Can we have a real thread about speed density?

[twdorris]

If we're going to get off on the topic of boost control, here are my thoughts...

Why not target volumetric flow rate through the turbo? I generally just want the most flow I can get from my turbo at the lowest shaft speed. That's basically "find the choke point and run that using an algorithm that targets minimal duty cycle". Now, IMO, *that* would be sweet.

Yeah, you might get 30psi one week and 40psi the next as measured at the manifold. But you know you'll be maximizing your setup under current atmospheric conditions and that's what I'm all about.

Thomas Dorris
ECMTuning, Inc.

 

[the_mork]

Build it and then when I'm out of school I'll probably buy it from you if I don't sell the car and switch platforms first.

 

[twdorris]

when I see the word pressure I see a map sensor measuring a cylinder, so please call it peak cylinder pressure (static reference) or cylinder combustion pressure, unless you actually mean the air inside the cylinder with valves open or down and has nothing to do with compression.

Like I said, I can't explain it to you any more clearly, so I'm not going to try any more.

Thomas Dorris
ECMTuning, Inc.

 

[Rob10_99]

from page 2:
Hakcenter, Yeah you get the idea, but SD is more than just rpm and VE.
TPS and load acount for a LOT of variables!

Edit: Doh my bad hak, I didnt see your Jackal tag, I'm sure you know this z

 

[jrohner]

You may get a slightly higher boost psi, but you always get the same amount of air in the cylinder(g/rev), which makes tuning easier.

I'm thinking that would make you more consistent at the track too -- your HP output shouldn't change as much as the weather changes.

 

[tkelly27]

I'm on my own computer now, was on speedaddict's.

Your boost control works well for you, and I'm not attacking you, but it's not something I would trust. Example, the temperature goes up. Now your turbo has to make more boost to get the mass flow through the inlet and to make up for the lower efficiency of the intercooler. Now with even higher boost your are at a lower efficiency. So you have the same mass flow rate, but you have significantly higher temps and pressures at the inlet manifold. This is bad news.

It is dumb to have more turbo than fuel. It always will be : D It's a newb thing to do man, that bigger turbo is a liability without the fuel to keep it happy.

G/rev is as accurate a description of pressure post valve shutting as the map sensor is. With pv=nrt, knowing g/rev gives us the number of molecules (grams) in a given volume (swept volume of one revolution). Using temperature, pressure is known. It will be the same as the map sensor value times the ve scalar. It has to be for SD to work. Both values are only truly known through the use of a wideband, or maybe an airflow sensor.

In my opinion pressure is the easier of the two to work with. It doesn't change with temperature, atmospheric pressure, volumetric efficiency, etc. G/rev will. Ultimately, in a speed density system, the only way to get g/rev is to get it through calculations based on the pressure sensors, so it can't really offer any more information than those sensors can get you anyway. In specific applications it can be useful, or it can be a total pain. In the boost control scenario, there are good things and bad things.

I think that when Hal brought up the differences between what people want out of a tuning system (setting up a VE table and changing targets vs. tuning directly in the VE table) is a really good point (post #54). This is where the g/rev vs. map discussion should be split, not what they can tell you about what's going on in the engine. Through a small amount of math either one is going to be telling you the same information.

I could see where the people at link would find g/rev convenient in how they have organized the tuning, user friendliness, and for weening people off the maf. I'm not sold, but I'm probably not the target audience.

It seems like we got off topic with the boost controller thing.

 

[twdorris]

NOTE: mVE below is manifold pressure relative VE. This will have a typical "max" around 100% regardless of boost. It indicates how much fill a cylinder gets relative to how much it could have gotten had it been pressurized completely to manifold pressure before the valve shuts.

G/rev is as accurate a description of pressure post valve shutting as the map sensor is.

Sure, but only once mVE is factor into the MAP value. As I think you said here:

It will be the same as the map sensor value times the ve scalar.

To get the same indication from MAP as you get from airflow per rev requires scaling for mVE. And once you've done that, you're just working with airflow again. You have to take mVE into account when trying to calculate how much "stuff" actually makes it into the cylinder. How much "stuff" makes it in will determine pressure for a given temperature.

Let's look at it another way. If MAP were a clear indication of cylinder pressure post valve shutting, then why would we even need mVE? We already know swept volume. If we could also assume cylinder pressure equaled manifold pressure, we wouldn't need mVE at all.

