Page 3 of 5

Re: What exactly does tplate % represent in datalogs?

Posted: Wed Apr 27, 2016 3:19 pm
by Hondata
The S2000 is the only vehicle we've seen using the additional index table. The US 06 Si and FN2 don't use that table but the throttle table index has changed like the S2000. FN2 table:
fn2-th.png
fn2-th.png (40.89 KiB) Viewed 5880 times
I'm not sure why Honda has scaled to table in such a way for the S2000. It almost is like a mix up in units, since all the other ECUs have tables like the one above (they still use a different scale for the Pedal however). It could be confusion between percentage opening and degrees opening. It could be that it was designed that way and we don't expect to see a table like that.

Re: What exactly does tplate % represent in datalogs?

Posted: Thu Apr 28, 2016 6:39 am
by R-12914
Is it also possible, to calibrate the throttle dependent by the vehicle speed on a FN2?
Thats would be nice and better for city driving!

Re: What exactly does tplate % represent in datalogs?

Posted: Thu Apr 28, 2016 7:30 am
by SlowTeg
R-12914 wrote:Is it also possible, to calibrate the throttle dependent by the vehicle speed on a FN2?
Thats would be nice and better for city driving!
From hondata's post above:
"The S2000 is the only vehicle we've seen using the additional index table. The US 06 Si and FN2 don't use that table"

So the answer is, no.

Re: What exactly does tplate % represent in datalogs?

Posted: Thu Apr 28, 2016 9:36 am
by R-12914
I now, but is it possible to add this option for the FN2?

Re: What exactly does tplate % represent in datalogs?

Posted: Thu Apr 28, 2016 3:24 pm
by Hondata
We're not going to alter the DBW operation.

Re: What exactly does tplate % represent in datalogs?

Posted: Wed May 04, 2016 12:52 pm
by shind3
Very interesting. Thank you for tracing the code a little better. I'm disappointed it took so many years for you to take some of our posts seriously enough to look into it. I've disassembled close to a 100 relevant maps out of a late 90s BMW ECU so I know how easy it is to check the math on values used in tables with the help of IDA. But it takes time to verify. You guys just assumed the S2000 code runs similarly to other Hondas even while knowing that the S2000 and the TSX are the only weird DBW implementations (to my knowledge) out of the Honda line up to use an external DBW driver. Anyway, let this be a lesson learned and work with us on issues instead of taking a defensive or apathetic stance. And give us the technical nuts and bolts. Some of us can handle it and are not the kind of customers who require a 'walled garden' to operate in.

Anyway, just knowing the background table and math is extremely helpful even if we aren't allowed to adjust it. And yes, I agree from a liability standpoint of not touching the DBW code. Just adjusting calibration parameters ought to be enough.

Re: What exactly does tplate % represent in datalogs?

Posted: Wed May 04, 2016 3:55 pm
by Hondata
shind3 wrote:I'm disappointed it took so many years for you to take some of our posts seriously
As far as I know we were only informed about this two weeks ago. The original posts were a little cryptic, which is why the answers were narrow.

Re: What exactly does tplate % represent in datalogs?

Posted: Wed May 04, 2016 4:41 pm
by shind3
OK, fair enough. And in that case, thanks for your support. :)

Re: What exactly does tplate % represent in datalogs?

Posted: Thu May 05, 2016 8:58 am
by SlowTeg
Hondata wrote:Release probably will be next week.
Just a reminder that some of us are eagerly awaiting this release. It'd be great if I could give it a try at my next track day may 15th. I know things take time, but just a friendly reminder. No rush.. :D

Re: What exactly does tplate % represent in datalogs?

Posted: Fri May 06, 2016 2:19 pm
by Gernby
I'm looking forward to the new throttle table too! Any chance that you'll enable live tuning on one or both of these tables? It's really nice that the throttle table is live tunable on the Civic...

BTW, here is the first thread I posted about the target throttle issue back in 2010.
http://www.hondata.com/forum/viewtopic.php?t=10631

Re: What exactly does tplate % represent in datalogs?

Posted: Sun May 08, 2016 12:23 am
by shind3
Well, like I said, let this be a lesson learned. 6 years is a long time to wait for support.

I'll copy/paste what I said on S2Ki about not being able to tune the background table. So I don't believe it is that urgent for Hondata to include this stuff in the next release. You can tune for an adequate majority of the existing logic just fine as is as long as you stay vigilant of what the numbers actually represent.
As of version V2.3.7, Just use the first 10 columns of the throttle table and for the axis just equate 0 to 66.7% as actually 0 to 100% 'pedal position' . This would be 100% correct if not for the "VS vs Pedal" background table. Stock, That table is tuned to desensitize the throttle at lower VS (marginal effect) . Also, the actual Values in the main throttle table are off by the same scale. 0 to 66.7 actually equals 0 to 100% throttle plate. This is a hex conversion formula error for now.

Hence, Ignore the last 4 columns as Gernby instructed years ago.

And finally, what Hondata wants to rename from 'Pedal' to 'Index' should actually be called 'Vehicle Speed Compensated Pedal Position'.

Re: What exactly does tplate % represent in datalogs?

Posted: Sun May 08, 2016 10:15 am
by Hondata
Yes, there was evidence of this problem in 2010. However, the values used in the example explained the behavior at the time and I did not investigate further. I apologize. When this was bought up two weeks ago I think our response was better.

You can not reduce the behavior to a single table with a constant conversion factor, and exposing the underlying index calculation is the right way to do this.

Gernby: Sorry, no life tuning for DBW. We can't change the logic or behavior.

Next release will be Mon or Tue.

Re: What exactly does tplate % represent in datalogs?

Posted: Sun May 08, 2016 3:52 pm
by shind3
Yes, response this time was fantastic! Thank you. And also thanks for such quick timing on a fix!

Re: What exactly does tplate % represent in datalogs?

Posted: Thu May 12, 2016 7:58 am
by SlowTeg
I saw there's a Beta release that includes the throttle changes for the S. Just wanted to give folks a heads up and say thanks to Hondata!

Re: What exactly does tplate % represent in datalogs?

Posted: Thu May 12, 2016 9:20 am
by R-12914
The new feature is very faulty like this:
http://www.hondata.com/forum/viewtopic. ... 135#p87135