Accuracy of Delivered Torque PID; Torque mgmt
Logged this param today because I was trying to
play with the torque management settings (roll off
timing only on the very high end of operation, not
part-throttle). The torque management table I have
to work with is a torque reduction % vs torque. Not
clear on whether this is delivered torque or commanded
torque, as the axis. By the way, the "Desired Torque"
PID returned 377 all the time (even at idle... odd).
The values I got for Delivered ranged from about 20
(idle / low cruise) up to a little over 200 (lb-ft?).
This included some WOT pulls in 2nd, where I would
have expected to see more like 300lb-ft range of
flywheel torque (per the specs of the LS-1 motor).
What are the units of "Delivered Torque" and how
realistic should I expect the value to be vs measured
FWT?
So I'm wondering just how accurate, or not, this
PCM-computed Delivered Torque value is expected
to be. And, if the "Desired Torque" is just a static
value, what would be the PID that corresponds to
the index of the torque reduction vs torque TM table?
What I'm trying to do is just "crop the top" and leave
part-throttle shifts "unmanaged", and have a small
tolerable timing pullback at WOT.
Setting the torque reduction table to hold 300lb-ft
at all values above 300lb-ft (by fitting the reduction
%, increasing) seemed to eliminate TM pretty much
while setting it in the same manner to 200lb-ft made
is come back, a bit lower down than my goal. While
I'm sure I can binary-search my way to satisfaction,
the same can't be said of enlightenment.
Re: Accuracy of Delivered Torque PID; Torque mgmt
hi jimmy, the desired torque amount, is what ever you put in for MAX Torque tab, it is a set value that the pcm uses for tm stock is 350,"on ls1edit" it wont ever change unless you change it, most of us change that value to 600 or something to that size, so the pcm never hits that value and kicks in tm.. This number means nothing for performance. Just torque management, for pulling timing ect.. As for the delivered torque, i dont think it is accurate at all, but, somepeople have been able to just log it while making changes and if the number went up, then they keep going in that direction.. Autotaps scanner is for sure wayyyy off.. i checked on the last dyno, i hit 574 ft pounds of torque and autotap showed 698..
if you want to just crop the top,,, all you would have to do is just put zeros in all the rpms that you want on your tm tables...
Re: Accuracy of Delivered Torque PID; Torque mgmt
since this is an old'ish thread...
Is this more in the range you expect now? From memory we had a little units trouble on the Torque PIDs.
I just logged my own car and it was showing about 580Nm/430ft.lbs which is pretty close to the money for me.
Re: Accuracy of Delivered Torque PID; Torque mgmt
Yes, it appears much more sane (in the 1.0.0 release
I was seeing 210 lb-ft peak; now it's in the 350-400
range and the transmission one w/ torque mult figured
in is maxing over 500 lb-ft).
Have a question about the transmission delivered
torque PID. I guess it works by assuming the stock
converter's STR and torque mult-vs-slip-vs-torque
model. For those of us w/ aftermarket converters,
is there any way to change this "model"? My STR is
2.2 vs 1.8 stock, my stall speed is 3000 vs 1800,
my TCC slip seems to sit about 600-700RPM under
load @ midband, all quite different from the underlying
assumptions I think the trans DT calculation must be
using.
But, as to the original question, I'm feeling much better
now. The only thing that would make me grin is if there
were a flywheel dyno vs engine delivered torque PID
curve pair-up to gauge the absolute accuracy. But my
use for this is really just relative-to-last-tweak tuning
comparo, and my torque management profile fiddling,
so it's in the "nice, but not necessary, and not bloody
likely", that dyno bit.
Re: Accuracy of Delivered Torque PID; Torque mgmt
yeah there is a torque multiplier vs slip RPM table. I'll put it on the to do list.
Re: Accuracy of Delivered Torque PID; Torque mgmt
Can anyone explain the mechanism by which
the Delivered Torque is derived? I think it's off
the crankshaft position sensor's ripple /
acceleration, somehow, but how the PCM would
get rid of the down-shaft load (esp. on a hard-
coupled manual trans) kind of mystifies me.
I would like to know what kind of input & algorithm
is involved in producing the calculated torque
(at a fairly general level) just to know what kind
of assumptions or sensitivities might be embedded
and whether (say) if I change timing, to improve
torque/HP, but timing is part of the calculation, it
might be lying to me about improvements in the
torque output (or, is it an "impartial judge"?).
Re: Accuracy of Delivered Torque PID; Torque mgmt
it uses a combo of current AFR and Airflow to make a base torque estimate. Then allows for any cylinders that are disabled (usually 0), then accounts for parasitic losses such as friction, is the AC on etc. and then finally checks how the current ignition timing compares to the ideal timing for maximum torque (a spark table just like the HO and LO tables, genernally this is known as Mean Best Timing or something). If the current timing is less than the MBT value then it reduces the calculated torque value in accordance with a torque reduction % vs MBTdelta table (non linear). At the end of all this you get the final torque estimate.
Re: Accuracy of Delivered Torque PID; Torque mgmt
Hmm... interesting. Then I guess the Delivered Torque
is maybe not such a good "feedback tool" for checking
the performance as I'd thought, because it is really
"predicting" rather than measuring, and basing its
predictions on the very things I'd be playing with in
all likelihood. :-/
Re: Accuracy of Delivered Torque PID; Torque mgmt
well, it might not be so bad. The AFR has all the normal errors associated with open loop but its going to be pretty close. Certainly a relative change would be consistent.
Airflow is airflow, unless you are running SD and have a wildly out of whack VE table that you are compensating for on the fuel side then it will be fine no matter your mods.
Parasitic losses would be the same (likely).
As for the MBT proportional Tq loss, i'm not sure where that comes from but i assume it comes from some form of burn rate analysis on the cylinder head and piston design. Many things could change this but for the most part it should be okay.
From my perspective i run SD, with a 240+ cam and ported heads and the torque number here matches i'd say within 5% of my dyno number, notwithstanding the possible error in converting from Dyno rwkw to engine kw.
I think the airflow and the spark are the most important terms of this calculation and if you are running MAF and fuel with "normal" characteristics it should be ok.
Re: Accuracy of Delivered Torque PID; Torque mgmt
Quote:
Originally Posted by jimmyblue
Hmm... interesting. Then I guess the Delivered Torque
is maybe not such a good "feedback tool" for checking
the performance as I'd thought, because it is really
"predicting" rather than measuring, and basing its
predictions on the very things I'd be playing with in
all likelihood. :-/
Not to mention when I turn my AC on, the engine torque increases (as it should) but so does the delivered torque by the same amount. Now I know the AC compressor torque used did not make it to the trans.
GM can chime in anytime here :)
Ken
