Practical Reflash Tuning: Step 4: MAF Calibration
Watch This Course
$229 USD
-OR-
Or
8 easy payments of only $28.63
Instant access. Easy checkout. No fees.
Learn more
Step 4: MAF Calibration
46.25
| 00:00 | - The next step of our process is where we would traditionally handle either our MAF rescaling, our injector rescaling or potentially both of those together. |
| 00:08 | With our particular example here, we are still running the stock injectors so obviously there's no work to do there as far as our injector scaling goes or our injector characterisation, that's going to be stock. |
| 00:20 | Likewise, understandably we have removed the mass airflow sensor so there is no MAF scaling to be done. |
| 00:25 | We do however now have our volumetric efficiency table and this is what we're going to be working on in this step. |
| 00:31 | Let's just jump back into our software and we'll refresh our memory here, we can see that the main VE tables that we now have access to are filled with zeros so as this sits right now, if we flash this into the ECU, the engine isn't going to run. |
| 00:46 | So we need some way of adding numbers into our VE table and there's a few ways of going about this. |
| 00:53 | Once you're a little bit experienced and you've done a fair few tunes you're going to end up building a library of previous tunes where you're probably going to have a VE table that's going to be at least close enough to get you up and running and that'll be your starting point that you would use. |
| 01:10 | This is another reason why it is so valuable to add a little bit of detail in your calibration name so that you know exactly what the specifications for that tune were so you can choose something that's at least going to be broadly compatible. |
| 01:23 | Doesn't need to be perfect, we're still going to be optimising the VE table but should be enough to get you up and running. |
| 01:29 | Of course if you are brand new to tuning and this is your absolute first tune, you're not going to have that library to draw from so I'm going to show you how you can deal with that. |
| 01:39 | Conversely as well we will actually add the calibration we're using here as a upload to this worked example so that you can use this as a comparison if you're doing something similar. |
| 01:51 | Let's start with our basics though. |
| 01:54 | What we're going to do is dive back into our software and we'll close down the current calibration here and what we're going to do is open our existing stage 1 file. |
| 02:02 | So this was the calibration that we started with, VE table or virtual volumetric efficiency here was optimised for the stock or the modifications to the engine using the mass airflow sensor. |
| 02:14 | So what we can do here is actually grab this VE table and use it as our starting point. |
| 02:19 | So we'll come up to our edit menu and we're going to come down here to our virtual VE option and this gives us a graphical and numerical representation of the VE table based on the coefficients and the quadratic equation you use to calculate them. |
| 02:34 | So we've got the numbers there numerically, we can also view this as relatively conventional VE table, you'll notice that it is a little bit blocky which is part of the calculation, the quadratic equation that is used, we can't get complete control over every single individual cell which is why it looks the way it does. |
| 02:54 | Come back to it numerically though and we can use this and essentially paste it into our VE table in the calibration we were working from but there is one problem. |
| 03:04 | With the virtual VE table here in our editor we can see that we've actually got manifold absolute pressure on the vertical axis and we've got RPM on the horizontal axis. |
| 03:12 | Without going back to our speed density patch, these 2 axes are actually transposed. |
| 03:20 | So we have the around the other way which makes it difficult because we can't directly paste this information straight into that table. |
| 03:26 | Other issue that we've got here is our axes so we can see here for our manifold absolute pressure, we're going up to 105 kPa, we're going out to 6400 RPM. |
| 03:37 | If we go back to our speed density patch, we're going to find that by default the break points for our manifold absolute pressure go out to 175 kPa, it is a 2 bar speed density patch so it is set up for boost. |
| 03:52 | There's a couple of ways of dealing with this, let's start though at the very start and we'll see how we can get these axes around the other way. |
| 03:59 | So what we're going to do is copy our entire table here but we're going to use the copy with axes option. |
| 04:06 | That's important because it is also going to copy in our axis brake points. |
| 04:10 | Now to do this we are going to also need to use Excel so I've got that opened up down here and what we're going to do is just paste that in and we can do that with our control V option so actually we can see that's exactly what that's done there. |
| 04:25 | What we're going to do now, we're going to make sure that we are only highlighting here the actual data so we'll come out here, we don't need RPM and we don't need kPa. |
| 04:38 | So I'll make sure that we've highlighted all of that data, again I'll control C to copy that to the clipboard or you can right click and press control C, or right click I should say and choose the copy option. |
| 04:50 | We'll choose another cell down here and if this time we right click and we come up to paste special, click on that, we've got a range of options here but the one we actually want is to use this transpose option. |
| 05:02 | So we'll click on OK and low and behold that's exactly what that's done, it has just transposed the axes so we now have data that we can paste in. |
