AUDIS
Update 09.23.2008
Audi is my favorite field. Almost half of my units run on Audis.
Last weekend my first VEMS equipped Audi brings Crew to champions!!! in Rally-sprint 4WD n/a class. It was very tight fight: both main pretenders starts this event with equal standings and gives very close SS times. Finally only 3.2s separate both. Here is results: http://autosport.lv/ofiles/5522/docs/results-official/4WD.pdf
All my 5cyl Audis run on Auditrigger. One with strange 7A flywheel with Hall signal inverted. Maximal Rpms I ever try is little under 8000. Without errors. Almost all use 1.1.27 firmware.
What I like year later:
- Still responsive WBO2 control
- Boost control strategy with W-gate map (waiting for TPS scaling). Bulletproof strategy!!!
- Easy to tune using analyzing software.
- true VE map and separate map for lambda.
- Idle control close to perfect (looking forward to new IAT compensation)
- EGT logging.
- LCD for customer diagnostics. No need every time for me with laptop.
- soft rev limiter.
What I miss:
- limp-home for MAT, CLT sensors. WBO2 damage/open circuit did not switch off ego correction.
- CLT range higher than 102C.
- relative retard for launch control cause imprecise boost.
- support
Update 09.05.2007
Put 1.1.23 firmware. Looks it works. Yes without Hall sender is possible to start/drive!
I make fine tuning on this old engine. Must say I like to tune VEMS. Configuration in nightmare, tuning is almost pleasure!
Old engine produce good power. Somewhat 15-20Hp higher than stock 170Hp. Only modifications is loud exhaust and now restrictive AFM is in the garbage!
What I like:
- responsiveness of WBO2 closed loop operation. Looks great!
- idle control using ignition advance!
- VE learning (note below)
Cons:
- It is almost impossible to find description of some configuration values. So I stay w/o knock control and self learning is a bit on slow side. Renaming of many Megatunes value names would be very welcome. To be value name plain describe what it affect.
- no visualization of VE self learning. Every time need to hit Ctrl-F in Megatune.
- Idle control try to manage IAC PWM before revs becomes close to idle. I think there must be some Rpm margin for IAC ON.
- maximal temperature value is 102C in Megatune and 101C on LCD. Is it bug? Or have I something misconfigurated?
Questions:
- how to adjust/configure VE learning process? Wideband operation looks much faster and stable allowing learning process act faster.
- What is "scaling factors"?
- What is "Min. weight for modifying VE Entry"?
- What is "Speed of changes"?
- Is somewhere description of LCD pages? Looks very promising. But w/o knowledge much of this info is unusable
Flyback problem
Update: here is my solution for flyback: http://www.vems.hu/wiki/index.php?page=MembersPage%2FGintsK%2FFlybackForLowZ
7A engine use 2 Ohm injectors. ~315ccm. With resistors from factory. My mistake is install without resistors.
With flyback wire only it was completely impossible to tune. Injectors as soon as opened stays opened loooong time. Highest lambda possible 0.7
I install flyback board inside. And hope fix this problem completely. In reality things goes better, but stays far from good. This engine is for competition - no need for super-clean idle. Lambda 1.0, rough Ego control.
But without flyback (yes my mistake - two fried FETs at redline) control over injectors at idle is far far better. Idling even at lamda 1.10...1.15 !!!
08.02.2007
So reality is: my first Genboarded cars will be Audi 20V turbo and n/a.
Here is board modifications for cramhome signal (crankhome + Hall on camshaft)
- R91 * C103 = 10k * 1nF = 10usec
- C103 does some filtering (not strictly needed)
- for the secondary trigger, the HALL masks out the LM1815 output. There is much more than 55microseconds for the HALL edge to settle
- though only 55 usec between primary trigger teeth pulses @8000RPM, so you need R57=33k (instead of 75k. If you stuff a 75k on top of the original 75k, that is 75k/2 is OK, but the traditional 75k might prevent reaching 7000 RPM on this engine with 135 tooth trigger)
question:
I imitate Audi Trigger using WAW file: 270 pulses on right, 1 pulse on left chanel. 1.0.53 firmware.