Again, I'm not saying "MAP as load" won't work. Sure it will. It even has one advantage I mentioned above. But I'd rather have the advantages of using airflow per rev. The biggest of which is that the same "load" has the same meaning across any number of different cars.

Thomas Dorris
ECMTuning, Inc.

 

[tkelly27]

You have only given that one advantage, which is really only useful if you are converting from mass air, and never changing your VE. I could see how it would give a mass air type of feel to speed density. It is user friendly like a mass air system, but that comes with the bad parts of tuning mass air.

If you dramatically change your VE and inlet temps, your g/rev will be the same at a lower boost pressure and inlet temperature. Now, theoretically, with lower inlet pressure and temps, you can get away with more timing advance and potentially less fuel. Same load, but now you might be able to go faster with it, but you wouldn't know that just looking at g/rev. You can't have a g/rev value and say "I'm at a lower/higher inlet pressure/temperature" and know what that indicates.

As I stated before, if you have exceedingly high temperatures, you will also need exceedingly high pressures. You could be at autoignition pressure and temps at the closing of the intake valve and never know it with g/rev.

With g/rev, you retain the same targets, even after you've tuned your VE to make it match. So you are now tuning two different tables to get your end tune, with one table based off of a value calculated off of the values in the first table. I find this redundant when I use jackal, and don't even use the target AFR table. You won't ever admit to it's redundancy, but I point to the fact that you can't purchase a stand alone system that will give you g/rev, and it's not because they took the easy route.

Haha, I know you aren't saying map load won't work, because you've got a load table in your own program based off of a map load table.

 

[twdorris]

If you dramatically change your VE and inlet temps, your g/rev will be the same at a lower boost pressure and inlet temperature. Now, theoretically, with lower inlet pressure and temps, you can get away with more timing advance and potentially less fuel.

See, I disagree here. In fact, that's the very core of my disagreement with this whole "MAP is great for load" thing. If you've managed to get more airflow into the cylinder with lower boost, great! But if it's the same airflow per rev, then you've got the same cylinder pressure you had before. You're just doing it with less boost. You're probably making more power as a result just from reduced pumping loss if nothing else.

But because you've got the same cylinder pressure, you've got the same tuning limitations you had before. If you were using MAP as your load variable, you'd have to retune your table. But if you're using airflow per rev for your load, you do not...because you'd be getting the same airflow per rev (and same cylinder pressure) as you had before. It doesn't care how you got it. That's the whole beauty of using airflow per rev. And it doesn't matter if airflow per rev is calculated by mass air or SD. As long as it's accurate (and that's an assumption here regardless of which implementation we're talking about), it works as an appropriate load indicator. If it's not accurate, then you've got some tuning left to do.

Adjusting timing based on inlet temps is a completely separate function.

I think we're going to have to agree to disagree and leave it at that.

Haha, I know you aren't saying map load won't work, because you've got a load table in your own program based off of a map load table.

That's a misleading statement. The only MAP value in our system is used for what it should be used for...as an index into engine VE. It's used there because VE varies by MAP. So you use it to calculate a proper airflow value and use the airflow as a proper load else where. I would not use MAP for load. It's just not right.

OK, you've stated your position and I've stated mine. Hopefully this is done.

Thomas Dorris
ECMTuning, Inc.

 

[SpeedAddict62]

OK, you've stated your position and I've stated mine. Hopefully this is done.

Thomas Dorris
ECMTuning, Inc.

I sort of hope it isn't, I'm kinda enjoying this

But seriously, this isn't pointless, in fact I think it is quite useful. Many many people use forums like this everyday, and some of them are full of some really great information (some more than others). Sometimes, you have to sort through the BS to find the good stuff, but this isn't one of those threads, and here is why:

Lots of people browse this forum, and probably a good number aren't as knowledgeable about cars as they'd like to be - that is why they are here. There are definitely a lot of Link users around, but how many actually understand the product they use? How many people were just told "get DSMLink, it's really really cool and it'll make your car go fast"? Sure, maybe it does, but do they know why? Here we have a thread with a ton of information, it may not be all truthful, but there is a ton of it. It may be a heated discussion, but no one is calling names, it is simply an argument based on what people believe. Throw all the facts on the table, the more the better. Let the people who read this forum decide who is right. Tom, the way I see it, you should take this as an opportunity to get more business; you have a chance here to explain your system, prove why your implementation is the right choice, and convince noobs of the same.

