From: Torsten W. <tor...@fh...> - 2006-06-27 09:13:11
|
Hi everyone, Since this is my first mail to this list, sorry for my maybe newbie behavior. I noticed that more and more people asking for support of the USB based TI FET devices. I just wondering, since it seems to me that the FET device "just" create an additional COM-port how difficult it would be to verify this for Linux. FTDI for example supports Linux with there USB-RS232 Chips. If nobody knows already I think it might be worse to open my box and have a look to the USB-Chip inside. If the USB box really just emulates an additonal COM port, the communication between a USB and a RS232 FET box should be identical. CC. Is there no other way to install the drivers for the USB FET box without installing IAR ? There should be some drivers somewhere. Greetings Torsten |
From: Steve U. <st...@co...> - 2006-06-27 12:41:55
|
Torsten Wagner wrote: >Hi everyone, > >Since this is my first mail to this list, sorry for my maybe newbie >behavior. > >I noticed that more and more people asking for support of the USB based >TI FET devices. >I just wondering, since it seems to me that the FET device "just" create >an additional COM-port how difficult it would be to verify this for Linux. > >FTDI for example supports Linux with there USB-RS232 Chips. If nobody >knows already I think it might be worse to open my box and have a look >to the USB-Chip inside. > >If the USB box really just emulates an additonal COM port, >the communication between a USB and a RS232 FET box should be identical. > >CC. Is there no other way to install the drivers for the USB FET box >without installing IAR ? There should be some drivers somewhere. > > The TI USB FET tools use a TI TUSB3410 USB controller. This is supported by Linux, when used in its default configuration as a simple serial port. However, in the FET tool it is configured to look different from a standard serial port. If you look at the Windows drivers for the FET tool, they are special. The FET tool appears as a multi-port serial card in the Windows device manager. You actually install 2 device drivers during the Windows installation, which confuses people quite a bit. Regards, Steve |
From: Torsten W. <tor...@fh...> - 2006-06-29 15:28:08
|
Hi Steve, thanks for the info. > The TI USB FET tools use a TI TUSB3410 USB controller. This is supported > by Linux, when used in its default configuration as a simple serial > port. However, in the FET tool it is configured to look different from a > standard serial port. If you look at the Windows drivers for the FET > tool, they are special. The FET tool appears as a multi-port serial card > in the Windows device manager. You actually install 2 device drivers > during the Windows installation, which confuses people quite a bit. I just open my USB-FET Box and surprise .... the JTAG interface is controlled by an MSP itself. This means a MSP430 programs MSP430s. O.K. I will not start to ask the question what was first the egg or the chicken ;) Nevertheless, it might be interesting to know (if not known already) that the following application notes from TI use nearly the identical PCB Technical Documents MSP430 USB Connectivity using TUSB3410 MSP430 USB Connectivity using TUSB3410 Application Report. Don't know whether this helps to create driver for Linux. Furthermore, I checked the .INF files of this project. Fortunately they provide a "standard" and a "customised" version for the drivers. As you said, the one version loads an additional driver and done some configuration for that. The PID VID Number stuff is more a cosmetic stuff, most important to load the correct driver set for the device. A diff of the drivers for the USB FET with the "customised" version in this project shows many similarities. It might give some additional infos. However, I think it should be possible to get it working under Linux.... lets see ;) Greetings, Torsten |
From: Oleg V. <ol...@gm...> - 2006-06-30 21:43:28
|
Hallo. > Nevertheless, it might be interesting to know (if not known already) > that the following application notes from TI use nearly the identical PCB > Technical Documents > MSP430 USB Connectivity using TUSB3410 > MSP430 USB Connectivity using TUSB3410 Application Report. > Don't know whether this helps to create driver for Linux. > However, I think it should be possible to get it working under Linux.... > lets see ;) If you really want to do things starting to work, just get firmware sources (free from TI) and begin to try to code using latest linux ddk: <http://www.kernel.org/pub/linux/kernel/people/gregkh/ddk/> Also check this article and its refs: <http://analogzone.com/iot_1003.pdf> IMHO, today most of the work is done via firmware, and PCB and/or connections are just things to do once. That's why it's problem to get something to work. I'm willing to start developing, but my order of EZ430-F2013 is scheduled by TI e-store to september. Maybe someone may suggest some on-line store with more fast ship ? Also I haven't experience with ms-windos at all, and can't get all that shiny ready-to-work-get-the-fucks things and bring it to linux-power-user easily. So, Torsten, get your linux-friends and just start sending patches. Maybe you're that talent we've waited for ? Good luck ! -- olecom |
From: KIMURA S. <sk...@ma...> - 2006-07-01 13:59:39
|
On Jun 30, 2006, at 4:43 PM, Oleg Verych wrote: > > If you really want to do things starting to work, just get firmware > sources (free from TI) For those who have USB FET's firmware sources, I can give you still immature interface program source code for MacOSX. My program can 1. read target boards MSP430's memory 2. load ELF file to target boards. 3. hang up host Mac!! Porting BSD or linux should be easy if you know how to deal with USB on those machine. -- skimu |
