Practical Reflash Tuning: Step 5-A: Optimising the Tune
Watch This Course
$229 USD
-OR-
Or
8 easy payments of only $28.63
Instant access. Easy checkout. No fees.
Learn more
Step 5-A: Optimising the Tune
26.29
| 00:00 | - We left our last module after correctly scaling our mass airflow sensor as well as the load compensation tables and of course during this module if we had also changed our injectors we would have adjusted our calibration to suit those. |
| 00:14 | The point being, we're now moving into our fifth step, at which point we're going to be looking at optimising our tune. |
| 00:20 | And we're working on the basis here that our fuelling now should be under control. |
| 00:25 | What we're going to be doing is optimising our fuelling, remembering that in our last module we had our wide open throttle fuelling set to an across the board 11.6:1. |
| 00:34 | We're going to be using something a little bit more realistic than that. |
| 00:38 | We're going to be optimising our ignition timing so we're going to be diving a little bit more into how the knock control strategy works in the Subaru ECU and seeing how we can go through optimising that with the MegaLogViewer HD package. |
| 00:52 | We're also going to add into this module a little bit around the boost control strategy and we'll make some changes there. |
| 00:59 | Nothing particularly significant but just to show you some of the considerations you need to make when you are adjusting the boost control system. |
| 01:08 | To start with, we're going to ignore that we already know that we've actually got our boost relatively well under control so just to simplify our worked example we're going to break it up, we're going to start by focusing on our fuel and ignition, we're going to get our tune dialled in with the stock boost control tables and then we're going to have a look at those changes separately. |
| 01:27 | So let's get started by getting our fuel table a little bit more dialled in. |
| 01:32 | We'll jump into EcuFlash here and you can see this is where we left off with our table set to 11.6:1 basically any time we're above a load of 1.45 grams per revolution so again not particularly sensible, not really where we're going to be wanting to set this. |
| 01:51 | So to start with, what I'm actually going to do is we're going to load up the original file, the stock file and we're going to drag in the fuel table from that stock map so let's just scroll down and we'll find that, we're looking for our primary open loop fuelling table and we're going to start by simply copying and pasting the entire stock table back into our modified ROM. |
| 02:15 | You might be wondering why are we doing that, particularly when we've already highlighted that the air/fuel ratio targets from the stock file are a little bit excessively rich, particularly in the top end. |
| 02:26 | The reason for this is just to speed up our process. |
| 02:29 | Some of the transition areas, particularly are probably not too far away from where we want to be so rather than adjusting every individual cell I'm going to start from this, it's a personal preference thing and you can obviously use whatever technique you personally want. |
| 02:43 | Just to make sure that we are working on the right file, now that I've got the information out of that stock ROM I'm going to come over here, we can see we've got our two ROM files that are open, I'm going to click on that stock ROM file and I'm going to click on close ROM, just making sure that there's no chance of accidentally making a mistake here. |
| 03:01 | So what do we want to do here? Well we're going to be using some air/fuel ratio targets that are sensible and suit the load that's being placed on the engine. |
| 03:10 | And I've got some reasonable experience with this particular model of car so I kind of know where we're going to want the fuelling to be. |
| 03:16 | 11.6:1, particularly under high boost that we used in the last module, actually isn't a bad place to be. |
| 03:23 | But we don't necessarily need to be that rich during the transition area as we're coming up onto boost. |
| 03:30 | Now a good place to start is actually having some understanding about what the load axis actually represents so we've got our load axis again in grams per revolution but this doesn't strictly transition across nicely to manifold pressure so let's have a quick look at a log file in MegaLogViewer HD. |
| 03:48 | So we can see here that at the start of the ramp run we're at wide open throttle. |
| 03:52 | We've only got relatively low boost pressure though because the engine is only just starting to spool up that turbocharger so somewhere in the region of maybe a psi, two psi, thereabouts. |
