[FW MOD] ms2extra 3.2.1 beta gslender v2

This is a forum for discussing the development and testing of alpha MS2/Extra code. Documentation
(Runs on MS2 and Microsquirt)

Moderators: jsmcortina, muythaibxr

tpsretard2
Master MS/Extra'er
Posts: 662
Joined: Thu Feb 14, 2008 4:59 am

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by tpsretard2 »

i really like the idea of the dynamic dashpot
jsmcortina
Site Admin
Posts: 39569
Joined: Mon May 03, 2004 1:34 am
Location: Birmingham, UK
Contact:

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by jsmcortina »

Have you noticed any TPSdot etc. difference with the latest MS3 firmwares?

James
I can repair or upgrade Megasquirts in UK. http://www.jamesmurrayengineering.co.uk

My Success story: http://www.msextra.com/forums/viewtopic ... 04&t=34277
MSEXTRA documentation at: http://www.msextra.com/doc/index.html
New users, please read the "Forum Help Page".
gslender
Super MS/Extra'er
Posts: 1072
Joined: Fri Sep 16, 2011 5:29 am
Location: Brisbane, Australia
Contact:

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by gslender »

jsmcortina wrote:Have you noticed any TPSdot etc. difference with the latest MS3 firmwares?

James
James, are you able to give any kind of timeline expectation on when the ms3 xxxDot improvements will be passed down to ms2?

I'd much prefer to have your mods than mine, but with zero knowledge of how long/when etc, it kinda makes it hard to just sit in silence and wait (especially when Rob's worked, I've got a mod that does, and both improve the current situation)... so anything you can share would be great.

g
Mazda MX5 + MS3 Pro
jsmcortina
Site Admin
Posts: 39569
Joined: Mon May 03, 2004 1:34 am
Location: Birmingham, UK
Contact:

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by jsmcortina »

I was hoping to hear user feedback to see if they are useful.

James
I can repair or upgrade Megasquirts in UK. http://www.jamesmurrayengineering.co.uk

My Success story: http://www.msextra.com/forums/viewtopic ... 04&t=34277
MSEXTRA documentation at: http://www.msextra.com/doc/index.html
New users, please read the "Forum Help Page".
gslender
Super MS/Extra'er
Posts: 1072
Joined: Fri Sep 16, 2011 5:29 am
Location: Brisbane, Australia
Contact:

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by gslender »

jsmcortina wrote:I was hoping to hear user feedback to see if they are useful.

James
So no ms3 users willing to test, but we've got ms2 asking for it, but we're waiting on the ms3 users? :roll:

As mentioned before, happy to look at the code, do my own port and test - even within a restricted no public posting way if that's what ya concerned about.

G
Mazda MX5 + MS3 Pro
Greg G
Experienced MS/Extra'er
Posts: 311
Joined: Thu Mar 10, 2011 11:33 pm

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by Greg G »

You'd get user feedback in less than a day, here :)
1996 Mazda MX-5 1.6L NA6/ Mazdaspeed M45 SC/ BSP AW Intercooler/ Maruha F-cams/ 425cc RX-8 injectors/ DIYPNP

MS2/Extra test mule :)
tpsretard2
Master MS/Extra'er
Posts: 662
Joined: Thu Feb 14, 2008 4:59 am

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by tpsretard2 »

ok i have been driving for a bit today and all yesterday.

the tpsdot and mapdot in ,s3 are better than what is in the ms2, remarkably better. however the filtering option that was in the v2 of gslender code was still MUCH better than that of the ms3.

The new carb style AE is times better, but still not right.
i think i am going to have to find some time soon and write up on how some of the mega buck ecu's do their ae and see if you are willing to come up with a compromise between the 2 of them.
jsmcortina
Site Admin
Posts: 39569
Joined: Mon May 03, 2004 1:34 am
Location: Birmingham, UK
Contact:

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by jsmcortina »

tpsretard2 wrote:i think i am going to have to find some time soon and write up on how some of the mega buck ecu's do their ae and see if you are willing to come up with a compromise between the 2 of them.
Go for it. I'm interested to see what they do and why you feel what they does works better.

James
I can repair or upgrade Megasquirts in UK. http://www.jamesmurrayengineering.co.uk

My Success story: http://www.msextra.com/forums/viewtopic ... 04&t=34277
MSEXTRA documentation at: http://www.msextra.com/doc/index.html
New users, please read the "Forum Help Page".
muythaibxr
Site Admin
Posts: 8228
Joined: Thu Oct 14, 2004 12:48 pm

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by muythaibxr »

tpsretard2 wrote: The new carb style AE is times better, but still not right.
i think i am going to have to find some time soon and write up on how some of the mega buck ecu's do their ae and see if you are willing to come up with a compromise between the 2 of them.
How is it not right? What is not right about it? I don't particularly care how the other guys do it, I want to know how what we already wrote isn't working for you.

