Morgan, do i have the .75 sec delay ? how can i see this in the logs ?
Morgan, do i have the .75 sec delay ? how can i see this in the logs ?
Yours was a bit shorter. If you look at the log. You will see trans current gear change well before your actual shift point. This is based on the MPH in the tune. The actual shift can be identified by the Torque management kicking in at the shift. Also you will see the timing for current shift begin. It's not an exact science, but you will see the delay I'm speaking of.
What if, the shift has to be commanded (using rpm) to happen before, or at, the mph in the tune ?
It doesn't make sense, but the 3-4 shift issue doesn't make sense.
Notice how in the stock tune, the 1-2 commanded, happens 1 mile above the mph, but 2-3 and 3-4 are either below the mph, or at the mph ?
What if the delay is happening because the trans is expecting to reach the next mph commanded, based on the previous speed reached by a previous gear ?
ex: 2-3 in the stock tune, the shift is commanded at 62mph, but in the logs it starts at 60mph, and 3-4 starts at 98mph, right where the commanded mph is set in the tune.
In my modified tune, the 2-3 mph is commanded at 64mph but starts at 65mph, and the 3-4 is commanded at 102mph, but is starting at 103mph. the rpm is different though, but somehow rpm helps to reach the expected commanded mph.
Just a speculation, i'll give this try tomorrow lowering the rpm for 2-3 only and see what happens with the 3-4 shift mph. we still don't know what exactly tells the trans to perform the shift, but maybe we can help the issue telling the trans to start the commanded shift at the right mph using rpm to do so.
Hope it makes sense, but more importantly, hope it works
Last edited by bluegoat06; 07-31-2010 at 06:16 PM.
Alright I took it out for a drive.
Here's how I have the shift points set now.
1-2 21.6mph 4450rpm
2-3 46.1mph 5200rpm
3-4 75.0mph 5650rpm
Here's what it did on two different runs. First was a roll 2-3 3-4, the next was from a dig.
2-3 56mph 6307rpm .687sec (commanded at 48mph)
3-4 85mph 6235rpm 1.02sec (commanded at 77mph)
1-2 35mph 6790rpm .718sec (had to roll, 2nd gear commanded right at WOT at 25mph.)
2-3 52mph 5900rpm .621sec (commanded at 48mph)
3-4 84mph 6189rpm .765sec (commanded at 77mph)
So I really don't feel like I will ever get the shift right until we can identify why there is the delay and how to modify it. There are too many changing conditions. It doesn't seem like the RPM does anything other than adding another inconsistent variable.
Maybe I'll have D be a little lower and S be a little higher, lol
On a positive note there was no weirdness happening on the 3-4 shift. So I think what I am going to do at the track is keep upping the MPH and match the RPM to that MPH point until I get it to shift where I want.
Last edited by morgan72; 07-31-2010 at 06:16 PM.
Did you change the gear in the cal to match you new final gearing (3.70)?
sure did. I actually used the auto calc for everything initially too, but I re-did it manually. For some reason it would change the tire revs per dist.
When I ran at the track on Wed, I used the auto calc. Now I am using it with the manual calc. I'm going to go to the track again on Wed and give it another shot.
Well, the theory didn't work. i lowered the rpm from 6200 to 6100 for the 2-3 shift, and the shift started at 65mph, but now it looks like it delayed the 2-3 shift. i didn't feel it, but the log shows it.
It looks like she's not paying attention to the rpm, but yet, the shift tries to complete around the rpm, how can this be ??? this is exactly what i've seen at the track whe the 3-4 delay happens. shift starts around the mph, then it's like it wait for the rpm, exceeds the rpm, then wait for the rpm to go down again, then shift completes. in the meantime, the car doesn't accelerate anymore, and it feels like the rev limiter is reached around that shift rpm.
