Practical Reflash Tuning: Step 4-A: MAF and Injector Scaling
Watch This Course
$229 USD
-OR-
Or
8 easy payments of only $28.63
Instant access. Easy checkout. No fees.
Learn more
Step 4-A: MAF and Injector Scaling
23.40
| 00:00 | - The next step in our six step process is to adjust our mass air flow sensor scaling and also to adjust our calibration to suit our injectors. |
| 00:09 | Now as we've already discussed, in this case the way the Ford ECU deals with injector characterisation is quite comprehensive so the process of adjusting our calibration is really just as simple as taking the data from the injector manufacturer and then dropping that into the relevant locations inside our calibration file. |
| 00:29 | As you know, we've already got that done so this leaves us with our mass air flow sensor to calibrate. |
| 00:34 | Now with the Ford ECU we do have the speed density subsystem so we're going to break this particular step down into two parts. |
| 00:42 | Initially we'll deal with rescaling our mass air flow sensor. |
| 00:45 | Following on from that we will discuss the speed density system and we'll see how we deal with that as well. |
| 00:52 | So at this point we have got our calibration adjusted, ready to go. |
| 00:57 | But you'll be aware that we still haven't actually written that into the ECU, we haven't flashed that into our ECU. |
| 01:03 | So right now our vehicle is running as delivered. |
| 01:07 | And in this case because it is a tune that is already, it's set up for the supercharger kit, we are in a position where we can get a base run. |
| 01:18 | So we always want to do this, so we've got a base run that we can compare any gains that we make in terms of power and torque. |
| 01:25 | But also this base run is going to tell us a lot about what is going to be needed. |
| 01:29 | So let's get our engine up and running, we'll perform a run, we'll scan that, and we'll have a look at our data. |
| 01:59 | Alright we've got our base run complete there and on our dyno screen we can see that we produced 382 horsepower or 285 kilowatts at the back wheels. |
| 02:08 | Now the red line that we've got at the top here is our boost pressure. |
| 02:13 | So we were logging that and we can see that our peak boost pressure was around about 155 kPa so a little over seven psi of positive boost pressure. |
| 02:23 | So that we can have a little bit more of a detailed look at what was going on during that ramp run, let's jump into our laptop software and we will have a look at the VCM scanner. |
| 02:32 | Now there's a variety of parameters that I have logged here. |
| 02:35 | At the top in red we have our RPM, so at the start of this run, we're around about 1200 RPM. |
| 02:41 | At the end of the ramp run here you can see we got up to about 5700, 5800 RPM. |
| 02:47 | One of the first pieces of information we're going to look at here is our air fuel ratio data. |
| 02:53 | Down at the bottom here we've got three pieces of data being logged at the same time. |
| 02:58 | We've got our commanded air fuel ratio or commanded lambda at the bottom there in green. |
| 03:04 | So this is what the ECU's requesting. |
| 03:06 | Then we've got our yellow and our orange data coming from our wideband air fuel ratio sensors. |
| 03:11 | Now just because these are wideband sensors, doesn't necessarily mean that we can automatically trust the data. |
| 03:18 | So it is a good idea to compare the numbers in our scanner to a known wideband sensor. |
| 03:25 | In this case we've previously done this as we've compared it to the Motec PLM we use on our dyno. |
| 03:30 | And the data is exactly the same. |
| 03:32 | This means that we can trust the data in our scanner, and we can use this for the rest of our tuning. |
| 03:38 | So as we'd expect, because we are running in closed loop here, we can see that the data really closely matches the point that I've just jumped onto there right at the end of our run. |
| 03:47 | We can see that our target was 0.79 One bank was at 0.80 and the other bank was at 0.79 So that's great but it doesn't tell us the whole story. |
| 03:57 | Remembering because it is running a closed loop, we've got the trims doing a lot of work in the background sometimes to get us to that target. |
| 04:05 | So if we look straight above this we've got our short term and our long term trims. |
| 04:11 | And in this case the long term trims aren't actually doing anything. |
| 04:14 | But we can see at the same point that we've just looked at, the short term trims are pulling 8.6% from bank one and 9.4% from bank two. |
| 04:24 | So essentially what this means is our calibration of our MAF sensor at that point is around about 9% out. |
| 04:30 | It's over estimating the air flow by about 9% so we've certainly got a little bit of work to do. |
| 04:37 | The other aspect we can see here is our knock retard. |