| 04:04 | Not a huge amount of boost pressure. |
| 04:07 | We've got our manifold absolute pressure showing that here. |
| 04:09 | And this is of course in bar so we do need to be careful, and that's absolute. |
| 04:14 | We've also got our load being displayed there, it's our yellow parameter so we can see that at the start of our ramp run there, we're about 1.5 grams per revolution. |
| 04:24 | We peak up here around about 3.8 grams per revolution and then we drop back to 2.3 grams per revolution so let's head back to EcuFlash. |
| 04:35 | So we know that we're starting the ramp run down around about this area here. |
| 04:40 | We're actually ending up off our scale. |
| 04:43 | That particular ramp run that I just looked at, a little bit higher in our boost than we're actually going to end up running. |
| 04:51 | And you'll remember we've already rescaled our ignition tables out to account for the increased load that we're going to see. |
| 04:57 | And then we're dropping back to sort of this region out here. |
| 05:01 | So we're going to take a path through our table that's going to end up looking something like this. |
| 05:07 | And it's a good idea to understand this so we know where we're moving through this table and where we want to make our changes. |
| 05:13 | Alright so let's have a look at some of these changes and again I'm going to keep this pretty simple for the purposes of our demonstration. |
| 05:18 | The techniques that we're going to use here, you can of course adapt to suit your own air/fuel ratio target so for a moment what I'm going to do is essentially highlight the entire wide open throttle operating area, incorporating 2.35 grams per revolution and above and also 3600 RPM and above and we're going to set that to the existing target that we're using, 11.6:1. |
| 05:45 | Now some of this targeting down here, remembering we're starting at around about 1.5 grams per revolution. |
| 05:51 | I've actually got no problem with that being in the mid 13s, we definitely don't need to be 11.6 so I'm actually going to leave that as it is. |
| 05:59 | Now we would be pretty lean if we were up here 2.9, 2.8 grams per revolution, 13:1 but with the turbo spool response there's no way we're going to end up getting into that point so we can actually just leave that as a matter of course, it just doesn't matter. |
| 06:14 | What I'm going to do again just for simplicity, I'm going to highlight the 11.6 that I just entered there, we're going to pull that up there and we're going to be using our V key which is for vertical interpolation just to get a nice smooth transition as we move higher in the RPM there and we come down that table. |
| 06:31 | So this leaves us with a few weird areas in our table, we can see we've got this little idiosyncrasy here, 11.6:1. |
| 06:39 | And in fact if we press alt G and bring up our graphical view, this is a good way of actually seeing if we've got anything strange going on in our table. |
| 06:48 | Sometimes this shows up easier in a graphical format than it does when we're looking purely at numbers. |
| 06:54 | So you can see we've got this particular dip here which coincides with this area here. |
| 06:59 | Now we're going to simply fix that by using our interpolate again so we'll scroll down here to 1.45 grams per revolution. |
| 07:07 | We don't need to be too specific here, I'm just trying to get a general trend to our table and of course again, you can make whatever changes you feel are appropriate or what the engine actually responds to. |
| 07:19 | So we're going to use the H key here for horizontal interpolation and that'll just make sure we've got nice smooth trends in that particular area. |
| 07:27 | So we're going to follow through making some more interpolation and we'll clean up this table a little bit more. |
| 07:33 | Alright so at this point we've got a table that's a good starting point, we can flash that into our ECU so we're going to go ahead, we'll flash it in and then we're going to perform a full power ramp run and that's going to give us a bit of a baseline for our current ignition timing, plus we can have a little bit of an analysis around what our ignition timing's doing and what we're seeing from our knock feedback so let's go through that process now and flash this in. |
| 08:04 | Alright we've completed our flash, we've got our engine back up and running, we've got all of our temperatures stabilised, so let's perform a wide open throttle ramp run we'll log our results and we'll see what we've got to work with. |
| 08:31 | Alright we've got our dyno plot up there, again we've overlaid this with that very first run we completed. |
