PCV with OSX Lion
Keith Wong
I'm having trouble downloading data from my PCV to GC, and I believe
the issue is that I need a 64-bit PL2303 driver. I'll spare the
details for now of why I think that's the problem (or check
http://groups.google.com/group/wattage/browse_thread/thread/a3298d327d21e9e1),
but does anyone know if such a driver exists?
Thanks!
Keith
Gareth
http://www.prolific.com.tw/eng/downloads.asp?id=31
On Feb 6, 9:38 pm, Keith Wong <keith.t.w...@gmail.com> wrote:
> Hi all,
>
> I'm having trouble downloading data from my PCV to GC, and I believe
> the issue is that I need a 64-bit PL2303 driver. I'll spare the
Rainer Clasen
> Have you tried the one on Prolific's download page:
>
> http://www.prolific.com.tw/eng/downloads.asp?id=31
I *think* the prolific drivers show up as /dev/cu.usbserial ... which
isn't accepted by GC as valid devicename. This can be verified with
ls -lrt /dev/cu.*
As suggested on the wattage list, Keith can circumvent this with a
symlink for testing:
ln -s cu.usbserial /dev/cu.PL2303-0
If the 64bit drivers work more reliable than the 32bit ones, we can add
their device name to the whitelist regexp in Serial.cpp.
> On Feb 6, 9:38�pm, Keith Wong <keith.t.w...@gmail.com> wrote:
> > Hi all,
> >
> > I'm having trouble downloading data from my PCV to GC, and I believe
> > the issue is that I need a 64-bit PL2303 driver. �I'll spare the
> > details for now of why I think that's the problem (or checkhttp://groups.google.com/group/wattage/browse_thread/thread/a3298d327...),
> > but does anyone know if such a driver exists?
Rainer
--
KeyID=759975BD fingerprint=887A 4BE3 6AB7 EE3C 4AE0 B0E1 0556 E25A 7599 75BD
Mark Liversedge
I *think* the prolific drivers show up as /dev/cu.usbserial ... which
isn't accepted by GC as valid devicename. This can be verified with
ls -lrt /dev/cu.*
Rainer Clasen
> On Tuesday, 7 February 2012 19:03:58 UTC, rainer wrote:
> >
> > I *think* the prolific drivers show up as /dev/cu.usbserial ... which
> > isn't accepted by GC as valid devicename. This can be verified with
> > ls -lrt /dev/cu.*
> >
> It does support /dev/usbserial-xxxx where xxxx is a hexadecimal number.
> Does the device really just appear as /dev/cu.usbserial with nothing else?
My notes say so for the 32bit version.
> I can easily change Serial.cpp to add this.
I'd wait until we receive positive feedback, as the 32bit versions were
completely unusable.
Keith Wong
On Feb 7, 11:47 am, Rainer Clasen <b...@zuto.de> wrote:
> Mark Liversedge wrote:
> > On Tuesday, 7 February 2012 19:03:58 UTC, rainer wrote:
>
> > > I *think* the prolific drivers show up as /dev/cu.usbserial ... which
> > > isn't accepted by GC as valid devicename. This can be verified with
> > > ls -lrt /dev/cu.*
>
> > It does support /dev/usbserial-xxxx where xxxx is a hexadecimal number.
> > Does the device really just appear as /dev/cu.usbserial with nothing else?
>
> My notes say so for the 32bit version.
>
latest update (copied from Wattage):
Using the latest OSX 32-/64-bit drivers from Prolific (v1.4.0 at
http://www.prolific.com.tw/eng/downloads.asp?id=31) and creating a
symlink according to Rainer's directions the port shows up in GC
(every version I have). For version 2.x I get the error message
"Couldn't open device \dev\cu.PL2303-0: Host is down", and in version
3 I get the error message "Failed to initialized Powercontrol
communication: no PCV found at any baud rate."
Thanks for the help!
Keith
PS: It sounds like there are more people familiar with such issues
here than on the Wattage list (or it's all the same people...), so
I'll probably just keep the discussion here rather than in both places
from now on.