| 04:41 | So if we move up, we've got our red parameter here which is our knock retard and this is one of the reasons that we have got this vehicle in for some tuning work. |
| 04:51 | On our local fuel, this particular calibration is running too much ignition timing and we can see that a lot of retard is being pulled here. |
| 04:58 | Right at the start of the run we can see that our retard is around about 3.5 to four degrees. |
| 05:03 | And we can see that that very slowly tapers out as we go through our run. |
| 05:09 | So on face value one of the questions a tuner might ask here is if we've already got too much ignition timing, how are we possibly going to make any more power? Well as we'll see once we get a little bit deeper into this tuning exercise, we're gonna find that often we can actually make more power by removing ignition timing and preventing that knock retard from occurring. |
| 05:28 | Another aspect we're going to look at is speeding up the rate which that knock retard has removed, again we can see that from the point where knock stops happening, you can see there's very gentle drop in our knock retard. |
| 05:42 | So if we can get that knock retard removed a little bit more quickly, then we should be able to pick up a little bit of power in the top end. |
| 05:49 | Alright so that's our base calibration, that's our base run. |
| 05:52 | We're going to save that run on our dyno so we can use that as a comparison later on. |
| 05:58 | What we're going to do now is go through the process of flashing our modified calibration into the ECU, so we can start some tuning. |
| 06:06 | Before we can flash that calibration into the ECU we do need to disconnect our scanner and we can do that using the little disconnect from vehicle icon in our tool bar up the top. |
| 06:18 | Now we'll head back across to our VCM editor software and we'll simply click on the little write vehicle icon. |
| 06:24 | This will bring up our vehicle writer box and we'll click on write. |
| 06:29 | And we also get a warning box here telling us that we need to have disabled one of the diagnostic trouble codes in order to use the fast write. |
| 06:37 | We've done that so we can click on OK and the write will commence. |
| 06:42 | With our write complete, we can now get our engine up and running again. |
| 06:45 | What we want to do is allow the engine to run for a couple of minutes just to get rid of any chance of heat soak effecting the accuracy of our results. |
| 06:53 | And we'll be able to head across to our scanner and see what we need to be looking at and how we need to use the dyno in order to help us calibrate the mass air flow sensor. |
| 07:03 | So let's get it running now. |
| 07:06 | Alright we've got our scanner running and we can see that straight away while we're idling here, we've got some quite large short term fuel trims active there so this indicates that we've got some work to do. |
| 07:18 | But of course we may be still operating with some heat soak active there. |
| 07:22 | While we're allowing the heat soak to dissipate, we'll have a quick look at the histogram setup that we are using and I'll explain what I've done there. |
| 07:31 | So let's just drag our charts out of the way. |
| 07:33 | And we'll have a look at this histogram that I've set up here for our mass air flow sensor which is this one we've got at the bottom. |
| 07:41 | I've called it long term fuel trim plus short term fuel trim. |
| 07:44 | Now we can right click on this and we can have a look at our graph's layout to see exactly what this histogram is and what it's doing. |
| 07:52 | We'll click on the one we're interested in. |
| 07:54 | Obviously you can add a label here, you can make this whatever you want. |
| 07:57 | Now the parameter that we're logging into this is the important point here. |
| 08:01 | And you can select this by clicking here on the parameter. |
| 08:05 | So the one I'm using is an inbuilt parameter, it's a MAF function that is developed and provided inside the scanner which is a long term and short term fuel trim added together. |
| 08:17 | You can see it also says bank separated. |
| 08:19 | So essentially what this parameter is doing is for each bank it is adding the current short term and long term fuel trims together. |
| 08:26 | And then it's adding both banks together, dividing those by two, to average the results, so that's what we're looking at here, is the average between the two banks. |
| 08:34 | And then it's outputting that as an error. |
| 08:37 | So essentially if our long term and short term fuel trims are all sitting there at zero doing nothing, there's no correction required, then the number that's going to be logged into our histogram will also be zero. |