morgan72 the shift time isn't measured from when the shift is first commanded. What happens (I think you've mentioned some of this in your latest posts) is that when the shift is commanded the TCM starts to fill the oncoming clutch with fluid to get it ready to handle the torque load. Once its full then the torque is transfered from the offgoing to the oncoming clutch, then the oncoming clutch de-celerates the turbine and engine. The actual shift time is measured from when your rpms start to drop until they are finished dropping. All the stuff that happens before this isn't affecting the cars speed you are still accelerating while its happening. We don't have access to any of the parameters that control that stuff anyway.Originally Posted by morgan72
Looking at the log where there is the huge delay for the 3-4 shift it is very odd. The actual shift is done quickly but there is a big delay beforehand. It looks like the rpm stays constant for a while as if there is some form of torque management involved but the throttle angle and timing don't change so I don't know whats causing it. Maybe if all the PCS pressures are being logged while it happens it might give a clue.
VE Calais V 6.0L L98 V8 6L80E
Certainly very odd. Tq jumps up 30ftlbs or 45Nm with a jump in MAF and drop in AFR during this "lag" period. Long shot but maybe the oncoming is a little early.
carneb, That's the problem it's not the shift time it's the delay from the commanded shift to actual shift. That throws in a huge error opportunity that will affect the shift point based on a given condition. More power or gears will cause the car to run through the gear quicker which will raise the shift point with no other changes. So the question becomes, is the delay a mechanical limitation of the transmission or is there an additional programmed value that is added so the transmission is "not rushed".
I've attached my current tune and a log. The WOT run is at about 14:06
I haven't had a good opportunity to run a 1-2 2-3 3-4 from a stop yet, but the 1-2 and 2-3 shifts are just about perfect. I started from scratch and used the factory tune. Shift torque management is stock on the 1-2 and 2-3. It is off on the 3-4. Most of the shift improvement was through lowering the shift times down by modifying the inertia adder. I bumped up the Oncoming volume a bit on the 2-3 and 4-5 which seemed to help. I have not touched the oncoming and offgoing pressure presets or the shift pressures except for the max pressures. Also I have disabled any 4th gear lock-up for now, full and part. With the 3.70 gears, it's really not needed anyhow.
What I really like about the whole tune is my part throttle shifts are much better now too. In D it's a positive feel, and in S it's a bit quicker and firmer. I haven't really messed with M because of the delay from moving the stick to actual shift.
It seemed to take the transmission a few minutes of driving to adapt the shifts. I wish we could log those adapt values so I could adjust the tune to match the adapt, much like we adjust the fuel to zero the trims.
bluegoat06,
check out your stock tune log vs your current log. Pay attention to PCS1 and PCS 2. I think this is what we need to look at.
For 2-3 at the shift
Your stock Tune
PCS 1 - 48.9
PCS 2 - 53.0
Your last upload
PCS 1 - 63.6
PCS 2 - 55.9
My upload
PCS1 - 48.9
PCS2 - 60.7
With my quicker shift program it upped the PCS2 pressure, but my PCS1 stayed about stock. Your program has a much higher PCS 1 pressure. Not sure if this is Oncoming or offgoing but it is probably one of them.
So here is your stock 3-4 vs my 3-4 at the track with my problem shift
your stock tune
PCS1 - 64.8
PCS3 - 47.8
PCS2 - 25.1 (just before the zero frame)
PCS1 - 78.0
PCS3 - 54.5
PCS2 - 29.1 (just before the zero frame)
Whatever controls PCS1 in the tune may be part of the problem. More pressure may not be the answer.
The way I understand it is, the shift occurs at either the MPH or RPM, whichever is sooner. If you're changing one of the parameters and it still shifts the same it's because it's shifting off another parameter.
I also see the commanded shift delay in my C6. Once the shift is commanded it usually takes .7 to begin the shift. The PID for Shift Times only records the actual shift, it doesn't include that delay. I've only made one test but I reduced all the Sport shift time adders by 1/2 and saw a small decrease in shift times but that delay seems constant. I think if we find where the delay comes from, we'll finally fix the paddles. It's obviously not the actual shift times, it's that delay where nothing seems to be happening.
Basically, I reduced all my base shifts to .15 sec and all adders to 50% of stock. It's definitely better but I'd still like to shave off more on some shifts. My shifts are roughly 1 sec total.
