Practical Reflash Tuning: Step 4: MAF/Injector Scaling
Watch This Course
$229 USD $114.50 USD
-OR-
Or
8 easy payments of only $14.31
Instant access. Easy checkout. No fees.
Step 4: MAF/Injector Scaling
1.08.03
| 00:00 | - The next step of our process is to address the rescaling of the mass airflow sensor and/or the fuel injectors as required. |
| 00:08 | Now again we've really covered over what's going to be done here, we know that we've got stock injectors so there's going to be no need to address the injector scaling inside of our calibration. |
| 00:18 | Likewise our mass airflow sensor is also factory so technically we should have no requirements to do this. |
| 00:24 | This actually sets us up in a pretty good position because with this particular car, it means that we are able to perform a ramp run right and now before we actually flash our modified calibration into the ECU and this is always a good idea where we practically can do this. |
| 00:41 | It gives us a good idea on the condition of the existing tune, we're going to straight away see any short comings or any areas that we're going to need to address, plus just as importantly it gives us a baseline power figure that we can then compare our iterative changes to as we go through and optimise our tune. |
| 00:59 | There are a couple of aspects we do need to address here just to reiterate, from the body of the course. |
| 01:05 | Remembering that if you are going to be changing both your mass airflow sensor and your injectors, then it's really important if you don't have really solid data for one or the other as you move forward, that you do this change iteratively, making one change and then the other. |
| 01:21 | Otherwise you're not going to know where to address any errors that you've got in your air/fuel ratio so that's the first important thing to mention. |
| 01:29 | The other aspect here is there are a few idiosyncrasies with the Subaru ECU or more to do realistically with the way the fuel system operates. |
| 01:39 | And this creates some error that is a little bit hard to get around in our MAF scaling so this can be addressed in the aftermarket with hardware parts in the fuel system itself. |
| 01:51 | However for a stock car we are stuck with what we've got. |
| 01:54 | To make life worse, the 2008 model onwards tends to show this problem much more dramatically than the earlier models although it is an issue that plagues all of these Subaru engines. |
| 02:06 | And essentially what this does it it creates an area, generally somewhere in the region of about 2400 to about 3200 RPM, under light load, particularly slightly into positive boost pressure as well, where the MAF scaling alone would actually create errors so Subaru address this by incorporating a MAP compensation table. |
| 02:27 | Let's just jump into our tuning software for a second and we'll have a quick look at that. |
| 02:31 | So we'll find these tables under the mass airflow and engine load part of our tree and we can see we've got two components there that are labelled engine load compensation cruise and non cruise. |
| 02:42 | Let's just open one of those tables up here for a moment. |
| 02:46 | And what we can see is we've got a 3D table, this is a compensation for our fuelling, we've obviously got our RPM on the vertical axis and this time our horizontal axis is manifold pressure, you'll have noticed that basically all of the load axes so far that we've looked at have been load in terms of grams per revolution so derived from the mass airflow sensor. |
| 03:06 | This time the load is derived purely from manifold absolute pressure. |
| 03:10 | So in particular in the stock table you can see there are some quite large trims in here, particularly around 2800 RPM, we can see there under light load and vacuum we've actually got a 21% trim being applied. |
| 03:24 | This shows you the magnitude of the problem that Subaru were trying to tune around. |
| 03:29 | So the problem is, this is a bit of a hack approach if you like to fix the actual problem which is to do with the fuel system design and more specifically pulsations in the fuel pressure. |
| 03:39 | But this is what we're working with so it's important to understand the implications. |
| 03:44 | Now the problem with this is that generally when I rescale a mass airflow sensor I'm going to do it in two ways. |
| 03:53 | I'm going to try and replicate the way the engine's actually going to operate so I'll gather some data, looking at our measured air/fuel ratio and comparing that to our commanded air/fuel ratio under steady state conditions. |
| 04:04 | I'm going to fill in as much of the MAF scaling curve as I can in terms of the voltage range so we can access under those conditions. |
| 04:11 | From here, I'm going to them switch to doing some ramp runs, this will obviously move us into the open loop area of the table and we'll do the same. |
| 04:17 | However, if we try and gather data in this way, particularly through that area where we've got those really large compensations, we can end up with some pretty misleading data, particularly if that map comp table is not correct, which with a modified vehicle we can expect some problems with. |
| 04:34 | So how are we going to address this? Well there's no set way that we must go about doing this, the technique that I'm going to use and the way I'm going to demonstrate this is to gather our steady state data in two places. |
| 04:47 | Both below the RPM range where we see these dramatic compensation numbers coming in, generally around about 2000 RPM and then I'm also going to move up above that region, up around about 4200, 4400 RPM and gather some data there. |
| 05:02 | So that basically allows us to miss out on this big chunk of that table where we've got some really big changes going on. |
| 05:09 | We can see that in this case our maximum break point there, 3800 RPM and also 3.87 psi. |
| 05:17 | Above both of those points, there is zero compensation being applied by that table. |
| 05:21 | So the key question here of course is how do we know if we've got an error, where that error is coming from, is it coming from our MAF scaling or is it coming from the fact that the numbers in our map comp table are incorrect? And the way we're going to address this, and we'll see as we go through this module is we're going to be looking at the error in our air/fuel ratio based on a certain mass airflow voltage. |
| 05:45 | Let's say we're looking at the mass airflow sensor voltage at 2.45 volts. |
| 05:50 | Let's look at that at 2000 RPM steady state, so we'll get up to that voltage and we'll look at our compensation in terms of our, correction I should say, in terms of our air/fuel ratio and let's say at that particular pint our air/fuel ratio is within 2% of our target which is pretty good. |
| 06:06 | We can then check that again at let's say 4500 RPM. |
| 06:09 | Again we find that at the higher range we've got the same sort of compensation, maybe around 2%. |
| 06:15 | If we then check that particular voltage from our mass airflow sensor in the middle of that area of the map comp table where we do see large compensations, we're likely to see that there is now a big error and we can account for that in the map comp table as opposed to the mass airflow sensor scaling. |
| 06:33 | I know that might sound a little bit messy and a little bit complex but as we move through, we're going to get a better idea of how all of that works. |
| 06:41 | For now though, let's get our engine up and running, we can lay down a base run on our dyno and at the same time we're going to gather some data that's going to give us a better idea of what the condition of our tune is like. |
| 06:53 | Before we go ahead and perform our first run, we're going to start by setting up our data logger so that we can actually log what's going on inside of the ECU during our running on the dyno. |
| 07:04 | This is important, otherwise we're not going to know what the ECU's doing and what changes might be required. |
| 07:09 | There are a variety of ways of logging data on the Subaru platform and for our particular worked example, we're going to be using the RomRaider logger. |
| 07:18 | Now RomRaider can also be used for editing and viewing the maps and parameters, however since we still need to rely on EcuFlash or another alternative option for reading and flashing, this is why I prefer to use EcuFlash for the editing as well, it's a personal preference thing. |
| 07:36 | If we jump into our software, we can see here we've got our RomRaider up and running. |
| 07:41 | We've actually loaded up the ROM file here as well and I mean basically we can look through this, it essentially looks identical to what we've already been looking at inside of EcuFlash so no real differences there. |
| 07:54 | However it's the logger that we're actually interested in and we can access this from the drop down menu here and we're going to click on launch logger. |
