After a few tests and after Gintis advice, here is what i have now:
PID:0/0/0 Anytrim: OFF
Boost DC: 77% from 3000rpm---> gives 290kpa @ around 4400-4700rpm and holds. on 3d gear, on 4th it starts to become a bit higher on higher revs.
KPA target: at the time, set to 290kpa from 3000rpm onwards
Now my problem is the fact that the turbo can give me 310kpa easily on 3d gear from 4000rpm. So basically from 1.1.27 to 1.1.81 firmware i got more lag!
2 choices:
1) try using 85%DC from 2500-3500rpm then drop it to 77% to avoid spikes and see if i am getting rid of the lag?
2) Start tuning the PID--> I presume the P would allow for the solenoid to pick up signal to get to the target faster.Then the I will try and accelerate the rate of correction towards the error from the target, and finally the D will try and avoid the overshoot.
If that is the case, i usually use a large value on P and if it oscillates, i half it. Once its ok, i introduce the I with a small value, and if not ok, then bring in some D which usually is higher than the I.
Can i ask what is the most common PID settings that others have been using that seems to work? Esp if it is an audi!
thanks
vasilis
Here are some more results.
1) tried elevating the refDC higher, and it didnt work.
2) Introduced PID with various settings and settled for: 60/8/28. This tends to prevent overshoot and monitor boost as best as possible.
HOWEVER: on megatune, i was seeing boost much sooner still! and once i started increasing the I from PID, it started getting faster, but overshot! From 12 going to 14 on I, it wasnt working. Tried increasing D to compensate, nothing happened.
here are 2 photos from megatune and vemstune where you can spot the difference. Although in both occasions you see DC starting from 100 then tailing off, in megatune it comes up much faster, whereas in vemstune it doesnt require more than 78DC to keep 300+kpa after 5000rpm. which is ok, but the problem is the spool here.
can anytrim help on this matter? on megatune i was using boost offset of 8 in order to get my kpa target.
3d gear in both instances
VEMSTUNE
(http://i117.photobucket.com/albums/o59/vasp3690/vemstuneboost2.jpg)
MEGATUNE
(http://i117.photobucket.com/albums/o59/vasp3690/megatuneboost1.jpg)
something is diffrent now(mat is 5C higher).
Try to change inj angel, to 260-300deg.
how will that affect anything?
im a bit worried changing the injector angle. I have it from Marc Swanson as 430 on low revs and around 500 on the rest
can you explain a bit what it does and how it can help?
by the way, on megatune, MAT never affected the spool, only the target boost. It was always boosting less on hot weather.
after a few more tries, it appears that with elevating the P and keeping I and D stable at 7-8 and 28 respectivelly, the boost seems to be coming up better.
i would still want to know more about the injector angle though.
So boost comes later? There is not so much what we can do to make boost comes earlier or later. Mainly it is upon to hardware.
You can prove it this way: 1) set overboost protection close to max acceptable value 2) disconnect all hoses (hose coming from N75) from W-gate. and 3) try to log 5th gear WOT until boost hits limiter. You will get maximal spool-up what is possible with your hardware with no influence form electronics.
430...530deg inj angle means injecting on closed valve before it opens. It is how mainly OE does. It's OK.
Gints
Gintis,
what you are saying is not possible. The only thing i changed to the car this week was the ECU from megatune to vemstune. So no hardware change. It is not possible to have lost because of it 300rpm spool!!!!
So you got something impossible! :) And now you have to sort out why.
BTW it is possible to get sooner boost build with retarded ignition. But it doesn't automatically mean better performance.
what part is worse: build-up part or area where N75 starts to reduce something?
i think it is simply PID control / boost control in general of vemstune. It just needs different control.
on megatune i was able to have PID: 240/5/52 and offset of 8 and got the above results. A rapid increase of boost and no issues.
With this one i can keep easier boost towards the end of the rpm, but i cannot get it to rise faster. I have tried increasing the P gradually with some good results, but anywhere near my old PID setting and the boost oscillates
the last setting on PID that worked yesterday was: 90/6/28. Did not spike and seemed to come up faster. I will try and make a log of it today, if vemstune does not get stuck!
I'm watching this with interest since we have the same solenoid, I haven't played much with boost as I haven't got the roads locally to do boost runs.
I did a test on the weekend and my boost spiked to 2.2 bar over from 1.7 bar.
As I didn't use boost in 1.1.27 wasn't aware there was a lag difference.
What are you doing to make the logs? accelerating in 3rd gear? so I can log and recreate
Where in the UK are you Jamo?
