Pages: 1 ... 4 5 [6] 7
Author Topic: 1.8t: electronic SUV via N249 output  (Read 39985 times)
TijnCU
Hero Member
*****

Karma: +38/-2
Offline Offline

Posts: 648


flying brick


« Reply #75 on: November 05, 2015, 09:25:24 AM »

I havent logged the variables anymore, but what I can tell from the sound is that when I get off the throttle at ~0.3-0.5 bar I get a fluttering sound, and when I go into real boost it just sounds normal (pshhh  Roll Eyes) Could be the issue that I'm running it backwards though  Grin When I put on my new turbo, I'll reverse it. Not so easy because it connects on 2 different diameters, and I welded the larger diameter on to the TIP.
Logged

TijnCU
Hero Member
*****

Karma: +38/-2
Offline Offline

Posts: 648


flying brick


« Reply #76 on: December 14, 2015, 02:55:20 PM »

I am still running this setup, on standard APX values. Works great, valve is open on cruise unless requested goes above 1300mB.
Well, after driving home from work today, a quick thought passed through my mind... Did I interpret the maps wrong?
In that case, I could have mapped it for working only above 1.3 PR. I pressed the throttle, let boost come on to 0.2-0.3 bar and on liftoff I get surging sounds. Then I press the throttle full and let boost build, I lift off at 1.3 bar and sounds PSSSHH SH-SH-SH-SH (or something like that, you can make your own sounds if it suits you better).
Anyway, thats about correct thinking about what I put in those maps. I disabeled N249 control UNDER 1.3PR in stead of locking it open untill 1.3 (that was my plan). This will work okay when you have a vacuum operated DV offcourse, but with this electronical DV you're screwed with this settings  Grin
Its on 1.0 now, I'll log it this week to see if it works better now.

No updates from other users?
This is a Polo WRC with same Turbo and eSUV on the Cold Side like i build it in my car.
They have all to deal with the high peak boost at redline?
I'll log my spike as well. Have you tried to reduce the hysterisis for control and pedal % ? (GWPLDU, HDLDSUA) and maybe tweak the predictor PIDLDSUA. mine are stock APX
« Last Edit: December 14, 2015, 03:04:48 PM by TijnCU » Logged

aef
Hero Member
*****

Karma: +34/-12
Offline Offline

Posts: 943


« Reply #77 on: December 15, 2015, 12:54:57 AM »

Hi,

Here is the story:

I bought a eSUV to test and log because noone did it at this time.
I was hoping that the boost peak at redline would disappear and my turbo would have a safer life.
This high boost at high rpm will kill your turbo thrust bearing Sad

So i replaced my GFB T9301 diverter valve which is mounted between throttle and fmic (which is a big design flaw)
with a stock tfsi piston style diverter valve and with a GFB dv+.

surprise surprise... the old vacuum actuated valve will release the air much quicker which results in less boost spike.

first picture shows the moment where i release the pedal (red circle)
and all three diverter valves

second picture: old dv
third picture: piston style
fourth picture: gfb dv+

I believe the old dv is bigger (25mm like the stock s3/rs4) and thats the main reason.
I wonder if everyone with k04-064 upgrades on tfsi is facing this problem because they are running more boost and airmass like i did.

logs were done with 15hz, axis is NOT seconds

« Last Edit: December 15, 2015, 01:03:28 AM by aef » Logged
TijnCU
Hero Member
*****

Karma: +38/-2
Offline Offline

Posts: 648


flying brick


« Reply #78 on: December 15, 2015, 04:25:31 AM »

On the other hand, the S3 has this eDV and the k04-064 combo. Could it be the issue of the piston dv? Or different programming in the software possibly.. I mean our software is calibrated for n249 and not for a eDV. Mine is diaphragm oem S3, maybe there is a flow difference between those 2 as well.