| 08:01 | And that'll bring our logger up online. |
| 08:04 | Alright with our logger up and running, we need to choose the parameters that we want to log and there are a lot of parameters, if we look at our list here, that we can choose from. |
| 08:15 | And we don't want to be logging every single parameter, that might seem like a sensible option but if we do that, what we're going to end up with is slowing down our logging rate so that's going to mean that we're not going to get enough samples from the inputs that we're looking at. |
| 08:30 | So we want to be mindful of the number of parameters we're choosing and make sure that the ones that we are choosing make sense. |
| 08:36 | So we can choose the particular inputs or parameters from the list just by clicking the little tick box to the left and you can see that I've selected a range here that we're going to have a look at so for a start we've got our air/fuel ratio correction one. |
| 08:52 | So this is our short term fuel trim essentially based on our front O2 sensor. |
| 08:58 | This is the instantaneous response to an error in our air/fuel ratio. |
| 09:01 | Below this we've got our air/fuel ratio learning one. |
| 09:05 | So this is the long term fuel trim. |
| 09:08 | So this is the baked in response to a long term error that remains over time. |
| 09:11 | The idea here is that the long term fuel trim kind of increments and takes away the short term fuel trim, getting that short term fuel trim back to zero. |
| 09:20 | So we can see right now at idle we're sitting with 0% short term fuel trim and our long term fuel trim's sitting at -2.3%. |
| 09:28 | So everything's actually looking pretty good. |
| 09:30 | We've also got our closed loop to open loop fuelling so this tells us whether the ECU is operating in closed loop or open loop mode, a value of eight there is closed loop, a value of 10 is open loop and a value of seven is during the warmup procedure. |
| 09:44 | We don't really need to be viewing that so I'm actually going to remove that from our parameter list here. |
| 09:52 | Another aspect that I'll just mention here that we can log is our air/fuel sensor one. |
| 09:57 | So this is the front lambda sensor, we'll bring it up here and that'll be added into our parameter list now down at the bottom so we can see right now we're sitting at about 14.6:1. |
| 10:08 | We can use this to help us with our tuning and it is a wideband air/fuel ratio sensor. |
| 10:14 | However Subaru actually only use this really close to the stoichiometric air/fuel ratio where it is pretty accurate. |
| 10:22 | The further we get away from stoic, the worse the accuracy of the sensor becomes and particularly at full boost we can find that the sensor starts to read a little bit inaccurate so it's not the best idea to rely on this soley for our open loop tuning. |
| 10:37 | For this reason I'm actually bringing in an external sensor, we can find that here on our external sensor tab. |
| 10:45 | We'll click on that and the one I'm using here is an Innovate wideband AFR sensor. |
| 10:50 | So this is being brought into the software straight into the laptop via a USB to serial adaptor. |
| 10:58 | While I've got this already set up, if you are starting from scratch, if you click on the plug ins menu here and click on the drop down, you can choose the sensor that you're using and you can then choose the comport that that will come into the software on. |
| 11:12 | That's already set up and we can see that that's just sitting there directly above our AF sensor one parameter. |
| 11:20 | Getting back to our parameters though, let's have a look at our next one down which is our engine load. |
| 11:24 | And this is a key input because this tells us where abouts we're operating inside most of those tables. |
| 11:30 | This is the load in terms of grams per revolution derived from our mass airflow sensor. |
| 11:36 | We've then got our engine speed which is pretty self explanatory. |
| 11:39 | We come back down further to our feedback knock correction and there's three parameters here we're logging in terms of the knock feedback, our feedback knock correction, this is an instantaneous response to knock activity. |
| 11:51 | We've also got our fine knock learning which is what the response over time is, there's also a stored map here for fine knock learning as well so that if knock has been occurring consistently in one area the ECU will determine this and then add a retard into that area of the map so that this helps prevent future knock when the ECU goes through that particular area. |
| 12:17 | We've also got our ignition advance multiplier which we discussed in the last module. |
| 12:21 | Coming back up, we've also got in that list there our final fuelling base. |
| 12:26 | So this is essentially the air/fuel ratio target that the ECU is chasing or asking for and it's an important aspect to be logging the final fuel fuelling base target as well as our measured air/fuel ratio. |
| 12:40 | This gives us that input on how far away from away our target we're actually operating. |
| 12:44 | We've got our ignition total timing so this is the final timing being delivered based on the base ignition timing map, the knock advance map and basically any knock feedback that is going on in there so we can see exactly what's happening. |
| 12:58 | We've got our main throttle position and that's pretty self explanatory, we've got our manifold absolute pressure, also pretty self explanatory. |
| 13:05 | Then we move down, we've got our mass airflow as well as our mass airflow sensor voltage so these are two aspects that are important when we are rescaling the mass airflow sensor. |
| 13:15 | The mass airflow sensor voltage, the output from that sensor into the ECU in terms of a direct voltage so that we know where to make changes to that table and the mass airflow value in terms of grams per second. |
| 13:27 | This is the value that the ECU thinks is going in in terms of airflow into the engine. |
| 13:34 | Then we've also got our primary wastegate duty cycle so this is a parameter that we would use to help us with our boost control tuning. |
| 13:41 | There's a few other parameters that we would add in there to optimise our boost control but we can look at those a little bit later on. |
| 13:47 | For now, we're going to click on the green start file log button and this will start recording data so we can review it later on and we'll get our first run underway. |
| 13:57 | It is important to mention here that when you are running the car for the first time, particularly if it's an unknown quantity, you want to be mindful of monitoring knock, your boost pressure as well as your air/fuel ratio and be prepared to abort the run if anything is outside the areas that you're comfortable with. |
| 14:12 | Let's get up and running now. |
| 14:46 | Alright we've got our first run complete there, now that the run is complete we can click stop file log on our logger. |
| 14:53 | Before we look at our log though, let's have a quick look at the data on our dyno. |
| 14:57 | And we can see that we laid down 297 horsepower at the wheels and our air/fuel ratio plot in the middle here shows our measured air/fuel ratio. |
| 15:07 | You can see I've got a reference line in here just at 11.6:1 to give us a quick guide as to where we are. |
| 15:14 | We can see that we've got this lean spike as we start to transition up onto boost and we'll discuss that in a bit more detail shortly. |
| 15:20 | We can see that the air/fuel ratio actually ends up tracking very rich, we're down around 10.5:1 but you'll remember that's pretty much on point with the target air/fuel ratios we had inside of our table so nothing too surprising there. |
| 15:35 | We've got our boost pressure and we can see that we're peaking at around about 17 or 18 psi so our boost pressure's doing pretty much what we expect it to do. |
| 15:44 | What we're going to do is we'll save that run there and this will allow us to use that as a reference. |
| 15:51 | We'll just call this base run. |
| 15:54 | OK so that's going to stay there on our dyno screen for later on and we're going to be using this in particular in the next step of our process when we are optimising our calibration. |
| 16:03 | Now let's dive into our data and we'll see what that can tell us. |
| 16:07 | And in terms of analysing the data logs I am going to be using MegaLogViewer HD. |
| 16:11 | It's really versatile and it makes it easy to analyse data from a wide range of different packages that produce a .csv file so let's load that file up now and we can have a look at it. |
| 16:24 | Alright we've got all of our data loaded up and we'll just zoom in a little bit so we can see a little bit more detail. |
| 16:30 | Now the sort of parameters that you want to view here is going to be dependent on exactly what you're doing. |
| 16:35 | I've got some basic ones set up that I'm going to be using for a lot of my tuning. |