Rainer Clasen
> On Feb 7, 11:47�am, Rainer Clasen <b...@zuto.de> wrote:
> > Mark Liversedge wrote:
> > > On Tuesday, 7 February 2012 19:03:58 UTC, rainer wrote:
> >
> > > > I *think* the prolific drivers show up as /dev/cu.usbserial ... which
> > > > isn't accepted by GC as valid devicename. This can be verified with
> > > > �ls -lrt /dev/cu.*
> >
> > > It does support /dev/usbserial-xxxx where xxxx is a hexadecimal number.
> > > Does the device really just appear as /dev/cu.usbserial with nothing else?
> >
> > My notes say so for the 32bit version.
> >
>
> It showed up as /dev/cu.usbserial with no number for me. Here's the
> latest update (copied from Wattage):
>
> Using the latest OSX 32-/64-bit drivers from Prolific (v1.4.0 at
> http://www.prolific.com.tw/eng/downloads.asp?id=31) and creating a
> symlink according to Rainer's directions the port shows up in GC
> (every version I have). For version 2.x I get the error message
> "Couldn't open device \dev\cu.PL2303-0: Host is down", and in version
> 3 I get the error message "Failed to initialized Powercontrol
> communication: no PCV found at any baud rate."
Hmm, sounds as if the 64bit version is behaving similar to the 32bit one.
Though... If you keep trying, you might have luck every now and then...
say 1 out of 50 attempts... I seemed to get better results when waking up
the PCV by hitting "mode" right before starting the download... Sorry,
just kidding. This drove me nuts debugging srmio until I tried another
driver.
It's actually the same on windows... the original prolific drivers fail to
provide reliable communication for *all* pl230x usb2serial adapters I've
tried or crashed the system. Same adapters work flawlessly when used with
the driver shipped by SRM (it's not restricted to the PCV download cable,
only :D )
Keith Wong
anyone else is in a similar situation, the best solution I've found is
to download data using Trainingpeaks Device Agent in OSX with the 64
bit Prolific driver and then export to GC. I haven't tested this
extensively yet, but it worked once at least. If it works ok, it's
more convenient than rebooting in 32 bit mode or running Windows.
Cheers,
Keith
Rainer Clasen
Uhm... ok, if Device Agent is getting along with the prolific driver, then
there's a chance we can get this going, as well...
I guess having the prolific kext loaded prevents DA from doing the raw USB
access itself (say libusb...). So it's likely just using different termios
settings... would be quite interesting to trace invocations to tcsetattr,
tcflush, tcdrain and tcsendbreak - including all passed arguments.
Hmm, is DA working with the 32bit prolific driver, as well?
Jamie Kimberley
While It would be great to have srmio work with the official
prolific drivers we may also be able compile the opensource drivers
with 64-bit support on mac 10.7.
from a comment on the project page:
( http://sourceforge.net/projects/osx-pl2303/ )
Compiles and works under Mac OS X 10.7.2 with an ATEN usb-to-serial
adapter. Start a new IOKit Driver projekt within XCode, copy over
osx_pl2303.cpp and osx_pl2303.h. Remove all calls to
KUNCUserNotificationDisplayNotice(). Replace all casts of void* to
UInt32 with casts to UInt64. From the original Info.plist, copy over
the IOKitPersonalities node. Under OSBundleLibraries, replace all
three references of com.apple.kernel* (6.9.9) with com.apple.kpi*
(10.0.0)
I would gladly give this a shot but I don't have a machine that
runs Mac 10.7.
Anyone want give it a shot?
Jamie
"You don't use science to show that you're right, you use science to
become right."--xkcd
__________________
Jamie Kimberley
Postdoctoral Fellow
Department of Mechanical Engineering
The Johns Hopkins University
Office: 410.516.5162
Mobile: 217.621.8272
Fax: 410.516.4316
E-Mail:jamie.k...@jhu.edu
Jamie Kimberley
After a little more looking someone has patched to compile a
universal binary kext that includes 64-bit support for os 10.6.
http://sourceforge.net/tracker/?func=detail&aid=2952982&group_id=157692&atid=804837
I believe this should run native in 64-bit on 10.7.
Keith, if you want to give this a try and need help installing the
.kext hit me up off list.