MS3 pre-1.1 alpha 18 released

Testing and development of Megasquirt 3

Moderators: jsmcortina, muythaibxr

y8s
Master MS/Extra'er
Posts: 478
Joined: Wed Jan 07, 2009 3:21 pm

Re: MS3 pre-1.1 alpha 18 released

Post by y8s »

Noticing a decimal place display issue that original surfaced in alpha 17. Under the table view for the new AE system, the graph increment can be raised and lowered by up/down arrow in 0.1% increments. The table only shows whole numbers.
muythaibxr
Site Admin
Posts: 8228
Joined: Thu Oct 14, 2004 12:48 pm

Re: MS3 pre-1.1 alpha 18 released

Post by muythaibxr »

OK, that's easy enough to fix. A 1 character change in the ini most likely.

Ken
Megasquirt is not for use on pollution controlled vehicles. Any advice I give is for off road use only.
hoveringuy
Experienced MS/Extra'er
Posts: 172
Joined: Fri Aug 29, 2008 8:52 am

Re: MS3 pre-1.1 alpha 18 released

Post by hoveringuy »

James,

Love the new loader. No issues.

Motor spit and coughed and ran like crap after I flashed 18. Fixed it by turning off exhaust VANOS. Did anything change there?

Thanks for updating Traction Control! I'm confused by the term Mininum MAP(%). Is that the same as my Fuel Load(%)? I think of MAP in terms of KPa.

Steve
ragepower
Experienced MS/Extra'er
Posts: 280
Joined: Fri Dec 17, 2010 9:13 am
Location: Leiria - Portugal

Re: MS3 pre-1.1 alpha 18 released

Post by ragepower »

This morning, i flashed this alpha code, and it is working fine.

I had a issue with the new loader, i didn't work, but i think to know why. I think that my latency values on serial controller aren't the best, and this new loader, shows the message, "barley reply", or something like that, i used the old loader and it worked, but this process token about 3-4 min.

Im using the CL idle activated by VSS, so i really like to try the CL ilde initial value, But i had some question about the table.

x-axis is the target idle rpm
y-axis is the MAT or CLT
table values are the duty of the idle valve

My question is why is a table? Something like "PWM idle cranking duty" style wasn't enough? Because the target idle rpm is dependent of CLT, that relation is defined in the "CL Idle settings" and if you are using the CLT is the y-axis, so you don't need a table, a vector is enough and you can save memory. Or there are some functionality or situation that im not seeing?
muythaibxr
Site Admin
Posts: 8228
Joined: Thu Oct 14, 2004 12:48 pm

Re: MS3 pre-1.1 alpha 18 released

Post by muythaibxr »

You should use MAT. As MAT changes so does the duty required to give you a particular RPM, so a table is required to fully cover all conditions.

Ken
Megasquirt is not for use on pollution controlled vehicles. Any advice I give is for off road use only.
wrenchdad
Master MS/Extra'er
Posts: 614
Joined: Mon Feb 14, 2005 5:30 am
Location: Natchez, Mississippi

Re: MS3 pre-1.1 alpha 18 released

Post by wrenchdad »

Guys,

I have a question on the changes on AC control, I see that gslender work has been front-ported to MS3. I was reading his pdf file on the features as in below..

AC disengagement
To further improve the shut-off of the AC and avoid an idle surge (as load is suddenly removed and the
PID needs to quickly react and remove duty) I’ve added a backwards counter to leave load remaining for
a small period to allow the AC compressor load to gracefully bring down the idle RPM before the AC
compressor is shut down.
This is configured as follows - when the AC is turned off by switch or by thermostat, the added AC idle
duty adder is removed and idleupcnt is instantly put to whatever value is set for AC off idleupcnt
countdown. It then counts backwards to zero by the amount set in AC off idleupcnt rate (so if you set
rate to 1 it will goes backwards at the same rate as it counted up, if you set it to 5 it will count
backwards 5x faster).


SO I am wondering if MS3 has this adjustable feature for turning off the AC clutch? If I understand gslender's write up, his changes allow for holding the AC clutch engagement "after" the ICV has reduced to non-AC load setting for either the same amount of delay that happens at turnon OR up to 5 times faster.
Just asking if we have this feature, where is the setting for it? I looked but couldn't see it anywhere.

