Help me understand my engine map (1.9L ALH) (Or I intend to do stupid things)

robnitro said:

X values are Kelvin. 273K= 0 C. So you take X/10= get kelvin. Then -273 gives you C. Many other maps run this way, but sometimes vagedcsuite does the recalculation for you and shows C. In those cases, right click and choose edit axis. See how the values are in the K format!

Click to expand...

It slipped my mind that the factor was a division thing, I was maybe thinking in terms of an offset. Was thinking "gee, 500 degrees difference (c or k), sure doesn't seem like a plausible range"

So, RPM, degrees BTDC, and temperature.

Cool and thanks!
My maps were calling for up to two degrees more timing than the limiter would allow no wonder I wasn't seeing much change. I've been adjusting the maps very cautiously, but also with little fear, as looking at every other diesel, their timing is much more advanced than ours. On all of the LD diesel pickup engines I see it is around 8 degrees at idle, on up well into the thirties at 4k RPM.

Some nutso cummins guys are running two bosch P pumps with one set up normally at 20 degrees or so, and the other at 40 degrees, with the governor set up so it starts delivering fuel around 3k RPM, kind of a crude timing advance where their engine had no provision.

Wow, I only now realized that my brain was automatically reading 2729 as 272.9 without any mental triggers flying up while typing the conversion.

Absolutely right, divide by 10 first.

[486] said:

It slipped my mind that the factor was a division thing, I was maybe thinking in terms of an offset. Was thinking "gee, 500 degrees difference (c or k), sure doesn't seem like a plausible range"

So, RPM, degrees BTDC, and temperature.

Cool and thanks!
My maps were calling for up to two degrees more timing than the limiter would allow no wonder I wasn't seeing much change. I've been adjusting the maps very cautiously, but also with little fear, as looking at every other diesel, their timing is much more advanced than ours. On all of the LD diesel pickup engines I see it is around 8 degrees at idle, on up well into the thirties at 4k RPM.

Some nutso cummins guys are running two bosch P pumps with one set up normally at 20 degrees or so, and the other at 40 degrees, with the governor set up so it starts delivering fuel around 3k RPM, kind of a crude timing advance where their engine had no provision.

Click to expand...

Some PD engines' maps I see with 30 degrees or more, but without knowing all the maps involved (as in a complete flow chart) I don't think we would know if there's more "variable event" correction maps. I guess a non technical way would be to do logging again and again.

PD does soi differently, from what I've read. Plus on the dyno someone has mentioned that there was no more power past 17 or so BTDC on VE. A bit higher is fine, because it reduces EGT.

Someone on ecuconnections (kentdi's post here- and more info later on -- http://www.ecuconnections.com/forum/viewtopic.php?f=2&t=16272&start=1850 ) found the best power while tweaking the soi on his PD. He found that more advance for less load is actually beneficial. Some had concerns on PCP, but really, PCP isn't excessive at part load, but full load- which is unchanged from my existing map. It's been working well for me. A bit more diesel sound at idle and part load, but response feels better than the stock which is very low (theorized due to NOx emissions).

robnitro said:

Plus on the dyno someone has mentioned that there was no more power past 17 or so BTDC on VE.

Click to expand...

That's right around the limiter map's values, did they know that that limiting map is there?

No idea, I think malone said it too and he definately knows the limiter.

Also other tunes go to around 18-19 max, and one european tune which was highly recommended to be strong and smokefree (tuv testing), maxed at 17.6.

Oh and btw, the soi map I posted has an oddly common way of timing like the older mechanical diesels: flat timing at low/mid load and rpm, but more at higher rpm. The idea is that low fueling has a smaller injection window, but more ingnition delay (less heat).e

You don't always have to use 100% of every map. In fact it is good practice not to.

Look at the altitude based boost limiter (3d limiter) for most applications other than factory ALH. It cuts into the boost basis map pretty substantially at higher altitudes. That's the proper use of a limiter.

Unfortunately most tuners are too lazy to calculate those things so they just do stuff like make the torque limiter so big that it never comes into play. Or bump the torque limiter by some flat percent, or to make a "nice shape". This can cause some strange things to happen because the factory may NOT have calibrated the whole entire map(s) that the limiter effects, only the active part of it (under the limiter!) (example: look at EDC15 SDI pump volts vs the torque limiter)

So, I just realized that I'm going to be at a dyno day in a couple weeks. This means car doesn't move but gets to have a full load applied to the engine. If possible, but no promises since I'll be wrenching my car while on the dyno, bu I'll try and drag my 'scope along to do a data capture of the 3rd injector from 1500 RPM to 4500 or 5000. In order for this data to be useful, I'll need to see if the shop can give me RPM output scaled over the duration of the run and time stamped else I won't be able to correlate RPM to the waveforms captured.