| 08:49 | If on the other hand, we're adding too much fuel and the trims are negative, pulling fuel out, then we'll end up with a negative value in our histogram. |
| 08:58 | Now we'll just have a quick look here at the options we've got. |
| 09:02 | So if we click on parameters, just so you can see where to find this, we'll move down to our MAF channels which are down the bottom. |
| 09:09 | You do have the freedom here as well to create your own user channels. |
| 09:14 | The ones we're looking at here are fuel trims. |
| 09:18 | I'll just move these down so we can see all of them. |
| 09:20 | And we've got the options depending on exactly what you're trying to do here. |
| 09:24 | So in this case we're looking at our long term plus short term fuel trims bank separated. |
| 09:30 | OK just to make this really nice and easy to see visually as well I always like to have some colours for this histogram so we can see at a glance whether our trims are positive or negative. |
| 09:40 | So we can see we've associated a colour of red with a trim of plus 5%, where the closed loop system is adding fuel in. |
| 09:48 | And we've got a colour green associated with it trimming 5% out. |
| 09:52 | Now the other important aspect here is we need to set the break points for this axis. |
| 09:58 | So the idea here is that we set this histogram up so it's got exactly the same break points as our mass air flow sensor calibration table. |
| 10:06 | Let's see how we can do that. |
| 10:08 | We'll head back across to our VCM editor, we're going to head across to our air flow and we are on our MAF calibration here. |
| 10:16 | So what we wanna do is look at our MAF air flow versus period. |
| 10:20 | And we can right click here, and what we want to do is go down to row axis and copy labels. |
| 10:26 | And what that's going to do is it's going to copy all of the break point labels here. |
| 10:32 | Then all we need to do is head back over to our scanner software, and here under row axis, you can see these numbers are already in here, we just need to paste those numbers in and this will give us the same break points, this makes it really easy to apply the changes to the correct point in our air flow meter calibration table. |
| 10:52 | Alright now we'll have a look and we'll see how we can use the dyno and this histogram to help us scale our mass air flow sensor. |
| 10:58 | And this is going to be done in two ways. |
| 11:00 | First of all we're going to use the dyno in steady state. |
| 11:03 | And we're going to drive the truck really smoothly, being very very careful with our throttle input so that we don't bring in any transient enrichment. |
| 11:11 | And we're going to move through as much of this table as we can get to, right up to the point where we transition into our power enrichment. |
| 11:19 | So once we've gathered that data, we're then going to apply those changes into our mass air flow sensor calibration table. |
| 11:26 | Then we're going to swap our technique slightly and we're going to go through and perform some wide open throttle ramp runs. |
| 11:32 | And we're going to use that to fill in the higher air flow values of our table. |
| 11:36 | The reason we can't do all of this in one go is because as we transition from our target of lambda one, through to our power enrichment lambda target, this is going to necessitate a step in our table. |
| 11:50 | The actual air fuel ratio can't track that change instantly. |
| 11:53 | So it gives a pretty ugly little step in our histogram that's not very realistic, it's not very real. |
| 12:00 | And if we try and apply that to our mass air flow sensor scaling, it's going to end up giving us some problems later on. |
| 12:06 | So let's perform our steady state tuning now. |
| 12:09 | OK we've got our engine running now on our dyno. |
| 12:12 | And we can start our scanner to start gathering some data. |
| 12:17 | So we can see at the moment we are sitting down with a mass air flow sensor period of 408 microseconds. |
| 12:25 | So what we wanna do is just look at the numbers we've got in our table here in our histogram. |
| 12:30 | We can see we've got an 11% error there where the closed loop trim is adding fuel. |
| 12:35 | So what I want to do is just very smoothly and slowly, increase our throttle position. |
| 12:40 | You can see we've jumped up to 389.4 microseconds. |
| 12:44 | We're filling in that zone there quite nicely. |
| 12:47 | Once we've got a relatively consistent number in the cell, we can open our throttle slightly further and move up into the next cell. |
| 12:55 | And basically it's just a rinse and repeat of this process. |
| 12:59 | We wanna try and stay in each cell long enough for the number to become stable. |
| 13:04 | So I'm just going to continue through this process right up to the point where we transition into power enrichment. |