I'm in Birmingham Rob
I would prefer 4. gear logs (if it is possible of course) since there is more load and you likely have more spikes there.
im doing 3d and 4th gears as i have found that sometimes 3d gear can spike faster
what PID are you using Jamo?
you wouldnt be able to compare if you hadnt tried to 1.1.27
the difference comes after 1.5bar,where in megatune i got a climb of 0.2bar every 50-100rpm wheras in vemstune is much later
I've tried various PID I was using 50/20/50 recently after my old 200 odd setting was brutal.
Are you using 21 or 31 hz? I found better response with 31 hz
i had a word from turbosmart initially and they had told me to use 10Hz. i have used anything from 10-12 so far, without any significant change in the reaction. Ill try 31hz as well. They said the higher the hz the most likely to get oscillations. but ill check it.
thanks
but mine anything more than 12 on I gets me a high overshoot
Interesting, I'll load in your settings tonight and see at the weekend if I can find somewhere to do a run
ill see if there is any use of the 50/20/50.
any oscillations on that one?
ill try increasing the P as well to see how high it can go.
obviously different engine size and cams will have different affect as well
I wouldn't bother with my settings for the moment, I've tried so many however it's only recently i've gone into high boost on regular driving and it is now I'm seeing i'm getting spikes
ah i see
look at my post on s2forum as well.
i feel positive that i may sort the boost control out today.
i looked through all my previous logs, and i have realised that there are 2 possible ways to keep boost sort of stable;
1) High PID values but fast moving solenoid---> this creates highs and lows on valve DC but the boost stays pretty much stable. I had one try on this before with a 0.3bar flactuation, which i assume can be improved. This gave me as fast spool as megatune used to give me, so good stuff. on this the DC target is set low though otherwise it will overshoot. eg: i had 220/70/60 PID and 60DC
2) Higher target DC values on areas of the turbo that is not very responsive to takle as fast spool as possible (reduce the lag a bit) and then reduce the DC and let PID take control. In this instance, PID needs to be much lower, otherwise you overshoot all the time, BUT the actual DC does not fluctuate therefore giving you a more steady boost. eg: 45/10/25 on PID and about 85DC on lower revs and 76-77 on the rpm that boost usually spikes. On this one, i did as well get an equally fast spool as i did in megatune.
So i have done a diagram of things to try and luckily as my most tests are done on 3d and 4th gear, once those 2 gears are set, then the rest are ok as well.
Today i will do more tests on these findings and im positive i will find the answer.
I've also usually configure boost target little over real maximally reachable in areas where w-gate is closed. So that target boost curve copies real boost + some KPA. I mean area before 4000 in your case.
As result PID don't go crazy before boost control really do something.
i see
thanks for that info.
boost control seems more logical now once you get your head around it. it's just that megatune with the boost offset feature had an easy solution for most applications
well i am at a loss...
i tried all combinations, i tried compinations which previously gave better results and today all i got was boost kicking in after 4000rpm. This cannot be explained. Maybe IAT was a bit higher today, but other than that, i cannot see why with IATs of 20C you would get boost later...
with DC being 100% and boost not climbing i can think of 2 things: either the valve is not working properly, or there is a boost leak somewhere? any other guesses?
anyone knows what the equivalent to the boost referrence is? which was evident to megatune? i was using 4/5 there and it was one of the reason which made the solenoid very fast.
IMHO 4/5 meant 4/5*100=80% DC
i see
after multiple tests today, a stiffer spring (0.8) and the turbosmart valve, i came to the conclusion that the boost can built up faster but unfortunatelly the turbosmart is not able to maintain boost on the higher rpm. Therefore an electronic boost controller would be the only solution, unless someone from vems developers, explain to a better extent how the PID should be used as well as the anytrim control
Are you using an internal wastegate or a external?
If external then plumbing it to the top and bottom would yield the same results as a stiffer spring.
i see
i am using an external and i have it plumbed to top and bottom. Stiffer spring only required less DC for the boost to be reached and not spike, but the overall reaction was the same.
Quote from: gunni on November 14, 2010, 08:35:10 PM
Are you using an internal wastegate or a external?
If external then plumbing it to the top and bottom would yield the same results as a stiffer spring.
Realy?
What is difference between atmosphere/atmosphere versus boost/boost in both sides of membrane: situation corresponding to solenoid 100%DC ???