| 08:36 | And again we've cheated a little bit, we haven't got a completely realistic result here just simply because of the boost response on our initial factory run so it's a little bit of a shame because it does make it look like we're doing a lot better through this mid range here than realistically we are. |
| 08:53 | What we can see straight away even on the dyno is our air/fuel ratio is nicely under control. |
| 08:59 | We've got a nice gentle trend where it transitions down to our rich target of 11.6:1 so you can see much much leaner than our factory calibration. |
| 09:09 | We don't have this lean dip in here as well. |
| 09:12 | So let's dive into our logger and actually see what happened during that run. |
| 09:16 | And really what we're interested in initially is just seeing is our mass airflow sensor calibration, our load compensation table working? And you can see from our third row down here with the parameters we've got logged, yes it is. |
| 09:29 | Remembering this yellow parameter that I've got here set up is a math channel, our air/fuel ratio error for Subaru. |
| 09:35 | And we can see both our green trace which is our target air/fuel ratio and our red trace from our wideband, are sitting really nicely together and we do have a little bit of error here as we transition up. |
| 09:48 | We're never going to get anything that's absolutely perfect but we can see that basically as I click through this, our error's never more than about 1% so we're pretty much on the money there and again if you're at this stage and you aren't seeing good control over your air/fuel ratio, then you're going to need to head back to that last step and put in a bit more work there. |
| 10:07 | We should at this point be within around about 1% of our target.. |
| 10:10 | It's also worth mentioning at this point that if you do have one or two little areas where you've got a touch of variation, now obviously the correct approach here is to address this in our mass airflow sensor scaling or our load compensation tables but occasionally if you've got one area that could actually be difficult to correct with our mass airflow sensor scaling without adversely affecting other areas, we can actually manipulate that slightly with our open loop air/fuel ratio target tables. |
| 10:38 | So let's have a quick look at what I'm talking about there. |
| 10:41 | So we look at this particular area right here and we can see that our target is of course 11.6:1, we're actually slightly richer here, 11.4:1. |
| 10:51 | Now that could also just be the case of a run to run variation. |
| 10:53 | I'm not personally too worried about this for the case of our demonstration but let's see how we could address that. |
| 10:58 | So what we need to know is our RPM, we can see that that's 5200, we also need to know what our engine load is and that's here, 2.95 grams per revolution. |
| 11:09 | So let's head back to our EcuFlash table. |
| 11:12 | So what we want to do is come down to our primary open loop fuelling table and the point that we're interested in is 2.95 grams per revolution and 5200 RPM. |
| 11:23 | Just make sure that we have got the correct zone. |
| 11:25 | So that's this one here so obviously it's 11.6:1 no big surprises here. |
| 11:29 | Now if we wanted to lean that air/fuel ratio out that we're actually receiving at that particular zone, let's say in our example, we're 11.4:1 so we're 0.2 richer than our target. |
| 11:42 | What we could do here is just set this to 11.8:1. |
| 11:46 | So we've leaned out our target by 0.2 Now I want to be really clear, yes we are fudging the results here so the correct technique is to correct our mass airflow sensor scaling. |
| 11:56 | But when we're very very close then this is a technique that you can use just to make fine adjustments if we're not quite where we want to be. |
| 12:04 | In our case though, we're going to set that back to 11.6:1 and we'll jump back into our logger and continue looking at what we've got to deal with. |
| 12:12 | Now that we've got our fuelling under control and our air/fuel ratio targets are about where I'd expect them to be for this engine, we can concentrate now on our ignition timing. |
| 12:21 | And this is where our parameters in our third graph down come in, we've got our ignition advance multiplier, we've got our feedback knock correction and we've also got our fine learning knock correction so the interaction here is that the feedback knock correction, that's an instantaneous response to a knock event. |
| 12:38 | So if we've got continual knock occurring at the same point, that will then become a fine knock learning correction. |