Ken
Megasquirt is not for use on pollution controlled vehicles. Any advice I give is for off road use only.
gslender
Super MS/Extra'er
Posts: 1072
Joined: Fri Sep 16, 2011 5:29 am
Location: Brisbane, Australia
Contact:

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by gslender »

tpsretard2 wrote: the tpsdot and mapdot in ,s3 are better than what is in the ms2, remarkably better. however the filtering option that was in the v2 of gslender code was still MUCH better than that of the ms3.
I'd be curious to know what specifically we are targeting to correct here... :? as having a clear understanding of the problem ensure a correct solutions is being developed. James/Ken - what do you think you're fixing?

For me, I have tpsADC that spikes wildly around a base value, shown by this example range when TPS is at 0%

109,98,113,107,113,106,107,108,112,111,109,104,110,106,92,114,102,109,108,98,111,111,105

No amount of rewiring and gnding of the TPS sensor fixes it - so I have to develop a filter for it so that it normalises the outliers.

I believe I've done that and have a working median-sliding-window filter that corrects this (this is different from the previous sliding window implementation that I posted in earlier code). So far it does a great job of removing the outliers (the < 100 and > 110 from the above sequence), and yet still allows for almost immediate response to tps ADC changes.

James, I'm still testing, but would really like to also compare the MS3 implementation mentioned above - I hope I can appeal to you in letting me view this source code and test.

PS - might be a good time to break off the last few posts into another thread on the forum. This MS3 test/feedback doesn't seem the right place on the back end of a MS2 MOD released by me??

G
Mazda MX5 + MS3 Pro
Red_Baron
Experienced MS/Extra'er
Posts: 328
Joined: Wed Mar 05, 2008 2:05 am

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by Red_Baron »

I'm running this code, (60-2, 4-cyl french crap)

Well, the one with the TPS smoothing code anyway

I'll test almost immediately for you too...

I love idle, I throw it into neutral on down hills now just to watch it idle spot on, I don't even care if it uses more fuel, it just makes me smile :mrgreen:
gslender
Super MS/Extra'er
Posts: 1072
Joined: Fri Sep 16, 2011 5:29 am
Location: Brisbane, Australia
Contact:

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by gslender »

Red_Baron wrote:I'm running this code
Which code? MS3 alpha17 or [FW MOD] ms2extra 3.2.1 beta gslender v2 ?
Mazda MX5 + MS3 Pro
robs
Master MS/Extra'er
Posts: 564
Joined: Sun Jan 17, 2010 4:26 pm
Location: Sydney, Australia

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by robs »

gslender wrote: I'd be curious to know what specifically we are targeting to correct here... :? as having a clear understanding of the problem ensure a correct solutions is being developed. James/Ken - what do you think you're fixing?

For me, I have tpsADC that spikes wildly around a base value, shown by this example range when TPS is at 0%
[snip]
I agree that a clear statement of what is to be addressed would be a good thing. IMO, that the TPS value itself jitters around isn't much of a worry since few things care about the actual value of TPS. That's why I put my focus on supporting a jitter-free but responsive tpsDOT, but perhaps there is a good reason to make a bigger change.

Have fun,

Rob.
gslender
Super MS/Extra'er
Posts: 1072
Joined: Fri Sep 16, 2011 5:29 am
Location: Brisbane, Australia
Contact:

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by gslender »

robs wrote: I agree that a clear statement of what is to be addressed would be a good thing. IMO, that the TPS value itself jitters around isn't much of a worry since few things care about the actual value of TPS. That's why I put my focus on supporting a jitter-free but responsive tpsDOT, but perhaps there is a good reason to make a bigger change.
Spot on! 8)

I've focussed on getting rid of the noise in the ADC, but as you point out, the only issue is tpsDOT reacting incorrectly to very small changes.... But the root of the issue is the Incomming signal and how it's converted to digital values.

G
Mazda MX5 + MS3 Pro
robs
Master MS/Extra'er
Posts: 564
Joined: Sun Jan 17, 2010 4:26 pm
Location: Sydney, Australia

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by robs »

gslender wrote:But the root of the issue is the Incomming signal and how it's converted to digital values.
I think the root issue is that tpsDOT itself is a woolly term. Mathematically, it would be the instantaneous rate of change, but in the discrete real world, it must be something measureable, so it must be an average rate of opening over some period of time. But what period is important, and what change is significant? Hence my approach -- just driven by the maths. Other concerns come to mind when I try thinking about the physics, but I'm on much shakier ground there.