Why do you want 3rd injector data? Just log with vcds?

Digital Corpus said:

This is why

Click to expand...

I understand now...

Just did the hammer mod without even opening my IP.

Was having a surging return to idle ever since putting in pp764s, it was idling at 1.6-2 mg/s, knew what I had to do, just didn't want to do it.

Tweaked the pump voltage map so that it idles in the middle of the "idle" section of it. It loped something fierce when I changed it way too far to begin with (it was jumping between 0mg/s and 25mg/s). Got it dead on at 4 mg/s on the third try.

Also been scaling the right side up some to get more fuel. You can get a lot more outta a 10mm pump by doing that, and if you do it in a somewhat linear manner you can compensate in the smoke maps to keep the smoke down. Right now my smoke maps are "set to" 30:1 or so with a light haze wide open.

Is the N75 hysteresis map known?



I adjusted my boost request maps such to get exhaust braking from the vanes, and it works and all, but now I'm fighting the vanes staying closed on light throttle. Dip into 0mg/s fueling for a moment in traffic and the computer requests 2 bar abs. of boost as it should, but roll into the throttle gently and as the fueling comes on the EMP goes up around 15 PSI, IMP up around 1.5 bar absolute and stays there until you give it enough throttle to reach the previous request of two, then it opens the vanes wide and continues along with whatever the thing makes with the vanes wide open (the compounds together make around 7 psi EMP and 5 psi IMP at cruise with the vanes wide open).

I've even tried screwing with the N75 map to get it to open the fawkin' vanes without boost being made, but it still does it.

Make special note of the X axis, when there's any fuel going in, it should be opening up the vanes, but still it'll scream along at 15 PSI EMP, 7psi IMP until you stab the throttle enough to reach the request that was made for the zero fueling which has long since passed.

There's gotta be a hysteresis value or map somewhere that I've gotta get with the times and speed up.

Yes. There are N75 hysteresis map(s). In your ALH it is the IQ switch between the using the pure N75 map passively and active regulation. Below the hysteresis, you will use the N75 map ONLY to determine the vane position. Above the hysteresis, it is a blend of N75 and PID gains to determine the vane position. (oversimplification: In reality there are multiple operating regions/modes but good enough for purposes of this discussion.)

On your 2002, the turbo hysteresis maps are at 0567D6 and the next similar map. Axis is RPM, values are mg/r. Make it smaller numbers for big injectors, if you wish to use active boost regulation at lower loads. Matters more as the injectors/pump get bigger. If you are using only sprint 520 with 10mm you probably don't care. In fact, you can probably get by without considering this unless you have issues like those mentioned. Also one reason why you don't want to set idle IQ too high, or raise the idle too much.

For max usable pump voltage, it's good to have 4 iq or so use 1.4 volts or so on the map. Then hammer mod to make 4 iq have that 1.4 volts.

I don't think requesting boost at 0 is helpful even for keeping boost between shifts. On a single turbo VNT17 it hasn't hindered between gear boost to set 0 the same boost as the next column. Too closed vanes at 0 iq might cause a choke and actually impede flow, but I am not sure of your setup.

Here are the maps that I have collected Ein and aus is start and stop for PID use iirc. Max and min are max vane min vane position that pid can use.

EDIT: If your ecu doesn't have the maps here, you can do something to find out for other ecu's of the same family (PD/ALH).
Set vagedcsuite or winols to show addresses in hex.
Use a hexadecimal calculator (or windows calc set to programmer mode) to find the difference between the map you want to find and a known common map. Ex: n75@ 5b28c and you want to find aus 5b20a.
5b28c-5b20a= 82 (hexadecimal)
Now let's say your ecu uses 6f02a for N75. Take this difference (82) and apply it to your address.
6f02a-82=6efa8 - which you can look for a similar map around that address. Usually vagedcsuite will have the unknown maps down at the bottom.

