Practical Reflash Tuning: Step 5-B: Boost Control
Watch This Course
$229 USD $114.50 USD
-OR-
Or
8 easy payments of only $14.31
Instant access. Easy checkout. No fees.
Step 5-B: Boost Control
21.12
| 00:00 | - So far, for this worked example, for the most part we have been ignoring the boost control side of the tuning and of course this is quite an important aspect that we do need to be mindful of. |
| 00:10 | There's two reasons why we've ignored it so far. |
| 00:13 | First of all we wanted to simplify the worked example so we can focus on the core aspects of fuel and ignition as well as getting our math sensor scaled correctly, injectors, if that's what we're doing etc and these are really the core fundamentals of the reflashing process. |
| 00:29 | The other aspect for our worked example is that the boost control settings that we have from the stock file, actually worked for our example on the dyno. |
| 00:37 | However they would be problematic once we got out onto the road or the racetrack and of course you may find that you're in a situation where you do need to make some changes so we're going to go over some of the core concepts here. |
| 00:48 | We're not going to go into too much detail, we're going to keep it fairly brief and to the point. |
| 00:54 | If you do want a more thorough understanding of the boost control strategy and how to go about tuning that from scratch, you can check our boost control tuning course, it goes much more in depth than we can do here. |
| 01:05 | So for a start, let's have another look at our core tables that we're going to need to deal with and we start here with our boost control target. |
| 01:13 | So we'll bring that up here and we can see we've got our table we've already looked at, we've got our RPM on our horizontal, sorry vertical axis there and we've got our requested torque on our horizontal axis. |
| 01:26 | So the first aspect is understanding where abouts in this table we are actually accessing, so where abouts is the ECU looking up the values in this table? And we already looked earlier in the worked example at where those requested torque numbers come from but for simplicity we can add that in as a parameter that we're logging so that we can see as we go through the ramp run exactly where abouts the ECU is pointing inside of that table so let's jump across to MegaLogViewer HD and get a big of an idea of what that looks like. |
| 01:56 | So in this particular set of graphs here we've got our requested torque parameter along with our engine load which we've already been looking at. |
| 02:03 | So we can see that at the start of our run here, around 2000 RPM, our requested torque's sitting at 330 newton meters, that ramps up so that by the time we're at 2800 RPM, we're at 450, that holds flat through to about 4500 RPM and then we see a gradual reduction again so that we're down at 305 newton metres. |
| 02:24 | Let's have a quick look back over at our EcuFlash table so we know that we're starting the run somewhere around about this point here, then by the time we're at about 2800 RPM we were up at 450 newton metres so we're off this table here. |
| 02:44 | We hold that through to about 4500 RPM and then we're dropping back down to around about 300 newton metres so our trajectory through this table looks something a little bit like this. |
| 02:57 | Obviously roughly but that's the idea, you kind of can see where we're going to be accessing, it's really important to know because otherwise you don't know where to make your actual changes. |
| 03:06 | Now that we've got a better idea of how this table works and where we're accessing inside of this table, let's jump back into MegaLogViewer HD and we'll have a look at how our boost control has been working so far. |
| 03:17 | And this is a ramp run from one of the last runs we did in the last part of this module. |
| 03:23 | Basically with our stock boost control parameters. |
| 03:26 | So let's see what's going on. |
| 03:28 | I've simplified the view here so we're really only looking at the things that are important to us right now. |
| 03:33 | In our third graph here we've got our turbo dynamics proportional and integral, so these are the proportional, integral elements of our closed loop control system. |
| 03:41 | Basically how much that's affecting the final wastegate duty cycle and then the yellow parameter there is the final wastegate duty cycle being sent out to the wastegate. |
| 03:50 | Then on our fourth graph we've got our boost pressure in psi and our target boost. |
| 03:57 | So basically here we're comparing how close we are to our target. |
| 04:00 | Now while we were find with the boost pressure that we were running, in reality you can see now that we weren't doing a very good job. |
| 04:07 | In particular down here in the low RPM range where we're spooling up the turbo, we can see that our green parameter which is our target, right at this point here, 3200 is requesting 18.5 psi and we've basically got about seven psi so we're undershooting quite dramatically. |
| 04:25 | We also see that once we get up to the point where the turbo can reach full boost we are only asking for about 18.8, let's call it 19 psi. |
| 04:33 | We've got almost 21 so we've got this overboost here so we've exceeded our target. |
| 04:38 | And let's see why that's the case. |
| 04:40 | So the turbo dynamics parameters is what we want to look at, remember these are influencing the final duty cycle. |
| 04:46 | So in this particular point here, we've got our boost target of 18.5 psi, we've only reached seven psi. |
