TSE EFR NB2 Build Thread
#342
Here's what I do. Wow, this wound up being longer than I intended to type:
The goal is to find "MBT", or the amount of advance that produces the maximum amount of torque. You don't necessarily want to run at MBT, but you want to find it because that's the basis for the tune. MBT varies with RPM and with MAP, so you need a 3d map to represent it. You find MBT by turning up the timing until it knocks, then backing off. This is safe, so long as you're careful, paying attention, and not being an idiot.
Things to do before you go to the dyno:
1) Nail down the AFR fuel tuning on the street beforehand, you don't want to spend $150/hour to tune fuel.
2) Nail the closed loop boost control so that it's very stable and consistent. This is critical because you're going to be looking for differences in power of a few hp, and if the boost control isn't good enough then you can wind up losing that signal because the car's running 5-10 kpa higher MAP on one run than it is on another.
3) re-bin your spark map so that the horizontal rows are the ones you're going to use in the dyno session. Set one of them at the waste gate pressure, another one at the maximum boost level you want to tune at, and scatter the other ones around in reasonable places. You want to do pulls at exactly the MAP that's in a horizontal row so that you only have to mess with one row of numbers at a time.
4) start with a conservative spark map that shouldn't knock at all.
5) Get everything *else* properly sorted on the car first, no leaks, everything tight, etc. You don't want it to break from some random other cause while it's on the dyno. Check the base timing with a timing light to validate that the timing delay config in the MS3 is correct. Make sure you remember to set it back to normal timing (i.e., not fixed timing) after you do this -- I once wasted 15 minutes on the dyno trying to figure WTF the timing changes made no difference to power when I had it in fixed timing mode.
When you get to the dyno:
1) Set boost to the minimum value, do a pull. Monitor AFR & knock, if it looks squirrelly then abort the run and fix it, don't keep pushing. Monitor intake air temps and coolant temps, ideally you want these to be the same at the start of every run. Consistency is key, anything that changes run-to-run can throw off your data.
2) If it knocks but AFR is where it's supposed to be, take out timing in those cells to make it stop. (if AFR is wrong, fix that too)
3) After the run, assuming it didn't knock, add 2 degrees of timing to every cell and do another run. Compare the two pulls, you should see a healthy power increase from 2 degrees of timing. Repeat this step.
4) At some point the amount of power it gains will start to drop, you're getting close to MBT. Start going up only 1 degree at a time.
5) if it starts to knock at any point during this process, then remove timing in the knocking cells until it stops. You may need to add timing in some cells and remove it in others, remember that it's a 3d map.
6) Once it stops gaining power with added timing, you've found MBT. Stop here, there's no point in going higher. You won't make any more power (at some point you'll starting losing power), but it'll drastically increase the load on the engine.
7) Reset your boost to the next higher row and repeat the process for that one.
8) Once you've found MBT everywhere, save that map for future reference then back the entire thing off by 1-3 degrees for safety.
9) Go in and manually smooth out the sharp edges in the map.
If you're starting with a base map that's close (such as one from Reverant), then it's likely that the curves in the map are already close to being the right *shape*, you'll just need to adjust the overall map up or down a bit. You can thus speed up the process by applying any timing changes you make to the current row plus all of the rows above it as well. So if you're adding 2 degrees to the 160 kpa row, you'd also do that to the 180, 200, and 220 (or whatever) rows. Be aware that this makes it more likely to knock the first time you do a pull on the next row up.
Generally speaking, the timing curve is going to be U-shaped, it'll want higher timing at lower revs (say 3-4K), lower timing at the torque peak (5K), then back to the higher values at higher revs (6-7K). Those rev ranges are for stock cams, aftermarket cams will be different. Also, the higher the MAP, the less timing it'll want.
Another useful technique I've learned is to take copious notes while you're doing this. Have a line of notes for every run, include the run # (match it up with the one in the dyno operator's log files, so that you can go back to it), the reason for the run, the boost level, and the results & other impressions. Datalog every run on the megasquirt, save it with the same run number as on the dyno. Save the tune to a unique file name at least every few runs, if not every single one. You're paying $2.50 per minute to use the dyno, you do not want to discover halfway through that you're lost and you've just wasted 20 minutes to learn nothing.
Cars have a tendency to be more prone to knock on the street (or especially track) than on the dyno, so don't be surprised to discover that it would hit MBT at a certain boost level on the dyno, but if you try it on the street then it pings. This is because dyno runs are relatively fast compared to RPM sweeps on the street, so the motor spends more time in the ping-prone cells, accumulating heat.