I meant cause he´s upped the spring, he could have put boost on top of the wastegate as well as bottom to force the valve shut for longer and then control boost with the solenoid. This way I have been able to use weak springs make steady strong boost.
in a few weeks time im going to try a 4port solenoid to use in the same configuration as turbosmart eboost. if this doesnt work as well, then im not sure what else i can try, other than going to an electronic boost controller.
by the way, which is the setting that controls rpm on vems? i want to make sure this is not something silly like rpm signal being read wrong and therefore the log is off on the rpms....
A boost controller that handles boost very very well is going to cost close to your vems ecu cost.
it´s about configuration and setup. Something you´d have to do on a electronic boost controller anyway.
Keep at it. The key is in the details.
can someone tell me what do i have to do with the I and D when i need to use higher P values so that to allow for the boost to build up faster?
to me it seems a bit odd, because no matter which PID settings, when looking at the logs the DC is always 100% and then once the boost level is still on 200kpa and target on 300kpa it starts to come down from 100%DC and so i think this is the problem.It should stay on 100% more in order to reach the target value faster.
thanks
Quote from: gunni on November 14, 2010, 10:39:22 PM
I meant cause he´s upped the spring, he could have put boost on top of the wastegate as well as bottom to force the valve shut for longer and then control boost with the solenoid. This way I have been able to use weak springs make steady strong boost.
Can you clarify how resulting force to W-gate valve stem changes if pressures on both sides of membrane are equal?
In 20VT OE case it atmospheric pressure in both chambers. In old MC variant - boost pressure in both chambers.
Real way is to use two opposite solenoids or 4-port solenoid. Then it is possible to keep upper chamber pressurized, but vent bottom chaber to atmosphere.
Another way might be using very soft spring. And connect just upper chamber. I never seen such setting.
im going to go with a 4 port in a few weeks to see results.
this week ill try making that area a bit richer and ignition timing a bit less to see if it will pick up faster.(more exhaust gasses going through on lower boost levels from 170-150kpa)
Quote from: gunni on November 14, 2010, 10:55:24 PM
A boost controller that handles boost very very well is going to cost close to your vems ecu cost.
it´s about configuration and setup. Something you´d have to do on a electronic boost controller anyway.
Keep at it. The key is in the details.
Do you mean controller what uses speed of turbine shaft as additional reference??
AVP sent me logs with anytrim applied (seems buggy) where different boost targets (0.6...1.8bar) was kept I would say just close to perfect! Main problem usually is to find right settings for VEMS BC.
which is something i can see not many people other than Gintis and gunni are offering as info at the moment.
some tried settings for audi cars or any cars regarding boost control will be helpfull
Quote from: AVP on November 15, 2010, 08:07:39 AM
can someone tell me what do i have to do with the I and D when i need to use higher P values so that to allow for the boost to build up faster?
to me it seems a bit odd, because no matter which PID settings, when looking at the logs the DC is always 100% and then once the boost level is still on 200kpa and target on 300kpa it starts to come down from 100%DC and so i think this is the problem.It should stay on 100% more in order to reach the target value faster.
thanks
I saw from your files refDC is just 70%. It is like base position. When boost comes to target, PID adjusts DC to ref DC with ease. Try rise it!
If rising of DC cause problems. I would try to reduce D value. D value works opposite to rapid boost change. Like it is at spoolup area close to target.
you can see from the logs that the refdc is 85 on those range and goes to 77 on higher revs. if i get them higher, the boost spikes at 4000+rpm
so reducing D means that the correction the PID does is sharper and allows for more DC at spool areas?
i have to send you in the afternoon a log with my current setup which seems to work well once boost is in, but its just late
my pid on that is 110/10/28
AVP you mix RefDC with resulting DC. No?
In RefDC table you have 70% for all boosted rpm range.
that was on one of the runs with anytrim on.
is that what you mean to change? while using anytrim?
I mean Boost Control->Boost RefDC vs Rpm curve. Set it higher!
With no anytrim as it seems buugy.
I will try boosttrim on bench one more time with other config and then report.
without anytrim i have it set to 85 up to 3600 then 80 to 3900 then 76 to 4300 and 77 thereafter. this curve gives me a stable boost even though later spool.
i have tried 90dc or more and it spikes
Quote from: GintsK on November 15, 2010, 09:33:10 AM
Quote from: gunni on November 14, 2010, 10:39:22 PM
I meant cause he´s upped the spring, he could have put boost on top of the wastegate as well as bottom to force the valve shut for longer and then control boost with the solenoid. This way I have been able to use weak springs make steady strong boost.
Can you clarify how resulting force to W-gate valve stem changes if pressures on both sides of membrane are equal?
