If anyone hasn't donated something to Steveo, now's the time. There's an ebay auction for some repair parts his car could really use.. ending in the next 18 hours. If you can throw $50, 20 or even $10 his way, do it soon.
If anyone hasn't donated something to Steveo, now's the time. There's an ebay auction for some repair parts his car could really use.. ending in the next 18 hours. If you can throw $50, 20 or even $10 his way, do it soon.
Definitely deserving!
1994 LT1/4L60E Formula
can you confirm that this is still the case with latest beta?TCC indicator is backwards when selected manually, but correct when on auto.
so if you click: On the converter is locked, but the indicator is OFF?
and if you click OFF converter the is unlocked but the indicator is ON?
or is it something else
im flying blind with this one
That's exactly what I saw yesterday.
I will try to confirm asap
1994 LT1/4L60E Formula
appreciate it buddy, thanks. im hoping 94 and 95 don't react differently, can someone with a '95 trans/ecm confirm too?
on a side note, been desperately trying to get this working on nix. I've posted on qt forums and posted on qtbug for qtserialport.
here's some of the conversation:
So even though it hangs with that debug output, it's not an error.I'm setting to 8192, I get the error:
"Baud rate of serial port /dev/ttyUSB0 is set to 8193 instead of 8192 :divisor 2929.687500 unsupported"
Yet it works in windows, and in emulation under linux, so the hardware does work at this baud rate.
If I set it to 8193 baud i get this error, which is strange:
Baud rate of serial port /dev/ttyUSB0 is set to 8193 instead of 8193 :divisor 2929.329834 unsupported
Baud rate is set to 8193 instead of 8193
I'd really like to get this working
thank you
Denis Shienkov added a comment - 7 hours ago - edited
@Brian Labbie ,
in your case it is not an error! it was already explained to you at a forum
in this case your device got 8193 baud instead of 8192.
UPD: You can also try 5.6.0 branch.
Brian Labbie added a comment - 1 hour ago
Not sure how it was explained to me...
I got no real answer on this.
"in this case your device got 8193 baud instead of 8192. "
then why if I set to 8193 do i get the error:
Baud rate of serial port /dev/ttyUSB0 is set to 8193 instead of 8193 :divisor 2929.329834 unsupported
it makes no sense!
Denis Shienkov added a comment - 1 hour ago
> Baud rate of serial port /dev/ttyUSB0 is set to 8193 instead of 8193 :divisor 2929.329834 unsupported
it is just incorrect output of the debug information, where the fractional part of real baud rate is lost.. should be like:Baud rate of serial port /dev/ttyUSB0 is set to 8193.blablabla instead of 8193 :divisor 2929.329834 unsupportedInstead of:
if (serial.custom_divisor * baudRate != serial.baud_base) {
qWarning("Baud rate of serial port %s is set to %d instead of %d: divisor %f unsupported",
qPrintable(systemLocation),
serial.baud_base / serial.custom_divisor,
baudRate, (float)serial.baud_base / baudRate);
}
should be:
if (serial.custom_divisor * baudRate != serial.baud_base) {
qWarning("Baud rate of serial port %s is set to %f instead of %d: divisor %f unsupported",
qPrintable(systemLocation),
(float)serial.baud_base / serial.custom_divisor,
baudRate, (float)serial.baud_base / baudRate);
}
and I repeat again: this debug output does not matter to your issue, because the method does not returns an errors!
I'll just keep rebooting into win7 to log and flash, unless you have some idea.
EDIT:
the latest:
"Maybe the linux drivers of your device do not support using of custom divizors... try 5.6.0 branch, where are used the new 'termios V2' feature instead of old 'serial_struct', maybe it will work.."
So I'll probably muck around with it tonight.
would be so awesome to have a native linux app for this
Last edited by fbody_Brian; 12-30-2015 at 07:05 PM.
1994 LT1/4L60E Formula
we can do it. do you use debian/ubuntu or something else? i need a distribution to make an official port/package for to start with.
qt is a real clusterfluck sometimes. it's the most gigantic library ever. sometimes it's like using a 10 pound sledgehammer to push a tack into a wall.
anyways qt's serial driver probably isn't the issue, it's well tested on linux.
what kernel are you running, what serial driver and chipset do you use?
the ultimate solution in linux or freebsd (which i used in my old aldl-io program) was to completely disable the kernel's serial driver and use libftdi to just send/recieve raw usb commands directly to the chip from userland. a bit more overhead, but that way you don't have a universal serial driver that needs to interface with ttys and junk telling you that a baud rate is no good, you can command the chip to do whatever it's capable of doing.. i found it rediculously stable.
unfortunately qserialport does not support libftdi, but if i had to write a version that DID use libftdi, it would be a no brainer, i could just port the serial layer from aldl-io, as the api is fairly similar.
edit: also what version of QT are you building against?
Last edited by steveo; 12-30-2015 at 07:48 PM.
i read the termios patch for QT and it seems like it might solve the problem. 5.6 is due for release soon too so we could make it a requirement for linux support..
qt5.5.x at work right now so I can't be super precise, but it's definitely 5.5
I've written a program using QT before, it was gigantic at qt3, I can only imagine it's just gotten bigger.
The thing is, I can connect using winflash, datamaster, tunercats, all using wine in linux, so I know that it can work in linux. Not sure how wine talks to the serial device. I'm using the ftdi arduino setup. the $10 setup.
That's what I'm hoping for
1994 LT1/4L60E Formula
i think there's only 5.5.0 and 5.5.1qt5.5.x at work right now so I can't be super precise, but it's definitely 5.5
i read all the 5.6 changelogs and eehack should build against it with no problems. qt is pretty good with backwards compatibility.
me either but i betcha it's either using termios (which has been around for ages) or it just knows how to deal with bitchy divisor complaints in a better way than me and qserialport do.
good, same here.I'm using the ftdi arduino setup. the $10 setup.
it does depend how you build it. 5.4 bundled with qt creator was built with internationalization and webkit junk and required like 20mb+ of libraries for your average program.I've written a program using QT before, it was gigantic at qt3, I can only imagine it's just gotten bigger.
getting qt to recompile on windows can be a nightmare, so i just left it alone, also it's GPL so even changing some #defines means technically i'd have to host the entire source for it.. at least if it's unmodified, i can just link people to the qt windows source if i get called on it.
5.5, though, seems slimmer in it's default windows binary form, it doesn't seem to require linking against internationalization shit, which is just awesome. that's how i managed to get eehack+libraries+installer down to around 6mb (uncompresses to like 12mb, no big deal)
i should mention every single other developer i talked to before writing my own aldl software (and that includes quite a few) told me that setting the serial baud rate to 8192 in a way that works on most configurations can be the hardest part of the entire development. im glad qt did all the nasty work for me on windows.
Latest beta:
Driving with torque converter set to auto:
torque converter works as expected, indicator works as expected.
set torque converter to off:
torque converter stays unlocked, indicator light stays off until the normal lockup mph is reached then the light turns on, even though torque converter is unlocked.
drop below the normal speed and indicator goes off, above the speed and it comes on. Light has the same behavior as if it's set to auto.
set torque converter to on:
torque converter is locked, works in any gear, light never indicates a locked condition
1994 LT1/4L60E Formula
so the indicator light doesn't function at all with mode4 tcc enable. i can live with that.
so what i'll probably do in this release is just force the light to be on/off to not confuse people, then in the next release i'll try to patch it so the mode4 override also sets/unsets the status bit.
Bookmarks