good chance that this would be a one off event, and if we run the engine through again we may find that it isn't repeated. |
| 46:39 | Okay, we now know we've got good control over our air fuel ratio, we've got our measured air fuel ratio meeting our commanded air fuel ratio right through our open loop area of operation. |
| 46:53 | We also know we've got good control in closed loop, and again we could expect that because our car is completely standard. |
| 47:01 | We've advanced our timing up, we've gained very very small amounts of power, and the lower area of our map we've gained around about a kilowatt, in the top area of our map we know we aren't suffering from any knock. |
| 47:14 | So in this case our task on the dyno here is complete, our job's done, and our next process would be to go out and confirm and test this calibration out on the road, make sure that everything that we're seeing here on the dyno stacks out with what we see out in the real world. |
| 47:31 | It's also going to give us the opportunity to test some of the transient areas of our map, make sure that everything is working and correct in those areas as well. |
| 47:41 | Now I'll just make a point on that the power that we've seen there, obviously we've seen very much the same power through our runs on the dyno. |
| 47:49 | And in some cases that's to be expected. |
| 47:53 | In this particular situation we had a relatively good factory calibration, we haven't really made much change to our air fuel ratio under wide open throttle, and we saw that we were already very close to our optimal timing in that factory calibration, so there really hasn't been a lot of room here to move in order to improve our calibration. |
| 48:16 | In some instances however, we're going to see the opportunity to make quite dramatic changes to the power, over and above our factory calibration, it all really depends on what our starting point is. |
| 48:29 | And this is why it's so important if we do have the opportunity to do a baseline run with the car as delivered, that we do that so that we have something to compare against. |
| 48:39 | All right, so our dyno work is complete, let's move on. |