| 05:11 | We do need to be a little bit mindful though again just due to the axes that we're using and there's a couple of ways that we can deal with this so just again before we jump back into our calibration, we can see that this particular table here, 10 kPa out to, if we come out to the right hand side, maximum value here, 105 kPa which makes a lot of sense for a naturally aspirated engine, we wouldn't expect to be over 105 kPa. |
| 05:36 | Let's head back to our editor now and what we'll do is we'll close down this file and we'll open up our speed density patch again which is what we're working from. |
| 05:45 | Now if we click on one of our VE tables, we can see that our axes actually span for manifold absolute pressure from 15 kPa out to 175 kPa. |
| 05:55 | Now how you're going to deal with this is really personal choice. |
| 05:59 | And you can choose to either reduce the range for that axis which may make sense of course with a naturally aspirated engine, we're not going to be out in positive boost. |
| 06:09 | Or you can leave it as is which still provides absolutely plenty of resolution. |
| 06:14 | I mean if we look here in the areas that we're actually going to be running in. |
| 06:18 | We've still got resolution every 5 kPa which is absolutely plenty. |
| 06:23 | And we also are now spanning out to 6800 RPM which is 400 RPM further than our base file. |
| 06:31 | So how are you going to go about this? If you are going to copy and paste your existing file, well it will make sense to modify your break points here and what we can do here is if we right click and we choose column axes, and we choose edit, you can actually edit the break points so to closely match the ones that we're using in the sample virtual volumetric efficiency file, we'd simply choose 10 kPa out to 105 kPa, highlight the entire lot and then we would use the interpolate function there and that's going to essentially do exactly what we'd expect there. |
| 07:07 | Let's just bring that back up to the front and now we can copy and paste that information from our Excel spreadsheet in, we'll go ahead and do that now. |
| 07:17 | Now it's not going to work perfectly. |
| 07:20 | We'll actually see that the number of break points is different so this is going to present a little bit of a problem which I'll just show you now once we've highlighted all of this data. |
| 07:29 | Control C on that data, we'll come back across to our editor and control V, we'll paste that in. |
| 07:37 | Alright so what we can see first of all, we've got no data at all for 6600 and 6800 which is completely to be expected. |
| 07:43 | No problem though, we can simply copy our 6400 RPM row and extrapolate that up using control C, control V. |
| 07:53 | These numbers aren't going to be perfect, it's just a starting point so we don't need to be too stressed here. |
| 07:59 | Likewise our break points here for our manifold absolute pressure, they don't perfectly match the table data that we've just copied in but again we're expecting here some pretty wholesale changes so the chances of this being perfect are pretty unlikely. |
| 08:14 | It's going to be close enough to get us up and running so that's how you can get yourself a base volumetric efficiency table if you've got 0 data to work with. |
| 08:23 | We could also choose to make some broad changes to this table based on our expectations. |
| 08:27 | In particular with our large cam profile, we'd expect that down here, probably up to about 2500, 3000 RPM, low RPM we're expecting the VE to be reduced, the efficiency of the cam won't be as good as stock so we could make a 10 or 15% reduction in VE down in that region. |
| 08:45 | Likewise what we're hoping with our bigger cam is that probably from about 4500, 5000 RPM and above, our VE should be massively improved because that's where the benefit from the cam's coming from so again we could potentially choose to make a 10 or 15% increase in there, maybe some just smoothing to get our table at least some semblance of order before we start optimising, again we don't need to stress too much here, it's not that critical. |
| 09:12 | The table will be wrong, that's the process we're going to go through within this module. |
| 09:17 | So that's one option there if we've got nothing to work with. |
| 09:19 | What I'm going to do here is just go back to our column axes and we'll edit those back to what they were because the sample file that I am going to be using here does use the standard spanning from 15 kPa out to 175 so we'll just get ourselves set up again on that spanning. |
| 09:41 | Alright so we've done that, what I'm going to do now is load up a compare file and again for your own benefit you'll be able to use the base file that we're creating here as your compare file and we've got this file here, GM L98 HS108D cam so we're going to load that up. |
| 10:00 | So once we've done that, we've now got the ability to use these options here to view our base data which of course, when you're starting out, would look exactly like this. |
| 10:10 | We've then got our compare file which is the one we're going to be using. |
| 10:14 | We can view that graphically as well. |
| 10:16 | And what we're going to do is simply click in the top left corner, control C to copy all of the data and then control V, that'll paste it into our VE table. |
| 10:27 | Alright so once that's done though we can also go ahead and copy that data into our other 2 tables, so we've got 3 VE tables here just a matter of good housekeeping here, we want to make sure that all 3 of those tables are always the same. |
| 10:43 | Once we've done that we can go to our compare, we can close our compare file and it's a good time to also save our file again. |
| 10:51 | So we've made our changes that now should be enough to get us up and running. |
| 10:55 | We do have one last task left though before we can start the engine which is to flash that calibration back into the engine control module. |
| 11:03 | Remembering here that our very first flash does require a write entire function so we're going to go ahead and do that now. |
