Page 5 of 8

Re: Trigger hysteresis using avload instead of load?

Posted: Fri Mar 27, 2015 10:58 am
by Bernard Fife
Addicted,

Here is new 5.099zviii code that has one small change; it makes sure that the new load hyst value is the greater of the instant load or the average load (the average load is normally set to the instant load if it is higher, but this wasn't done previously during a shift, so they could get out of sync). Combined with the delay, this might help.
Monitor_5099zviii.abs.s19
Uses previous INI
(83.16 KiB) Downloaded 615 times
Lance.

Re: Trigger hysteresis using avload instead of load?

Posted: Fri Mar 27, 2015 12:55 pm
by Addicted
Thanks Lance,

It looks like our thinking is starting to converge.. or maybe I'm understanding better ;)

You remember a few weeks ago I asked you to alter the code to allow the hysteresis to use avload instead of instant? that was because I felt that spikes in the load when shifting were turning off the hysteresis, my hope was that using avload would smooth it out a little as can be seen in this image:
Image

As we know, it didn't quite work. But if I was missing points where hyst=0 because of the granularity of the datalog it makes a lot more sense. I'm assuming on the bench your instantaneous load isn't quite as 'spikey' as my realworld load either..

I've tried quite diverse settings for the load smoothing, in fact I was under the impression that it WAS set to 1000, but you're right its currently 3000.

I'm itching to get out and try your new code now... but the weather outside is blocking me.. :cry:

Re: Trigger hysteresis using avload instead of load?

Posted: Sat Mar 28, 2015 8:29 am
by Addicted
OK, tried the new code.. not good, see log attached :lol:

avload is either 100kpa or 0kpa... I tried to drive it, it nearly killed me.

Re: Trigger hysteresis using avload instead of load?

Posted: Sat Mar 28, 2015 12:52 pm
by Bernard Fife
Whoops, sorry! I think I know why (the load average is only updating during shifts....), but I need to re-think this obviously. Lance.

Re: Trigger hysteresis using avload instead of load?

Posted: Sat Mar 28, 2015 2:19 pm
by Bernard Fife
Martin,

Were the above logs with 5.099zvii or 5.099zviii?

If this was mine, I would revert to the code previous to whatever you generated the log with, but try to shift the load more to the TPS. If that helps, we can play with the TPS /RPM/load table a bit too.

Lance.

Re: Trigger hysteresis using avload instead of load?

Posted: Sat Mar 28, 2015 5:12 pm
by Bernard Fife
Addicted,

I had another thought about this; it's related to the decel mode. I believe this, and the way that the average load increases with the instant load, will help explain the issue quite well. So what I am going to do is roll back the code a few versions, and implement fixes to the decel mode and load averaging. This may take a day or so, I will post back here when I have something for you.

Lance.

Re: Trigger hysteresis using avload instead of load?

Posted: Sun Mar 29, 2015 1:33 am
by Addicted
Lance wrote:Martin,

Were the above logs with 5.099zvii or 5.099zviii?

Lance.
Hi Lance, those two logs are with the latest 5.099zviii

Re: Trigger hysteresis using avload instead of load?

Posted: Sun Mar 29, 2015 2:44 pm
by Bernard Fife
Okay, thanks Martin. I have some ideas based on your logs. I need to implement these and test them a bit, then I will post something here (hopefully tomorrow). Lance.

Re: Trigger hysteresis using avload instead of load?

Posted: Mon Mar 30, 2015 1:39 am
by Addicted
Lance, can you confirm something for me?

If I set TPS load to 50% does that mean 'load' becomes the average of TPS(%) and CAN load (KPA)?

So if TPS is a steady 30%, the TPS/REV table cell is 100 and CAN load is 70KPA then 'load' will be 50 (KPA)?

and if TPS is a steady 30%, the TPS/REV table cell is 70 and CAN load is 70KPA then 'load' will be 45.5 (KPA)?

All my early tuning was with TPS load at 50%, but after a while I reverted back to 100% KPA... might be time to go back again..

Re: Trigger hysteresis using avload instead of load?

Posted: Mon Mar 30, 2015 7:08 am
by Bernard Fife
Martin,

The way it works is like this:

- first the raw TPS is passed through the table (with the rpm) to get the adjusted TPS.
- then the adjusted TPS is combined with the MAP (kpa) on a percentage basis to get the Load.

If user really wants to use raw TPS for load (with no adjustments), they can set the table so the the cells in each TPS row are identical to the bin value for that row, then the output from the table will be the raw TPS.

There is more here: http://www.msgpio.com/manuals/mshift/V5tune.html#gh
If I set TPS load to 50% does that mean 'load' becomes the average of TPS(%) and CAN load (KPA)?
It becomes the average of adjusted TPS and the MAP.
So if TPS is a steady 30%, the TPS/REV table cell is 100 and CAN load is 70KPA then 'load' will be 50 (KPA)?
Assuming the TPS/MAP ratio is 50% and the adjusted TPS (current TPS/REV table cell value), the load value would be: 100%TPS x 50% + 70kpa x 50% = 85%
and if TPS is a steady 30%, the TPS/REV table cell is 70 and CAN load is 70KPA then 'load' will be 45.5 (KPA)?
Assuming the TPS/MAP ratio is 50% and the adjusted TPS (current TPS/REV table cell value), the load value would be: 70%TPS x 50% + 70kpa x 50% = 70%

I am working on the code, and making progress; hopefully I can post something soon.

Lance.