| 04:53 | So what the ECU is doing is the closed loop strategy is driving that wastegate duty cycle higher and higher, as much as it can, in order to try and get to the boost pressure that it's targeting. |
| 05:04 | Of course it can't get there, there's absolutely no hope of seeing 18.5 psi. |
| 05:09 | So if we look at what our proportional and integral elements are doing up here at the same point, we can see that our proportional gain which is our red element, that's sitting at 7.5% so it's adding to the wastegate duty and then our green element which is our integral gain is sitting at 8%. |
| 05:27 | So in terms of those two elements, the proportional gain is an instantaneous response to the error, the rate of change of error. |
| 05:35 | The integral is a slower moving compensation for error over time. |
| 05:40 | And what we get in a situation where we're targeting a boost pressure we can't reach is a term called integral wind up. |
| 05:47 | And essentially the integral gain can't get us to the target so it consistently increases until it maxes out at its upper value and then when we get to a situation where all of the sudden the turbo is in a range where we can reach our target, the integral elements is wound up so positive that we get this overshoot and that's exactly what we see at this point here where we exceed our target. |
| 06:10 | So let's look at what the integral and proportional gains were doing just prior to that. |
| 06:15 | We've got 8.4%, 8.39%, 7.5% and we see at the point where we get this crossover our integral in particular in green there is still sitting at positive 10 so it's adding 10% to the wastegate duty cycle. |
| 06:30 | And we see this sort of sudden correction here, realise that it's gone too far and then it goes completely the opposite way. |
| 06:36 | So this is going to cause overboost. |
| 06:38 | Now we didn't really see this on the dyno but particularly on the road on gear shifts this is going to be quite noticeable and quite dramatic. |
| 06:44 | Could also be potentially quite dangerous and we may likely end up with much more boost than the 21 psi we've been seeing on the dyno. |
| 06:51 | Now the flipside of this, as we come past the RPM range where the turbo now can create maximum boost pressure we can see that our green element, our integral gain as dropped down. |
| 07:03 | It's now pulling -6%, proportion gain, -2.8, 3% let's call it. |
| 07:08 | And we can see we actually end up with our integral gain completely off the scale there, on our graph we're down at minus 33% so it's pulling all of the duty cycle out. |
| 07:17 | So what we can tell from this is two things, first of all our boost targets are not realistic and it's important to make sure that our boost targets are something that the turbo can actually achieve. |
| 07:30 | Otherwise the closed loop control strategy is going to come into play and is very likely to cause us overboost particularly on transients or gear shifts. |
| 07:38 | So that's the first thing, target boost pressures that are realistic, that we can actually get. |
| 07:43 | The other things we want to do then is adjust our wastegate duty cycle tables so that the closed loop strategy isn't doing all of the heavy lifting to correct any errors in there so in the perfect world what we'd like to see is that our proportional gain and our integral gain are sitting relatively closed to zero through the ramp run. |
| 08:02 | This means that our open loop duty cycle tables are pretty close to the money. |
| 08:07 | Alright now that we've got a bit of an idea of what's going on in there and where our problems lie, even though we've completely overlooked those thus far in our tuning, let's jump back into EcuFlash and we'll have a bit more of a look at the tables. |
| 08:19 | So we've got our boost target table, we've discussed that one there so we'll close that one down for the moment. |
| 08:25 | The next two tables that we will focus on here is our initial wastegate duty cycle and our max wastegate duty cycle. |
| 08:31 | So as we've already discussed earlier in the course, these two tables are essentially the same and they retain a split of 15% between the two tables. |
| 08:43 | So you can see that the initial wastegate duty cycle just in this particular area there 50%, and then in the max duty cycle table, 65%. |
| 08:52 | Pretty self explanatory there, the max duty cycle table is simply the maximum duty cycle that can be applied regardless of what the closed loop control strategy is doing. |
| 09:01 | The initial, what happens here with the initial table is that when we first make a transient change, maybe a geatshift or something like this, this is the table that the ECU will revert to, the duty cycle the ECU will revert to initially before the closed loop strategy comes into play. |
| 09:18 | So what we want to be doing is adjusting these two tables to get us close to the target without a lot of proportional and integral closed loop control being required. |
| 09:30 | Alright rather than going through this entire process, I've actually gone ahead and made a set of tables that are going to get us pretty close so we can start to see how this all works. |
| 09:40 | So let's head across to our tune tables and we'll have a little bit of a comparison. |
| 09:46 | So in our tune table, first of all we're going to come back to our target boost table and see what we've done with that particular table. |
| 09:55 | So that we can see it side by side, let's bring up our original boost table and what we can see is for a start, I've pulled a lot of boost pressure out, or target out in the lower RPM region where we know that the turbo can't respond, can't make boost. |