| 11:11 | In order to do that, what we're going to do is come up to our vehicle writer icon and we'll click on that. |
| 11:17 | Now here we've got the option to write to our engine control module which is this option up here. |
| 11:24 | Because we've got a 6 speed automatic transmission, we've also got the ability to write to the transmission control module, we don't need to do that here so we're going to save ourselves a little bit of time and select do not write. |
| 11:34 | If we come up to our engine control module, we want to select the write entire option. |
| 11:39 | We can now key on and we can click on the write option. |
| 11:43 | This is going to take around about 5 or so minutes so we're just going to speed up that process now. |
| 12:00 | Once our write's completed we can close down our write box and we're going to be ready to start our engine for the first time. |
| 12:06 | Before we try and start the engine for the first time, we will just come over to our scanner and we can just click on the start scanning icon on the top toolbar so we can start gathering some data. |
| 12:16 | Now we do need to understand here that when we first try and start the engine there's a good chance we may need to use a little bit of throttle to keep the engine idling. |
| 12:23 | There's a pretty good chance that our calibration won't be dialled in perfectly for idle right at . |
| 12:29 | the start If that is the case, don't stress though, we will cover the process of dialling in our idle as we go through the worked example. |
| 12:36 | So let's see if we can get our engine up and running. |
| 12:41 | Right just used a little bit of throttle there to actually get the engine to start but as I back off the throttle, I'm completely off the throttle pedal at the moment and we can see that our idle seems pretty good. |
| 12:52 | We are a little bit high at the moment with our speed sitting at around about 1000 RPM but given that so far we haven't actually let the engine warm up, we are only at about 55°, given the condition of our idle tuning a well, I'm actually pretty happy with that. |
| 13:09 | What we can see if we look at our measured lambda vs our commanded lambda, we are sitting pretty lean, that yellow data point there's coming in to our scanner via a wideband that is wired up to our CAN bus. |
| 13:23 | So we can see we are a little bit lean here but at the moment that's not too much of a concern, we'll just bring the RPM up and we can see that our lambda actually does settle back down so at the moment we're probably a little bit leaner than we'd like to be at idle but everything else is looking OK. |
| 13:39 | So this gets us into a position where we can actually start doing some tuning. |
| 13:43 | Now while the engine is warming up, what we'll do is just have a look at how we're going to go about this. |
| 13:48 | So we'll just slide our chart logger out of the way here and we'll have a quick look at one of the histograms that I've already created and I'll talk you through what's involved in this so you can create it for yourself if it doesn't exist. |
| 14:00 | So the one I'm going to be using for this is our equivalence error SD. |
| 14:05 | Let's right click here and we'll go to our graphs layout. |
| 14:08 | This allows you to add, remove or modify any of our existing histograms and we'll come down to our EQ error SD histogram. |
| 14:16 | So creating one of these from scratch we can simply use the add chart little icon here and that'll give us basically a histogram that's completely blank. |
| 14:25 | The first option here is to add a label for our particular chart or histogram, basically we can label this anything we want, it doesn't really matter, just as long as it's something that's easy to reference at a later point. |
| 14:36 | Obviously we've called it EQ error SD. |
| 14:39 | Next is our parameter so this is what's going to actually be logged into our histogram. |
| 14:44 | So we want to use here the math channel called EQ ratio error and this is just looking at the difference as a percentage between our commanded lambda and our measured lambda so basically how much error we've got. |
| 14:57 | So if we click on this, because by default there will be nothing here, we can then just type in what we're looking for so in this case, EQ ratio error and that will show right down the bottom we've got this math option here which is EQ ratio error. |
| 15:12 | So we'll click on that and that will give us that parameter. |
| 15:15 | We can choose the number of decimal places to be displayed here as well and generally one decimal place is sufficient precision for that. |
| 15:24 | Coming down further we also have the ability to add some shading. |
| 15:28 | Now this isn't entirely essential but what it does is it gives us a pretty good visual glimpse as to whether or not we're lean or rich and we'll just pull this out of the way here, as we've been talking here you can see we've filled this table up with a lot of red data which means that we are leaner than our target. |
| 15:43 | Obviously particularly at high load that can be potentially dangerous. |
| 15:46 | Then we've got some green values here where we were richer than our target so just gives a good visual cue and I've set those to plus or minus 10% but basically within these here you can choose whatever suits your own personal preferences, there's no right or wrong option there. |
| 16:03 | The key points here are our column axes and our row axes so what we want to do here is make sure that these are the same as our VE table so let's head back to our editor here, we'll go back to our airflow and we'll go back to our VE table so we can see the column axes here will be manifold absolute pressure, of course the horizontal axis there, sorry the vertical axis will be our RPM. |
| 16:27 | So we can also go one step further here and if we right click here and we go column axes and we click on copy labels, what that's going to do is copy all of the break point data so this makes it really easy, we'll come back to our scanner. |