| 13:43 | Alright we've gone through there and we've gathered some data. |
| 13:46 | We've gone from 408 microseconds all the way up to 159.9 microseconds. |
| 13:52 | So what we wanna do first of all is have a look and see if we've got any obvious trends in this graph, and also if we've got any outliers in our histogram data. |
| 14:01 | So what I'm talking about here in terms of outliers, we should be expecting a relatively smooth trend or error in this histogram data. |
| 14:09 | So if we've got something that is a massive change from one break point to the next, this would probably indicate that we've got some bad data and we'd wanna go and revisit that site. |
| 14:19 | In this case everything we've got looks relatively sensible. |
| 14:21 | We can see right down here at low air flow we've got quite a large positive error, so the ECU is adding an additional 11% fuel. |
| 14:29 | Whereas right at the end of our test there we can see that our numbers are starting to go negative where the ECU is pulling fuel out. |
| 14:37 | So what we want to do now is use the data that we've just collected in our histogram to help us change our mass air flow sensor scaling. |
| 14:45 | So what we want to do is basically apply this error onto our mass air flow sensor. |
| 14:50 | There are a couple of ways we can do this. |
| 14:52 | What we can do is use the copy and paste special option. |
| 14:57 | Or we can manually make these changes. |
| 15:00 | So lets show you how we can use the paste special option. |
| 15:05 | So what we'll do is we'll highlight all of our data and we can either right click and click on copy or control C will do exactly the same thing. |
| 15:15 | We can then head across to our VCM editor again, and what we can do now is right click and use the paste special function. |
| 15:25 | So this allows us to paste that error or apply that error directly onto our mass air flow sensor calibration data. |
| 15:32 | So we've got the option here to multiply by the percent. |
| 15:35 | So this will make the exact change that we logged into our histogram. |
| 15:39 | Or alternatively we can make a slightly less aggressive change. |
| 15:43 | In this case by half, so we'll apply half the error. |
| 15:46 | What we'll do for our first round here is I'm going to click on multiply by percent. |
| 15:51 | Now there is a little bit more to it than that though. |
| 15:54 | Let's head back to our scanner. |
| 15:55 | And we'll see that we only ended up gathering data from 408 microseconds and above. |
| 16:02 | So what I wanna do is extrapolate the values that we were seeing down here, I wanna extrapolate those down so we're being a little bit more complete. |
| 16:10 | On the other hand the data that we're gathered at the top of this table here, we don't want to extrapolate this upwards because we're going to be using our dyno again in a second to perform some ramp runs and that's going to help us fill out the data that we haven't seen so far. |
| 16:26 | So what we wanna do here is make a change of about 11%, let's call it 12% because that is increasing as we move down this histogram from 435.8 microseconds and below, so let's do that. |
| 16:40 | We'll find our point which we have made no adjustments to, 435.8, and we're going to highlight all of those cells below, and in order to add 12% I want to multiply it by 1.12 So we'll press multiply, that's made that change there. |
| 16:59 | So now we've got those changes applied into our mass air flow sensor calibration table, let's head back to our scanner and we're now going to perform a ramp run on our dyno and we're going to be looking at filling in the rest of this table. |
| 17:14 | Let's get our engine up and running again now. |
| 17:19 | Again in order to make sure that our data is accurate, we want to make sure that after we've started the engine, we allow it to run for a couple of minutes just to make sure it's not being affected or influenceed by heat soak. |
| 17:31 | Once we've done that, we can get our scanner up and running and we can log a ramp run, so let's go ahead and do that now. |
| 17:55 | Alright so we've got some data in our histogram now so let's have a look at that. |
| 18:00 | Now you'll remember from our steady state test, we'd managed to get the data up to 159.9 microseconds. |
| 18:07 | So what we're interested in for this particular test is the data that we've gathered above this particular point. |
| 18:13 | And we can see that we've got the trend that we've already talked about where as we continue out through that ramp run into higher air flow values, the trim becomes increasingly negative, so it's pulling more and more fuel out because the mass air flow sensor is over estimating. |
| 18:29 | So in this case what we're going to do is see how we can apply these changes manually rather than the special paste function that we've already looked at. |