A piece of my log today, its not as high def as yours but my spike seems much less. This is offcourse a much smaller turbo I'm comparing.
But what I can see, is that it is actuated well before the throttle plate is physically closed, so any spike that comes should be the effect of flow restriction. I think my charge side of the valve is also restricted (oem 180degree bend silicone, might be a bit restrictive) but the valve itself is not smaller than my old OEM vacuum DV
« Last Edit: December 15, 2015, 11:25:12 AM by TijnCU » Logged

golfputtputt
Full Member
***

Karma: +2/-1
Offline Offline

Posts: 229


« Reply #79 on: February 01, 2016, 07:29:33 AM »

I did this mod a while ago when I discovered this thread and am still playing around with things. I have all the maps related to SUV control defined in my 1.8t xdf and which pertain to ldsuas and ldsuad. I still have to fully study the FR on how they actuate and how the flip/flop resets so to be honest, I should'nt be messing around with it just yet....
but I could'nt keep my mitts off it. It seems i've activated ldsuas...pretty correctly... because my SUV is open most of the time and closed only under medium-large pressure differentials (boost not pictured). SUV closes somewhere between 4-5.25 PSI (and obvs WOT). Plus ldsua seems to follow ldsuas more closely (not pictured)
Having this going is good because it appears ldsuad is def not doing what it should be.

As you can see from the screencap, ldsuad becomes true only under two out of the three pictured gear shifts (throttle plate slams closed=DV should be open, ldsuad should be true)

Once I fully understand the SUV FR I think I'll get what I need it to do but it def doesn't seem right as off right now.

FYI, took a loose look at mpg's with my elm327 over a long trip and I was getting about 28 which isnt bad. gotta deactivate ldsuas (SVD map) once i get ldsuad doing what its supposed to so I can measure mpg differences.

Edit: screencap part throttle driving

How does B_ldsua determine if it uses B_ldsuad or B_ldsuas as the profile it follows? The FR says B_ldsua can be set if either B_ldsuad OR B_ldsuas is true

Also, when determining variables over time, such as: (current boost error-previous boost error) what is the time lapse for determining these two points from the ecu's perspective? .1 second? less?
« Last Edit: February 04, 2016, 11:10:54 AM by golfputtputt » Logged
golfputtputt
Full Member
***

Karma: +2/-1
Offline Offline

Posts: 229


« Reply #80 on: February 07, 2016, 08:31:52 PM »

 Smiley success

I think this thread is dead as fuck but whatever, still gonna reply. If this is all redundant and is all common knowledge that I'm repeating, my apologies.

So, I found this: http://forums.vwvortex.com/showthread.php?6881579-Why-you-should-keep-the-N249&p=84617263&viewfull=1#post84617263

It's the original post for all my gathered information outside the FR.

I also condensed it into a pdf spread so if the pics ever die, it's on hand if anyone ever wants it. I understand how it operates and how it can be applied to an eDV (eSUV) for the most part. I'm still messing around with it so I don't quite fully know what all the parameters are capable of yet. I will find out though.

Ok, here we go:

I outfitted the TFSI DV to my 1.8t, which to my knowledge can be done also to the 2.7t (a pair of DV's and n249s -->2 eDV's) which means this mod can be for the s4 guys too. (aside from the difference in software) It just makes a hell of a lot of sense to me because you can now monitor exactly and precisely when your DV opens and closes instead of relying on a partial electric signal and a partial pneumatic signal, where you cannot record the pneumatic signal in a log.

So as most of you know, in ME7.5 B_ldsua (DV actuation) is set true by either B_ldsuas or B_ldsuad. B_ldsua WILL follow B_ldsuas over B_ldsuad if both are set true or false at separate, conflicting times. I found that out through experimentation. I used this to my advantage as B_ldsuas is much more of a safe option. It's conditions set to true the majority of times and resets the flip-flop only under certain conditions. This way, your turbo is always bypassed unless you're under WOT (under the tt225  SVDLDUVS values, all over 1). This gave me a chance to truly play around with the main operating condition- B_ldsuad whilst B_ldsuas remained my safety net.

The n249's main operation is B_ldsuad, this is considered MAIN because under most files, B_ldsuas is always false and does not come into play. It is also considered MAIN because its conditions set the DV to closed most of the time and open under certain conditions, the opposite of B_ldsuas.

On a 1.8t or 2.7t, if you are removing your n249('s) and pneumatic DV and replacing it with an eDV from a TFSI, it will not operate as it should! The eDV will not open and close at the exact times that the pneumatic DV does. This is a problem as every time you shift and the throttle plate slaps closed due to deceleration, the DV needs to relieve back pressure from the turbo. If it does not, the turbo will "surge."