| 10:11 | In particular here, at 3200 RPM, we're only targeting around 10 psi whereas in the stock table where you can see that we're up at about 19 psi, 420 newton metres, obviously the targets change a little bit. |
| 10:26 | So that's the first change. |
| 10:28 | Now I've actually gone ahead and targeted a slight increase in boost through this mid range here once the turbo can come on boost. |
| 10:35 | So you can see we're targeting around 20 to 21 psi, whereas we've got a flat 19 psi. |
| 10:40 | So small increase in boost, really the same boost pressure we were actually seeing. |
| 10:44 | Now making these adjustments is simply a case of looking at our logging and seeing what sort of boost pressure was achievable so for example if we look here, 3200 RPM, we know that we were targeting 18.5 psi and we had around about seven psi. |
| 11:00 | So let's have a look back at our table and 3200 RPM, you can see, as I've already said, set that to 10. |
| 11:07 | So I've set it slightly above our target, we do need to be a little bit mindful, a little bit careful here because if you set the boost pressure exactly where you are, or what is achievable, you're going to find that you're going to artificially reduce the respoinse. |
| 11:19 | So it is a bit of a fine line here and it's a case of just testing to see what gives you the best response without overboost. |
| 11:25 | Alright so now that we've got our boost target table dialled into something a little bit more realistic, and I actually will also mention that due to our fuel system limitations I have left the upper boost targets exactly as stock, we just simply can't get anymore boost into the engine because we are out of fuel. |
| 11:43 | And now we've got that dialled in, let's have another look at our wastegate duty cycle tables. |
| 11:50 | And we'll bring up the wastegate duty cycle, we'll look at our initial wastegate duty cycle table from both of our maps here and see what the differences are. |
| 11:59 | Alright so on the right hand side we've got the table that we have been running up until this point and for a start we can see that at higher RPM we've got quite high duty cycle values in there, 65%. |
| 12:12 | And we can see what the actual duty cycle was back in our log, let's jump back over there and our duty cycle, our wastegate duty cycle is this yellow parameter that we can see dropping consistently as we go through our ramp run. |
| 12:26 | So particularly we can see right at the very high end, right at 7000 RPM at the end of our ramp run, the wastegate duty cycle is actually sitting at zero so we can't get any reduction in boost pressure from that point on. |
| 12:38 | So we can then have a look back over at our table and we can see that 7000 RPM, that particular point, the base duty cycle was sitting, actually we are a little bit further down, we're about 300 newton metres at 7000 RPM. |
| 12:53 | We're sitting, 40, 50% duty cycle depending on exactly where we are in this table. |
| 12:58 | So clearly we're a long way ahead and this is where we do get ourselves into problems because that initial duty cycle table, that's what the ECU will revert to when you make a gearshift or you come on boost initially. |
| 13:11 | So we're going to be in a situation where we're very likely to have quite a massive overboost before that closed loop can catch us so we don't want that happening. |
| 13:19 | So what you can see here is in that same area of the table, down to 300 newton metres, I've set that all to 10% duty cycle so a significant reduction, a bit more realistic a bit closer to the actual targets that we are going to be seeing. |
| 13:34 | Also if we look at the spool up areas, particularly at 2800 through to about 4000 RPM, we can see that the stock map is asking for 49 to 50% duty cycle and again that's more than we need. |
| 13:49 | We've dropped that down there to between 30 and 35% duty cycle so again if we just look at a point 4000 RPM for example, where we can actually reach our boost target or thereabouts, let' have a look back and see what we're getting. |
| 14:03 | And if we go down to 4000 RPM here, we can see that our wastegate duty cycle is 60% but we've got this heading towards quite a significant overboost so we need to be a little bit mindful of that as well. |
| 14:18 | This is, at the moment we're just below our target but we are also about to just go and exceed that target here as well. |
| 14:26 | Remembering at this point we've got our integral and proportional gains adding to that, we've got 10% being added to our wastegate duty cycle from our integral gain and another 2% from our proportional. |
| 14:40 | So it's all about matching the wastegate duty cycles to actually what the engine needs in order to meet our target boost and if we can get that done, our closed loop control's going to have a lot less work to do. |
| 14:51 | Alright let's head back over to EcuFlash and we can make some of these changes. |
| 14:56 | So we're going to copy the numbers here from the table that we've just looked at and we're going to paste those into our actual map, the one that we're going to be using in our vehicle. |
| 15:08 | So before we go and modify our maximum duty cycle table, I am just going to be a little bit more thorough with my changes here. |
| 15:15 | You can see down in this area we've still got some large numbers that if we drop down into these requested torque values are going to potentially result in some unwanted boost response. |
| 15:25 | So again I'm just going to quickly go through here and we'll smooth out these numbers so we've got something that is a little bit more representative so let's do that now. |