ALH CP
D 5adf4
DT1 5aed2
P 5B1D2
MIN 5AF16
MAX 5B1D2
SVBL 5AE7C
svbl 5AD92
TargetBoost 5af80
N75 5b28c
ein 5b236
aus 5b20a


ALH GN ecu
I 563d4
D 564A0
DT1 564E4
P 5679E
MIN 56418
MAX 5645C
SVBL
TargetBoost
N75 56858
ein
aus

KERMA said:

Yes. There are N75 hysteresis map(s). In your ALH it is the IQ switch between the using the pure N75 map passively and active regulation. Below the hysteresis, you will use the N75 map ONLY to determine the vane position. Above the hysteresis, it is a blend of N75 and PID gains to determine the vane position. (oversimplification: In reality there are multiple operating regions/modes but good enough for purposes of this discussion.)
On your 2002, the turbo hysteresis maps are at 0567D6 and the next similar map. Axis is RPM, values are mg/r. Make it smaller numbers for big injectors, if you wish to use active boost regulation at lower loads. Matters more as the injectors/pump get bigger. If you are using only sprint 520 with 10mm you probably don't care. In fact, you can probably get by without considering this unless you have issues like those mentioned. Also one reason why you don't want to set idle IQ too high, or raise the idle too much.

Click to expand...

Wow, thanks for the information.
Looking at the maps, looks like below 1200 RPM it won't open the vanes more than 25% no matter what boost you make (if it is truly only using the N75 map) In the area I'm having trouble with it seems as though it should be using the N75 map exclusively, but the thing still has the vanes shut even when the map should command 60% duty cycle...
If it were the other way around where the N75 map is ignored below the fueling number in the hysteresis map and PID control is used it would make a bit more sense to me, as below 1200 RPM the n75 map isn't filled in with anything other than 25%, making me think that that region just isn't utilized so they didn't bother to fill in values. That or they would prefer more vane travel in that rev range and wished for them to be able to close further...
Set it up with values around 4-5 (hundred, without the factor) in the troubled range and stil it has issues though they are reduced by quite a bit.
Going to have to watch it a bit more, just took it for a quick test drive.

robnitro said:

For max usable pump voltage, it's good to have 4 iq or so use 1.4 volts or so on the map. Then hammer mod to make 4 iq have that 1.4 volts.
I don't think requesting boost at 0 is helpful even for keeping boost between shifts. On a single turbo VNT17 it hasn't hindered between gear boost to set 0 the same boost as the next column. Too closed vanes at 0 iq might cause a choke and actually impede flow, but I am not sure of your setup.
Here are the maps that I have collected Ein and aus is start and stop for PID use iirc. Max and min are max vane min vane position that pid can use.
EDIT: If your ecu doesn't have the maps here, you can do something to find out for other ecu's of the same family (PD/ALH).
Set vagedcsuite or winols to show addresses in hex.
Use a hexadecimal calculator (or windows calc set to programmer mode) to find the difference between the map you want to find and a known common map. Ex: n75@ 5b28c and you want to find aus 5b20a.
5b28c-5b20a= 82 (hexadecimal)
Now let's say your ecu uses 6f02a for N75. Take this difference (82) and apply it to your address.
6f02a-82=6efa8 - which you can look for a similar map around that address. Usually vagedcsuite will have the unknown maps down at the bottom.
ALH CP
D 5adf4
DT1 5aed2
P 5B1D2
MIN 5AF16
MAX 5B1D2
SVBL 5AE7C
svbl 5AD92
TargetBoost 5af80
N75 5b28c
ein 5b236
aus 5b20a
ALH GN ecu
I 563d4
D 564A0
DT1 564E4
P 5679E
MIN 56418
MAX 5645C
SVBL
TargetBoost
N75 56858
ein
aus

Click to expand...

The idea is to choke things up on overrun to act as an exhaust brake. It is making the programming wig out and chokes things up when it shouldn't though, so more tuning is required.
Thanks for the further info, I'm going to look into it now and edit into this post what I learn.

ETA: my file has many of the maps you specify in the bottom set, the P, I, D, D1 and from my research these maps all seem to have axes in fuel consumption liters/hour, with a factor of .01 and then the values are just the control values which don't really have a unit.

Reading up on the PID control of these motors is going to take quite a while to sink in, meanwhile I'm probably just going to turn off the exhaust braking boost request where I set my cruise (below 3k RPM) and do some more learning on the PID control once I have my turbos swapped out with something more permanent (larger ones have been on my shelf for months) so that I can start down the line of making a proper N75 precontrol map and working out from there.

install a vacuum gage to watch the N75 vacuum to the actuator. thats how I have my beetle set up. 3 gages, boost, egt, vacuum. very educational.

Then know where your stop screw is set in relation to vacuum and you will know the relative position of the vnt in relation to the stop screw.

