MS3 pre-1.4 alpha 1 released

[jsmcortina]

Another way to look for the passthrough stuff is to disable all polling and broadcasting on the MS3 and realtime data polling in the project. Make sure the CAN bus is quiet.
Then enable the IOx realtime data. At this point the only thing on the CANbus should be the passthrough requests and responses. Take a CAN dump of it.

I'll likely expand that piece of C to analyse the packets, I only wrote it today (and updated the MS3 CAN tech page.)

What CAN capture device are you using?

Thanks for those ini fixes, I have applied them. TS is too tolerant of parsing malformed ini files.

James

[MWPau]

Another way to look for the passthrough stuff is to disable all polling and broadcasting on the MS3 and realtime data polling in the project. Make sure the CAN bus is quiet.
Then enable the IOx realtime data. At this point the only thing on the CANbus should be the passthrough requests and responses. Take a CAN dump of it.

Ill give it a try.

What CAN capture device are you using?

This: http://www.crystalfontz.com/product/CFA921TS with added MCP2551's.
The iMX287 ARM SoC has two CAN interfaces built in which the Linux kernel has drivers for.

Thanks for those ini fixes, I have applied them. TS is too tolerant of parsing malformed ini files.

No problems.

[thokes82]

Can Broadcast works great for me. Just one question: I have got a problem with Fahrenheit units. My Race Technology Dash cannot convert them properly. It is always first multiplying, then adding. This cannot be changed. Is it possible to have this configurable to be sent as Celsius?

[jsmcortina]

Can't you work around it?

Typically we say C = (F - 32) * 5 / 9
But also that can be written as C = (F * 5 / 9) - 17.7

James

[jsmcortina]

Already on to 1.4! You guys waste no time.

Is this the stream meaning 1.4X where we might see launch control and nitrous not being mutually exclusive?

This can be looked at later. I'm trying to stick to the MS2 and MS3 roadmaps at the moment.

James

[thokes82]

Can't you work around it?

Typically we say C = (F - 32) * 5 / 9
But also that can be written as C = (F * 5 / 9) - 17.7

James

how could I not have thought about this?

[Peter Florance]

Had some trouble at dyno with this and VVT. This was not a car local to me, BMW M54, and I had to revert back to pre 1.3 alpha 3 that it has run on since last year.
I suspect I have missed some changes in VVT setup but could not zero out the cams. They just read the lower of the range numbers and did but not seem to move in test mode.
Would like to retest with local car; any suggested roadmap on which codes to try?
Thanks

[jsmcortina]

Had some trouble at dyno with this and VVT. This was not a car local to me, BMW M54, and I had to revert back to pre 1.3 alpha 3 that it has run on since last year.

Yes there were changes. Of course they weren't _supposed_ to cause a problem for existing users.
The change was around the way the angle were calculated because one engine came along that didn't work with the previous code.
Setting min and max to zero should still allow the test mode to work as it did before.

James

[Suprazz]

I ran on 1.4 alpha1 for over 1000km without any issues and I implemented boost control and it everything was working very well.

[muythaibxr]

I ran on 1.4 alpha1 for over 1000km without any issues and I implemented boost control and it everything was working very well.

Do you have logs and msq for the boost setup? I want to take a look to see how the new code is performing.

Ken

[nathaninwa]

I've got a full 1/4 mile run I could post. Tbh I'm not sure how far off the bias table is but each of my targets were nailed flatline within about 5 kpa

I can post tomorrow

I even upgraded intercoolers with tremendous drops in temps and didn't fool with boost control at all and it was still spot on

[Peter Florance]

Had some trouble at dyno with this and VVT. This was not a car local to me, BMW M54, and I had to revert back to pre 1.3 alpha 3 that it has run on since last year.

Yes there were changes. Of course they weren't _supposed_ to cause a problem for existing users.
The change was around the way the angle were calculated because one engine came along that didn't work with the previous code.
Setting min and max to zero should still allow the test mode to work as it did before.

James

I'll check. Thanks

One of the challenges is I have not worked out any easy way to bench test VVT

[jsmcortina]

One of the challenges is I have not worked out any easy way to bench test VVT

Agreed - it is tough.
One method I have used is with two Megasquirts from a single tach source. I use the spark output from one as the cam input to the other. I can then use "Fixed Advance" to move the spark signal which the second Megasquirt sees as the cam moving position.

James

[muythaibxr]

I've got a full 1/4 mile run I could post. Tbh I'm not sure how far off the bias table is but each of my targets were nailed flatline within about 5 kpa

I can post tomorrow

I even upgraded intercoolers with tremendous drops in temps and didn't fool with boost control at all and it was still spot on

Looking forward to seeing logs!

Ken

[nathaninwa]

Heres the logs Take note of the MAT. I upgraded from an Ebay CXRacing IC to an ETS core and went from 10psi pressure drop to just over 2psi at 28psi of boost. The old IC would start a run at 80ish degress and run out at 140 or so. Not any more

I was still dialing in fuel from the meth delete and new IC so fueling will be off just a bit, but its close. Anyways both logs are clean pulls the first was the 1/4 mile drag run with lower boost, then the other was a pull onto the highway later at night at a higher boost rate. My second better pass the track had me put the laptop int the trunk and the usb came unplugged so no data.

Overall the car runs amazing! thanks guys for all the hard work

[Peter Florance]

One of the challenges is I have not worked out any easy way to bench test VVT

Agreed - it is tough.
One method I have used is with two Megasquirts from a single tach source. I use the spark output from one as the cam input to the other. I can then use "Fixed Advance" to move the spark signal which the second Megasquirt sees as the cam moving position.

James

Good idea. I'll give that a shot

[thokes82]

Hi,
something is strange in here... This is in the ini. The number of options do not fit in water_freq

Code: Select all

      water_freq_on     = bits,     U08, 823,[0:0], "Off", "On"
      water_freq        = bits,     U08, 823,[1:4], "INVALID", "INVALID", "250Hz", "225Hz", "200Hz", "175Hz", "150Hz", "125Hz", "100Hz", "78Hz", "39Hz", "26Hz", "19.5Hz", "15.6Hz", "13Hz", "11.1Hz" "INVALID", "78Hz", "39Hz", "26Hz", "19.5Hz", "15.6Hz", "13Hz", "11.1Hz"
      water_freq_type   = bits,     U08, 823,[5:6], "Off", "Fast", "Slow", "INVALID"

Edit: Ok, found it...
water_freq = bits, U08, 823,[1:4], "INVALID", "INVALID", "250Hz", "225Hz", "200Hz", "175Hz", "150Hz", "125Hz", "100Hz", "78Hz", "39Hz", "26Hz", "19.5Hz", "15.6Hz", "13Hz", "11.1Hz" "INVALID", "78Hz", "39Hz", "26Hz", "19.5Hz", "15.6Hz", "13Hz", "11.1Hz"
The red part should be deleted.
I found a couple of other warnings, which I corrected.

[jsmcortina]

Those are ini errors that older versions of TunerStudio didn't notice.

They are fixed for the next alpha.

James

[thokes82]

I found a Problem with my pure Alpha N Setup.
I am taking Alpha N as primary load for fuel and ignition and taking the new setting mapport_t=Off. The Engineload does not change anymore when I am depressing the gas and engine is dying. If I enable a voltage or frequency in mapport_t, the engine works with alpha N as expected. (I do not have a map sensor connected).

[jsmcortina]