| 12:45 | And this means that when the ECU goes through that particular zone again it will preempt the fact that knock occurred last time by pulling timing out ahead of schedule so we do want to be mindful of that, make sure that we don't have an adversely advanced timing table and we're just fixing that with our fine knock learning, and we'll just see how we can do that. |
| 13:08 | For a start though we're actually looking like we're in pretty good shape here so we can see we've got no activity, we've got our IAM which is in red, sitting at 1.0 right the way through this entire ramp run. |
| 13:20 | Our fine knock learning and our feedback knock correction are also sitting at zero so this means that the ECU is not detecting any knock so it's not pulling any timing, so we're in good shape. |
| 13:32 | So if we've got this sort of situation, generally the first move I will make is to advance our entire table by 2° and see the effect of that change that we're going to make. |
| 13:43 | Before we make that change though, let's just head back to our logger and we'll go to our tools and we'll go to our learning table values and this will just pull in all of the current learning values so we can see what we've actually got going on in there. |
| 13:58 | So this is the section that we're interested in here which is our fine knock learning. |
| 14:02 | So this is exactly what we'd like to see here, the table's filled with zeros, meaning that the ECU is not pulling timing at any point and it responds to knock. |
| 14:12 | If we don't see that, if we've got some negative values in there where the ECU has some pre learned knock values, we can then reset our ECU and that will clear those and it's really important just to keep an eye on that and make sure that that's not occurring. |
| 14:26 | Particularly if you are out on the road or the racetrack and you've got a lot of data where the ECU has more of a chance to go about its fine knock learning. |
| 14:33 | So in this case, that's all sitting at zero so that's fine. |
| 14:37 | We'll head back to EcuFlash now and we'll see how we can make our timing adjustments. |
| 14:42 | Alright so you'll remember that we've got four base timing tables here and in our stock ROM these were all identical so the changes we make to one, we're again just going to simply copy across to the other three tables, making sure that they're all the same. |
| 14:56 | And we also have our knock correction advance max table. |
| 15:00 | Remember that we've manipulated that in the third step of our process. |
| 15:05 | We're not going to make any changes to that and remember that allows up to an additional six degrees of timing to be added in by the knock control strategy. |
| 15:13 | Alright so let's go to our base timing primary cruise. |
| 15:17 | From our log file, we know that we started our ramp run at about 1.5 grams per revolution and what I'm going to do is just split the difference here, we'll come up to 1.6 grams per revolution, highlight the entire table and then we will add two degrees to that particular section. |
| 15:34 | Now at lower RPM as well I'm going to add a little bit more timing in here at 1.45 grams per revolution. |
| 15:41 | What I'm trying to do here is just maintain a relatively smooth trend so this time I'll just use the increment key here just to add a little bit of timing here. |
| 15:51 | So I'm just keeping a nice trend, nice and smooth in here. |
| 15:54 | We can see we haven't created any adverse steps in our table. |
| 15:58 | Again we can use the alt G and bring up our graphical view. |
| 16:02 | So this is a pretty broad change that I've just made just 2° across the board but when we've got no knock activity showing like this is a pretty good place to start. |
| 16:11 | Now that we've made our changes to our first table, we're of course going to copy and paste those into our remaining three tables and we can just do that using control C to copy, we'll open up the next table, click into the top left, control V to paste. |
| 16:27 | We'll go through that for our remaining two tables. |
| 16:30 | With our changes now made, we're going to flash that map into the ECU and we'll get our car back up and running ready for our next ramp run. |
| 16:39 | Right we've got our flash complete, our car's back up and running, we've got all of our temperatures stabilised so we're ready to perform another ramp run on the dyno. |
| 16:46 | Now this ramp run will be overlaid directly with our previous run, so we've removed our reference run from the stock calibration and we're just looking now in real time as we perform this ramp run at the effect of our ignition timing so let's start our scanner and get our run underway. |
| 17:29 | Alright looking at the results on our dyno there you can see that we did pick up some power from that timing, we've gone up to 238 kW at the wheels. |