PID essentially controls the gain applied to the "correction signal" so to speak. Negative deviation applies a positive gain, and vice versa. There are three gains, large gain + large gain - and small gain. Small gain is when you are close to the boost setpoint. You can set the distance from the boost setpoint for each of these, for all the PID components separately. Study what the factory does with the various turbos/power levels/etc, and consider what things affect the turbo response in a relative sense as you do.

Have fun going down the rabbit hole!!!

KERMA said:

install a vacuum gage to watch the N75 vacuum to the actuator. thats how I have my beetle set up. 3 gages, boost, egt, vacuum. very educational.

Then know where your stop screw is set in relation to vacuum and you will know the relative position of the vnt in relation to the stop screw.

PID essentially controls the gain applied to the "correction signal" so to speak. Negative deviation applies a positive gain, and vice versa. There are three gains, large gain + large gain - and small gain. Small gain is when you are close to the boost setpoint. You can set the distance from the boost setpoint for each of these, for all the PID components separately. Study what the factory does with the various turbos/power levels/etc, and consider what things affect the turbo response in a relative sense as you do.

Have fun going down the rabbit hole!!!

Click to expand...

I've got a water well pressure gauge on the EMP, a SUN diagnostic vacuum/pressure gauge on the interstage boost, and a chinese junkyard boost/vac gauge on the IMP. So, there's already three gauges semi-permanently shoved into the cupholders above where the radio used to be. haha, I've been meaning to get a couple more plumbed in there for the interstage exhaust pressure, and another for the N75 vacuum, then hook up the fluke to the N75 to see real time duty cycle, but just haven't quite got around to it yet. I'm pretty sure the atmospheric turbo's exhaust side is somewhat undersized, but that's neither here nor there.

From what I'm reading it isn't so much gain on all three, it seems to be more or less large gain, small gain, and smoothing for the P, I, and D bits respectively. Do the P first, until it gets to where adding more gives diminishing returns, then go for the I, until it hunts around the value you wish for, then bring on the D until it doesn't hunt around the target.

Ein and aus maps will show you at what IQ the pid will be in use. They are close, so change them similarly. ex: 30/29 to 35/34 or what not
If let's say the fueling limit at 1200 is set to 35/34, anything above 35 IQ will use PID, and in that case it goes by the boost request map, which may command the vanes to close to meet boost level. The problem is to find this fine line for your "exhaust brake" setup, so that it can transition smoothly to the low fueling numbers.

What do you have set up for boost request at the lower columns for the problematic rpm/loads?

Question... What EXACTLY does IQ control, especially at idle? Since it's a fuel throttled engine adding fuel makes it go faster, removing fuel makes it go slower. No boost at idle so no way to alter the amount of air. All you can really change is timing?

At idle, the IQ number is just showing how many volts needed for the ip to supply fuel to keep idle (quantity adjuster). It's not really direct, you can make IQ higher or lower w hammer mod or adaptation, but still same fuel needed for idle.

As you give throttle, now you ask for a specific fueling which means more volts to qa, more fuel.
If there is not enough air smokemap will limit fuel.

Without boost, like an sdi engine, you can run up to around 30 mg/s IQ.

[486] said:

it isn't so much gain on all three, it seems to be more or less large gain, small gain, and smoothing for the P, I, and D bits respectively. Do the P first, until it gets to where adding more gives diminishing returns, then go for the I, until it hunts around the value you wish for, then bring on the D until it doesn't hunt around the target.

Click to expand...

you are misreading what I posted. SO I guess i need to take the time to be more verbose.

There are Three "sub-components" to each PID component (P, I, D, DT1 are what I'm calling components).

There is a directional as well as a magnitude attribute to each of these. Positive direction, negative direction. Small gain, large gain.

Whether it's a small or large gain is determined by how far apart actual boost is from requested. Then once the decision is made whether it is a large or small gain, and if it should go in the positive or negative direction, the appropriate set of k factors, etc are chosen for the actual calculation of the instantaneous gain.

1) large positive factor (when actual boost is << requested boost. PID wants to push it higher, harder = larger gain factor in the positive direction to strive to push boost higher, get it?)

2) Large negative factor (when actual boost is >> requested boost. PID wants to push it lower, harder = larger gain factor in the negative direction)

3) small factor (when actual boost is very close to the requested boost)

SO you have to "tune" these 3 conditions separately and independently.