| 16:41 | So we've chosen manifold absolute pressure as our column axis and we can see here this would be default be blank so we'll just get rid of that and we will paste in those pieces of data. |
| 16:54 | So our break points now match. |
| 16:55 | I'll do exactly the same for our row axis which is engine speed and this way the break points when we close this down in this particular table here, these will match what we're seeing in our VE table, meaning that the errors that we're gathering in this table, we can then copy and paste those errors directly into our VE table. |
| 17:13 | So that's the process we're going to go through here and this histogram is going to help us with our calibration. |
| 17:19 | Now that we've got our engine up to temperature we can go ahead and actually start gathering some data so the idea is we want to gather this data in 2 ways. |
| 17:27 | We've obviously got our closed-loop operation, even though we're not actually running with our short term fuel trims enabled at the moment but under closed loop we're obviously targeting lambda 1, 14.7:1, then we've got our power enrichment which is when we go over our throttle transition point where we're going to be targeting our 0.87 target. |
| 17:46 | So what we want to do is to be a little bit mindful because as we transition from closed loop to power enrichment, there understandably will be a bit of a step in our air/fuel ratio where there's inevitably going to be some error. |
| 17:58 | Our measured air/fuel ratio is never going to track perfectly onto our target in an instant like that so we're always going to see generally a lean error so what I want to do is try and gather some data under closed loop target and then I'm going to gather some more data under power enrichment. |
| 18:14 | We're going to do this in steady state for our closed loop and then doing ramp runs for our power enrichment. |
| 18:19 | There's a bit of interpretation with this data as well but you'll see that as we go along. |
| 18:22 | So what we're going to do here is use our dyno and we're going to get ourselves up and running first of all, we'll get our fan going so our engine isn't going to overheat and we're going to use manual mode here on our 6 speed automatic transmission and we're going to be using 4th gear. |
| 18:37 | Alright so I'll just get us up and running here and what I'm going to do is actually just disable the scanner for a moment, turn that off and I'm going to then start scanning again. |
| 18:48 | The reason for this is we've gathered a lot of garbage data there while we've been idling and that's going to affect the accuracy of our actual data. |
| 18:56 | And it is really important as we learned in the body of the course that when we are gathering data like this we want to be mindful of the way we drive the car. |
| 19:04 | Very smooth on our throttle inputs and getting as much data from as many cells as we can. |
| 19:08 | Now at the moment we're sitting here at 1200 RPM and we'll see how much further we can get down in the RPM. |
| 19:16 | The automatic transmission doesn't give you as much control as a manual but we will be doing some extrapolation of the data we gather. |
| 19:22 | We can see that we're sitting with basically a 11% error where we are 11 leaner than our target and we can see that if we look down the bottom, commanded lambda, 1.0, measured lambda, 1.1, 1.11, moving around obviously. |
| 19:37 | So we do need to be mindful of this because when we're first getting set up with our calibration, if we are quite lean, particularly under high load, then clearly we need to be a little bit sensible here and it's not going to be a very good idea to stay under wide open throttle or high load with a very lean mixture. |
| 19:52 | That's OK, we can simply take what we've learned, we can back off, go and make some adjustments to our VE table and then come back to this. |
| 19:59 | It is an iterative process and there is really no one size fits all answer for you here. |
| 20:04 | How we do this step is going to simply depend on how far out our mixtures are and it may be that we start the engine up, it's so far out of the ballpark that we can't even get it running and we're going to need to straight away make some changes to the VE table before we can get to where we are now, that's absolutely fine though, there's no need to stress about it, look at the data, understand what it's telling you and make the appropriate changes. |
| 20:26 | Alright so what we're going to do as I've mentioned, we'll try and get a little bit lower in our RPM here just using our dyno and under steady state conditions, I'm just going to see how far down I can get into our manifold pressure. |
| 20:38 | There will be a point where, as I slow down, basically we're not making enough torque to keep the dyno running and it looks like probably I might be able to get down to 55 kPa, maybe not. |
| 20:48 | No we can't really get down there so 60 kPa at this point is about as low as we can get. |
| 20:53 | So I'm just going to continue gathering some data here and we'll use the dyno to keep our RPM as low as we can and what I'm going to do is come up to about 80, 85 kPa, just want to make sure that we're not actually transitioning into power enrichment. |
| 21:08 | Being very smooth with my changes here. |
| 21:11 | Again just being mindful of how lean are we, am I happy at this point with the load that I'm applying so we're still about 10% lean here. |
| 21:21 | So we'll see if we can come up to about 80 kPa. |
| 21:25 | You can see that because of the automatic transmission as I apply more load on the dyno here, we do tend to see the RPM flare a little bit so that is a problem with the automatic transmission, we're not going to be able to accurately access every individual cell. |