The reason I ask is because I am seeing a large drop in idle rpms after AC clutch disengagement and at the lowest point in the drop before the rpms start to turn around is at the same amount of delay time I am set to at 700ms? Just looks like the delay is working on both ends but on the turnoff end is pulling the idle rpms down? Bug maybe? I am attaching a screen shot of AC on and then off, you can see the icv steps in the bottom chart going up and the delay getting the rpms up before AC turnon then you can see the huge drop in rpms after the icv drops back down. Just doesn't seem right for revs to drop off that much with the load coming off?

thanks for all the hard work and long hours you guys put in on the code writing !!!!
wd
ps I can supply a log if needed but the one from this screen shot is like 25meg long !

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

Re: MS3 pre-1.1 alpha 18 released

Post by gslender »

wrenchdad wrote: I see that gslender work has been front-ported to MS3. I was reading his pdf file on the features as in below..

AC disengagement
To further improve the shut-off of the AC and avoid an idle surge (as load is suddenly removed and the
PID needs to quickly react and remove duty) I’ve added a backwards counter to leave load remaining for
a small period to allow the AC compressor load to gracefully bring down the idle RPM before the AC
compressor is shut down.
This is configured as follows - when the AC is turned off by switch or by thermostat, the added AC idle
duty adder is removed and idleupcnt is instantly put to whatever value is set for AC off idleupcnt
countdown. It then counts backwards to zero by the amount set in AC off idleupcnt rate (so if you set
rate to 1 it will goes backwards at the same rate as it counted up, if you set it to 5 it will count
backwards 5x faster).


SO I am wondering if MS3 has this adjustable feature for turning off the AC clutch? If I understand gslender's write up, his changes allow for holding the AC clutch engagement "after" the ICV has reduced to non-AC load setting for either the same amount of delay that happens at turnon OR up to 5 times faster.
Just asking if we have this feature, where is the setting for it? I looked but couldn't see it anywhere.
This is where the implementation between MS3 and MS2 differs.

As MS3 has more space and more inputs, they didn't need to rely on an idleupcnt output channel that is tied to the Output Conditions like it is in MS2. The MS2 implementation was done that way as it avoided lengthy code and reused existing functionality.

In my testing I saw the same problems as you when switching off, so I made sure we could count down the idleupcnt and start from different positions (creating an off slope that was different from the on slope). As you point out, without this you get an sudden surge in RPM as the AC clutch (load) is disengaged (with the duty removed) leaves the engine with more air than it needs for the given load - and so the small rpm surge is seen. It is a minor issue, but in getting it perfect you need to have that duty off, then wait a few tenths of a second before cutting the AC clutch (load).

I'm sure Ken will probably implement something similar for MS3 as it wouldn't be too hard to do so.

G
Mazda MX5 + MS3 Pro
wrenchdad
Master MS/Extra'er
Posts: 614
Joined: Mon Feb 14, 2005 5:30 am
Location: Natchez, Mississippi

Re: MS3 pre-1.1 alpha 18 released

Post by wrenchdad »

Yes, I understand what you are saying but my issue is that I am not getting a surge of MORE rpms but a very large drop in rpms like 400-500 drop. That was why I was asking if there was a setting to adjust the off timing. It may just be my tune that is giving problem but I think I have a pretty good setup on the AC turning on, just odd that I get such a big drop in idle when the load comes off. Trying some spark timing adjustments plus the rpm delta table to see if that will help BUT if you look at the screen shot you can see where spark timing took a big jump but rpms still dropped.

thanks for the reply and your support work too !!
wd
ragepower
Experienced MS/Extra'er
Posts: 280
Joined: Fri Dec 17, 2010 9:13 am
Location: Leiria - Portugal

Re: MS3 pre-1.1 alpha 18 released

Post by ragepower »

muythaibxr wrote:You should use MAT. As MAT changes so does the duty required to give you a particular RPM, so a table is required to fully cover all conditions.