In 20VT OE case it atmospheric pressure in both chambers. In old MC variant - boost pressure in both chambers.
Real way is to use two opposite solenoids or 4-port solenoid. Then it is possible to keep upper chamber pressurized, but vent bottom chaber to atmosphere.
Another way might be using very soft spring. And connect just upper chamber. I never seen such setting.
If you have the same pressure on both sides then you have zero pressure acting on the diaphragm + spring pressure holding it down.
Like so.
(http://www.perrinperformance.com/shared/images/jeff_pics/boost_control/proexternalrestv2sm.jpg)
So and where then is difference between plumbing like 20VT OE versus plumbing on this drawing? Both will give same pressure.
BTW MC 10VT top side was fed from manifold. May be it gave little help for boost reduction on partial throttle.
what gunni is suggesting is that when the top chamber is pressurised and you have a stiffer spring, it will be the spring itself adding to the improved spool because it will allow more gasses to go through the turbo.
But it is no matter how to plumb 2 or 3 port solenoid: to top or to bottom. When DC is at 100% the only forces comes from backpressure to valve area and spring to opposite direction. So max boost is limited by spring. Not by connection type.
At least until bottom chamber has no significant leak trough valve stem. This is why Audi put big port N75 comparing to MC engine.
well im off to try a few more things today.
a bit of timing reduction on the spark map, as well as possibly some anytrim playaround.
No need to try anytrim. It do not work also on default config. I just wrote report in IRC.
set it to 100% and temporary forget about it.
ok, thanks
i just wrote to marcell as well.
im going to try the ignition table reduction, as well as the speedsensor which marcell explained a few things more about it.
Back from the tests
i managed to get the boost up to 270ish kpa at 4100rpm on 3d gear which is a big improvement as its only 30-40kpa away from the goal which is the 305kpa i had on megatune
i used pid: 110/10/28 and tested increasing P and decreasing D without any better results.
I have to try more on the P value though.
i also decreased the timing as well as increasing the fuel in those areas which i think also helped.
when i used someone's values of 220/70/60, the boost spiked all the time even with reduced DCref.
here are the 2 turbosmart options of connection:
1) this is what i use now
(http://i117.photobucket.com/albums/o59/vasp3690/connection2.jpg)
2) this is what i can try
(http://i117.photobucket.com/albums/o59/vasp3690/connection1.jpg)
this last connection apparently is for maximal boost, but does not say if it is for faster as well.
does IAT have anything to do with spool and lag as well?
Quote from: AVP on November 14, 2010, 10:49:29 PM
in a few weeks time im going to try a 4port solenoid to use in the same configuration as turbosmart eboost. if this doesnt work as well, then im not sure what else i can try, other than going to an electronic boost controller.
by the way, which is the setting that controls rpm on vems? i want to make sure this is not something silly like rpm signal being read wrong and therefore the log is off on the rpms....
This is my setup Vas, 4 port mac solenoid to upper and lower wastegate.
For Gints it's not the same as putting in a stiffer spring, the boost response is better with the top and bottom configuration when you use a 1.9 absolute spring over the 1.4 absolute spring
so you are using 4 port and 1.9bar spring?
i must say i did not see any spectacular improvement regarding the 1.9spring a couple of days ago
Quote from: AVP on November 16, 2010, 08:22:56 AM
here are the 2 turbosmart options of connection:
1) this is what i use now
2) this is what i can try
this last connection apparently is for maximal boost, but does not say if it is for faster as well.
Are you sure about you use method 2?
It looks funny! :)
Connection can't work as controller: There is no way to bleed pressurized chamber somewhere out while turbo gives boost! It should give hard overboost and then go to high boost as we usually see at 100% DC.
Because at first top chamber being pressurized from turbine housing (means hard overboost). And each solenoid cycle transport some portion from top chamber to bottom. Finally both chambers are just filled with no any sense.
All working kinds of connection reserve way to bleed chamber to atmosphere during regulating. Not this. Think about it!
GintsK : the solenoid vents to atmosphere, as written in the picture so it works as you want it to.
"method 2" is the method used in 99% of cars with external wastegates that I'm involved in. In some cases we use 0-2 bar air pressure regulator, instead of an electronically controlled 3-way solenoid.
I can't be agree! Sorry! :o
On 100% production cars "method (2)" is not used!
There is no connection to atmosphere:
port 1 - top chamber.
port 2- pressure source
port 3- bottom chamber
that's all!