2007 3LT Z51 A6 Coupe, Kooks Headers, NPP Exhaust, Vararam, Ported Intake & TB, TBS, 160*Stat,
B&M Trans Cooler, HPtuners, Line-Lock, Racemesh, Z06 Rear Guards
Best 1/4 ET: 11.87 @ 120.02
1998 Convertible - Car used in the filming of "Mr Deeds". SOLD
Yes, exactly. Same on downshift. Shift occurs fast but recovering timing advance takes over a second on the 5-1 or 6-2 downshift. The 4-1 only takes .4? seconds. I feel somewhere in the spark tables is creating this upshift and downshift delay. Of course on upshift you can disable and shift encounters zero timing reduction. I dont want zero reduction but rather control over the amount of reduction.
Which pattern are you having the trouble with? Normal, A or B?
Your B is all zeroes for part Throttle and WOT on speed so assume its Normal or A.
I looked at it, but i don't know where this PCS pressures are coming from
There is nothing in our tunes (at least that i can see) where this exact values would be getting referenced at. The only thing i did as you probably saw in my tune, was to add more pressure in the Max line pressure from D1 through D4, matched Max Pressure with Max pressure B values, and upped the Shift pressure upshift for X,Y and Z profile (by the same amount).
I've also modified the inertia profiles and shift times, TM times and inertia profiles.
BTW, the way i have things setup in my tune would make the 3-4 shift to happen as it should, or last least, that's what i've experienced at my last test&tune where out of 10 runs, the last 2 passes were clean, This is the one tune i've been wanting to test again, but haven't have the time to do so at the track.
Bill@HPT or anybody @HPT, it would be awesome if you guys can elaborate more on this commanded to actual shift delay and what we can do to control it.
Last edited by bluegoat06; 08-02-2010 at 03:33 PM.
The timing delay is the only reason I'm running with TM turned off. I couldn't find enough parameters to get rid of the timing pull. It's just wayyyyyy too much timing pulled for wayyyyyyy too long for a sports car. I got tired of the car falling on it's face every time I lifted. My TM is basically controlled by my right foot which is tricky with an auto.
Now that I have the shifts and flare solved, I'm open to suggestions on how to get the TM turned back on and get the timing under control.
Last edited by C5pilot; 08-02-2010 at 03:33 PM.
I am pretty sure the PCS values are based off the oncomming and offgoing lookup tables along with any learn the TCM has done. It would be nice to know exactly what the y-axis 0 1 2 means though.
C5pilot, based on my testing a WOT shift is initially based off of MPH only. There is then this preset timer of about .75 sec that we don't seem to control. Then it will reference the RPM shift point to insure that the RPM has been reached. Right now the way my shifts are wet up, it is 100% MPH and then this preset delay. Once we get control of this delay we can in theory precisely control the shift off of RPM.
Did these runs shift fine into 4th? Reason I just ran the car on the street and I had the 3-4 hang up. About a tenth of a second prior to the actual shift PCS2 went to 0 psi. On your logs PCS2 goes to 0 right at the shift. So this may be the identifier of the problem.
I think anything we do is just a bandaid though. We are still needing certain parameters that control these triggers.
I'm on a roll, lol. Here is my latest file. PCS2 still did the same thing, but no lag. I upped all the oncoming volume presets. Excellent, excellent... plus a better part throttle shift. I then copied the 3-4sp oncoming and offgoing pressure presets from the CTS-V as well as the Shift initial and Shift final transition % 3-4sp.
I also tweaked all the shift points based off of MPH only and increased the shift RPM the equivalent. It is all nearly perfect. I plan on raising the 3-4 shift a bit more, but I really need a track. I just don't have that many test spotsThere was no lag and the shifts really pushed the car forward, even with 100% tq management on 1-2 and 2-3.
Can someone try out all the parameters, including the torque management, but leave all your shift points the same?
Nice work Morgan, i can test it but i have to wait for the weekend, i only have from 6:00-7:00 am Sat-Sun to test safely.
Answering your previous question, yes, in the tune i posted the 3-4 didn't have any lag that i could feel or see in the logs.
Reply With Quote