| 16:39 | To start with up the top we've got our RPM and that's pretty straightforward, we start the run at about 2000 RPM and you can see that we ran the engine all the way through to about 7200 RPM. |
| 16:51 | In our second set of graphs we've got our manifold absolute pressure, our mass airflow sensor voltage as well as our engine load. |
| 16:57 | Now the boost pressure, we want to have a look at that. |
| 17:00 | We can see on the left hand side it shows us our maximum values here, we hit 2.39 bar, just a touch under 20 psi so around about where we want to be. |
| 17:08 | We can see that that drops down at the higher RPM range and we're sitting just above 2 bar or about 15 psi. |
| 17:15 | Nothing untoward there so we're happy with our boost. |
| 17:18 | Our airflow sensor voltage, we do want to be a little bit mindful of this and in the top left corner here we can see that our maximum voltage was 2.64 volts. |
| 17:27 | Not necessarily a problem but remembering that this is a 0-5 volt output sensor, we are getting very close to the maximum range of that airflow meter. |
| 17:36 | If we wanted to run any more boost we would definitely need to consider an upgrade to a larger MAF tube. |
| 17:42 | In our case though we are fuel system limited so we're not going to be raising the boost any further and it's right at the higher RPM range that we are seeing that maximum value there. |
| 17:51 | We also want to have a look at our yellow parameter here which is our engine load and we can see that that hits 3.42 grams per revolution. |
| 18:00 | You'll remember that we've already extended out our ignition tables to 3.6 so we're comfortable at the moment we're within that range. |
| 18:08 | We've got another couple of parameters down here in our third graph, we've got our ignition advance multiplier, we've got our feedback knock correction and our fine knock learning. |
| 18:17 | And essentially what we want to see with a good quality tune here, we always want to see our ignition advance multiplier sitting at 1.0. |
| 18:24 | You can see right at the start of the run here it actually does increment up. |
| 18:27 | The factory calibration always starts with the IAM value at 0.5 and then advances or increments that if no knock is occurring. |
| 18:36 | And we've got no knock feedback occurring so again no detonation sent during that run. |
| 18:42 | We're happy with that. |
| 18:44 | It's our third set of values here, third and fourth sets of value here that we are most interested in though so let's have a look at what we've got here. |
| 18:52 | First of all, the red parameter here is our Innovate wideband air/fuel ratio sensor so that's our wideband air/fuel ratio meter coming into the laptop. |
| 19:02 | We've also got our final fuelling base in green, so remember this is our target air/fuel ratio. |
| 19:09 | We can see that this starts at the beginning of the run, 12.8:1 and during the run we see that it drops down to 10.3, 10.2 and it does move around a little bit based on the targets inside of that table. |
| 19:21 | It's really a good idea when we are setting this up as well to make sure that the scaling of our wideband as well as our target air/fuel ratio are the same, otherwise it can be quite off putting or misleading when we're looking at these values. |
| 19:36 | We can do this by clicking on the three little lines to the right hand side of the parameter and we can come down to our scaling. |
| 19:42 | You can see at the moment I've selected manual scaling with a minimum of 10:1 and a maximum of 15:1. |
| 19:48 | We can edit that though, let's say because we're getting quite close to the bottom there, we change that 10 down to 9:1, now you can see that because I've rescaled the wideband input but not the target air/fuel ratio, we now look like we've got quite large discrepancies in our air/fuel ratio. |
| 20:05 | Let's fix that by going to our final fuelling and we'll change that as well. |
| 20:10 | I'll just edit that so that the ranges are exactly the same. |
| 20:14 | So this is a good idea to do before you really analyse your data. |
| 20:18 | Now you can see that in particular through the mid range here, we're actually pretty close, we're not too far off the mark at all. |
| 20:24 | What we can see however is particularly in this region here around about 3300 RPM through to about 4000 RPM, we've got this lean area and I just want to focus on this for a second. |
| 20:37 | Let's first of all look at what's happening under the start of our run, 2000 RPM where we're setting the car, under these conditions our measured air/fuel ratio is actually richer than our target, our target is 12.8, our measured is 12.2 so we're a little bit rich here. |
| 20:53 | We've actually got this yellow parameter I'll discuss in a second that is a nice math channel that tells us exactly what that error is. |
| 20:59 | So at the start of the run we're a little bit rich. |
| 21:02 | Then if we look at what happens at the end of the run here, we find that at the end of the run, we're actually pretty much on the money, we are, if anything very very slightly too rich but pretty damn close to our target. |
| 21:17 | However the problem we've got is this mid range here which is where we are quite lean as the engine transitions up onto boost. |
| 21:24 | And this is a problem you will find quite a lot with Subaru engines and this comes down to the fuel system issue that I mentioned earlier. |
| 21:33 | Now this is what we're going to need to be correcting in our compensation map. |
| 21:38 | So we'll have a look at how we can do that. |
| 21:40 | If we tried to fix this just in our MAF scaling, we're going to end up with a really ugly offset into our MAF scaling, we're actually going to cause more problems that we fix with that. |
| 21:51 | Now that we've discussed that, I just want to talk about our math channel that I've created here, the yellow parameter which is our AFR error Subaru. |
| 21:59 | Now this is quite helpful to create a math that just instantaneously gives us the error between our measured and our target air/fuel ratio. |
| 22:07 | We can use this to help us with our MAF scaling. |
| 22:10 | So we can do this by going up to our calculated fields drop down menu here. |
| 22:16 | And we can create a custom field. |
| 22:19 | Now I've already gone ahead and done this and this is the channel that I've created here. |
| 22:22 | If you're starting from scratch, you'd simply go to add custom field, you can give that custom field a name and then you can create your math channel. |
| 22:32 | Basically you can bring in any of the parameters that you're interested in by holding down the control key, pressing the space and then you can choose the parameters from that drop down list. |
| 22:43 | I've already done that so let's just have a look at how I've set that up. |
| 22:46 | We'll bring that up under edit and we can see what we've got here, we're just using the calculations that we've already learned in our understanding air/fuel ratio course. |
| 22:57 | If we want to correct our air/fuel ratio, we look at our measured air/fuel ratio, in this case from our Innovate wideband, we divide that by our target or desired air/fuel ratio which in this case is our final fuelling base parameter. |
| 23:10 | That's going to give us our error but in order to represent this as a percentage which makes it a lot easier to work with we want to subtract one from this and then multiply it by 100 so that's what that math channel is there. |
| 23:23 | Not going to edit so we'll cancel it. |
| 23:25 | So we can see that that math channel, we know what it means. |
| 23:28 | At the start when we're settling, we have an error there of -4% which means that we're 4% richer than our target. |
| 23:36 | We see, during this transition area here, our error's quite massively positive. |
| 23:41 | We need to actually add 17% in order to get our air/fuel ratio back under control. |
| 23:46 | Then as we move through the rev range, through the run we can see that our error moves around but generally it's within about 1-2% of our target. |
| 23:55 | And generally that's what I'm trying to aim for is 1-2% or better. |
| 23:59 | The closer obviously we can get it the better as wel. |
| 24:02 | So that covers off our air/fuel ratio parameters that we're going to be using for the rest of this particular module. |
| 24:09 | Lastly though I also just want to talk about the bottom graph that we've got here. |
| 24:12 | We've got three parameters in here, we've got our air/fuel correction one, which is our short term fuel trim, our air/fuel correction, sorry air/fuel learning one which is our long term fuel trims. |
| 24:23 | Now obviously these aren't active in open loop so we can see that those are just sitting at zero right the way through as we'd expect. |
| 24:31 | One other parameter that we will just mention here that I've got in that last set of graphs is our mass airflow value and we do want to be mindful of this, we can see that the maximum value that is shown here, 295 grams per second. |