Common method is this:
(http://www.vems.hu/files/GintsK/SinglePort.png)
actually im sorry, the pictures are mixted up. the method i use now, is the bottom picture. The top picture is the one i am questioning if i should try.
word from turbosmart was that it wont change the characteristics though...
Quote from: AVP on November 17, 2010, 07:12:51 AM
actually im sorry, the pictures are mixted up. the method i use now, is the bottom picture. The top picture is the one i am questioning if i should try.
word from turbosmart was that it wont change the characteristics though...
I knew that!
With "(2)" your boost problems should been much harder than it is now.
gintis, i am now plotting down 2 excact curves between 1.1.27 and 1.1.81 on my old laptop and trying to see differences.
it seems that on 1.1.27, the car picked up faster even from 3000rpm, it went much faster to 140kpa which is my spring strength and therefore moved much easier to 170-180-200 and therefore all the way to 304kpa on 4057rpm.
the Boost DC on that run was 100% then dropped to 82 on peak boost at 4047rpm and held at 82-83 to give me 304-305kpa for the rest of the run. As mentioned before, the PID setting on that was 240/5/52 and i was using map as ref target, not DC(which was only put to 66 in mid range then up to 70 on higher revs, but im not sure if on 1.1.27 that makes any difference anyway.
so what could be the reason why when the TP is 100% open and the refDC is 100%, the car is not picking up as fast as it was???
Have you possibility to post both logs?
i will send them to you shortly
you should have the files now
i tried to do the excact identical boost settings from 1.1.27 to 1.1.81 again, copying all refDC and map targets.
obviously, there are no boost offset features here as well as DC ref position so that will need to be seen how it will work.
You have very advanced ignition in 100KPa and low boost areas on new firmware. It can be that cause far less energy to turbine wheel. Instead of spool turbo energy goes to heat up engine.
Try ignition map from 1.1.27
can we also see the logs?
i rechecked today and i saw that the ignition timing map is identical to the 1.1.27 map. on vemstune i have 120kpa and then 170kpa, as the table is smaller and i cannot use the 140 value. Could interpolation be causing this?
i now reconfigured the table and im going to try it out today, but i usually do have high timing up to 120kpa as on cruise on light pedal press to make a small acceleration it seems to work better.
i also decreased the timing on the last runs by far on these areas, even less than 1.1.27 on the 170kpa line, but it didnt make much difference.
billman, you already have logs like that, but i can email them to you again
Really identical?
Advance seems is the only huge difference between both logs.
Usual 100KPa advance is shifted to 170KPa.
in a few hours, ill send you the msq file i was using on megatune.
what i can do, is instead of using the 120kpa line on vemstune, change it to 140kpa and keep the 170 as well and try and use interpolation between 100-140 or something. but im pretty sure it is identical. on megatune i had 100-120-140-170 lines and what i did on vemstune was to use 100-120-170 and keep the 170 line the same in both firmwares.
on the megatune file, you need to be aware, that as the boost is coming in much faster, its hitting areas of kpa much sooner and therefore the spark on that area is indeed less, but if it didnt, then it would be higher.
on vemstune, as it is not hitting those areas fast enough, it seems that it has higher timing
Can be.
Bad thing about these both logs: 1.1.27 is not WOT at all and 1.1.81 is WOT only from 3xxx Rpm :(
yes, that is because i have a quite high tps vs kpa table. So you see that even with part throttle the boost goes high up. It may be also a bit faster if it was on full throttle!!
i forgot to mention that my timing map with it being much lower from 100kpa and upper, i had noticed that the car was loosing lots of torque on those areas, which is why it is so high. I dont seem to get any ping, but it also feels more torquier( not tested on a dyno yet though).
i will try getting it really low today to see if there is a difference, but assuming that the maps are identical it shouldnt be the reason
It would be nice to see full range WOT runs on both settings. And, yes, msq too.
If there is difference it might be valuable to sort out cause.
ill try and find old logs of wot to show you
just got back from trying
unfortunatelly nothing i tried today worked.It seems that up to 200kpa there is very slow buildup of boost and then it comes alive.
i did find out however how anytrim works.
anytrim requires the refDC to be zero completelly in order to be taken into account.
so i tried a run with refDC 0 on all rpm and my pid settings to : 240/5/50. without anytrim i had about 230-240kpa full boost.
once i went to anytrim and dialed it in for 160-180----> i got 300kpa
at this point it would be VITAL from someone of the developers and come up and tell us how the BOOST SETTINGS affect each other so that we can stop guessing and start taking proper steps.
how can i have boost with map target only and PID on operation? how is that done? why does it go to 230kpa instead of my 160kpa spring? who decides how it will go?
looked at the logs with more time now and i can only find as a difference that other than the fact that the PID may become a bit more accurate to reduce fluctuations, the only other difference i can find is that at the beggining of the runs, around 3000rpm when the car is slowly going into the 130kpa, the lamda values are fairly low. I have seen anything from 1.12 to 1.03 on those areas.