| 21:37 | Now you can see now that we're up to 80, 85 kPa, I'll try and put a little bit more throttle on. |
| 21:42 | Our error is significantly lower, we're down to 2 and 3% so I'm not really too worried about this. |
| 21:46 | We can just get up to 85 kPa here and at this point our error is actually pretty close to 0. |
| 21:53 | So what we're going to do now is try and fill out as much of this table as we can, it's going to be a case of manipulating our throttle position and then also manipulating our dyno set point. |
| 22:03 | This is an easier process if we are dealing with a manual transmission compared to an automatic transmission but generally what I'm trying to do here is just manipulate my foot position on the throttle and my dyno set point just to get through as many of these cells as we can so we'll just bring our RPM back up here. |
| 22:27 | Everything's looking pretty good there, might actually be able to get up to 90 kPa, looks like we got some data there. |
| 22:34 | Probably if I go much further than that given our atmospheric conditions here as well, we will end up in the power enrichment area so I don't want to go that far. |
| 22:42 | So I'll just bring up our RPM a little bit further and now I'm just going to back off the throttle and we'll see how far we can get down here. |
| 22:49 | So we're down at 70 kPa, 1600 RPM. |
| 22:52 | And what I'll do now 'cause we can see we've got no data in the 1400 RPM, 70 kPa cell so I'm just going to slow the dyno down a little bit here, see if we can fill that cell in. |
| 23:02 | So it is a case of working the dyno and the engine to get us the data that we want. |
| 23:09 | So let's continue now and we'll fill in as much of this data as we can under steady state conditions. |
| 23:24 | Each time we go into a cell, when we are in a position a little bit higher in the RPM, we can sit quite nicely in the middle of a cell and as soon as we move into a new cell we can start seeing the number in that cell will move around and ideally what we want to do is stay in that cell until that number has stabilised because when we move into a cell to start with, we won't be perfectly stable in that cell and we may see the air/fuel ratio move around a little bit so we want to stay in the cell as long as we can until we've essentially achieved equilibrium there and the number isn't moving. |
| 23:55 | That's going to give us a pretty solid idea of what the actual error is in that cell and also the longer we're in a cell, the more hits we get so the more we can rely on that data. |
| 24:03 | We'll explain a little bit about that as we go so let's continue now. |
| 24:07 | Yoy can also see that as the RPM increase, it does become a little bit easier for us to stay or control exactly where abouts we are in the individual cells compared to when we were trying to run the engine down at 1000, 1200 RPM. |
| 24:21 | Again, just an aspect of the cam coupled with the automatic transmission. |
| 25:15 | Alright we're going to stop there and before I back off the throttle I'm just going to press the stop scanning button and that's just going to ensure that we're not going to end up scanning further data while the engine is slowing down because this can introduce some problems with deceleration fuel cut off affecting the accuracy of our data. |
| 25:32 | Now one thing that I should have mentioned there before we go started, it is important just to keep an eye on both your engine coolant temperature and intake air temperature so you can see there at 3800 RPM which is as high as we got in the rev range, we're getting up to 103°C. |
| 25:47 | Now not necessarily concerning but definitely something that we do want to keep an eye on and watch. |
| 25:54 | If our engine coolant temperature gets too high, obviously we can end up risking damage. |
| 25:58 | It's also going to potentially affect the accuracy of our results. |
| 26:02 | So let's shut our engine down now and we'll have a look at what we've got and see what we can do with it. |
| 26:07 | Now obviously the first thing we can see is we haven't been able to fill out the entire table, particularly down here in this region, we simply haven't been able to get down into that area. |
| 26:19 | We've purposefully stayed away from the 95 kPa and above zones as I've mentioned, we'll deal with those when we get into our wide open throttle part of our calibration. |
| 26:30 | We've also purposefully cut everything off here at the moment above 3800 RPM. |
| 26:34 | And we do need to be a bit mindful of how far we're actually going to go here. |
| 26:38 | My general rule of thumb is around about 2/3 of our engine rev limit. |
| 26:42 | So 7000 RPM, I might like to come out a little bit further than this, maybe 4000, to maybe 4500 RPM but we do need to be realistic, we're not going to end up running our engine under steady state conditions at 6000, 6500 RPM and 55, 60 kPa, we're only going to be transitioning through those areas so the accuracy or the required accuracy there isn't quite as critical. |
| 27:05 | So we've got some good data though in the meat of the area that we're actually going to be running in so the areas that we're going to be cruising in in particular. |
| 27:13 | And we can see that while down at lower RPM we are looking a little bit lean here, we can see that once we actually got up above about 2000, 2200 RPM, everything is actually pretty good, we've got a few exceptions but other than the area, I'll just circle it here, around about 2600 RPM, 2400 to 2800 RPM I should say, we're actually pretty good. |
| 27:37 | Generally my rule of thumb here, I'm looking for somewhere about plus or minus 2-3%. |
| 27:41 | Obviously if we had zeros in here this would be great but realistically we're not going to be able to always get zeros in here so rather than chasing your tail trying to achieve the impossible, it is important to be realistic about your data. |
| 27:56 | OK the next question is how are we going to actually use this data to help us? And there are a couple of options here. |
