03-02-2020, 09:50 PM | #1 |
Major
1404
Rep 1,466
Posts |
Lag between shifts...
I can't remember who posted it, so I apologise for not giving credit where it's due, but I'd heard it suggested that the lag between shifts with the 6MT is a software thing, and not true turbo lag.
This bothered me quite a bit, mostly because it went against established thought, including my own thoughts on the subject, but it could explain why this lag between shifts a) doesn't improve by making engine modifications that in theory should b) doesn't happen when shifting above 6,000 rpm. I'd also heard that this phenomenon also happens with the M2C. The S55 is fundamentally very similar to the N55, but the induction system is different enough that I'd expect a different characteristic between shifts. So I started investigating this a little deeper by looking in to some logs. These logs were captured with 100% accelerator pedal input, shifting from second to third but before 6,000 rpm. I'm no expert at reading logs, but it appears that the Waste gate Duty Cycle trace shows the waste gate opening after I've got back on the gas, for a period of 1 second. This prevents boost from building until it again goes to 100% duty cycle and boost again climbs. I have some questions then about this finding:
__________________
2018 ///M2 LCI, LBB, 6MT...
Current Performance Mods: CSF FMIC, ER CP, Fabspeed Cat, Aquamist WMI, GFB DV+, NGK 97506, BM3 (Stage 2 93 OTS), CDV delete, UCP, M2C/M3/M4 Strut Brace, M3/M4 Reinforcement Rings Last edited by Nezil; 03-13-2020 at 12:40 AM.. |
03-02-2020, 09:57 PM | #2 | |
Banned
1448
Rep 1,414
Posts |
Funny, I started a thread with a similar question, but yours is much better.
Subbed for info Quote:
|
|
Appreciate
0
|
03-02-2020, 10:21 PM | #3 |
Major
1404
Rep 1,466
Posts |
So the chart above was a shift at 5,800 rpm. I tried on the way home today logging a shift at 6,600 rpm:
To my surprise, the WGDC still drops for 1 second, but the turbo lag was definitely less noticeable. To be honest, I had to brake immediately after the very short pull, so I need to test again on a bit more of an open road. I'm not sure this clears anything up... The WGDC drops lower than my other log with the earlier shift, but it still drops and slows the boost building. One other thing that I think is worth noting, is that in the earlier shift, the boost drops to as low as 1.6 psi, whereas when I shifted at 6,600 the lowest the boost went was 4.5 psi. The duration I was off the gas for the shift was slightly longer in the lower rev shift - 0.12s @ 5,800 vs 0.07s @ 6,600.
__________________
2018 ///M2 LCI, LBB, 6MT...
Current Performance Mods: CSF FMIC, ER CP, Fabspeed Cat, Aquamist WMI, GFB DV+, NGK 97506, BM3 (Stage 2 93 OTS), CDV delete, UCP, M2C/M3/M4 Strut Brace, M3/M4 Reinforcement Rings Last edited by Nezil; 03-13-2020 at 12:40 AM.. |
Appreciate
1
243Racing1447.50 |
03-02-2020, 11:54 PM | #5 |
Major
1404
Rep 1,466
Posts |
Since you have DCT, I think it would be helpful to see logs of fully floored 100% throttle upshifts at < 6,000rpm and at > 6,000 rpm.
In fact, just thinking about it, you would know if there was lag between shifts. 1 second feels like an eternity when you're upshifting, and while logs would be useful, subjective feeling is actually more what I was asking for.
__________________
2018 ///M2 LCI, LBB, 6MT...
Current Performance Mods: CSF FMIC, ER CP, Fabspeed Cat, Aquamist WMI, GFB DV+, NGK 97506, BM3 (Stage 2 93 OTS), CDV delete, UCP, M2C/M3/M4 Strut Brace, M3/M4 Reinforcement Rings Last edited by Nezil; 03-13-2020 at 12:40 AM.. |
Appreciate
0
|
03-03-2020, 12:12 AM | #6 | |
Banned
757
Rep 2,149
Posts |
Quote:
What you want is in those logs though. Towards the end of the pulls is where I would be shifting below 6000rpm. Just don't have the motivation to look through them myself. |
|
Appreciate
1
///M TOWN22355.00 |
03-05-2020, 12:04 PM | #8 |
Colonel
3179
Rep 2,577
Posts |
WG% does not indicate lag. It simply means how much exhaust flow is needed to quickly build boost to target or stay on target. When it shows 100%, you're off boost that's all.