Now along with low advance, i have been told that more fuel will also help spool up, and on megatune i had 0.96 and then 0.94 on those areas.
Im not sure how much difference that can make, but im hoping it does, as i changed my lamda table a bit to compensate.
as for the anytrim, ill try it since i have it now and hope for the best!
with such high lambda you are not supplying your engine and turbine with energy.
yes i suppose not, so i have reduced that and ill test it tomorrow... fingers crossed...
IMHO, you need to be more scientific about this. With all of these changes, and you are surprised the spool-up is different ?
You will fix this and understand better what works well with your setup, I'm sure.
i may fix this in the end mattias,
but it seems that VEMS developers are not bothering or have forgot to add help to the boost controls, as well as inform all you guys when all features are actually working or not.
i dont like trying to solve an issue with my car and at the same time trying to realise how all these boost settings interact. For example, you can have boost control with PID and rpm/kpa values only. refDC all to zero. Did you know that?????
you can add anytrim to the above setup and get boost increase.
Gintis found out that if you have redDC values + anytrim, it doesnt work! did you know that as well????
you can have boost control with no PID and just refDC and rpm/kpa values...
can you have boost control with no PID and no rpm/kpa values, and just use refDC ? and then possibly add PID?
i mean all these SHOULD have been known to users, and not us finding out the hard way. If i knew that i had all these different ways to control boost, then i would plan my action and try them all. It seems though that when changes are implemented and people are unaware of them, even tuners who are involved with the development, it is not easy to give advice thereafter.
just saying to someone: try harder and you'll find it is not something you should put on the F1 help options next to boost...
AVP point about be more scientific is good! By playing with BC so much, it is surprising how you did not catch refDC purpose and idea. Even with no helpfile. It is the same as for idle!
I just try different refDC settings on the bench. It do not influence boosttarget by single bit. And do not influence buggy anytrim (http://www.vemssupport.com/forum/index.php/topic,1490.0.html). Anytrim cause disturbed boostarget as previous. Just in same amount with any ref DC.
Please try to be more analytic! Look at your logs. Try to change detail by detail and after each change look to your log. incluiding bosttarget.
on the car, yesterday, anytrim being on 160 and PID 240/5/50 gave minimal oscillations and boost close to my target. refDC was all zero
when i was using any kind of refDC anytrim didnt work, esp. when i was using fluctuating values rather than the same value on all areas.
i will retest tomorrow, but to im trying to explain to everyone that to be scientific, i need to know how all these different aspects work with eachother otherwise im just trying things blindly and going by the final effect.
I dont have a bench tester so i can only test on the road.
try on the bench to go only with PID alone and target kpa, then add anytrim and leave refDC all to zero, and let me know what happens.
Just to clarify a few things in my mind regarding boost control, before being scientific in my options:
boost control options:
1) PID: 0/0/0 + refDC/rpm= zero ---> use only: map target/rpm
2) PID: 0/0/0+ map target/rpm=zero---> use only refDC/rpm
3) PID--> in use with refDC/rpm= zero AND map target/rpm ONLY
4) PID--> in use with map target/rpm= zero AND refDC/rpm ONLY
5) PID: 0/0/0 + maptarget/rpm + refDC/rpm
6) PID: in use + maptarget/rpm + refDC/rpm
7) PID in use + maptarget/rpm + anytrim (refDC all zero)
8 ) PID in use + refDC/rpm + anytrim (maptarget all zero)
which one of the above are right and which wrong?
In short:
Pid 0/0/0 means closed loop control is completly OFF.
In this case works just refDC - you have open loop boost control.
Target table has zero influence on process.
===========
When you start to add small PID values, your BC start to tend to be closed loop. But at small PID values resulting DC will be close to refDC. BC tends to correct initial refDC to target side. Same as for idle. Now BC looks for value from target table. But it not means that target will be reached.
More PID values means less refDC inluence. Due to this you will be able to reach target using refDC table filled way off values.
Try playing with idle - it is the same. Just target comes from two values interpolated inbetween.
Quote from: gunni on November 17, 2010, 10:31:55 PM
with such high lambda you are not supplying your engine and turbine with energy.