| 28:02 | What we can do is highlight all of the data there and we can use control C to copy that data so these are percentage changes that are necessary and we can come back over to our VE table and we can then right click and we can use our paste special. |
| 28:23 | And what we can do here is use the multiply by percent and what that's going to do is make the percentage change from our histogram directly into our VE table. |
| 28:32 | I'll just do that for a moment and we'll have a look at our results. |
| 28:35 | So we can see that that has made those changes that we've just looked at. |
| 28:40 | Now that can be a good way of going about it, particularly if you're a long way out of the ballpark but there are a few caveats that we need to understand with using this technique so I'll just reverse that for a moment. |
| 28:51 | So for a start what we can see here is we're only going to be making changes in the area that we've actually got data. |
| 28:57 | So for the sake of completeness and making sure that everything's going to operate nicely, we would need to extrapolate our results out into these areas where we really haven't got any data. |
| 29:08 | Can also be a good idea if we're going to do this in 2 goes, and flash in between, can also be a good idea to extrapolate what we're seeing out into the higher load zones because if we're lean at let's say 2600 RPM 90 kPa, chances are we may also be lean at 95, 100 and 105 kPa so it'll just get us probably closer to the mark. |
| 29:30 | So the other problem with this is if we do have garbage data in our set here, it's going to just simply blindly apply that to our VE table and that may not give us the desired result. |
| 29:42 | While on the dyno we can generally get pretty good data, particularly if you're doing this out on the street or on a racetrack, you're going to be affected by your throttle inputs, transient response etc so I always start by having a quick look and seeing if there's any clear outliers and in this case we've actually got pretty good data, everything makes a little bit of sense. |
| 30:02 | By an outlier I'm meaning anything that has a big jump from the surrounding cells. |
| 30:06 | And actually in this instance here our 2200 RPM row of data is actually a little bit unique because we can see that either side of that we've got quite large positive errors whereas the 2200 RPM row is actually pretty good. |
| 30:22 | That may be the case but if we've got outliers, it's a good idea to go back and review that data. |
| 30:27 | The other way we can also get a sense of how good our data's likely to be is by using our little controls up here. |
| 30:34 | So at the moment we're looking at, we've got the little A or average values box enabled. |
| 30:39 | If we go to C this will give us our count or how many hits we've had in each of these cells. |
| 30:45 | So the longer we've been in a cell, for example here, 1200 RPM, 60 kPa, we've got 5300 hits so we could be pretty confident that that data's solid. |
| 30:56 | On the other hand we've got a few of these cells there we've sort of had a bit of a flash in the pan, we've got just a couple of hits and that data I would be much more sceptical of relying on so this is just going to give us a sense of what that data actually is. |
| 31:13 | Either way we do this we can apply the corrections that we've just seen there using the paste special function or we can do some hand blending and hand smoothing. |
| 31:22 | Personal preference here and you're going to need to understand both techniques. |
| 31:27 | What I'm going to do here is start by using our paste special function so we'll go control C, we'll go back into our table here, right click, use our paste special, multiply by percentage. |
| 31:38 | Another little trick here is we also have the paste special multiply by percentage half. |
| 31:44 | So when you're really a long way out of the ballpark, I would stay by using the multiply by percentage but once you've done a couple of iterations of this, you're probaby going to be pretty close. |
| 31:53 | You may want to use the multiply by percent half, particularly if you do have some maybe questionable data in your set there, that's going to limit the effect of that. |
| 32:02 | Anyway we've made that first change there but what we're going to do now is do a little bit of hand blending. |
| 32:07 | I'll show you one area of this and then we'll speed it up. |
| 32:10 | So let's head back to our scanner, So let's see here, so basically the first pieces of data we had were around about 800 RPM and we know that at idle as well we were lean so we're seeing 18 to 21% there so what we can do there is go back to our editor. |
| 32:26 | And what I'm going to do is highlight down from that data and I'm also going to highlight to the lower RPM columns and what we're going to do is make a percentage change of 1.2, or a multiplication change of 1.2 which is the same as adding 20%. |
| 32:42 | So we've done that there, we will also just head back to our scanner, we don't really have data in this area here, we can see though, looking at what's happening elsewhere we can see that error does reduce a little bit so we can make a bit of a guess here, we're not really going to be down too often in this high load area though so let's just highlight those cells out here, we'll go out to in this case 90 kPa and let's add 15% so 1.15 and multiply, we will also make that change out here, 1.15 and multiply. |
| 33:21 | So it's just a case of looking at our data, making some educated guesses and extrapolating that data out, we'll do one more row here which is 1000 RPM, let's head back over to our scanner. |
| 33:31 | Look at our 1000 RPM data, we can see that we've got a bit of variability here, at higher load we're up around 11, 12, 15% and down in the lower load areas we were 8 to 10% so I'm going to extrapolate 8% down further. |