| 15:43 | Now that we've got our initial duty cycle table somewhere close to where we want it to be for our subsequent runs, we're going to bring up our maximum duty cycle table and we're going to make the required changes to that. |
| 15:55 | Remember we're trying to keep a 15% split between these tables, we'll start by just copying and pasting the entire table across. |
| 16:01 | And then we can highlight the areas where we do have values and we're just going to use the add to data option to add 15%, we'll just go through and do that to all of the data points now. |
| 16:24 | OK at this point we've got all of our changes made, we're going to flash the new ROM file into the ECU, we'll get up and running and we'll see what the results look like on our dyno as well as our logging. |
| 16:34 | Alright we've got our flash complete, our engine's up and running, we've got all of our temperatures stabilised so let's perform a ramp run and see the effect of those changes to our boost. |
| 16:56 | So on face value there it doesn't really look like we've achieved too much. |
| 17:00 | We've got almost the same peak boost, essentially identical boost response, we've actually dropped the boost very slightly through the rev range so we've actually seen there a very small decrease in our overall power down to 235 kW, remembering we were at 238. |
| 17:17 | At this point I'm not really chasing power with these boost changes though, we want to just see how that's affected our boost control strategy so let's have a look at that in MegaLogViewer HD. |
| 17:27 | Alright looking at our log file, straight away if we look at our boost target versus our measured boost pressure, we can see that the two are now much closer together. |
| 17:36 | We do still have a little bit of overboost present here which we'll talk about in a moment and our boost pressure is slightly above our target at higher RPM. |
| 17:45 | Now some of this, particularly the point I've got labelled here or noted here, we can't do anything about this, we look at our yellow parameter up here this is our duty cycle to our wastegate. |
| 17:56 | It's sitting at 3% and essentially that's as good as the wastegate being at its minimum duty so the problem here is that we are targeting a boost pressure that we can't get to, we can't reduce the boost pressure anymore. |
| 18:07 | So in this higher region here, what we could do, if we want to be fussy is go into our boost target map, adjust those boost targets up so they're actually matching what the turbo can provide. |
| 18:19 | So that'll clean up the top end. |
| 18:21 | We've still got the slight overboost here and let's just talk about how that's occurring, why that's occurring and in order to do that we look at our proportional and our integral gains so coming up to this group in our third graph here. |
| 18:36 | Remembering the green parameter here is our integral gain. |
| 18:40 | Now we can see that at the start of this run, our boost pressure is still well below our target, we've got our target boost pressure still probably set a little bit optimistic here through this region at about 8.5, nine psi whereas we're only seeing at that point about three psi. |
| 18:57 | So what this has resulted in at this point is our integral and our proportional gains are still quite positive, we've got about 10% from our integral gain and that continues because it's not until this point right here that we actually reach our target. |
| 19:12 | So the point where we reach out target, we can see that our integral gain is still wound up to positive 10%. |
| 19:18 | Proportional on the other hand has responded quite nicely, that's sitting down around 0%. |
| 19:24 | So it takes a little bit of time which is this portion here for our integral gain to get back under control at which point our boost pressure then drops back down to our target. |
| 19:34 | So what we're trying to do with this is essentially tune our boost targets first of all so that they're realistic and that we are able to hit them. |
| 19:43 | So that's our first thing. |
| 19:45 | Once we've got our boost targets to something that's actually reasonable then we're going to tune our wastegate duty cycle tables so that our proportional and integral gains through the ramp run are relatively close to zero. |
| 19:56 | So we can see this sort of oscillation and swing here we've got in our integral gain. |
| 20:01 | It's not ideal but we're still getting pretty close to the mark here. |
| 20:05 | Our proportional gain, our red parameter on the other hand, you can see is doing quite a nice job sitting there pretty close to zero. |
| 20:12 | So from here we've got a little bit more fine tuning to do, particularly in that spool up area and that will help reduce this overboost by getting rid of this area here where our integral gain is wound up to its maximum. |
| 20:26 | Like a lot of our tuning, the boost control tuning is an iterative process so it's unlikely you're going to get this dialled in in one run. |
| 20:34 | It's also important to mention that while getting the boost control dialled in on the dyno is important, we also need to make sure that our transients work out well on gearshifts out on the road or the racetrack. |
| 20:46 | So this has only been a brief overview of the boost control strategy inside of the Subaru ECU and as mentioned if you do want to learn more about boost control tuning in detail, you can check out our boost control tuning course. |
| 21:00 | For now though, we've finished with step five and we can move onto the final step of our process and check all of our work out on the road or the racetrack. |
