Ed, never got the logs but no biggie… Jran posted what I need to see and looks similar to the other logs… I’m no tuning expert but I’d prefer timing retard over boost bypass especially if the car is always bypassing boost… I also wonder if you continued to flog the car and got IAT’s way up there would it pull timing, dump more boost, or detonate?
Revo claimed that you would see timing pulled if IATs got high enough. I don’t know what that threshold is, but obviously much higher than stock/APR/GIAC.
The log was after we had done a couple of runs and the engine must have been hot, also drove about 10 miles to reach the deserted highway I use so temps were up. Meth injection and hood intake may have helped too?
Posted this on AZ, but wondered if you guys might have some ideas:
So what are the alternative signals we can look at to understand knock? Can’t we look at the raw knock sensor voltage? Maybe do some frequency domain analysis (FFT) or filtering? We have always been relying on ignition timing pull via knock (IGA_AD_1_KNK[1]) which could be viewed as the output of the factory protection algorithm. It should be easier to post-process than it is to run in real-time, since the implementation must be a causal system.
Would be kind of cool to look at the raw signals, since the Revo engineer who posted said they need to filter out all of the normal vibration and other noises the microphones pick up.
drob, Not sure… I do know you can log misfires at the cylinder level or in total… I think there are others but I’d have to search… I’d assume a misfire is filtering out any “noise” and is trye knock so to say? Maybe I’m missing something though… FYI… the values below are older from a log I have so not sure of current values but those are the names… I’ll look tonight or this weekend…
476 Misfire detection; cyl. 1
477 Misfire detection; cyl. 4
478 Misfire detection; cyl. 3
479 Misfire detection; cyl. 6
480 Misfire detection; cyl. 2
481 Misfire detection; cyl. 5
482 Misfire detection; misfire totaliser
jran, just read your write up on the coolant system… Nice work… Post it over here so others who don’t frequent/don’t comment on vendorzine…
Well that was a total fail on my part.
Here is an interesting overview of knock detection: http://www.ti.com/lit/an/spra039/spra039.pdf (there are tons of papers online about this stuff, another example is http://digitalcommons.mtu.edu/cgi/viewcontent.cgi?article=1562&context=etds)
My idea is massive fail…lol…knock frequencies are on the order of 5-20 KHz, which would require sampling at 40 KHz…lol…we get VCDS data at around 5 Hz in turbo mode it looks like.
The ones you bring up are interesting, wonder when they start counting or if they can be reset.
with the use of meth i seriously doubt skywagon will have any issues with detonation.
i do agree about the boost bypass vs timing. in this instance skywagon would see 10-15 more hp with full boost if there isn’t any knock
Will do. Just haven’t got around to it yet. I’ll have some better logs after going to the track on Friday also.
drob, yes they can be reset… I guess here is how I think of it but I think you have a much better understanding and I haven’t read the articles but will… The timing retard fields are based on a pre-determined value and once hit it pulls “x” degrees and keeps calculating and pulling or not based on those pre-determined values… I think the misfires basically say shit even though we tried to prevent it the car still detonated… You’ve seen the old Revo logs right? They had several hundred misfires on one 1/4 pass… My assumptions was since Revo was suspected of disabling the protection the car couldn’t do anything to protect itself… This is why I did a bunch of logs with these fields and no matter how hot it was and how much I flogged the car I never got more than 1 or 2 “misfires” since the car could protect itself so to say… Also, one time I threw a misfire code and it was still only a couple misfires and I used vag to reset the code and counter… I would guess it resets itself after “Y” period of the car being off, but not sure… There are some vag com fields that track how many driving cycles have occured before the car memory was reset… Sorry, if I’m way off base here…
Here is the only other misfire one I see…
397 Emissions-harming misfire detection; status
Brian, Good point on his meth setup… Just hope it doesn’t run out during a pass in 90+ heat when IAT’s are way up there because I doubt dumping boost will save the motor, but who knows… Hopefully, sky doesn’t find out… I also agree the extra boost would give him more power assuming no timing pull… So the real question if we assume no differences in safety is what method allows for more power?
I mentioned this on the other site, but on the dyno the Revo car I logged hit 80C IAT, and it still did not show any sign of timing pull (that was also with 21-22 degrees of timing). 93 octane gas/map, no meth, etc.
Thanks for the reply prime, I’m not even close to knowledgeable about this, hence the noob questions. Anyways, I just pinged a friend who does engine controls (more on controls side than engine side) and he said that changing pressure and spark ignition are both means to help mitigate knock, however, spark ignition is the more direct and controllable input. Mostly because of the bandwidth. You won’t be able to change pressure in the 2 ms or so before the next cylinder fires. So my guess is Revo is doing some kind of open loop strategy where they bleed boost with the RBF a priori, and then pull timing if knock detection gets severe? We’ll have to see more logs from the summer months I suppose. I have no idea which is a better strategy in terms of power, but it would seem closing the bypass valve gets the most air/fuel in there, and should allow you to get closer to MBT assuming
conditions are optimal.