Everything works OK until ~6000Rpm. Then starts trigger errors. After 7000 - no sync at all.
It seems things goes even worthier with new 1.1.11 firmware.
Q: How to sort out what is cause: sound card resolution or something with Genboard?
Disquiet is because 2 of 3 Audis must rev up higher than 7000!
- Audi have restricted turbo by rules ~300hp max. I think ~500Nm@~3500 and ~300Hp@~5000
- is project with big turbo and target power >500Hp
- is n/a engine with deep modificatons for rallying ~250Nm@~5200 ~230hp@~6900
here is my waw files:
Audi http://www.vems.hu/files/GintsK/Audi_270&1_4690Rpm.wav
Audi http://www.vems.hu/files/GintsK/Audi_270&1_6840Rpm.wav
60-2 http://www.vems.hu/files/GintsK/60-2.wav
36-1 http://www.vems.hu/files/GintsK/36-1.wav
I use http://www.goldwave.com/ Goldwave sound software. They have evaluation version: free and very sutable: simly loop .waw say 8000 times. And under tools is easy to control playback speed evenly.
Use 96000 samples/sec wav files for this 135 tooth pattern (at least if you go high RPM), not 44100 samples/sec.
third question:
Q: C30 cap between primary 135tooth VR input and ground? Is't it too big??
R56 * C30 = 18k * 1nF = 18 usec
18 usec is less than 55 usec/2, and even with a 27 usec time constant, the signal would reach the LM1815 input, just slightly lower amplitude (still much higher than needed to trigger). So this will not prevent reaching 8000 RPM.
However, if you want to reach 8000 RPM very quickly, like from 4000 .. 8000 in 0.4 sec or faster, than the adaptive hysteresis arming threshold might not be reached and trigger errors might be the result. A 7..8000 RPM revlimit in neutral gear is not a bad idea anyway.
This document:
http://www.vems.hu/wiki/index.php?page=MembersPage%2FTonciLoncar%2FQuestionsAndAnswers beside of all is mentioned:
...if there is noise C30 (which helps shape the input signal) may be broken, try replacing it if you don't have a scope 1nF is will give you a 16kHz filter frequency, you could easily have a bit more here for most applications....
But 16kHz is 7111Rpm for audi 135 teeth Wheel!
+++++++++++++++
Update mid 09.02.2007
Another observation:
This situation I got using balance button on already "started engine": it works on 135(270)tooth signal only! No trigger errors. But reality - no second trigger signal at all.
---
and this is with normal mid balance and C30 soldered back
Both tests 1.0.53 and Audi_270&1_4690Rpm.wav
Keep an eye on MembersPage/MarcellGal/PowerAndTraction, we purchased a 5cyl audi (engine should jump to bay on 2007-03-09) to have all the details hammered out so installation and diagnostics becomes smooth (which it is not yet) on these nice engines.
+++++++++++++++
Update 26.04.2007
My first Audi finaly runs on 1.1.22 firmware.
It is n/a 7A 20V engine. I have not time for bench tests for high revs. But in real life
- Trigger errors counter stays 00 up to 7200 (tested on freewheeling not @load).
- Synchronization at cranking is fast enough
- Resynchronization at working engine is fast enough (ingnition switch off/on)
- Advance looks precise. As I remember I use 1.1.22 published settings but with trigger angle 62 0r 63 decimal.
BTW 7A engines crankhome sensor seems already right polarity for LM1815 input. Note that (at least AAN, but likely 3B version as well) motronic seems to run great with the crankhome-VR reversed (where they have to be reversed for our LM1815 to work).
Thanks!
Q: Do you have some pinout, or other info for the 7A engine (+harness) ? Is it similar to ABY/AAN/ADU or 3B?
A: I have not pinout to publish. But 7A use management from hitachi with 2 own shape connectors. Nothing similar with ABY/AAN/ADU or 3B in this plan. But trigger patern is the same.
+++++++++++++++++