| 24:43 | Now that is dangerously close to our airflow limit so before we flash our modified calibration in, we are going to make a couple of changes. |
| 24:54 | Let's head back over to EcuFlash and what we're going to do is go through to our MAF limit here and we can see that that is set to 300 so we want to give ourselves a bit of headroom there, let's increase that to 330 grams per second. |
| 25:09 | There's also one other change I'm just going to make here before we flash this into the ECU, it's not really related to our mass airflow sensor calibration but it's something we should have done in our last step here so we'll address it now. |
| 25:22 | We'll bring up our ignition timing knock control and we want to come down to our advance multiplier initial value. |
| 25:29 | So we'll open that up here. |
| 25:31 | And we can see that that is set to 0.5 which I've already discussed. |
| 25:35 | So this is the default value. |
| 25:37 | Now what this means is that initially after we restart the car that IAM value starts at 0.5 so we're not going to get the full value of the ignition timing from the ECU, it needs to learn over a period of time that no knock is occurring and it will increase or increment that value up to 1.0. |
| 25:55 | Again because we're tuning this properly, we're going to be assuming that no knock will occur so I'm going to set this to 1.0. |
| 26:02 | Alright we're going to close that down and we're now going to go through the process of flashing the ECU. |
| 26:08 | In order to do this, we need to turn off the car and we need to also disconnect our scanner so we'll go back to our scanner here and we can do this just by using the little control buttons here, we'll click on the disconnect button. |
| 26:21 | We now need to connect our green two pin connector at our ECU and we can go through the flash process. |
| 26:29 | Once the two pin connector is connected, we can go back to EcuFlash, we can click on the little write vehicle button and we can simply follow the instructions on screen, just like when we're reading from the ECU, we want to be mindful of our timing between turning the vehicle on and pressing OK so let's go through the flash process now. |
| 26:52 | Once the flash is complete, we can click on OK, we can shut off the ignition and also disconnect our diagnostic connector. |
| 27:00 | We then want to obviously start the vehicle. |
| 27:02 | You'll find that after the reflash, it will actually crank for an extended period of time before it starts up so let's go through that now and get our engine up and running. |
| 27:12 | Once we've got the engine back up and running, we do also need to understand that we need to leave the engine running for a minute or so before it will operate as normal, otherwise you're going to get some pretty unusual results straight after a reflash if you jump into doing a ramp run. |
| 27:26 | What we're going to do here is gather data in three ways and this is a slightly complex way of scaling our MAF and unfortunately our job is just made difficult by that weirdness created by the Subaru fuel system. |
| 27:37 | So we're going to gather some steady state data, as we've mentioned, at 2000 RPM, some steady state data at around 4000 RPM and we're just trying to bridge the range of our mass airflow sensor voltage, essentially all the way from idle through to maximum range. |
| 27:52 | Also trying to gather that data under the conditions the engine is going to run, then we're going to, for our third piece of data we're going to gather that data doing a wide open throttle ramp run. |
| 28:01 | Now when we're gathering this steady state data, there's a few things we need to keep in mind, first of all we want to make sure that the engine is at a normal operating temperature before we do these runs, so we don't want any heat soak, we don't want any cold start enrichments to be affecting our results, so that's the first thing. |
| 28:18 | We also want to be very smooth with our throttle input, we don't want any transient acceleration enrichment coming in that will play havoc with our fuel trims. |
| 28:27 | So long runs, smooth on the throttle and smooth on any RPM changes, that's going to be the key. |
| 28:34 | It is easier to gather this data on a dyno but not impossible to do it out on the road if you're careful. |
| 28:40 | The other thing we're going to do is then perform a wide open throttle ramp run and that's going to gather our data for the higher range of our mass airflow sensor voltages and the reason we want to do this under a ramp run condition is that when we're at full throttle and we're really at the top end of our mass airflow sensor scaling or flow, we're obviously going to be accelerating so we're just trying to gather the data in a way that actually makes sense while skiting around that area between about 2800 and 3400 RPM give or take. |
| 29:09 | So let's get going, we'll have a quick look in our scanner for a moment and look at the key pieces of data that we want to take note of. |
| 29:16 | So we've got our AF correction one and our AF learning one values. |
| 29:21 | We can see right now, sitting at idle that they're actually looking pretty good, we've got our short term fuel trim essentially skirting around zero and we've got a slight negative long term fuel trim with AF learning one. |
| 29:32 | Also logging in our AF sensor so this is our front wideband O2 sensor and of course we've got our Innovate wideband sensor down the bottom, small discrepancy between those. |
| 29:42 | The other parameter that we want to take note of for this particular section is our closed loop and open loop fuelling. |
| 29:48 | So we want to make sure that while we're gathering this data that the ECU is actually operating in closed loop so as soon as we switch into open loop we're going to stop our log, we're going to basically ignore any data when we are in open loop. |
| 30:00 | I'll just mention here as well that if you haven't disabled the rear O2 sensor, then you're also going to want to log in AF correction three as part of this. |
| 30:10 | So what we can do is start our log file here, we'll just gather some data initially at idle. |
| 30:15 | We're going to bring our dyno RPM set point right down and because we're only looking at the low flow scaling here of our airflow meter, probably going to be going up to 2.4, maybe 2.5 volts or thereabouts. |
| 30:28 | I'm actually going to use second gear and the reason this, that I'm going to do this is it gives a little bit more torque multiplication, allows me to actually run the car on the dyno with less throttle input and this will allow us to do a better job of bridging between our voltage that we've got now at idle, if we have a quick look at our mass airflow sensor voltage right now, we're sitting, you can see at around about 1.2, 1.25 volts. |
| 30:53 | So that's going to let us bridge that gap between the lowest point we can maintain at maybe 1500 to 2000 RPM. |
| 30:59 | So let's go ahead and get that data now. |
| 32:35 | Alright we can see that our open loop, closed loop fuelling has just changed from eight to 10 in its status, 10 being open loop so we can stop our log file there. |
| 32:44 | So that's our first piece of data gathered there at our lower RPM range. |
| 32:50 | Instead of looking at this data, what we're going to do is just go ahead and collect all our data and then we can analyse our three different log files all at the same time so we're going to bring our RPM up now to about 4000 RPM, maybe just a touch above, we're going to bring our throttle back to the lowest point where we can maintain RPM on the dyno, going to be doing this in fourth gear this time and then we're going to go ahead and make another log file, so we'll do that now. |
| 33:15 | Now before we start recording, we do want to take note of our mass airflow sensor voltage again. |
| 33:20 | So what we're trying to do of course is span the entire range there so we've gone up to about 2.4 volts with our first test. |
| 33:27 | Of course we want to try starting from around about the same point, maybe with a little bit of overlap with our next test so I'll just bring the throttle back down and see if we can get a little bit lower than the 2.46 and we can get down to the 2.3s. |
| 33:39 | So we'll click on start our log file and we'll go through the same process, increasing our throttle position until we switch into open loop. |
| 34:21 | Alright so we've just switched to open loop there so we'll stop that log file, come back to idle. |
| 34:25 | Again making sure that we pay attention to our air temperatures and our coolant temperatures, making sure that that all stays under control as we're going through these tests. |