For example, a laggy big turbo car, floor it. WG instantly open to 100% and stay there for like 5-6 sec, before anything happen. The channels you really need to look at, are 1) pedal input %, 2) torque request, 3) load target and 4) MAP. Pedal input is driver's intention. Torque request is how ECU react to your input. Load target is the calculated load that's actually going on, not the target of load literally. MAP is the actual boost the cylinders are seeing. If you monitor above channels closely, I believe you're smart enough to figure out where the lag comes from...
__________________
Lemania 2320
|
Appreciate
2
Yellowflash21369.50 Poochie9107.50 |
03-06-2020, 02:16 PM | #9 | ||
Major
1404
Rep 1,466
Posts |
Thank you so much for stepping in @SeanWRT, I know you're a master of interpreting logs.
Quote:
Quote:
From what you've said, it looks like the Torque request might be the missing piece of the puzzle, and if this truly is causing this 'lag', then it's a software issue that could be potentially improved... Can you think of any reason why the DME isn't allowing boost to build correctly for 1 second after a shift in 6MT?
__________________
2018 ///M2 LCI, LBB, 6MT...
Current Performance Mods: CSF FMIC, ER CP, Fabspeed Cat, Aquamist WMI, GFB DV+, NGK 97506, BM3 (Stage 2 93 OTS), CDV delete, UCP, M2C/M3/M4 Strut Brace, M3/M4 Reinforcement Rings Last edited by Nezil; 03-13-2020 at 12:41 AM.. |
||
Appreciate
0
|
03-06-2020, 02:55 PM | #11 |
Major
1404
Rep 1,466
Posts |
Good call, but I've already done that. UCP too.
__________________
2018 ///M2 LCI, LBB, 6MT...
Current Performance Mods: CSF FMIC, ER CP, Fabspeed Cat, Aquamist WMI, GFB DV+, NGK 97506, BM3 (Stage 2 93 OTS), CDV delete, UCP, M2C/M3/M4 Strut Brace, M3/M4 Reinforcement Rings |
Appreciate
0
|
03-06-2020, 10:15 PM | #12 |
Major General
7394
Rep 7,340
Posts |
|
Appreciate
0
|
03-06-2020, 10:53 PM | #13 | |
Colonel
3179
Rep 2,577
Posts |
Quote:
WG is a not a direct control of boost. It's the DV dumping boost and throttle opening angle that dictate how much MAP cylinders see and get. When WG delay to close to 100%, as part of the boost control decision ECU make to handle torque/load to pedal input, turbo shaft speed is not so off, given current and anticipated torque demand. You really need to turn on torque request to see how it's delayed to "ask" the system to serve you...After that, for a turbocharged car, MAP(IAT and timing being equal) got a lot of say on how much power it's really making. So...follow pedal input, torque request, MAP/load target, would be my suggestion. In my market there is no manual BMW today. I don't have experience of it...
__________________
Lemania 2320
|
|
Appreciate
0
|
03-07-2020, 12:08 AM | #14 | |
Major General
7394
Rep 7,340
Posts |
Quote:
|
|
Appreciate
0
|
03-07-2020, 12:20 AM | #15 | |
Major
1404
Rep 1,466
Posts |
Quote:
It seems that 6MT is far more popular in the US than in other parts of the world, and even here it is the minority choice. I have a feeling that in part, this characteristic may have been overlooked because many tuners have DCT cars and not 6MT. I did capture some logs with Torque data this afternoon though, so I'm going to dig them out now and see what we find... stand by...
__________________
2018 ///M2 LCI, LBB, 6MT...
Current Performance Mods: CSF FMIC, ER CP, Fabspeed Cat, Aquamist WMI, GFB DV+, NGK 97506, BM3 (Stage 2 93 OTS), CDV delete, UCP, M2C/M3/M4 Strut Brace, M3/M4 Reinforcement Rings |
|
Appreciate
0
|
03-07-2020, 12:24 AM | #16 |
Major
1404
Rep 1,466
Posts |
Unfortunately the logs didn't save. I'll try again over the weekend if it's not raining.
__________________
2018 ///M2 LCI, LBB, 6MT...