 

[dsm-onster]

Ok I'm an idiot. And this question is for the other idiots who are reading this to learn what I'll learn from you guys.

What's is the deal with needing to lable the two forms of VE? Why would you have a table referenced in the memory for total VE, anyway? To me, that would make more headaches in tuning. If one were to simply up the boost, "tVE" (total VE) would change. For two reasons: the boost pressure puts more air in the cylinder so the volume flow referenced at ambient pressure would go up (tVE goes up). And also the wastegate would close more and/or sooner (would cancel a small amount of that potential tVE increase), so it would be something that could NOT simply be corrected for with referencing the manifold absolute pressure and generating a correction factor, right?

Mainly, I would be concerned about upping my boost. And since tVE goes up when raising the boost and the table I'm altering in such a speed density setup is a reference if tVE, then I'd have to alter my table every time I change the boost. I'm not one to set and forget my boost. I like to up the boost slowly and pull back timing and tweak for more pumpgas power. That means I'd have to re tune my VE table every time I wanted to up the boost, not jsut re tune my timing table and pissibly my fuel table, so that I could use pumpgas. To me, finding the most power for a setup is also in finding the best boost setting. That means that this variable will vary frequently through out my tuning session.

Hell, I've had my setup spike at 2-3psi DIFFERENCE when running in different ambient conditions and in both times really pushign a turbo. I can't imagine how random the fuel flow would be at the tip-in an peak torque points when you have a significant difference in spike when the ecu is referencing a total VE table.

 

[the_mork]

Which system are you asking about? I think neither link nor evoSD use tVE at all so you would just keep playing with your boost as much as you want and as long as your flow doesn't go off the end of the chart you'll be fine. If it does you need a bigger map sensor and you'll need to rescale load (for evoSD) which is a process of a few minutes that you would need to do anyway if you were flowing beyond what the stock tables could handle.

 

[SpeedAddict62]

Yeah...Jackal uses a VE table referenced to rpm and boost. As long as your entire map is accurate, you can adjust the boost whatever you like and you'll be just fine (up until, like the mork said, you outrun your MAP sensor, in which case you'll have to rescale the map and get a new sensor). That will lose a bit of resolution though.

 

[dsm-onster]

^^^This is what confuses me. Does the ds-map table refer to total VE? Total VE changes with boost. So the table will show VE going up as boost goes up? Wouldn't most rather find their VE (engine ve) and have it done with? Everytime you up the boost with ds-map you have to determine the VE for that boost and plug it into the table. That can get a little dangerous when youre running an twinscroll hx40 and getting to the +35psi columns. Id like my ecu to see the new MAP input and know already how much fuel to inject.

I'm not knocking dsm-map /Jackle at all. I'm trying to understand why that method was chosen.

Yes and no. VE is an integral part of our tuning tables. Hak's use is total VE and the usage for link and evoSD is engine VE.

Which system are you asking about? I think neither link nor evoSD use tVE at all

So ds-map references total VE.

 

[toofast82]

Mainly, I would be concerned about upping my boost. And since tVE goes up when raising the boost and the table I'm altering in such a speed density setup is a reference if tVE, then I'd have to alter my table every time I change the boost. That means I'd have to re tune my VE table every time I wanted to up the boost, not jsut re tune my timing table and pissibly my fuel table, so that I could use pumpgas. To me, finding the most power for a setup is also in finding the best boost setting. That means that this variable will vary frequently through out my tuning session.

I know at least for my setup and when I look at DS-mapv2. There is a soft and hard VE table. The soft is a simple VE number based on boost only which will adjust the whole curve. The hard ve is to fine tune per rpm and boost level. The two multiply off of each other with the A/F table i think in some formula to come up with the amount of fuel to be injected. Once your hard ve is dialed in pretty consistent, you could just tune off of the soft ve by making one adjustment if needed
Switching over from maf I found the car much easier to tune on the speed density for some reason and it acts more consistent.