| 34:33 | So for our third and final piece of data that we're going to gather, we're going to perform that wide open throttle ramp run and hopefully between those three pieces of data we're going to be able to span that full range of our mass airflow sensor scaling. |
| 34:46 | So far the sort of numbers I've seen with those trims actually aren't too bad. |
| 34:49 | Now one thing we do need to be mindful of here, if we're going to perform a wide open throttle ramp run, obviously we're going to be in open loop mode. |
| 34:57 | So traditionally what we're going to be doing there is we're going to be looking at the wideband air/fuel ratio sensor and we're going to be looking at the air/fuel ratio error between that and our command level target air/fuel ratio. |
| 35:09 | However our long term fuel trims, our learned trims will actually be carried over into our open loop operation. |
| 35:16 | So that's a little bit of a trap we need to be mindful of. |
| 35:18 | There's two ways of dealing with this, what we can do is that we can incorporate the long term trim that's been applied into our calculated error, otherwise if we don't do that we're going to be not accounting for the full error that's actually present, or alternatively what we can do is reset the ECU and reset those learned trims and we'll see exactly how we can do that. |
| 35:40 | So in our RomRaider logger, if we go into our tools menu here, if we go down to our learning table values and click on that it'll ask if we want to stop the logger and actually look at our learned trim so we'll do that, it'll load these up and then we're be able to view those. |
| 35:57 | So what we've got here is in particular, our air/fuel ratio trim so these are our learned trims, our long term fuel trims, so they are separated into different airflow ranges. |
| 36:08 | We can control that in the ROM file itself, we can decide what airflow ranges we want to use but these are stock, we're going to leave that as is and basically what we can see here is once the airflow range, the mass airflow is above 40 grams per second, there is a learned trim there of - 3.1%. |
| 36:27 | That's going to be applied in open loop operation. |
| 36:31 | This is one of the reasons we really want to pay attention to getting our trims as close to zero because otherwise it can affect our open loop in a negative way. |
| 36:39 | So what we can do here as well while we're on this page is we can see this is our fine knock learning so any knock trims that have been learned, obviously at the moment we've got zero which is great. |
| 36:51 | So what I'm going to do here is I am going to reset the ECU, we can do that by going to the tools menu and we can just use the reset engine control unit option, or you can use the F7 hot key. |
| 37:03 | It'll ask you if you want to do that, we'll click yes and then we need to turn the ECU off, turn the engine off and restart it. |
| 37:10 | And then we can go ahead and gather some data, so let's go through that process now. |
| 37:15 | We'll get up and running and we can start doing our full throttle ramp run. |
| 37:19 | Alright we're back up and running now, we've allowed our engine to run for a moment or so and everything to stabilise. |
| 37:26 | And what we can see here in our logger before we get started, our trims at the moment are sitting at zero and the reason for this is that right now we are actually in mode seven or status seven which is warmup so our wideband O2 sensor, factory wideband O2 sensor currently hasn't come online. |
| 37:42 | That's absolutely fine because those trims are going to remain at zero so we can get up and running and perform our ramp run, we'll start our log file and we'll get it underway. |
| 38:06 | Alright so our first run complete there, at the moment clearly we're not really that interested in our power although it is interesting that already we're a little bit above and beyond where we were power wise with our stock run. |
| 38:18 | In particular you can see we've gained a little bit of power in the top end. |
| 38:21 | We can also see that now we've got a little bit more boost than we had. |
| 38:25 | Now this is really not a case of any tuning changes we've made. |
| 38:29 | This is more a case of one of the idiosyncracies with the Subaru. |
| 38:33 | Because of the long exhaust manifold with the turbo position, we do find that the boost response on the dyno and on the road and track for that matter is very susceptible to differences in the heat in the exhaust manifold. |
| 38:45 | What I mean by this is if we just run the car on the dyno and do a single power run, we quite often will find that the turbo is a little bit lazy to come on boost and that's really indicated by our first ramp run there that we've used as our base. |
| 38:57 | If we'd done another run with the heat in the manifold we're likely to get that quicker boost response which we're seeing there so no real tricks to our tuning per se because we haven't made any changes to that area. |
| 39:09 | We've got all of our logged files now so let's jump into our logger and we'll see what we can learn from that data. |
| 39:15 | Alright so here's the data from our first test, so at the top we've got our RPM, we can see obviously we started at idle and we increased our RPM up to around about 2000 RPM. |
| 39:26 | At the same time we were increasing our throttle position to span the range of our mass airflow sensor voltage. |
| 39:32 | Now our mass airflow sensor voltage is shown here in our second set of graphs in red. |
| 39:38 | We can see at the start, at idle, we are sitting at around about 1.2, two volts as we've already discussed. |
| 39:45 | We've got a reasonably smooth and consistent increase in that mass airflow sensor voltage as we increased our RPM and throttle position and that's exactly what we want, nothing particularly erratic in there. |
| 39:55 | So we ended up just before we switched into open loop mode there at 2.44 volts which again is exactly what we saw in our data. |
| 40:05 | On our fourth group of graphs here, we've got our white parameter here which is where we can monitor our open loop and closed loop trims, what mode we're operating in so again eight being closed loop, so this is the point where we switch right here to status 10 where we go into open loop so you want to really disregard the data beyond there. |
| 40:25 | So how are we going to analyse this data? What we're going to do here, because we are in closed loop mode, we're going to disregard for a moment our wideband air/fuel ratio input and instead we are just looking at those fuel trims. |
| 40:36 | So down in our bottom group here we've got our air/fuel correction one and our air/fuel learning one, our short term and our long term trim. |
| 40:45 | Now if we're trying to analyse these together it can get a little bit messy, it's a bit tricky to actually figure out exactly what's going on. |
| 40:51 | So what I've done is I've created another channel which is this yellow channel here called total trim, pretty easy math channel, we'll just see how I did that. |
| 41:00 | We'll go up to our calculated fields, come down to our custom fields and then we're going to have a look at total trim. |
| 41:08 | So let's have a look at editing that. |
| 41:11 | So we've obviously got the name there, we can call this whatever we want. |
| 41:14 | And simply put, we are using a formula here that adds the AF correction one and the AF learning one parameters together so very very simple. |
| 41:24 | If you were using the rear O2 sensor as well, you'd also add in here, AF correction three into this. |
| 41:32 | So that's all it is, very very simple to create, if you do need to create that from scratch, you come to the calculated fields and you'll go into custom fields, you'll add custom field there and this is where you basically do exactly what you just saw, give it a name and then you can add your parameters in here into the formula. |
| 41:49 | You can bring up the available parameters by pressing control and space and choose from that drop down list. |
| 41:55 | Really really simple. |
| 41:56 | So now we've got a useful piece of data that we can plot to show us exactly how much error we've got going on in here. |
| 42:03 | So we can see this yellow line here and for the most part it's actually not too bad, we've got a few little idiosyncracies, this one particularly here, probably going to disregard this, this is as we're coming off idle, we're applying a little bit of throttle and actually getting the car moving so I wouldn't necessarily take too much notice of that and we've also got this area up in here where we are a little bit lean, we're adding around about 5-6% there with our corrections to get that back on point. |
| 42:34 | Then as we click through here we actually see that we start to go negative, or about -3% before we switch into our open loop mode. |
| 42:41 | Now we can use this data to directly help us with our mass airflow sensor scaling however in it's own right, it's still a little bit messy so what I like to do here is use our histogram view. |