Current Performance Mods: CSF FMIC, ER CP, Fabspeed Cat, Aquamist WMI, GFB DV+, NGK 97506, BM3 (Stage 2 93 OTS), CDV delete, UCP, M2C/M3/M4 Strut Brace, M3/M4 Reinforcement Rings |
Appreciate
0
|
03-10-2020, 11:39 PM | #17 |
Major
1404
Rep 1,466
Posts |
OK guys, sorry it took so long to get this log. I had issues where BM3 wasn't creating logs, and I was able to fix it by force updating the OBD agent.
So here we have a log with Torque Target as suggested by SeanWRT. This was a second to third gear shift at 5,600 rpm, with the accelerator pedal floored. As we saw before, there is WGDC dip as soon as the clutch is released for a period of exactly 1 second. Boost still climbs, but much slower than when the WG closes and what I'm trying to understand is why it doesn't just close instantly? Why is it preventing boost from building when the Accelerator pedal and Torque target both suggest that boost is needed? This really does look like a software 'feature' of 6MT, and it's robbing us of power and response between shifts unnecessarily. Your thoughts SeanWRT?
__________________
2018 ///M2 LCI, LBB, 6MT...
Current Performance Mods: CSF FMIC, ER CP, Fabspeed Cat, Aquamist WMI, GFB DV+, NGK 97506, BM3 (Stage 2 93 OTS), CDV delete, UCP, M2C/M3/M4 Strut Brace, M3/M4 Reinforcement Rings Last edited by Nezil; 03-13-2020 at 12:41 AM.. |
Appreciate
0
|
03-11-2020, 01:40 AM | #18 | |
Colonel
3179
Rep 2,577
Posts |
Quote:
__________________
Lemania 2320
|
|
Appreciate
0
|
03-11-2020, 09:43 AM | #19 |
Major
1404
Rep 1,466
Posts |
Here's the link to the log SeanWRT: http://www.bootmod3.net/log?id=5e684a57d10b43215e7f315b. Ignore the stuff at the start, it's only the part at the end that's relevant of course.
__________________
2018 ///M2 LCI, LBB, 6MT...
Current Performance Mods: CSF FMIC, ER CP, Fabspeed Cat, Aquamist WMI, GFB DV+, NGK 97506, BM3 (Stage 2 93 OTS), CDV delete, UCP, M2C/M3/M4 Strut Brace, M3/M4 Reinforcement Rings |
Appreciate
0
|
03-12-2020, 01:23 AM | #20 | |
Colonel
3179
Rep 2,577
Posts |
Quote:
It's clear and my intepretation is no issue from BMW tuning. No delay whatsoever in reacting. From the log, the shift takes 1.4s to complete, from pedal lift to it 100% back on. And in it, there's 1 full sec with 50% or lower pedal angle. What does ECU do on pedal lift? 1) Throttle closure, which pretty much follows pedal input, as a direct and most effective manifold boost (MAP) control measure. 2) The throttle closure makes the backpressure to the compressor (the cold end of turbo). That's when DV comes in (log doesnt show unfortunately) and release the charge pressure (the boost in the charge pipe, or Pre Boost). To avoid mechanical failure. See there is a short time in the log (shown below) when MAP is cut but Pre boost is significantly higher until they're closer? That's what it is. 3) Wastegate fully open (to 0%) to release the pressure on the impeller. When DV's done with boost control, WG reacts to boost target (or torque target which ultimately is pedal input) with correct opening %. As said earlier, WG is a lagging measure for boost, and only a reflection of turbo shaft speed compared to boost target. A simple put - Lag is from too much off pedal time. In this particular case, when you go back to 100% pedal, the boost is already 11.8psi off (MAP=Preboost=3.4psi, rel="" target=15.2psi). See the big boost delta area. It takes a whole lot of time (another 1.3 secs) for boost buildup to target. Revisit my conclusion - human shift is too slow, ECU has no choice but dump boost for mechanical safety. And boost takes time to recover. Hence the lag. I guess stick shift fun comes at cost LOL Hope it makes sense. Sean
__________________
Lemania 2320
Last edited by SeanWRT; 03-12-2020 at 01:39 AM.. |
|
Appreciate
3
|
03-12-2020, 07:43 AM | #21 |
Lieutenant
476
Rep 451
Posts |
So don't lift off the throttle?
|
Appreciate
0
|
Post Reply |
Bookmarks |
|
|