| 17:37 | What we could see is that as we overlaid that run in real time we picked up power almost everywhere. |
| 17:45 | We didn't really see much of a gain down around the start of our ramp run, down sort of around the 2000, 2500 RPM mark. |
| 17:51 | But we saw a marked improvement right the way through. |
| 17:54 | Let's have a quick look into our logger and we'll see what actually happened during that ramp run. |
| 18:01 | Again what we're most interested in here is the three parameters in our third graph, our ignition advance multiplier, our feedback knock correction and our fine learning knock correction. |
| 18:10 | So we can see everything's looking really good, our IAM's sitting at 1.0 the whole way through the ramp run, our two knock feedback corrections are sitting at zero so the ECU again detecting no knock. |
| 18:22 | We can see also I've added in an additional parameter into our fifth graph here which is our ignition timing total. |
| 18:28 | So this is something that is worth having a look at, comparing from run to run. |
| 18:32 | We can see what's actually happening and what's being fed to the engine. |
| 18:36 | So at our peak load here we can see that our timing is sitting at around about 8.5° total and we can see this nice ramp up in our timing here which is what we'd expect. |
| 18:47 | And we're sitting at around 28, 29° at the end of our run so we're probably pretty close to the mark for our pump fuel but because we have seen a gain with that particular change here, we're going to go a little bit further. |
| 18:59 | Now remembering the process here, if we've seen a gain everywhere then of course we would be free to advance the timing further everywhere. |
| 19:07 | In this case we didn't really see too much of a gain until we starting ramping up in our boost pressure so we're going to apply what we've learned there with these changes that we're making. |
| 19:17 | So let's head back to EcuFlash here, we'll come back to our base timing primary cruise and what we want to do here is highlight our ignition timing values here and I'm going to come down in this case to 2.0 and we're going to come out to about 3600 RPM so the area where we're starting to ramp up towards peak boost. |
| 19:40 | So what we're going to do now is just add an additional 2° to that particular area and just to make sure everything's nice and smooth here, I'm just going to use our horizontal interpolation and we can also use our vertical interpolation here just to again make sure that we've got no funny trends in our table. |
| 20:01 | With the changes made to our first table here, we're of course going to copy and paste those into our remaining tables and then we can flash that into our ECU and carry on. |
| 20:12 | Alright so we've got our temperatures stabilised again, we're ready to perform another run, we've started our data logger. |
| 20:17 | Again we are overlaying with that last run so we're again looking to see if that additional 2° since our last run has made a difference. |
| 20:26 | And it should go without saying here that of course when we are advancing the timing or for that matter any time we are running the car on the dyno, we do want to be very mindful of watching out for detonation. |
| 20:36 | And typically when I cam tuning as opposed to presenting a worked example like this, I'll also be using audio knock detection in order to just back up the information we're seeing from the logger so it's really important if you do notice any detonation as you're advancing the timing, don't stay in the ramp run, abort the run, clutch in, come back to idle, make some changes, get rid of that knock and then try again because otherwise you could risk damaging the engine. |
| 21:00 | Alright let's get our next run underway. |
| 21:18 | Alright so what we can see there is our final figure is almost identical, 238 kW at the wheels. |
| 21:24 | While we were ramping up in real time we could see that a lot of our power run essentially overlaid almost identically straight over the top of our last run. |
| 21:32 | We did see through our mid range there was a very mild power gain. |
| 21:36 | It's actually been eliminated, the gain was so small, just by the thickness of the line that's been drawn there. |
| 21:42 | However we can dive in a little bit deeper than that if we want to and see exactly what's going on, so let's save our run and we'll have a look at this on our analysis screen. |
| 21:51 | Alright analysing our last two runs here, we've got our purple run which was our previous run, our red run, the one that we've just completed with the additional 2° of timing. |
| 22:01 | So while they are very close, we can see in particular from around about 4800 RPM here, through to about 6000 RPM we have seen a mild improvement from that additional timing. |