| 42:52 | So we'll go over to that and see what we've got available. |
| 42:55 | Now with our histogram view, what we can do is set up a histogram to log any parameters that we like in a 3D table. |
| 43:03 | So what I've done here is set up a table where our RPM is our X axis, we've also got our mass airflow sensor voltage on our Y axis. |
| 43:13 | The reason is that this is the parameter we really are interested in, we want to know what our error is relative to our mass airflow sensor voltage. |
| 43:21 | Typically we would actually do this with a 2D table because RPM shouldn't really come into it, however the Subaru with the fuelling issues that we've discussed, we do need to consider our RPM and we'll see why that's important as we progress through this module. |
| 43:36 | Then the actual parameter we are logging into our histogram here is our total trim, remembering our combination of short term and long term fuel trims. |
| 43:43 | We can also set how many decimal points, or the precision of the display here to suit, we've got two decimal places on both of those parameters. |
| 43:53 | We can set up our table size to suit, we've got a 20 x 16 table, really doesn't matter too much. |
| 44:01 | We want a reasonable amount of accuracy at least in our mass airflow sensor voltage scaling because that table is relatively tight. |
| 44:08 | And down here in our data details we can see that it's recorded 692 hits while we've gone through that log file. |
| 44:17 | So what we're trying to do here is not be particularly specific. |
| 44:21 | We don't want to try and manipulate every single voltage point on our mass airflow sensor curve. |
| 44:27 | Instead, what we're trying to do here is look for trends in this data. |
| 44:32 | So if I look at this, basically the majority of this table we've got some negative trims down here below around about 1.8 volts. |
| 44:42 | And we've also got some negative trims in here around about 2.2 volts and above. |
| 44:48 | In between we've got this little spot here where we've got some positive trims. |
| 44:53 | Now I'm not necessarily going to take that on face value, again I feel that this is probably more to do with that transition off the idle that I already mentioned. |
| 45:04 | So basically if we look at this data, excluding that area there, our trims are probably going to be around about -2, 2.5%. |
| 45:13 | Now we don't have to be absolutely perfect here as well. |
| 45:15 | You're going to end up wasting a lot of time chasing your tail if you tried to get this absolutely perfect on zero, it's just not going to happen. |
| 45:23 | Generally if we're within about +/- 2-3%, we're actually looking pretty good and that's already where we are for the majority of that table. |
| 45:31 | So that's only our first piece of data there so let's load up our second log file and see what we've got going on there. |
| 45:37 | Alright we've got our second log file open now, we can see this time our RPM sitting around 4000 RPM and pretty well stationary right through this log file. |
| 45:46 | If we look at our mass airflow sensor voltage in our second lot of data, we can see we're starting at about 2.44 volts as expected and right before we transition into open loop we are sitting at about 3.1 volts so we've now spanned all the way from 1.2 volts up to 3.1 volts. |
| 46:05 | That's a fairly good chunk of our mass airflow sensor scaling. |
| 46:07 | And again we're looking in closed loop mode here so we're looking at that total trim parameter and we've got that here. |
| 46:15 | We can see that basically at the start of that log there, we are around about -5%. |
| 46:21 | It's a reasonably smooth change and we end up right up the top, we're actually not too bad, we're closer to about -1.6, maybe 1.0% so pretty good, again let's have a look at that in our histogram for a better idea of what's going on here. |
| 46:38 | Alright so again looking at this, basically just confirms what we've already seen at the start of the log file, around about -5% and then up the top of the log file, around about -1, 1.5%. |
| 46:49 | You can see this little jump here, around about 3.1 volts and above on our mass airflow sensor scaling. |
| 46:56 | Now we need to understand what's going on there and you can see it goes to about -4%. |
| 47:00 | Let's jump back into our log viewer and we'll see what happened there. |
| 47:03 | We can see that that is because our parameter here for our open loop, closed loop status has gone into open loop mode so we want to disregard that particular piece of information so realistically at the moment we are only going to be going up to about 3.1 volts and we're going to take the data from this area here where we were still in closed loop so around about -1, 1.5%. |
| 47:28 | So at this point we've got a good picture of what's going on in the closed loop areas of our mass airflow sensor scaling. |
| 47:34 | Let's load up our third piece of data now and we'll see what that can tell us. |
| 47:39 | Right we've got our ramp run loaded up here so we can have a look through that. |
| 47:42 | Of course we've got our RPM up the top there, our ramp run between about 2000 RPM and 7100 RPM. |
| 47:49 | However we are most interested in our air/fuel ratio data here. |
| 47:52 | So in green we've got our target air/fuel ratio and you can see that across the entire ramp run there we are operating in open loop and our target is that fixed 11.6:1 that we programmed in at the start so this is just to help us getting our mass airflow sensor scaling dialled in. |
| 48:10 | Our red parameter there in that same group of graphs is our wideband air/fuel ratio. |
| 48:15 | Again we can see this lean hole that we're going to fix with our map comp very shortly but for the moment we're not going to be too worried about that. |
| 48:22 | We can see that at the start of the run here we were actually a little bit lean and then as we move through the rev range and we get up onto boost we are actually sitting a little bit rich. |
| 48:34 | We're sitting down at about 11.0 to a target of 11.6. |
| 48:38 | So at the moment our yellow parameter which we've previously used, our total trim, unfortunately isn't actually that much use to us because our closed loop trims here, our air/fuel ratio correction and learning, they are all sitting at zero which is because we are in open loop and we reset those learnings. |
| 48:57 | So instead of using the total trim parameter, we're going to come back and we're going to use our other calculated parameter here which is our air/fuel ratio error Subaru, which remember that's out measured air/fuel ratio compared to our target. |
| 49:10 | So this is what we're looking at now and this shows us our error so to start with we can see that the beginning of the run, we have a positive error of around about 3, 2.5%, then as we move through the rev range we can see that that error actually goes negative and we're closer to about 4-5%. |
| 49:26 | Again rather than nitpicking through the time/distance graph there, what we're going to do is jump across to our histogram and at the moment we're still set up with our Z axis being total trim so we want to simply just change that over to our other calculated parameter here and now we can see our errors. |
| 49:44 | So what we can see essentially from this is that our error, in particular at the high end of our mass airflow sensor scaling, we're going to call that, let's say about -4, 4.5%. |
| 49:58 | Always better to underestimate the amount of trim required and this is always going to be an iterative process so we're not going to get it absolutely perfect on our first go. |
| 50:09 | So now that we've got a better idea of what's going on in the top end, we can also have a look at our mid range and we can see that we've got some weirdness going on here we've got these massive positive errors and again this is this transition area, we're going to ignore that for the moment because it is happening, you'll notice here around about that 2800 to 3800 RPM so this is again just being caused by the weirdness with the fuel system pulsation so we're going to ignore that. |
| 50:35 | Down at the lower end of the range we can see that we've got some positive trims in here. |
| 50:41 | We are still transitioning up into that 2800, 2400 RPM range so again what I'm actually going to do is ignore that. |
| 50:48 | So what we're going to do is head across to EcuFlash and we'll bring up our mass airflow sensor scaling and we're just going to make some broad changes here. |
| 50:55 | Remembering when we are adjusting our mass airflow sensor scaling we want to be mindful of keeping a consistent shape. |
| 51:03 | So if we're starting to get an irregular shape with peaks and troughs in our mass airflow sensor scaling then this is a good indication that something else has gone wrong and you need to address that. |