Ken
Ok, i will use MAT. Just one more question, the higher MAT, more duty is required, right?
gslender
Super MS/Extra'er
Posts: 1072
Joined: Fri Sep 16, 2011 5:29 am
Location: Brisbane, Australia
Contact:

Re: MS3 pre-1.1 alpha 18 released

Post by gslender »

ragepower wrote:Ok, i will use MAT. Just one more question, the higher MAT, more duty is required, right?
Yup - as the air is thinner/less dense and more volume is needed to compensate.
Mazda MX5 + MS3 Pro
gslender
Super MS/Extra'er
Posts: 1072
Joined: Fri Sep 16, 2011 5:29 am
Location: Brisbane, Australia
Contact:

Re: MS3 pre-1.1 alpha 18 released

Post by gslender »

wrenchdad wrote:Yes, I understand what you are saying but my issue is that I am not getting a surge of MORE rpms but a very large drop in rpms like 400-500 drop. That was why I was asking if there was a setting to adjust the off timing. It may just be my tune that is giving problem but I think I have a pretty good setup on the AC turning on, just odd that I get such a big drop in idle when the load comes off. Trying some spark timing adjustments plus the rpm delta table to see if that will help BUT if you look at the screen shot you can see where spark timing took a big jump but rpms still dropped.

thanks for the reply and your support work too !!
wd
Mmmm, do you have any target rpm being added in your tune? You may need it as PID will be thinking the rpm target needs to be idle (say 900) and with AC on the duty is forced higher (above target) and pid is fighting it trying to always drag it down. So when you take ac duty away, PID sets the IAC to much lower than it should be and therefore you get a little dip before it corrects and raises again.

Just a thought?
Mazda MX5 + MS3 Pro
muythaibxr
Site Admin
Posts: 8228
Joined: Thu Oct 14, 2004 12:48 pm

Re: MS3 pre-1.1 alpha 18 released

Post by muythaibxr »

wrenchdad wrote:Guys,

I have a question on the changes on AC control, I see that gslender work has been front-ported to MS3. I was reading his pdf file on the features as in below..
Wrong. For the most part the features started in MS3 where I implemented them and gave them to gslender to backport.
SO I am wondering if MS3 has this adjustable feature for turning off the AC clutch? If I understand gslender's write up, his changes allow for holding the AC clutch engagement "after" the ICV has reduced to non-AC load setting for either the same amount of delay that happens at turnon OR up to 5 times faster.
Just asking if we have this feature, where is the setting for it? I looked but couldn't see it anywhere.

The reason I ask is because I am seeing a large drop in idle rpms after AC clutch disengagement and at the lowest point in the drop before the rpms start to turn around is at the same amount of delay time I am set to at 700ms? Just looks like the delay is working on both ends but on the turnoff end is pulling the idle rpms down? Bug maybe? I am attaching a screen shot of AC on and then off, you can see the icv steps in the bottom chart going up and the delay getting the rpms up before AC turnon then you can see the huge drop in rpms after the icv drops back down. Just doesn't seem right for revs to drop off that much with the load coming off?

thanks for all the hard work and long hours you guys put in on the code writing !!!!
wd
ps I can supply a log if needed but the one from this screen shot is like 25meg long !

.
gslender implemented the delay after turning off the AC on top of what I did on MS3. I have not done that mainly because nobody had asked making me think nobody was having the same problem gslender was trying to solve there.

Ken
Megasquirt is not for use on pollution controlled vehicles. Any advice I give is for off road use only.
muythaibxr
Site Admin
Posts: 8228
Joined: Thu Oct 14, 2004 12:48 pm

Re: MS3 pre-1.1 alpha 18 released

Post by muythaibxr »

So looking at the graph, RPM is dropping after you push the AC button to turn off the AC. Gslender's off timer is for those who have the opposite problem where the sudden decrease in load causes RPM to spike briefly. None of my MS3 testers have reported the problem gslender was trying to solve, so I never implemented anything like it. It also would not solve what you are seeing.

If you are using the increased target feature of AC idleup, that could cause the P term to respond to falling target by decreasing the valve duty more than what happens just from pressing the button resulting in a dip in RPM.

Ken
Megasquirt is not for use on pollution controlled vehicles. Any advice I give is for off road use only.
y8s
Master MS/Extra'er
Posts: 478
Joined: Wed Jan 07, 2009 3:21 pm

Re: MS3 pre-1.1 alpha 18 released

Post by y8s »

I can vouch for the A/C Idle Up working fairly well--at least on a Miata.

The one issue I might see with the code is that for a while I was using a fairly high duty cycle adder to get the RPM to jump up and when the A/C cut off, the duty cycle was removed and brought the short term duty for idle below where it was before the A/C kicked on and it had to slowly recover.

I've since tuned that mostly out and now I get maybe a fluctuation of +/-50 rpm in idle speed after the duty drops back to "normal idle" before settling.