| 22:12 | Culminating here about this point we've picked up about five horsepower, six horsepower at the wheels so definitely not insignificant. |
| 22:20 | It does also show us here in the spool up region that we've gained a little bit of power there, in this case around about four to five horsepower. |
| 22:28 | We do need to be mindful of this because depending on the exact conditions when we start the run, the heat in the exhaust manifold and turbocharger, we can always expect to see minor variations in the spool so if you are seeing an improvement like this, just check your boost response and make sure that it's not a result of the boost itself and we can see that that's exactly what's happening here if we look at our boost pressure over here on the right we can see that in the red run, must have had a little bit more heat in the manifold so we've actually reached peak boost a little bit earlier. |
| 22:59 | This means we can disregard this improvement down here. |
| 23:02 | Alright so now that we know that we did pick up some power, albeit only through the mid range, 4800 through to 6000 RPM, let's dive into our logger and see what that looks like. |
| 23:12 | Inside of our logger we can straight away see our green parameter here which is our feedback knock correction is now doing some work. |
| 23:20 | So we know we've gone too far. |
| 23:22 | In particular right here at the start of the run we can see that our feedback knock correction is pulling almost 3° out. |
| 23:29 | You can see that once the knock is removed, the timing is incremented back in but we can see that before it gets fully fed back in, we can see at this point here we've got some more timing being pulled, in this case almost 2°. |
| 23:43 | So what I would say based on this is that we've gone too far, we've got too much timing in there. |
| 23:49 | We also didn't really see any improvement up in the high RPM range, up around 6500 RPM where we've still got a little bit of knock retard in there so it looks like we were really walking a tightrope there with our last run, we were right on the very edge of where we could get our timing. |
| 24:07 | Now of course we've done this in one run or two runs here, we've added some timing that we're now going to go and remove. |
| 24:13 | But of course this can be an iterative process, it's just a case of creeping up on the timing, adding timing into the areas where we see gains in power with no knock retard activity. |
| 24:25 | If we add timing and we see no additional power of course we're going to pull that timing back out because we know we're already at MBT. |
| 24:33 | So at this point, for our worked example, our wide open throttle tune optimisation is essentially complete.. |
| 24:39 | We'd simply go back and remove that additional 2° timing from our map, reverting back to our previous run configuration. |
| 24:47 | We saw there that we ended up with 238 kW at the wheels. |
| 24:50 | Let's just overlay that now with our original stock run and we'll see what we've ended up with in terms of a final result. |
| 24:58 | Alright so we've got our stock file and our final tune loaded up here and we can see of course we've got that additional boost response which we've already dealt with. |
| 25:09 | However from the point about here, 4800 RPM and above, essentially our boost is exactly the same as stock. |
| 25:14 | Despite that we can see that we've got anywhere from around 10 wheel horsepower gain right through to, at this point we've got around about 40 horsepower at the wheels gain. |
| 25:26 | So this is pretty significant. |
| 25:27 | Remember those gains have come essentially from our fuelling and our timing. |
| 25:31 | Again we've got our fuelling down here that we've already looked at so we're sitting pretty close to our target of 11.6:1, our original run as we can see there, 10.5:1. |
| 25:42 | In the next module however we will dive into a little bit of detail around the boost control that so far we haven't made any adjustments too. |
| 25:50 | So at this point we're comfortable with our final calibration, we've got a good gain in power, we've got good control over our fuelling, our ignition timing is not resulting in any knock, so we're ready to get the car off the dyno and move onto the next step of our process where we can check everything we've seen here on the dyno, out on the road or the racetrack, making sure that everything stacks up. |
| 26:11 | In the next module however we are going to cover some of the aspects of the boost control strategy which so far we haven't touched and we'll show you how to make adjustments to that if you do want to adjust the boost pressure your engine's running. |