High lambda - lean mixture and high egt, typically improves turbine response, actually. At least in my experience.
I agree he is changing too many variables at a time. Develop a repeatable test - acceleration in a certain gear, form a certain start rpm to a certain end rpm, at WOT, and work from there.
well, just came back from more runs.
still with more fueling, it did not make much difference. I still have the rpm difference of about 150-200rpm esp. on 3d gear.
on all occasions, as gintis mentions open or closed loop control, it seems that the boost is only reaching the target on the way it wants. no sooner.
when PID is completelly off, if i am too aggressive with the refDC, then boost can be ok for 3d gear(as i use higher refDC from 2500-3900) but then on 5th and 6th gear, it will spike and overshoot as there is much higher load.
therefore closed loop MUST be used to avoid this.
now, either i use closed or partially closed loop, OR anytrim for that matter, i end up with the same figure, 100%DC as it should and then a drop off as i come close to the target values. However, even though i get similar values as i did on megatune as far as when the DC drops(around 3700rpm on 3d gear) i end up with much slower boost rise, even now that the fueling is much improved. SO i cannot explain it,other than to say that there might be wrong with the turbo itself? (turbo is only 6month old, and has never been abused).
All my tests are done methodically. I dont go changing 3 or 4 settings at the same time. All runs are done on 3d and 4th gear on the areas which are involved in the problem 2800-4500rpm.
Personally, i give up. I dont know what the problem is. I have read about PID configurations, i have managed to tune my idle fine, this has been a nightmare to tell you the truth.
Quote from: AVP on November 18, 2010, 07:51:01 PM
All my tests are done methodically. I dont go changing 3 or 4 settings at the same time. All runs are done on 3d and 4th gear on the areas which are involved in the problem 2800-4500rpm.
May be you do your tests methodically. But you do not give much chance to help you. Both logs was not from similar circumstances: one was not WOT, other was WOT just from 3000rpm.
i did send you an msq file with a 100%TPS from megatune i think.
i have so many logs which make it very hard to look for one with excact same circomstances, but the fact that i have faster spool even on non full throttle must mean something, that the vemstune settings are not up to the job.
the car seems to have an area from 110 to 160kpa which does not spool as fast as before. I know that is an area of the turbo which does not respond anyway, as the hotside is big, but now that the fueling and the spark is also similar to the previous maps(on 1.1.27) i have no doubt that the issue lies with the way VEMS is adjusting the solenoid.
one thing that i havent tried is the OEM style of connection, just to compare, but i doubt i will have any different response. I will give it a go once im back from Greece.
I don't have the specifics of your setup in my mind right now.
This is not so hard to verify. If you don't spool as fast as before in the same gear, then it's either different tuning or health of the engine, a mechanical issue with the turbo or wastegate, clogged catalytic converter, or other piece of equipment.
If you have a single solenoid top-side, set a reasonable fuel-cut then apply boost directly from the compressor outlet at all times, that should lead to the best spool-up you can possibly get with that setup - until you bump into the fuel cut limit. Blaming any loss of spool at that point on the software is pointless.
Move old logs to an archive, or categorize your datalogs. It would be bad if what you're comparing to are numbers from WOT runs in a higher gear.
1) I am not comparing WOTs of different gears. It is very easy to understand which gear i am in from old and new logs, as the characteristics are so different that you can tell immediately. On 3d gear i had 305kpa on 4100 and on 4th at 3780rpm. On higher gears, even sooner, so there is no way i can be mistaken there.
2) As far as tuning is concern, i think i have exhausted every possibility.Unless there are hidden/ unrealised by me at least features of vemstune which are contributing to the poor results. The mechanical setup is the same, and the maps are 1:1 transfer from the megatune files. The last bit of the puzzle were the lamda values which had appeared to stay lean and we thought this was the cause of the poor initial spool BUT as this was corrected and matched to the megatune files, spool did not improve
3) Mechanical issue? How unluckly could that be? Just as i changed ECU and a week before the car was flying, now all of a sudden there is a mechanical issue causing this? Car has metal cats, and have been there for a while,never had an issue with them, have done less than 90K kms. Turbo is 6 month old and as mentioned before never ran out of it's specs, downpipe is new, cat back is old, but stainless and custom made.All lines to and from the solenoid were recently checked for leaks and cracks and are OK.
4) What do you mean a single top solenoid? i really need to check this thing about the fastest mechanical spool possible, you are right.How else can it be done? if i take the hose from the turbo compressor side and plug it to the lower chamber of the wastegate(external) will it work? i can put a fuel cut at 250kpa or whatever.. yes if that turns out to be identical with the current spool i will probably have to look for something in the mechanical aspect.