| 33:47 | Again we might not be right here but if we can't get onto it in the dyno, onto those cells in the dyno, we're going to possibly struggle to get there in the street as well and we've still got our closed loop trims, once we reenable them to help us pick up the pieces if there are any errors. |
| 34:02 | So what I'll do now is just speed this up, we'll go through and basically extrapolate those changes exactly the same way as you've just seen. |
| 34:23 | Alright so that completes our first round of changes there to the section of the VE table that we have got access to. |
| 34:30 | Of course if you've gone a little bit further, you can just repeat that process out into whatever region you've got to. |
| 34:36 | This does leave us with our high RPM areas though or sorry I should say our high load and our high load high RPM areas that we haven't touched. |
| 34:45 | Now we can address this in one of two ways. |
| 34:47 | We can either flash these results back into our table, back into our calibration and repeat the process iteratively getting our VE table dialled in under steady state before we move on. |
| 34:58 | In this case to speed up the process though I'm going to now do some wide open throttle ramp runs and we'll just see what our data looks like there. |
| 35:05 | It is important here to just be a little bit mindful of monitoring for knock while we are doing this so I'll just bring our chart logger up here, it's a good idea to either listen for knock audibly or at least monitor our knock retard parameter and if we do have knock occurring, want to stop obviously there, back out of the run and we can then remove the timing as required from the areas of the table where knock is occurring and repeat. |
| 35:31 | OK so let's get our engine up and running and we'll get our temperature stabilised and we'll go through that process on the dyno now. |
| 35:38 | Alright we're back up and running on the dyno and what we want to do is just make sure that before we perform any ramp runs that we have our temperatures stable at the normal operating range, we don't want to be suffering from any heat soak because just like our steady state results, this will affect the quality of the data that we're gathering. |
| 35:54 | What I'm going to do is just stop our scanner for a moment here and we're going to go through to wide open throttle. |
| 36:01 | Get to a point where we are ready to start our ramp run and just before I start the ramp run, we will start the logger again. |
| 36:07 | Likewise before I back off the throttle at the top of the run, I'll also stop the logger so I'm only capturing data during the actual ramp portion of the run and this again helps with the quality of our data. |
| 36:19 | Also mindful here I've talked about the knock retard monitoring, making sure that we don't have any knock occurring. |
| 36:24 | Clearly we also want to monitor our air/fuel ratio and if we aren't in the ballpark of where we expect, just back off and you can go and make some changes to your VE table as required. |
| 36:34 | We don't want to run the engine in a potentially dangerous air/fuel ratio and end up doing damage just for the sake of gathering data, this is an iterative process, it'll almost never be right the first time that you go about this. |
| 36:47 | Alright so let's get our first run underway, we'll just gently go through to wide open throttle here and we can start our logger. |
| 36:57 | And we'll now get our ramp run underway. |
| 37:25 | Alright let's have a look at our results on the dyno, we can see we've laid down 445 horsepower which considering our tune still isn't dialled in, it's a pretty good place to get started. |
| 37:35 | We can see that our air/fuel ratio, I've got a reference line in here at 0.87, we're always right up until the very top of the run, always a little bit leaner than our actual target and typically I would have probably recommended here backing off but we've only gone as lean as about 0.92 so not a big concern but we do have some work to do. |
| 37:55 | Let's have a look in our logger and we'll have a look and see what we've got there. |
| 38:00 | So looking at the data we've gathered, we've got some pretty solid data, solely in our 95 kPa column here and we can see that we've got an error around about 2300, 2200, 2400, 5, 6% there, we've also got another area here where our error is up at about 5% as well, around about 3500 RPM. |
| 38:23 | The top of the run actually was really good which is obviously where it's most important. |
| 38:27 | We've got an error of only about 0% there so what we're going to do here is we will use our paste special function, so we'll highlight the entire data, use control C to copy it, come back across to our data here, control V or paste special I should say and multiply by percent so that's only made that change there at 95 kPa. |
| 38:49 | Now what we're going to do is a little bit of manual smoothing here as well. |
| 38:54 | We'll come back down to our data. |
| 38:56 | We can see that we got out to 6600 RPM here and we've got about 1.4% trim so what I'm going to do is just extrapolate that out here, so we'll make 1., let's make it 1.2% there at 6800 RPM and really what I want to do is replicate the sort of error that we saw at 95 kPa up above that so at 100 and above so let's have a quick look, we'll do this manually and I'll just talk you through one area and then we'll continue. |
| 39:26 | Actually before we do that, first of all see we only had data here from about 1400 RPM, got about a 6% trim there so I'm going to extrapolate that data up to our lower RPM range so let's do that now. |
| 39:40 | So we're at 6% there and we will just extrapolate that up so we can do 1.0, actually we've got data there, no we'll do the whole lot there. |
| 39:53 | 1.05 is going to be probably close enough for our intents and purposes there. |
| 39:57 | And we can also make that change again just for the sake of completeness from 100 kPa and above. |
