Does the knock retard graph picture below look like real knock or false knock?
Attachment 126894
Does the knock retard graph picture below look like real knock or false knock?
Attachment 126894
thats not retard that advance and you can limit this, post file to show
It appears to be picking up knock on the shifts. Are you set to per cyl retard?
How much low boost and what spark plugs and gap
This is not real knock, not even a slight knock on shift. This is all about advance rate and torque modulation braking the "chain".
-Knock mode is per cylinderQuote:
Originally Posted by small tuner
-Boost is 9-10 psi
-gap has been tried from .028 down to .018 (this pass)
It does not seem to be real knock because it does not really care about load, it is more common to happen on the 1-2 shift, but will not always happen on the 2-3 shift if I lower the shift RPM. It also has a very odd stepped recovery that takes way to long.Quote:
Originally Posted by veeefour
If you can see the whole log you can see that it is retard (curser was on an advanced portion of the log)Quote:
Originally Posted by SultanHassanMasTuning
Log and tune attachedAttachment 126919 Attachment 126920
Let me elaborate. Shift torque modulation is not causing this because shift torque modulation calculates timing as a function of shift torque, brake torque, mbt, and the S-curve without affecting knock advance. I thought it might be the instantaneous false load spike due to maf delay vs shift rpm drop triggering tip-in retard, as I’ve seen this reset the knock advance to 0 before, but yours crossed zero and went to knock retard. Many things can cause knock retard to go to 0, but only sensed knock can make it go positive. The sensed knock can be real or false.
That said, if you have it set to per cyl retard and SDLF set to 0’s then you can log individual cylinder knock and it might reveal a clue as to the source and/or minimize the effects.
for that boost level i think you gapped your plugs way too tight
Zapo, the extreme negative timing during the shift is caused by shift torque modulation. You are using Drag mode which is Shift Character 1 (table 20044). In Trans torque management Alt 1 for the 2-3 shift (table 18275), for instance, it calls for -96 ftlb of torque. This will result in it pulling max timing (tables 10092, 6662, and finally reverting to 42365 for a value of -30) to attempt to achieve the -96 ftlb. If it were me, I would change max number of characters to 2, change the character to 2 in Drag mode, and then lower the shift torque in Alt 2 to around 500 ftlb. This will result in firmer shifts and less torque (timing) cut during the shift. You might also have to later address the transient tip-in retard because your load rises rapidly on the shifts. I do still think that false knock is causing the knock retard to go positive on the 2-3 shift. You might even consider trying character 0 to see if the gentler shift helps prevent false knock.
Something else to consider is changing SDLF to 0 in table 45323, which would allow per cyl at high rpm. Then add all 8 cyl knock retard to your log. Speaking of, you have way too many channels logged, which is slowing down the data rate. It's not typically that beneficial to log all the mapped point stuff and trans stuff unless you have a specific problem there to solve. I would focus on spark related data plus torque management at the moment. Also, make sure your spark source channel is on the fastest logging rate possible.
OK there is a lot here. Let me break this down a little.Quote:
Originally Posted by engineermike
1. Do you think the trans shift torque modulation, extreme neg timing, is a problem? (I normally leave this on purpose because I want clutch's to live and I value that over every last .001 second in the 1/8th)
2. I think I do have two knock related problems
A. false knock on the shift (have some ideas on the exhaust and turbo mounting that I will look at).
B. It takes a very long time to get the timing back in after a knock event on the shift.
Do you agree with the above assessment?
Yes I agree, .018 is not necessary, but it made no detectable difference over .028 so I just left it. This was logs from a drag and drive this summer, I was amazed mileage did not even suffer.Quote:
Originally Posted by SultanHassanMasTuning
1. Personally I agree regarding sacrificing a tiny bit of time for parts life. However it’s going to -96 ftlb. I consider myself conservative and cut to 500 thinking it’s not much higher than stock. Many leave it at 738. One issue is that it’s max’ing out the timing retard at -30. Per the sai torque reduction logic, if it can’t get the full reduction from retard, it will move on to enleanment then cylinder cut. Cylinder cut will pump a moment of fresh air into the manifold. I think I’d at least raise it to like 400 and test. Also, going to character 2 will speed up the shifts and allow less slippage.Quote:
Originally Posted by Zapo
2. A. Definitely check it out.
2. B. The recovery rate is very easy to manipulate and speed up but this is just mitigating the effect not the root cause.
I have the TT mustang on the dyno this week as my shop is closed for Holidays, so I can just leave it on the rollers. Logs are below, they do have per cyl knock logging added and cleaned up some unneeded trans pids, tune is the same on all pulls, just trying to baseline. Air temp is at freezing so I all runs were done with 91 octane in the tank so I could find octane limit. Car is running meth injection. Dual exhaust system when open is 3" to dump, when closed it 2 1/2 inch through cats and stock mufflers. Boost controller was not adjusted on all 4 pulls, it was about 5-6 psi with full exhaust and bumps to about about 8 psi with open dumps. Engine is built with 10 to 1 compression and not the stock 12 to 1 that the gen 3 coyote comes with.