 

[dsm-onster]

Well when you up your boost, you have to determine what you total VE is in those columns in between, right?. . . I want to go to the limit. So I just have to take a guess when I get started in those levels? That's a gamble where the stakes are as high as the boost. I guess you could start off by putting the total VE in those cells at a rather high number to guarantee you won't have a problem and reduce them to what's proper when tuning your total VE there.

I can see now how this "boost as load" method would work alright for spikes at different ambient conditions as long as you tuned your VE for the higher boost in the spike. So that second concern of mine is somewhat remedied.

BTW, thanx Thomas and Hakcenter (don't know your first name ) for your discussion.

 

[SpeedAddict62]

Yeah, well, I really don't see you doing it any other way, this is speed density after all. Seriously though, you start with a base map, and up the boost little by little. Once you tune at low boost, you probably have a good idea what you're going to need at higher boost in the table. Like you said, you just make sure your fuel table is plenty rich (say, 10:1) while tuning. I really don't see how it makes a difference what VE you use? Even engine VE is going to change with boost. For example, maxing out a turbo with a small hotside (Mitsu, anyone?) is going to have crazy backpressure at high boost, which will inevitably affect VE.

EDIT: Better yet, why don't you come over to the DSMAP forums, become a member, and see for yourself? You could be all setup for around $200 in, why not give it a shot? We can always use another convert

 

[dsm-onster]

The other way is to have your engine ve figured out before you even jump into pushing your turbo

But I understand what you're saying. It's just that I don't need to wonder what the total ve is with 35psi boost with my V3 link. or 45psi. Engine ve does change when the turbo is being maxed out, yes. But removing the other variable of total VE surely makes me more comfortable. I'll only have a smaller amount of adjusting to do with the difference in backpressure seen. I won't have to wonder where the back pressure starts to really affect engine ve (consequently total ve) AND how that will alter what I would extrapolate where my total VE will go to with the higher boost. I don't know my new setup as well as I will 3 weeks after I've been playing with it.

. . . No offense at all. A working and effective SD system like ds-map is a real bargain.

 

[dsm-onster]

Total VE can change when engine VE does not. It's just a difference in boost well within the TURBINE's efficiency.

It is VERY important to know where I'm at on the compressor map. If I didn't I wouldn't know if I could get more out of my setup or if I could get more out of my turbo. I'll know if I need to look at other fuels or compression or if I have enough with this turbo to reach my goal based on the airflow I'm using to attain a certain trap speed or dyno number.

 

[tkelly27]

I still have a lot of questions.

Now lets say you have gone from a 2.0 with a small turbo, to a 2.3 with a big turbo. The 2.3 could potentially be at 0 (or limited) boost at the same g/rev that the 2.0 is at under boost. Personally, I like to have my tune more lean under no boost conditions to put more heat energy into the exhaust. I would then need to rearrange my VE table to match the load table values, then tune the targets in the load table, correct?

An even bigger problem I have with g/rev is the automatic limitation of airflow that the user is allowed to tune. Lets say your peak load is 2g/rev, what happens when you exceed that? What do you tune with? From what I can tell in the ECMlink wiki it seems like there is yet another table to let the ecu take a scaled guess based on boost pressure and RPM, but maybe I've got it wrong?

Also, maybe the peak is 4g/rev, that's what it looks like on the only screen shots I can access. The people flowing 2g/rev only get 1/2 a table to tune with.

You also already know airflow in the map only load situation. Like had described, airflow = fuel flow * air/fuel ratio. Because fuel flow is really only a few scalars away on the load curve, you are really looking at load in those cells anyway. Both systems (map and g/rev) rely on this fact to allow you to tune them.

 

[Ceddy]

An even bigger problem I have with g/rev is the automatic limitation of airflow that the user is allowed to tune. Lets say your peak load is 2g/rev, what happens when you exceed that? What do you tune with? From what I can tell in the ECMlink wiki it seems like there is yet another table to let the ecu take a scaled guess based on boost pressure and RPM, but maybe I've got it wrong?

Also, maybe the peak is 4g/rev, that's what it looks like on the only screen shots I can access. The people flowing 2g/rev only get 1/2 a table to tune with.

With the Flashable ECUs you can rescale the Load and RPM Axis to what ever you like.
And the axis increments do not need to be equal.

Generally with a Fuel Map, Load Axis you put smaller changes(more resolution) in the spool up area, because AFRs change more there. And less resolution at high Loads, because AFRs change less up top.