Keep in mind that when we see normal passes, and some ignition retard is requested…it seems to last for quite a while (in a relative sense)…you aren’t seeing drastic changes between VCDS samples. So my guess is the timing pull is kind of event/threshold driven…detect knock and then pull timing depending on how severe for the next ___ cycles or until the engine load conditions change (shift or no longer WOT).
Sounds like this measurement of total misfires is a really good indicator of how the motor is running. I’ve never logged this so I’ll make sure to try the next time I go to the strip. I don’t remember seeing the revo logs FWIW.
I wonder how the degree of boost bleed off compares to what skywagon saw.
drob, great points and kind of feel the same on boost vs timing from a response standpoint… Would be cool to see CV chime in with his thoughts… You have mail… 
Thanks for the mail. It would be interesting to re-quote what the revo engineer posted on AZ, it seemed to divulge quite a bit and was interesting.
It was a lot less. Never got over 4%. Most of the time it was 0 or below 2%. I didn’t think a whole lot about it until I saw Sky’s logs.
I copied the Revo response in a thread on here: http://audirevolution.net/forum/index.php?topic=2859.msg82339#msg82339
[i]Here is a response from Robin, one of our engineers who works on the Siemens projects:
The numbers would be higher if the car was allowed to rev-- with the DSG flash, 7000-7100RPM is possible and the car will continue to make peak power the higher it revs. We usually see another 15-20HP when the car is at redline vs when the standard DSG file changes gear. These dyno charts don’t show the actual beit of having the DSG controller flashed.
To properly interpret the data, you have to have an understanding of how the SIMOS ECU works. It has three different timing tables that it will interpolate from to arrive at a final ignition timing value after passing through many factors for cat temp, intake air temp, load, etc etc… Precise calibration of these tables is necessary to achieve a smooth and strong timing curve. The fact that the car wasn’t pulling anything out is GOOD, that means it’s running happily and not having to make many adjustments. After 4-5 pulls, it will have adapted to the load and conditions on the dyno and would have started pulling a degree or two out now and then. We usually find the cars make peak power after a few pulls, not on the first pull like most cars. If you ask for too much timing, the car will actually make much less than is actually achievable. Best results are achieved with tables that are actually conservative! The stock strategy is EXTREMELY conservative with how it responds to knock. When it picks up a little bit, it will pull out MUCH more timing than is necessary, creating higher EGTs and a sloppy curve. We modify the factory strategy in a way that keeps the factory safety routines in place, but changes the response loop so it will pull less timing and add it back in quicker. It’s not “turning off the knock sensors” as posted above. Obviously we don’t recommend this, but the car can still remove enough timing to safely run the 100 program on 93, or the 93 program on 91. The stronger timing curve helps torque, power, and keeps EGTs under control. A word that keeps getting thrown around in this thread is “desensitizing”… that’s not what’s occurring, we’re simply changing the OEM strategy to allow the car to run as we desire. The OEM engineers who tuned this car weren’t looking for peak power, or pretty curves. They were given a power level, CO2 and NOx emissions limits, and fuel consumption limits and they had to make everything work in a package that has to satisfy everybody. This is why we are able to make such good gains on these cars, because of the highly conservative factory logic and the fact that the motor runs at about 70HP off of it’s potential. From a reliability standpoint, we have thousands of 3.0Ts running all over the world in various platforms without any reliability issues.
The factory fuel strategy in the 3.0T is modified as well with our code. The stock car will run fairly lean until crossing a modeled EGT threshold of about 900C, then will enrich heavily at that point to cool things off. If this strategy isn’t modified, the fueling will be inconsistent. The car is always in a targeted lambda closed loop mode, it never runs “open loop” as stated in this thread. The ECU targets a specific lambda based on mass air flow and RPM. We go richer earlier, to keep EGTs in control and allow for more ignition timing. In this case, the Revo car started to enrich at about 6500rpm probably because it was starting to get hot. On the street this will usually not happen until 5th gear or so, but it can happen sooner on the dyno with less airflow and different loading. Keep in mind a tailpipe wideband sensor as used on this dyno will read about half a point leaner than the car is actually running due to the catalyst. The black smoke noticed on the other cars is most likely a result of knock. The car will enrich under knock to try and cure it, and it’s usually visible as small puffs of smoke under heavy load.[/i]
Cool, thanks for finding that, interesting that there is nothing about boost control. Also, how it will only pull “a degree or two now and then” after 4-5 dyno pulls.
Sounds almost like they dialed out the magnitude of the timing pull…but in the logs we’ve seen, there was no timing pull at all. Interesting, curious what others think.,
Wonder if my running Eurocode headers rather than stock manifolds could be making a difference in the boost in my car?
In theory the headers should reduce the amount of boost you see. Boost is just a measurement of how hard the blower is working to push the air in the motor. The more efficient the motor is the less work the blower will be under to push in the air. The blower speeds are not effected but the work or (boost) should be lower. That is if the headers work and the motor needs them over manifolds