On a less waffley subject, I think the approach of a median filter is a great way to clean up a signal, but not so helpful when it comes to detecting changes. Here are a few samples and their median(5) filtered values (assume 0 for all previous):

Code: Select all

sample 0 0 0 10 10 10 10
median 0 0 0  0  0 10 10
so it takes three samples to spot quite a significant change. Is that a worry? I don't know, but it is unquestionably less responsive than applying "the moment we get a change more than 3, it's changing". Trying to detect trends in other filtered data will also cost a delay.

Have fun,

Rob.
Red_Baron
Experienced MS/Extra'er
Posts: 328
Joined: Wed Mar 05, 2008 2:05 am

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by Red_Baron »

gslender wrote:
Red_Baron wrote:I'm running this code
Which code? MS3 alpha17 or [FW MOD] ms2extra 3.2.1 beta gslender v2 ?
I'm running the "ms2extra 3.2.1 beta gslender v2"

but I can run anything. I have a PnP loom adapter and I drive about with the stock ECU in the car and a spanner so it'll never hurt, just throw some code out and I can do whatever with it. Normally do about 40miles a day
pmbrunelle
Helpful MS/Extra'er
Posts: 103
Joined: Mon Jun 06, 2011 1:21 pm
Location: Montréal, Canada

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by pmbrunelle »

jsmcortina wrote:
tpsretard2 wrote:i think i am going to have to find some time soon and write up on how some of the mega buck ecu's do their ae and see if you are willing to come up with a compromise between the 2 of them.
Go for it. I'm interested to see what they do and why you feel what they does works better.

James
I run 2 squirts/alternating and hot idle is 800 RPM on my 6 cylinder. With a quick throttle opening at idle, there's a slight stumble before the engine will rev. Not enough to interfere with the driving experience, but it's noticeable.

I know there's the option to switch to 6 squirts/simultaneous during AE, but it's not the same as a true asynchronous AE.

Going to multiple short pulses magnifies injector dead time inaccuracies when some megabuck ECUs can deliver the asynchronous AE shot in a single longer pulse, minimizing dead time inaccuracy.

Just noted that you're talking about ms3 now (I think)...
gslender
Super MS/Extra'er
Posts: 1072
Joined: Fri Sep 16, 2011 5:29 am
Location: Brisbane, Australia
Contact:

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by gslender »

This firmware has been upgraded by this 2.2b release of my mod >> see http://www.msextra.com/forums/viewtopic ... 45#p307629
Mazda MX5 + MS3 Pro
muythaibxr
Site Admin
Posts: 8228
Joined: Thu Oct 14, 2004 12:48 pm

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by muythaibxr »

Those ECUs are also using sequential, so you can just tell the next event to be longer and for that cylinder it'll be correct. With batch injection you can't do that, so the only way to ensure that the next cylinder gets at least some of the increase in fuel, you have to squirt more often.

I don't think this can be fixed without using sequential injection.

Ken
Megasquirt is not for use on pollution controlled vehicles. Any advice I give is for off road use only.
pmbrunelle
Helpful MS/Extra'er
Posts: 103
Joined: Mon Jun 06, 2011 1:21 pm
Location: Montréal, Canada

Re: [FW MOD] ms2extra 3.2.1 beta gslender v2

Post by pmbrunelle »

muythaibxr wrote:Those ECUs are also using sequential, so you can just tell the next event to be longer and for that cylinder it'll be correct. With batch injection you can't do that, so the only way to ensure that the next cylinder gets at least some of the increase in fuel, you have to squirt more often.

I don't think this can be fixed without using sequential injection.

Ken
With batch injection, it is impossible to fuel all cylinders uniformly during a rapid load change.

I've still seen asynchronous AE on batch injection though - so rather than doing nothing and underfueling all cylinders which have had their squirt already and are "committed", the injector bank is pulsed again. The impossibility of uniform fueling still applies - but this time, some cylinders are underfueled, some have about the right amount, and some are overfueled. Somewhat closer to the ideal.

Preferable to 2 or more synchronous squirts per cycle per cylinder which will accentuate injector non-linearity, and ultimately limit maximum safe injector duty cycle at maximum power.

Some thoughts (with suggested numbers pulled out of thin air) if anyone wants to try:

Enable async shot for rpm < maximum (~2500 RPM)

Immediately upon tpsdot > async threshold (~250%/sec), begin a single async pulse of some minimum duration (~3 ms)

During the async pulse, latch onto the highest tpsdot value, and extend the async pulse if necessary according to an async PW (PW measured from the initial start of pulse) versus tpsdot table

Only re-enable async shot when tpsdot goes to zero again.

Only a single async shot per throttle stab is required, because after the first async shot, normal AE added to synchronous pulses will handle the accel fueling.

Outputs from the regular synchronous pulses ORred with the outputs from the asynchronous AE shot.
Post Reply