This method allows maximum use of a maps resolution.


I think this method could be implemented on the early 2G and 1G ECUs also, but I am not sure how the table axis lookup is done on them.


Edit: A table's resolution limitation doesn't have anything to do with g/rev.
What ever unit(g/rev, psi, etc.) you use for that axis will have the same resolution limitations.

 

[tkelly27]

Generally with a Fuel Map, Load Axis you put smaller changes(more resolution) in the spool up area, because AFRs change more there. And less resolution at high Loads, because AFRs change less up top.

Edit: A table's resolution limitation doesn't have anything to do with g/rev.
What ever unit(g/rev, psi, etc.) you use for that axis will have the same resolution limitations.

What g/rev do you spool up at? I typically start spooling at 0psi and keep it lean until 3 or 4 : D

That doesn't seem bad for a mass air system, because the mass air sensor is giving you g/rev. If I'm using a map sensor to determine g/rev load, when I start flowing more I have to change the load table axis, then retune the VE curve to fit that, then I can start tuning in the load table again?

That I know of, the 1g and 2g don't have axis scaling. Maybe they do, I could see it happen with the extended maps thing, only instead of scaling it by 2 you could scale it by some other fraction. Isn't there an impassible internal limit of 4g/rev though?

 

[Ceddy]

That doesn't seem bad for a mass air system, because the mass air sensor is giving you g/rev. If I'm using a map sensor to determine g/rev load, when I start flowing more I have to change the load table axis, then retune the VE curve to fit that, then I can start tuning in the load table again?

No, the VE, Fuel, Timing all have separate axis.

The VE and Timing/Fuel maps are completely separate.
You may have to fine tune the VE table if you increase boost , but I think it is minimal.

I don't run SD, so can't give any hands on info.

 

[pneumo]

That I know of, the 1g and 2g don't have axis scaling. Maybe they do, I could see it happen with the extended maps thing, only instead of scaling it by 2 you could scale it by some other fraction. Isn't there an impassible internal limit of 4g/rev though?

The 1G has some slight scaling in the timing map load axis. Load levels 1 through 5 have a difference of 0.125 G/rev between load levels. 6 through 12 have a difference of 0.19 G/rev between each load level.

Yes, it's a pain to retune the timing, fuel and VE tables each time you re-scale the map axis, which is why it's nice to have more load levels.

 

[twdorris]

An even bigger problem I have with g/rev is the automatic limitation of airflow that the user is allowed to tune.

There's no such automatic limitation. You have the same "limitation" with a MAP load. At some point, you end up on the "top" load level, whatever that is. If you need more, you extend the table. That's exactly what we did with the ECMLink tables. Here's an example:

http://www.ecmtuning.com/images/forums/v3peeks/v3preview-maxtmng.png

If those load levels become limiting, we can extend the table further. So far, nobody has complained, though. And that includes guys running 40psi+.

From what I can tell in the ECMlink wiki it seems like there is yet another table to let the ecu take a scaled guess based on boost pressure and RPM, but maybe I've got it wrong?

I just don't know how else to explain it. It's REALLY not complicated. There is only ONE table to deal with when tuning in speed density with ECMLink. And that's the VE table. Once you have the VE table dialed in properly (and there are several tools to help with that), everything else works automatically.

Because load is based on airflow per rev, it doesn't matter if you're running SD, GM MAS or a factory MAS. It's just airflow per rev and it works out perfectly as an approximation of engine load regardless of how you're metering the air.

I posted an example way back at the beginning of this thread that I'll bring up again. Consider two different engines, one with 80% mVE at 6000 RPM and one with 95% mVE at 6000 RPM. At the SAME manifold pressure, the engine with 95% mVE will cram more stuff into the cylinder before the valve closes than the engine with 80% mVE. That will result in higher total cylinder pressure and that's what you want to tune around and that's where using airflow per rev is ideal. It doesn't care how much boost was required to generate the airflow, it just cares how much air and fuel managed to get into the cylinders for this particular firing event.

Normalizing the tune around airflow per rev means I can share my map with anyone else, regardless of what they're running (GM MAS, 2G MAS, hacked 2G MAS, SD, whatever). It's all going to work out exactly the same for them as it did for me regardless of their engine's VE or whatever MAP they run.

Thomas Dorris
ECMTuning, Inc.