B_ldsuad's threshold needs to be lowered in order for operation to mimic what the pneumatic DV does; opens under every shift.
After some experimentation (and attempted calculation which didn't really go as well as I'd hoped) I found that lowering the entire table from mostly 200 hpa to 50 hpa accross the board did the trick for the most part. As you can see from the 2nd log PNG, B_ldsuad is set to true for every shift excluding 1. I'm still messing around with it to actuate for EVERY shift completely.

B_ldsuad is actuated if ALL of it's conditions are set true. In short, pedal gradient needs to be less than GWPLDU (in deceleration, negative value) and a boost error equation is greater than KFSDLDSUA. The third condition is always true. As a general rule, GWPLDU sets the sensitivity of B_ldsuad upon deceleration. KFSDLDSUA determines when B_ldsuad is set to true. This needs to be the focus for proper DV operation (if B_ldsuas is not to be used as the primary function)

I believe that you can tune SVDLDUVS, HSLDSUA, HSSLDSUA and LDUVRS together to utilize B_ldsuas in tandem with B_ldsuad so it actuates only under consistent cruise conditions and B_ldsuad operates for the majority of around town driving. I think this way, superior gas mileage can be achieved without sacrifice to part throttle power (via allowing the turbo to build boost).

The PDF attached has been pivotal to understand how the n249 (and subsequently, the DV if an eDV is used) is controlled. My synopsis is that the n249's purpose is to support DV opening function under certain conditions and not all conditions, thus for adaptation to an eDV parameters need to be modified.

Attached files:
B_ldsuad measuring: B_ldsuad (B_ldsua subsequently) missing shifts
B_ldsuad measuring-6: B_ldsuad (B_ldsua subsequently) operating for all (99%) of shifts

Once again, if all of this is common knowledge and I'm just spouting redundancies, please, let me know and I apologize.
Hopefully this will help someone out who is attempting this mod.
« Last Edit: February 07, 2016, 09:16:33 PM by golfputtputt » Logged
aef
Hero Member
*****

Karma: +34/-12
Offline Offline

Posts: 943


« Reply #81 on: February 10, 2016, 07:19:01 AM »

Hi,

great post, thank you very much.

So if i understand you correct, you only changed one of the three inputs which is the 200hpa thing map?
Logged
golfputtputt
Full Member
***

Karma: +2/-1
Offline Offline

Posts: 229


« Reply #82 on: February 10, 2016, 10:48:07 AM »

yes, correct. only KFSDLDSUA (200 hpa table)

 I will be changing GWPLDU pedal percentage to check it's effect on B_ldsuad with KFSDLDSUA threshold changes today. I will also be logging to make sure B_ldsua is indeed following B_lsdsuad to the tee when B_ldsuas is deactivated, forgot to confirm that.

I am also getting a bit of flutter at particular times during gradual deceleration so there is still backpressure not being relieved. I'm going to start by lowering the KFSDLDSUA 10 increments until it dissapears. GWP may help it go away too. I'll isolate the variables with scientific method to make sure I don't mix up results.

edit: I lied, thought about it, viewed logs, listened to the car. Duration of B_ldsuad needs to widen also
« Last Edit: February 10, 2016, 05:00:19 PM by golfputtputt » Logged
golfputtputt
Full Member
***

Karma: +2/-1
Offline Offline

Posts: 229


« Reply #83 on: February 10, 2016, 05:45:28 PM »

ok, more success.

I lowered KFSDLDSUA 10 hpa more and that seemed to be the magic number for catching all the shifts. (40 hpa accross table)
Another thing that is of note is the duration that B_ldsuad is set to true which is controlled by the reset condition of the flip-flop.

This is what people interested in performance should be concerned about as this controls how quickly the DV closes before the shift into the next gear, after boost fully ramps down. Once boost accumulation is desired again, you want that DV closed as quickly as possible. I found that I had a faint sound of fluttering at the tail end of a shift: backpressure.