Mattias, AVP switched back firmware. And calibraton file which is different. Looks like spool-up on 1.1.27 setup is better.
But we do not saw clear WOT datalogs. It is hard to help then.
im sorry gintis, i thought i did send you one with WOT
ill do that this afternoon so that you have an identical WOT on the 2 firmwares
Plug the boost signal into the top chamber and leave the bototm chamber unplugged. Boost will forcibly hold the wastegate closed and minimize the potential for wastegate creep etc. You MUST have you boost cut set appropriately though of course as the wastegate will not open!!!!!
yes, i will do this now and report back in about an hour
i will use a fuel cut at around 270
spot diagnosis:
on the rev limiter, and fuel cut @ 300kpa
3d gear: just before 3000rpm
4th gear: well before 4000rpm
ON LOGS::::
both appear to be 200-300rpm later than what my car's rev counter showed!!!
is there a lag in the log? where is the setting of the vemstune regarding rev counter of the car?
edit: after closer look i can see that there is only minor difference early on regarding how fast the boost starts early on, but then the whole boost curve seems to be identical to the vemstune boost curve. What i dont understand is how is it possible to have that when my rev counter shows to cut much sooner than what vems is reading.
so i can say that the final outcome of this test is that either i have some kind of increased backpressure, OR maybe there is no good correlation between what vems is reading as rpm signal and what it actually is true.
Which is the actual? the mechanical rev counter from the car? or vemstune? should i check it's calibration? and where is that?
VEMS is the most accurate RPM, theres no doubt about that at all - or the spark would be random.
ok
but doesnt vems need to be calibrated first?
i know that on megatune there was a calibration at 3000rpm
doesnt vemstune have anything similar?
Here it is!
You have wrong Rpm reading on 1.1.27 caused by wrong constant for Rpm calculation. It should be 12000/cylCount=2400
From 1.1.27 msq:
(http://www.vems.hu/files/GintsK/RpmConst.png)
On 1.1.81 Vemstune reading is OK because user do not calculate this confusing constant any more.
i see. i remember changing that a long time ago on megatune as at the time it seemed that the rpms matched the rev counter better when i used 2300 rather than anything else.
well, it may have been that then and therefore the boost acquisition is ok. I wasnt expecting that though.
does that mean that the rev counter on the car is not accurate then?
Quote from: AVP on November 19, 2010, 11:05:52 PM
does that mean that the rev counter on the car is not accurate then?
It´s a mechanical device trying to keep up with the input into it. the motor driving the tacho is not strong enough to keep up.
Such is the case with almost all tachos.
The end of this is so unspectacular??? Just a wrong setting in megatune?? nice....i think there is no mechanical problem then at all.
AVP send me now please a 4. gear full throttle run.
You also have unusual divider for tacho in 1.1.27 - 17
Theoretically 16 is right one. But this value is allowed to change. It influence just tacho, not all calculations in ECU.
In new firmwares set “Boost valve off below this pressure†to 10 and see if that helps. I think there is a bug.
Peter
i had this actually to begin with but was told to try back to 120kpa.
what bug is there noted?
@Billman: when megatune and rev counter is identical for the last year+ i shouldnt have any reason to doubt them. Once i made a vemstune upgrade, thats when all this was noted. So although it seem dumb as a solution to the rpm difference, i must say that i wouldnt have believed it if mechanical boost wasnt there.
If i didnt use vems and used motronic, seeing that the rev counter still shows that my full boost comes sooner than 4000rpm on 3d gear, its only after i vemstune which is more accurate that it appears that the spool of this turbo is @ 4200-4300 instead of 4100. and on 4th gear 4000 instead of 3780-3800rpm
if all tuners/developers suggest that vemstune is more accurate then i do believe them.
Yes what bug is there?
All the configs I have seen there show 120-130kpa.
AVP I am with you there. If vems is correct then it should be fine.
I bench tested 1.1.85 a week ago and boost looked strength. Then I tested on my car (1.1.81) and it seamed that it spools up faster but you can easy make a test and see if you get the same results.
Peter
you mean by using 10kpa instead of 120kpa peter?
i dont like the valve to work from 1000rpm, so if use 10kpa as boost off below that pressure, and use refDC up to 2000rpm, will that have any negative effect on spool?
You can try 40-50 and dc ref table = 0 in the low bins. It´s difficult to say if it will wok on your car but you can easely test it. it will take you five minutes to due so.