| 51:13 | So what I'm going to do, remembering that our first pieces of data were around about 1.2 volts at idle so we're just going to go from 1.2, or right from 0.9 volts from the start of our scaling graph there, we're going to go all the way up to 3.1 volts. |
| 51:30 | Now there were some ups and downs in that but again I'm just going to be making a broad change here and in general we were around about -2, 2.5% so what we can do here is use the multiply symbol and to reduce our fuelling, or our mass airflow sensor scaling by 2.5%, we want to multiply by 0.975. |
| 51:51 | OK so that's just going to reduce that area of the table. |
| 51:55 | The other problem we had is that up around about 4.5 volts, we were somewhere around about 5% too rich so what I'm going to do is just scale that back a little bit to about 4.3 volts and we're going to multiply by 0.96 so we're going to actually start by just taking 4% out there. |
| 52:17 | Always better to creep up on it rather than trying to correct it completely in one hit. |
| 52:21 | So we've done 2.5% and then we've done 4% so just to keep things nice and smooth here, what I'm going to do is now highlight those cells in between that I currently haven't changed, as you can see, this has created a little bit of a step now which is not what we want in our mass airflow sensor scaling. |
| 52:38 | So what I'm going to do is basically split the difference and let's take our -3%. |
| 52:42 | So we're going to multiply by 0.97. |
| 52:45 | So that's our first round of modifications to our mass airflow sensor scaling. |
| 52:49 | Let's flash that into the ECU, we'll get back up and running and we'll see how well that's worked out. |
| 52:56 | Once we've got the engine back up and running of course, we're simply going to go through the exact same process that we've just looked at, gathering data in those three ways. |
| 53:04 | We're not going to repeat this because you've already seen how to do that. |
| 53:06 | Instead, we'll jump ahead and look at the data I've gathered after that first round of changes we've made to our mass airflow sensor scaling. |
| 53:14 | Again I need to reiterate, this is an iterative process so we're not expecting to get this perfect in one go. |
| 53:19 | So let's first of all jump into our logger here and we've got our first piece of data that is gathered as we can see here, from our mass airflow sensor voltage around about idle 1.3 volts all the way up to the point where we've switched into our open loop operation around about 2.4 volts, so just a repeat of our first pieces of data. |
| 53:39 | And we can see straight away from our yellow trace here which is, remember, our error, we've still got a little bit of error, particularly in the point I've just noted there, but we can see that we are, for the most part, really close to zero. |
| 53:52 | Again we'll jump across to our histogram to get a better idea of that. |
| 53:55 | So important of course to make sure that we've switched back to our total trim parameter here. |
| 54:01 | And we can see that for the most part, we are pretty good. |
| 54:04 | We've got a couple of idiosyncrasies here, we've got a 2.3% trim here and we've got this little area here, 3-5% which I am again putting down to the fact that this is just a transition area. |
| 54:18 | You note that this happened, this 5% trim happened in the 1.6-1.7 volt vicinity. |
| 54:26 | However once we got stabilised up at that exact same voltage, you see that we are back down to about 0.5%, 1% trim so I've got absolutely no problem with that, more than happy with that. |
| 54:37 | Now again we're not trying for perfection here and in all reality, 2% here which is the main amount of trim that we're seeing is absolutely fine. |
| 54:45 | We could go another iteration and see if we can get that a little bit better. |
| 54:49 | For now, let's open up our second piece of data and look at our higher RPM range. |
| 54:54 | Alright straight away, looking at our higher RPM range and our yellow parameter here for our total trim we can see that it's sitting really really nicely, really close to our 0%. |
| 55:04 | We've got a small increase here with about a 2% trim just before we switch into open loop mode which we've done here and again as soon as that happens we go negative so we're going to ignore that data. |
| 55:15 | We'll jump across to our histogram and have a look at all of that and again everything is really looking quite nice there, right up to about 3.1 volts. |
| 55:24 | So in reality, we could call this done, we're pretty damn close to our target. |
| 55:29 | If you want to be fussy, you could go one iteration again and just make a very small adjustment to get that a little bit closer but you can see just in one set of changes there we've got rid of the error, albeit we were pretty close to the mark straight off the bat. |
| 55:43 | And of course that's totally what we'd be expecting here given that we have got a standard mass airflow sensor. |
| 55:48 | Let's load up our last piece of data from our ramp run and we'll have a look at that. |
| 55:53 | Alright with our ramp run loaded up we are now going to be looking at our air/fuel ratio error calculated parameter. |
| 56:00 | So we've switched that out from our total trim, remembering we are operating here in open loop, we can again look at our bottom set of parameters here, we're looking at our short term and long term fuel trims, making sure those are both set to zero before we go any further. |
| 56:14 | Now we can see at the very beginning of the run here where we're settled, we're pretty close to our target, we're within about 1-2% there. |
| 56:21 | However we can see we've still got this lean ugly spot through the mid range. |
| 56:25 | Don't worry, we're going to get into that in just a second. |
| 56:27 | However by the time we get up to, in this case about 4.5 volts from our mass airflow sensor. |
| 56:32 | We can see everything's nice and stable and right through to the end of that run we're pretty much right on target right on our desired air/fuel ratio of 11.6. |
| 56:42 | We'll jump across to our histogram and just have a better look at that. |
| 56:46 | Remembering before we can really make any use of the histogram we need to change that total trim value there to our air/fuel ratio error Subaru trim and again it's all looking pretty good. |
| 56:57 | Obviously we've got this ugly spot here, we've got these positive trims in that transition area but for the most part, particularly when we're higher in the rev range once we're up on full boost we can see that we are pretty much right on zero. |
| 57:11 | And again this shows us, we're sitting at 4.6 volts so we're right on the limit of that mass airflow sensor but we're within the realms of what we can achieve. |
| 57:18 | We'll also have a quick look over at our dyno screen and we can see that ramp run overlaid again with our first reference run sitting at 317 horsepower. |
| 57:26 | Still not really looking for more power, more interested in the air/fuel ratio. |
| 57:30 | We can see that we've got this lean area that coincides with our logging. |
| 57:33 | We've got our reference line in here that we're sitting pretty much right on. |
| 57:37 | We can see that we've leaned the air/fuel ratio our significantly from our 10.5, 10.3:1 factory calibration. |
| 57:44 | And that has given us this nice improvement in our power so we're sitting up around 237 kW so without really touching the calibration other than leaning out our air/fuel ratio we've actually already picked up a good chunk of power there. |
| 57:57 | So now we need to look at how we can fix that ugly lean spot through the mid range and we're going to be doing this in our map comp table. |
| 58:05 | Now if you're lucky enough that you've got an earlier Subaru that doesn't exhibit these fuel system issues to this degree, then at this point your job is probably done but unfortunately for us we've got a bit more work ahead of us. |
| 58:17 | So let's look at how we can do that. |
| 58:19 | We'll start by having a look at the data we gathered under steady state conditions and if we look at it graphically here in the MegaLogViewer HD software it doesn't really make a lot of sense, what we really need to do is jump across to our histogram to get a more useful view of that data. |
| 58:34 | And what I've done is set up this histogram to replicate the table that we're interested in, that load compensation table from EcuFlash. |
| 58:41 | We'll jump across quickly and look at that. |
| 58:43 | What we want to do is replicate the load axes as well as the break points and we unfortunately just need to manually copy and paste these in or insert these as required. |
| 58:53 | We'll head back to MegaLogViewer HD. |
| 58:55 | We start here by selecting the axes that we want, so again just replicating how that table is set up, we've got our manifold pressure relative, we've got RPM and in closed loop mode of course we are logging in our total trims there. |