Quick run through 4th gear peaked at 5 psi boost, exhaust closed
Attachment 127002
Shifting gears on the dyno, can see some det come on the 3-4 shift. Exhaust closed
Attachment 127003
6th gear pull Exhaust closed (about 600 RWHP)
Attachment 127004
6th gear open exhaust, det really coming on nearing redline when I can usually feed some more timing in, but it did not want it. (about 770 rwhp with peak at 6500 rolling off as timing fades.)
Attachment 127005
I took a quick look and have a few comments:
Still logging way too many channels. Get rid of most or all of mapped point weights unless there is some specific mapped point issue you are chasing. The lift and GDI pressure fine, so that's 4 more that can be removed. DI blend mode will be warm every pull, so it's not needed. The DI injector PW is useless. There are easily a dozen more PIDs that can be removed. As a result, the per cyl knock data is much too coarse, as you are only getting 3 or so updates per cylinder during the pull.
Your DI blend seems a bit low. You should be able to maintain at least 75% blend with the stock pump and that boost level.
It's definitely picking up what it thinks is knock. I don't know if it's real or not, but on the street pulls it happened only on the shifts but on the dyno it's happening across all cylinders during the entire pull. I tend to think it's false for that reason, but that's not conclusive. You could always try something like 10 deg timing at min boost and if it still does it then it's 99% chance false knock.
MAF translation curve could use a little work. It's pulling 10% fuel at WOT. I like to get this below 5 or even less.
1. I will reduce more channels, I had the Mapped point channels all active as I was till playing with IMRC functions at light cruise and although everyone deletes it with boost, I was trying to figure out if it had any value.Quote:
Originally Posted by engineermike
2. I will feed a little more DI blend in and see if I can maintain pressure.
3. The knock I see on the dyno runs, even with shifting, does not really seem the same as drag racing knock, but I might need to add a little more load to the roller.
4. The MAF curve has not been retuned for Meth injection, IT is very close with meth off, because the meth does not come on a quickly as the real time demands exist I tend to not trim fuel for it, not sure if that is wise.
2. I believe it best to just set the gdi warm blend to 90% and address whatever other issue you have through specific control. If you are concerned about losing gdi pressure, then lower the max injection angle until pressure stays where you want it. Use each control for its purpose and you’ll get a better end result.
4. I understand now. That being the case, if the fuel trims are correct with meth off I would probably leave it alone, but perhaps disable the LTFT.
Spent some more time on the dyno today, Turning meth injection on and off seemed to prove knock as observed on the dyno is real. Played with DI and was able to get it to 75% and make more power, also played with VCT to see if intake or exhaust had a notable effect on knock, it did not not. The most interesting thing I found was analyzing Cylinder knock data with Meth on and meth off.
Attachment 127030
On this pass with meth on, the individual knock data was pretty good minus cylinder 1, I question dry intake manifold meth distribution from a single nozzle. Also I am not sure how it is using individual knock data, does it actual tune each cylinder or just global tweak based on the worst cylinder?
Attachment 127031
This pull is obviously without meth, knock is worse and all cylinders power is down a bunch and boost is down about .6 psi without adjusting the boost controller. I have seen meth add a little boost, probably just that.
Attachment 127032
This is the what I ended the day with, still happy to get over 800 hp on 91 octane junk fuel and meth, but this would definitely not be a good real use tune, just dyno hero pull on a cold day. Really wondering why my combo just does not want to make power above 6500 rpm (dyno chart RPM is a little off, RPM on log appeared to be about 100 RPM higher than dyno chart). Some of it is timing for sure, and it does have a 1 psi boost roll off, that I was attaching with the boost controller and you can see the power just start to pickup before run ends. I need to get a boost sensor for my dyno, was just video logging my boost gauge.
Attachment 127033
That's great power at that boost level, but why is it tanking above 6500 rpm?
The GDI pressure is lagging in the middle and up top the blend isn't increasing as more flow is available. You can command 90% blend and then use max injection angle to preserve mid-range pressure. This would cut the blend in the middle to below 70% but then allow 75 or more up top.
The knock data scares me a bit. If it's real knock I would be limiting the timing to values that result in little or no knock reductions. It seems like the timing is a bit low at that load, especially with meth, but it could be normal for a turbo setup.
On the cylinder knock retard, that is the retard applied to just that cylinder. Though, your SDLF is set to 3, which I don't know what causes. I only have experience using 0 and 4. Stock and Whipple use 4 at high rpm, while Roush and other Fords use 0.
On another note, your cam timing isn't really cooperating. It's not switching to OP until 4000 rpm which is probably slowing spool up. There are various things that could cause this, but one thing that can be done that helps other things also, is to have the highest load row of the Drivability cam timing to match OP. That way even if it doesn't switch to OP it still uses the right cam timing.