Image
wrenchdad
Master MS/Extra'er
Posts: 614
Joined: Mon Feb 14, 2005 5:30 am
Location: Natchez, Mississippi

Re: MS3 pre-1.1 alpha 18 released

Post by wrenchdad »

muythaibxr wrote:So looking at the graph, RPM is dropping after you push the AC button to turn off the AC. Gslender's off timer is for those who have the opposite problem where the sudden decrease in load causes RPM to spike briefly. None of my MS3 testers have reported the problem gslender was trying to solve, so I never implemented anything like it. It also would not solve what you are seeing.

If you are using the increased target feature of AC idleup, that could cause the P term to respond to falling target by decreasing the valve duty more than what happens just from pressing the button resulting in a dip in RPM.

Ken
Ken,

I am not cycling the AC with the dash button but allow it to do normal cycling by cabin temps, shouldn't matter I would guess. Yes I am using the increase target feature (add 50rpms) under AC idle controls plus under fan controls adding another 25 rpms. BUT I am not see icv duty fall off but instead go up trying to return rpms to target. I think some of my issue is coming from too rich of an idle AFR, so have cut down on fuel and I am seeing better response. Seeing some progress have about cut my rpm cutoff dip in half.

Couple of questions, under idle advance, are the numbers that are in the curve added to the spark map OR do they become the new map values so to speak? Also what about the correction curve? Does it add to the idle advance?

Last questions what does this new feature do under AC idleup----- Idleup CL Max decel load adder (%) ?

thanks wd
muythaibxr
Site Admin
Posts: 8228
Joined: Thu Oct 14, 2004 12:48 pm

Re: MS3 pre-1.1 alpha 18 released

Post by muythaibxr »

WD: In your posted log screenshot, I can clearly see duty (and RPM) drop significantly followed by a rise in duty and a corresponding rise in RPM since rpm dropped below target.

The idle advance feature replaced the table advance. The idle timing RPM assist feature works on top of whatever is setting the main timing (idle advance or table).

The max load adder is for the MAP lockout since turning on AC increases load.

Ken
Megasquirt is not for use on pollution controlled vehicles. Any advice I give is for off road use only.
wrenchdad
Master MS/Extra'er
Posts: 614
Joined: Mon Feb 14, 2005 5:30 am
Location: Natchez, Mississippi

Re: MS3 pre-1.1 alpha 18 released

Post by wrenchdad »

Thanks Ken, looks like I have some more log studying to do and some tweaking on my setup.

later wd

edit, Ok Ken, now I see, I was thinking about CL control and not the step down from AC idle control !! Yes, I can see what you are saying now about maybe the "P" term might get confused with the differences in target idle rpms. I will try out some different settings and see which combo gives me the best results.

thanks again, wd
krisr
Master MS/Extra'er
Posts: 799
Joined: Wed Aug 17, 2005 1:17 am
Location: Sydney, Australia

Re: MS3 pre-1.1 alpha 18 released

Post by krisr »

I put about 100mi on this code over the weekend and have not found 1 problem with my particular install. My SD card is temperamental though, as I reported in the beginning the card was giving me error code 5, then working, and i've noticed that on various occasions (sporadic) it'll give me error code 5 on a restart. I'll try a new card and see what that does and also play with the SPI speeds.
Sydney, Australia
1971 Holden Monaro HQ
MS3X Sequentially fuelled 400 Pontiac
UnaClocker
Super MS/Extra'er
Posts: 1933
Joined: Fri May 07, 2004 12:59 pm
Location: Tacoma, WA
Contact:

Re: MS3 pre-1.1 alpha 18 released

Post by UnaClocker »

Feedback on the new Mac firmware flasher, it's functional, but didn't save any of my MSQ values between alpha 17 and 18, despite me saying yes to the prompt (went with the defaults on all but serial port). A feature request/suggestion, when I fire up the flasher app, have it skim the /dev directory looking for tty.usb* and put them on the list, my usb always comes up as something a bit odd, something like tty.usbA013494223 or something like that..
I saw mention of alternator control. I'll be happy to try that out when it's ready. :)
Brian
'84 Dodge Rampage
muythaibxr
Site Admin
Posts: 8228
Joined: Thu Oct 14, 2004 12:48 pm

Re: MS3 pre-1.1 alpha 18 released

Post by muythaibxr »

That is odd, I have not built a MacOS version yet.

Ken
Megasquirt is not for use on pollution controlled vehicles. Any advice I give is for off road use only.
Post Reply