| 40:05 | Right let's head back to our data now, we're out to 1200 RPM there and we can see that we've got 6% and then 3% so let's make those changes, 6%, 3% and then 1%. |
| 40:20 | So we are at 1400 RPM, 1.06. |
| 40:29 | And then for our next couple of cells I think it is, no just 1, we are 3% so we'll highlight that out here. |
| 40:40 | And then we had a couple of rows there with about a 1% error. |
| 40:47 | Alright so this is the process, basically we're extrapolating what we saw at 95 kPa out into our higher regions there just for the sake of completeness. |
| 40:53 | Likewise if we've got big errors down here we can extrapolate those back out to the left as well, remembering again we don't need to be quite so accurate in these regions because we're not going to be spending any time there so I'm going to speed this up and we'll just go through and complete the rest of that process. |
| 41:24 | Alright we've got our process complete there, our first iteration between using the paste special function, a little bit of hand blending, you get a bit of a sense of what that VE table looks like now graphically and as usual we can see there's a little bit of ugliness in this table here so for at least our first iteration while we're getting up and running here what I will do here is just use the smooth function and just one or perhaps 2 clicks of that smooth function will do a little bit of a better job of getting that in some semblance of order. |
| 41:54 | We do need to be a bit mindful about this, a lot of tuners use this to give a nice pretty graph, obviously that's not what we're trying to achieve here. |
| 42:02 | But we are starting from a VE table not suited to this engine so at least while we're getting up and running, our first set of changes is likely to be quite coarse and we may end up with some big steps in there using the smooth function just gets rid of those, for our next iterations we're not going to necessarily need to use that. |
| 42:20 | So we've got our first table complete there and that is our IMRC closed table, remembering that for our purposes here we're going to copy and paste the results from there to our other 2 VE tables and once we've done that, we can then get ourselves back up and running. |
| 42:38 | So to do that, we'll head back across to our scanner, just disconnect from our scanner for the moment and we can use our write vehicle icon. |
| 42:47 | Again we're not going to be writing to our transmission control module, let's write the engine control module now and we'll get back up and running and see the results of our modifications. |
| 42:57 | OK so our write's complete here so we can close down our write window. |
| 43:00 | And we'll get ourselves up and running again, get our fan on and allow everything to reach operating conditions and we'll see what our results look like now. |
| 43:10 | Alright so straight away at idle we can see that our results are much closer to the mark. |
| 43:14 | Albeit we're still probably a little bit heat soaked so I wouldn't pay too much attention to it but straight away with one iteration of changes we can see how much closer we were compared to the 15 to 20% corrections that we were seeing to start with. |
| 43:27 | That's the power of using the histogram and this iterative approach. |
| 43:31 | We're not going to get it perfect in one iteration of changes but we're going to be a lot closer to the mark. |
| 43:36 | So what we'll do now is we'll just get ourselves up and running again in steady state conditions and we'll have a look and see what our results actually look like. |
| 43:45 | Alright so we're not going to go through the entire table here, we'll just have a bit of a snapshot at a few of these zones. |
| 43:51 | So we're at 1400 RPM here just slowly increasing that throttle opening and we can see that where we were needing a positive trim we were leaner than our target, we're now a little bit richer than our target which isn't a bad place to be. |
| 44:03 | Again we're not going to get our results perfect in that first iteration but we can see that everything is looking pretty good. |
| 44:10 | We'll just increase our RPM up a little bit here and we'll have a look and see what everything's looking like out at about 1800 and 2000 RPM. |
| 44:18 | We'll stop our scanner there and just have a quick look. |
| 44:21 | So we've got a couple of outliers here in the 5-6% vicinity but for the most part we're within about 2-3%. |
| 44:28 | Again being reasonable or realistic with our expectations here, if we're within about plus or minus 2-3% I'm going to generally be pretty happy. |
| 44:36 | Given the results we're got here, I'd probably go ahead and apply one more iteration of the steps we've just seen. |
| 44:42 | I'm not going to go through that here, you've already seen the process, it's just a rinse and repeat. |
| 44:47 | Let's jump on though and we'll have a look at our results and see what they look like under wide open throttle ramp run conditions. |
| 44:52 | Alright we've got rid of our heat soak here and we've got our engine running in 4th gear, let's get our run underway. |
| 45:23 | Alright so looking at our dyno plot we can see straight away that our plot is a lot closer to our target. |
| 45:29 | We've still got a lean area here around 2500 RPM and there's a little bit more work to do but it's pretty close. |
| 45:36 | Let's have a quick look down in our scanner and we'll have a look at the same results. |
| 45:40 | So we can see realistically that everything here is within about 1-2% of our target. |
| 45:46 | We've got this little lean area here at 2200 or 2400 RPM and we've also got slight positive trim here as well around 3000 to 3500 RPM so again one more iteration of those changes should clean this up but again it's exactly the same as what you've seen so we're going to jump ahead now and we're going to move on with the next step of our process. |