Doing this should get you a reasonable map for use on the track, and you can do it on any dyno that will give you a repeatable power graph. There are other techniques you can use on a fancier dyno like a DynaPack, but that costs more money and takes more time.
--Ian
#347
2) Nail the closed loop boost control so that it's very stable and consistent. This is critical because you're going to be looking for differences in power of a few hp, and if the boost control isn't good enough then you can wind up losing that signal because the car's running 5-10 kpa higher MAP on one run than it is on another.
The spark map from the basemap is knocking. So I'm not sure how much more conservative I should go.
Mine just looks like a hill... Dunno.
Thank You for taking the time and effort into that post Ian, I'm sure it will benefit many others as they come across my build thread.
#348
I had a look at EBC a while back (Aidan probably remembers this) but I never got around to actually setting it up properly because my solenoid wouldn't respond to my MAX and Min DC% values. So I got discouraged and gave up.
The spark map from the basemap is knocking. So I'm not sure how much more conservative I should go.
How am I supposed to know if my AFRs are "wrong"? Do you mean if my AFRs are not matching up with my target?
Mine just looks like a hill... Dunno.
Thank You for taking the time and effort into that post Ian, I'm sure it will benefit many others as they come across my build thread.
The spark map from the basemap is knocking. So I'm not sure how much more conservative I should go.
How am I supposed to know if my AFRs are "wrong"? Do you mean if my AFRs are not matching up with my target?
Mine just looks like a hill... Dunno.
Thank You for taking the time and effort into that post Ian, I'm sure it will benefit many others as they come across my build thread.
It's possible for a base map to knock in some areas but have a ton of head room for more power in others. There's no way to know unless you dyno tune it. And yes, you'll want to fix those areas before you take it to the dyno (better yet, before you drive it again)
AFRs are "wrong" if they're not what you want them to be. If you did a run and saw it knocking, but AFR was at 14:1 instead of 11:1, you want to start by fixing the fuel rather than yanking timing.
--Ian
#349
You definitely need to get closed loop EBC working to apply this technique of tuning. Open loop won't do it, because the amount of boost depends on the energy available in the exhaust to spin the turbine, and the whole point to changing the timing to to put more of the energy from the fuel into the piston/rod/crank, rather than blowing it out the tailpipe. So you could advance the timing, improve the efficiency, but make less boost and thus see less power on the dyno if the EBC doesn't compensate by adjusting the wastegate duty cycle. You can theoretically fix this in the dyno software by dividing torque by MAP and displaying the normalized value, but I don't know of any software that does it.
It's possible for a base map to knock in some areas but have a ton of head room for more power in others. There's no way to know unless you dyno tune it. And yes, you'll want to fix those areas before you take it to the dyno (better yet, before you drive it again)
AFRs are "wrong" if they're not what you want them to be. If you did a run and saw it knocking, but AFR was at 14:1 instead of 11:1, you want to start by fixing the fuel rather than yanking timing.
--Ian
It's possible for a base map to knock in some areas but have a ton of head room for more power in others. There's no way to know unless you dyno tune it. And yes, you'll want to fix those areas before you take it to the dyno (better yet, before you drive it again)
AFRs are "wrong" if they're not what you want them to be. If you did a run and saw it knocking, but AFR was at 14:1 instead of 11:1, you want to start by fixing the fuel rather than yanking timing.
--Ian
As far as EBC. I think that is something that is just hard for me to understand. So I will either struggle for a long time before it "clicks" or I will pester someone to help me. And I always feel bad after nagging people for help. So I'm probably just going to struggle though it.
#350
Keep in mind Rev's base map was/is made for different quality of fuel. Our 91 is literally the worst and in some places I had to pull timing even from his base maps (and Scott's, and DIYAutotune's, for that matter). Ian is right on the money with that post, you should read it like 100x and then apply the knowledge
#356
Its very sketchy doing WOT pulls in 3rd and 4th in the RAIN. I almost died. I did it for science and race car
2017-02-04 v1 EBC.msl This a 3rd gear pull
v3 .msl 4th and some of 5th.
As you can see, knock looks wayyyy better. I only got one count of knock in the 3rd pull.
I did put shell premium in the car last night so I wonder if that has something to do with it.
2017-02-04 v1 EBC.msl This a 3rd gear pull
v3 .msl 4th and some of 5th.
As you can see, knock looks wayyyy better. I only got one count of knock in the 3rd pull.
I did put shell premium in the car last night so I wonder if that has something to do with it.
Last edited by icantlearn; 02-04-2017 at 03:17 PM.