| 59:10 | We've also got the precision set up so it matches again the break points in our table. |
| 59:16 | And in order to set those break points we want to start by setting the number of rows and columns. |
| 59:23 | This is as simple as counting the number of rows and columns in the EcuFlash table and replicating that here. |
| 59:28 | Once we've finished we can click resize. |
| 59:31 | Once we've done that we can click on this little icon here in the corner and that will allow us to then go through and manually enter each of these values so we're just copying these straight from that table. |
| 59:44 | Once we've done that we can then click on the apply X and Y axes values and we're done. |
| 59:49 | Anyway we've already got that set up so it matches, we've got our data so let's see what we can learn from here. |
| 59:55 | First of all we can see that down at 1800 RPM and to a lesser degree 2000 RPM, we can see that our air/fuel ratio's pretty much on point, we see very small amounts of trim. |
| 01:00:06 | It's exactly what we'd expect considering we've gone through and rescaled our mass airflow sensor. |
| 01:00:11 | Likewise we've got a similar affect happening up here at 3800 RPM which again is what we'd expect. |
| 01:00:18 | We've got a little bit of trim going on in there but not too much. |
| 01:00:21 | What we can see though is there's some areas where there's a lot of work to be done in this table. |
| 01:00:26 | In particular through the middle of this table, in particular 3200, 3400 RPM we've got some quite large trims up around 17, 19%. |
| 01:00:37 | We do want to take note of the number of hits we've had in each of these sites though before we really take too much notice of the values. |
| 01:00:45 | These are coloured based on their hit weight which basically means how many sample have we got for a particular bin in our histogram? We can see that we can choose our table colouring here. |
| 01:00:56 | And we've got that set on our hit weight so basically the green cells, we've got the most hits, as we go yellow, less data for that particular bin and then finally our white data here means we haven't really got much solid data. |
| 01:01:12 | So for this reason we want to focus our energies or trust the data I should say, more accurately in the cells that have the most hits. |
| 01:01:21 | For example these two cells here, eight and 10% trim, I'm not 100% sure that that's actually accurate, although we do see a reasonably consistent trend as we go through here. |
| 01:01:33 | That's OK though because remember we have gathered data under our wide open throttle ramp run conditions so that'll help us fill in those areas. |
| 01:01:39 | So the process here is to manually take some of what we've learned from this table and apply it across to our map compensation tables. |
| 01:01:47 | Now we don't want to be absolutely specific here, I'm looking for trends and I'm going to be trying to copy those trends across. |
| 01:01:55 | For example, if we look from 1.93 I've already mentioned that it might not be a trustworthy cell, the trend does look pretty consistent. |
| 01:02:03 | If we go all the way through to -9.67, we've got a variation from around 10% up to a peak of 18,19% and then dropping away again back to 13%. |
| 01:02:16 | So let's see how we can manually apply that so we know that we've got our maximum trims there around about -7 psi, 6.7 psi, -5.8 psi. |
| 01:02:28 | So let's jump across to our table and we'll look at that. |
| 01:02:32 | So we know that we are at 3200 RPM and we were -6.7 to -5.8 psi. |
| 01:02:42 | So at the moment the numbers in there are 6% so it's a 6% positive trim. |
| 01:02:46 | We needed to add 19% to that. |
| 01:02:49 | Now I'm going to go a little bit less aggressive on this, again it's an iterative process so for the moment I might make half of that change and we can come back and get some more data. |
| 01:03:00 | So what we want to do of course is add to the values in here. |
| 01:03:03 | So if we take our average at 18% across those two cells and we halve that, so we want to add 9% to that, of course six plus nine, we want to make that 15, we'll enter a value of 15 in there. |
| 01:03:16 | So on either side of this we saw our trims were a little bit less aggressive there so let's just make a change here, we'll go back to our data and we were trimming around about 13 to 15% so let's average that and say we're going to add 7% to these sites here. |
| 01:03:36 | So what we can do here, if we go to our edit menu, we can click on add to data and we're going to add seven to those particular sites. |
| 01:03:45 | We'll head back to our log file again and above the area that we've first changed, we see 17% down to 10% so again I'll make an average of those changes there, those pieces of data and let's go with somewhere in the region of 7% again. |
| 01:04:04 | So we'll highlight those cells there and again we can use our add to data and 7%. |
| 01:04:11 | So what I'm going to do now, we'll go through and speed up this process a little bit, but we'll go and make some wholesale changes to this entire table. |
| 01:04:19 | So let's get that done. |
| 01:04:30 | Alright we've got our first iteration of changes now made to our load compensation table under steady state conditions, we're going to head back to MegaLogViewer HD and we're going to load up our wide open throttle ramp run and have a look at what needs to be done under those conditions. |
| 01:04:44 | For this particular set of data we're going to be mainly interested in our information here from our 3.87 psi column. |
| 01:04:52 | We can look at those changes there that are necessary and we're going to go through and do exactly the same, applying those into our table. |
| 01:05:01 | Again trying to do a little bit of averaging there to smooth our results. |
| 01:05:05 | So we also need to pay attention to the fact that at 3800 RPM we can see we've got a 1.7% trim there. |
| 01:05:13 | Now if we add that into our table here, we'll just go back across to it, we're going to find that at the moment, the 3800 RPM row as well as our 3.87 psi column are filled with zeros. |
| 01:05:28 | It's important to understand that just like the other tables we've talked about during the worked example, as we exceed 3.87 psi or alternatively we exceed 3800 RPM, basically when we fall off the table, any values that are in that column or row will be held into those higher boost or higher RPM regions. |
| 01:05:47 | Now I'm not specifically too worried about maintaining the trims into the higher boost level, so sort of out in this area which we're probably not really going to be getting to. |
| 01:05:57 | This turbo's not really at full boost until sort of 4000 RPM and above so this is only going to a transition area but what I do want to be mindful of is if we have any values in this cell here at 3800, that is going to then be held right through our entire wide open throttle ramp run so we don't want that. |
| 01:06:13 | So we want to make sure that even though the table says that we had a 1.7% trim there, for our first round of modifications there, we're actually going to leave that set to zero so let's go through and we'll add those values in. |
| 01:06:29 | Alright so we've got our first round of modifications made to our load compensation table. |
| 01:06:34 | Before we flash that into the ECU we need to copy and paste that and make sure that we put that into our other load comp table as well so that they're both the same. |
| 01:06:44 | So now the process will be the same, we're going to flash that back into the ECU. |
| 01:06:48 | We'll gather some more data and then we'll have a look at how those changes have panned out and how much more work is required. |
| 01:06:55 | So let's jump ahead and see what that looks like. |
| 01:06:57 | Looking at our data now we can see straight away we're a lot closer to the ballpark we haven't got any of those big trims in the 16-19% vicinity. |
| 01:07:07 | Most of the trims in here are in the range of 0-3% which I'm quite happy with. |
| 01:07:12 | We do have a few areas that do still require some work though, specifically if we look at 2400 RPM here, this particular column we've got errors in the range of 3-6% so we'll touch those up as well as this particular area here, 3000 and 3200 RPM. |
| 01:07:30 | We've got trims in the sort of 5-8% vicinity. |
| 01:07:33 | But this is an iterative process just like our MAF scaling so we're not going to get this right in one go. |
| 01:07:40 | You can now repeat the process as many times as you feel necessary until you've got your load compensation tables dialled in and looking for those errors in the range as I've said, 0-3% I'd be pretty happy with that. |
| 01:07:53 | So at this point you should have your MAF scaling dialled in, you should also have your load compensation tables dialled in and we're ready to move on with the next step of the process. |