1) when requested boost is much less than actual (large underboost)
2) when requested boost is much more than actual (large overboost)
3) when requested boost and actual boost are very close (let's call it fine tuning)

or you can do like most tuners do and just leave all that stock, and fiddle with the N75 duty cycle basis map as the pinnacle of "the value of custom tuning" retune retune retune and wonder why it worked yesterday but not today.

robnitro said:

At idle, the IQ number is just showing how many volts needed for the ip to supply fuel to keep idle (quantity adjuster). It's not really direct, you can make IQ higher or lower w hammer mod or adaptation, but still same fuel needed for idle.

As you give throttle, now you ask for a specific fueling which means more volts to qa, more fuel.
If there is not enough air smokemap will limit fuel.

Without boost, like an sdi engine, you can run up to around 30 mg/s IQ.

Click to expand...

I guess where I get confused is when I see people talking about changing their IQ values to smooth out their idle or similar... Trying to understand exactly what is being changed.

It's just a reference at idle and when the pump is at the extreme end, it can make it difficult for the pump to quickly adjust fueling. Its a control issue when qa is too low there isnt too smooth of a transition.

Kerma, yeah, the min and max n75 values are where I reduced my spikes and used the pd150 p i d values.
Some alh tuners also don't seem to go past "51 mg" indicated (not actual, of course w nozzles/ extra voltage) , so the pid being based on fueling is off.

486, take a look at min/max maps too!

robnitro said:

486, take a look at min/max maps too!

Click to expand...

Looked, not all that usable for my application, it'd affect drivability elsewhere in the fueling range. I could switch the axis to RPM from fuel consumption (l/hr*.01) but then I'd have to redo all the calibration from scratch in the other maps.
Learning a lot on this stuff, but the more I read the more I see that I'll probably be going back to no exhaust braking. The options for tweaking it just are not acting right so far. That hysteresis map (567D6) should have it working purely from the N75 precontrol map if what's said here is how it is working, but it quite clearly is not with how it is acting.
You bring up the "ein and aus" maps, near as I can tell, the hysteresis map from earlier is "aus" and "ein" is directly after it at 56802. The values are very close, with one mg/str difference between them. I feel like that difference is a blending region where it takes something between the two values the pid controller is throwing out, and what the precontrol map puts out.

Anyways, the "ein" and "aus" maps you talk about weren't given in your list for the ECU I seem to have (the one at the bottom), but I'm fairly confident they are at 567D6 for "aus" and 56802 for "ein"

"aus" means something akin to "off, out, finished"
"ein" means one. One is "on" in binary terms (and on my power strip's switch )
so, in 567d6 something's going off, and in 56802 something's coming on.

Is that something PID control? I sure wonder because if it is, then it'd be moving the actuator based entirely in the precontrol map where I'm having issue. At the range of 2000-2500 RPM in the aus map, it is 8mg/str, and in the ein it is 9mg/str, but I'm having pid control of the actuator where I don't want it all the way back at 6mg/str

min max is the min and max n75 values for the pid control to use. Remember it's inverse of the map... a low value means vanes open, high - vanes closed. The n75 is like this too but vagedcsuite shows it "easy mode" where ex 25% is vanes closed and so on.

Maybe you should raise the ein and aus (keep them within 1mg like they were) for the range you have an issue in. It may be that you are very close to where the pid kicks in and it can't fully move to the precontrol n75 values

robnitro said:

Click to expand...

How are you estimating wheel-slip if you only have the VSS and RPM as inputs?

It seems you would need a different LC map for different surfaces based on trial and error. If you could input actual road speed from the rear ABS sensors then you could do real traction control....

QuantumRallySport said:

How are you estimating wheel-slip if you only have the VSS and RPM as inputs?

It seems you would need a different LC map for different surfaces based on trial and error. If you could input actual road speed from the rear ABS sensors then you could do real traction control....

Click to expand...

If you flash it with the BD or CP letter part number flash, you can have up to 3 switchable tunes, where you switch to that flashbank with VCDS coding.

As an option.

Yeah, you end up having to compromise with a set %. Plus the map is kind of slow to react if revs and speed climb fast in 1st, when I had it at a close 10%, I sometimes would get power cut in 1st at 4k+

It would be nice to have the abs tcs set up to allow slip but I havent heard anybody being able to do this. Im not sure how it could be modified, because the abs unit is what tells to ecu to cut power.

It doesn't cut power hard or apply the brakes (over 25kph iirc?) like stability control does, but makes almost no slip. It's much quicker to react than the lc map, though!