So I raised the reset time: THLDUVD by 1/2 a second (2 seconds OEM value). This holds B_ldsuad true for longer. If you want a quicker DV function, perhaps 2.25 s is for you, perhaps leave it at 2 s, depends on what you're willing to compromise.

THLDUVD is not complex nor is it pressure differential based, so, if you have higher than stock boost levels, this may need to be raised. If you are running a shorter intake tract, this may have to be raised. If you are running a different turbo than K03/K04, this will probably have to be raised. If you're running a ball or needle bearing small turbo, something that spools up and down very very quickly, this may have to be lowered....? It depends, take a look at B_ldsuad and its relation to pressure decrease, B_ldsuad should turn false when boost is flat. PNG example: THLDUVD not raised but you can see boost is still ramping down when DV closes.

update edit: I zero'ed KFS as I was still missing shifts. I widened my log route and threw in hills and other scenarios to make sure all shifts were covered. Things seem alright now but I don't like that KFS is zero'ed. At a later date, I'm going to play with the multiplier and hopefully repopulate KFS.
« Last Edit: February 10, 2016, 08:01:40 PM by golfputtputt » Logged
golfputtputt
Full Member
***

Karma: +2/-1
Offline Offline

Posts: 229


« Reply #84 on: August 08, 2016, 09:14:05 AM »

For anyone trying this. There was a really easy error with the hardware that i made.
The oem mk4 pneumatic diverter valve has the piston/diaphram facing the suction side of the turbo. This is how it was designed.
The oem mk5/6/7 solenoid diverter valve has the piston facing the pressure side of the turbo. This is how it was designed.

The inlet/oulet sides need to be flipped. It will work without flipping it, unless you try and pair it with Go Fast Bits DV+, the DV+ will not work if you don't flip it. The general check for the flip is: the solenoid body should be facing sideways, piston face facing the pressure pipe.
Logged
Cookiez
Newbie
*

Karma: +0/-0
Offline Offline

Posts: 21


« Reply #85 on: November 08, 2016, 05:15:17 AM »

Has anyone been running this over a period of time?

Would love to ditch my DV and go eDV
Logged
Colt45
Jr. Member
**

Karma: +0/-1
Offline Offline

Posts: 27


« Reply #86 on: November 25, 2016, 10:20:33 AM »

Are there any logs of running the eDV without modified firmware?

golfputtputt says it will open too slow, but that doesn't really make sense to me. I mean the whole point of existing pneumatic+N249 system is to open the DV as soon as throttle is starting to release (and not have to wait for vacuum to build in manifold), right? If not then what is the point of the N249?!

thanks for any comments. I was thinking I would install eDV in my as of yet stock 1.8T
Logged
Colt45
Jr. Member
**

Karma: +0/-1
Offline Offline

Posts: 27


« Reply #87 on: December 03, 2016, 12:36:18 PM »

bump )Smiley)
Logged
prj
Hero Member
*****

Karma: +214/-87
Offline Offline

Posts: 2912


« Reply #88 on: December 05, 2016, 01:04:11 AM »

So i replaced my GFB T9301 diverter valve which is mounted between throttle and fmic (which is a big design flaw)
with a stock tfsi piston style diverter valve and with a GFB dv+.

surprise surprise... the old vacuum actuated valve will release the air much quicker which results in less boost spike.

Just to bring some clarity to this thread. The conclusion is flawed, as is the testing method.
If your DV was post-cooler and then you moved it to pre-cooler of course on the pre-throttle sensor you will see higher pressure, but it makes absolutely no difference.
This is because it is not where you should be measuring it in the first place.

The important place where to measure is between turbo and intercooler.
Logged
Cookiez
Newbie
*

Karma: +0/-0
Offline Offline

Posts: 21


« Reply #89 on: January 24, 2017, 04:43:18 AM »

so its been a bit of back and forth, could someone elaborate what exactly needs to be changed and whats concluded to be the correct direction of install in regards to the valve.

Would love to ditch the vacuum reservoir and solenoid, but a delete isnt an option.
Logged
Pages: 1 ... 4 5 [6] 7
  Print  
 
Jump to:  

Powered by SMF 1.1.21 | SMF © 2015, Simple Machines Page created in 0.023 seconds with 17 queries. (Pretty URLs adds 0s, 0q)