| 18:39 | Now remembering that when we got out to 159.9 microseconds we were pulling around about 2% fuel out. |
| 18:47 | So we can see this time we've got around about a 1% trim there, so not much difference there, we're obviously likely to see slight variations from one test to another. |
| 18:57 | But what I'm going to do is continue that value out that we'd already seen. |
| 19:01 | In these two zones here that currently are saying minus one, this seems like it might be splitting hairs but I'm going to actually remove 2% there. |
| 19:09 | So let's head across to our editor, we'll highlight these two cells here, and we'll multiply these by 0.98 Let's head back to the scanner and we'll gather some more information. |
| 19:24 | We can see that we then remove, 3%, 4%, 5%, 6%, 7%, 8%, 9%, and 10%. |
| 19:32 | So it's basically removing 1% per break point. |
| 19:35 | So let's continue that and that's up to 105 microseconds. |
| 19:40 | So we want to remove 3% from our 141.1 microsecond break point. |
| 19:48 | We want to remove 4% from our 134.5 Then we want to remove 5% from the next break point up. |
| 19:59 | And we'll just continue this out until we have completed our table. |
| 20:13 | Alright so we've completed that change there and you've seen I've just extrapolated that last value of a 10% trim right up to the top of our mass air flow sensor calibration table. |
| 20:23 | So that's our first round of modifications made there. |
| 20:25 | So we've seen now how we can use the special paste function as well as manually blending these numbers. |
| 20:33 | It's important to have both of those skills because particularly if your data is a little bit noisy or you've got some outliers in that data that aren't very realistic, using the special paste function is likely to actually create problems with your mass air flow sensor scaling. |
| 20:50 | The other important aspect to consider as well is you're viewing this data graphically. |
| 20:54 | What we should have with a properly scaled mass air flow sensor is a nice smooth and consistent shape to that graph. |
| 21:00 | We shouldn't be seeing any irregularities in that graph. |
| 21:03 | Alright with that round of modifications made now we're going to save our calibration and we're going to send it through to the ECU and we're gonna have another test and see the effect of those changes that we've just made. |
| 21:17 | Alright we're back up and running and straight away we can see here at idle that the closed loop control is doing a lot less work. |
| 21:24 | Our trims are still slightly positive but only around 3.5% now compared to what we initially saw. |
| 21:31 | So we can see exactly how well that round of modifications worked out, let's get our truck back into gear and we'll perform another steady state run just to gather the same data that we've already looked at and see if it's an improvement over our first test. |
| 21:59 | So we've gone through there and we've gathered some data repeating our earlier steady state test, and straight away you can see that our results are much better. |
| 22:08 | We're much much closer to zero although we're not quite right. |
| 22:10 | We can see that particularly around about this area here, our closed loop trims are pulling a little bit of fuel out and likewise up around here we can see that our closed loop trims are adding a little bit of fuel. |
| 22:22 | Obviously at this point I have only repeated our steady state test. |
| 22:26 | We'd also want to see what was happening during our wide open throttle ramp run test. |
| 22:31 | Now it's important to understand what sort of errors are deemed acceptable here. |
| 22:35 | There's no need to try and get this entire histogram filled out with zeros. |
| 22:41 | And in fact it's going to be completely impossible. |
| 22:43 | Generally I'm going to be aiming for about plus or minus 2%. |
| 22:48 | If I'm within that region, then I know that my mass air flow sensor is calibrated really well, and this is going to give the closed loop system very little work to do. |
| 22:58 | So we're gonna get nice accurate control over our air fuel ratio. |
| 23:02 | So the process is simply a rinse and repeat of what we've looked at here. |
| 23:06 | We're going to gather data under steady state conditions then under ramp run conditions. |
| 23:11 | We're going to then apply the data from our histogram into our mass air flow sensor calibration. |
| 23:16 | Now if we're doing everything right we should be able to dial in that mass air flow sensor in no more than perhaps two or three iterations of that process. |
| 23:24 | Once we've got our mass air flow sensor calibrated accurately, we're happy with our closed loop trims, we can move on and in the next step, we'll have a look at how we can calibrate the speed density subsystem. |