From: kavaler <ka...@di...> - 2006-07-04 05:44:02
|
I have been able to get the UIF to work under linux, but it wasn't straight-forward. The approach I took was to write my own driver and API for the UIF, loosely based on the official TI API. This approach is not great because I only have one version of the UIF (1.3) and a couple of chips that I use and I don't use gdb, so my development is of limited use. Also, I lose all the workarounds that TI put into their code, whatever they might be. The plus is that I am willing (assuming TI does not have a fit) to put my code into the public domain under GPL so others can add to it. I believe that an object code version of the gdbproxy will be available shortly from someone else. I am not sure the best way to post the code and instructions to get the UIF to work. Suggestions would be appreciated. -- View this message in context: http://www.nabble.com/USB-FET-for-Linux-tf1854102.html#a5162220 Sent from the MSP430 gcc - Users forum at Nabble.com. |
From: N. C. <ni...@nc...> - 2006-07-01 19:27:23
|
At 21:43 30-06-06 +0000, you wrote: >Hallo. > >> Nevertheless, it might be interesting to know (if not known already) >> that the following application notes from TI use nearly the identical PCB > >> Technical Documents >> MSP430 USB Connectivity using TUSB3410 >> MSP430 USB Connectivity using TUSB3410 Application Report. > >> Don't know whether this helps to create driver for Linux. > >> However, I think it should be possible to get it working under Linux.... >> lets see ;) > >If you really want to do things starting to work, just get firmware >sources (free from TI) and begin to try to code using latest linux >ddk: ><http://www.kernel.org/pub/linux/kernel/people/gregkh/ddk/> >Also check this article and its refs: ><http://analogzone.com/iot_1003.pdf> > >IMHO, today most of the work is done via firmware, and PCB and/or >connections are just >things to do once. That's why it's problem to get something to work. > >I'm willing to start developing, but my order of EZ430-F2013 is >scheduled by TI e-store to september. Maybe someone may suggest some >on-line store with more fast ship ? > >Also I haven't experience with ms-windos at all, and can't get all that shiny >ready-to-work-get-the-fucks things and bring it to linux-power-user easily. > >So, Torsten, get your linux-friends and just start sending patches. >Maybe you're that talent we've waited for ? Is the TUSB3410 doing anything special on that FET board? If no, would it help to load new firmware into the TUSB3410 which makes it act as a standard serial port? TI has tools available which allow to modify the contents of a EEPROM connected to a TUSB3410. Nico Coesel |
From: KIMURA S. <sk...@ma...> - 2006-07-01 20:13:45
|
> > Is the TUSB3410 doing anything special on that FET board? No nothing special. You can load TUSB3410's firmware from USB port as described in the manual. I am using firmware-0.1.tar.gz in http://sourceforge.net/projects/uticom with minor and trivial modification. -- skimu |
From: KIMURA S. <sk...@ma...> - 2006-07-01 22:00:07
|
http://homepage.mac.com/skimu/fetuif-0.0.tgz BOOT FET430UIF's TIUSB3410. 0. Files boot3410.c : boot TIUSB3410 from USB port boot430.c : boot MSP430 boot430.h firm.flat : flat binary for TIUSB3410's firmware hex.c : Intel hex utility hex.h titxt.c : TITEXT->Intel hex libusbcom/ : Library for USB communication. tusb3410/ : Sources for firm.flat libfet430/ : Empty because this is using TI's IP. main.c : 1. How to play with. When you plug FET430UIF into your PC (w/o TI's VCP driver) or Mac, it has only one bulk output port. Then it is the time for you to load USB3410's firmware with this program. This program will put firmware into the port according to Section 11.7 of ``TUSB3410/I Data Manual'' (SLL519D). If you successfully load firmare to USB3410, FET430UIF will have both Bulk Input and Bulk output port. Basically you can assume that these I/O port is MSP430's serial port. (In another words, USB3410 is now transparent) On MacOSX, just do, $ make boot3410 $ ./boot3410 On other system (BSD or Linux) 1. Write code as needed in libusbcom/usbcom_libusb.c 2. What I can do after loading firmware? You can either, 1. Get FET430UIF's source code from TI, and write host program to deal with FETUIF. 2. Develop your own FET430UIF's firmware. As far as circuit diagram from TI goes, FET430UIF newer than Rev1.4 has ability to bootstrap by USB3410 (i.e, host computer through USB). Good thing for 1. is that you do not have to modify anything on FET430UIF and let TI work to improve firmware. Bad thing for 1. is that your can not distribute your program as free software (like me), even binary form (violation of GPL). Good thing for 2. is that you know everything what's going on inside. Possibly you get much better gdb interface. You can distribute in binary form without violating GPL. (if you link GPLed library such as BFD.) Bad thing for 2. is your FET430UIF is no longer work with TI's official environment. (you have to write back TI's firmware.) There is a good chance that you have to deal with each device/family's quirky behavior. References SLAU138C SLAA276 SLAA149A SLLS519D SLLC139.zip --Awk script to split SLLC139.zip BEGIN { file = "README.txt" } /^File/ { file = $3; print file } !/^File/ {print >> file} -- 7/1/2006 Shigenobu Kimura <sk...@ma...> |
From: LvA <lau...@cs...> - 2006-09-26 07:47:19
|
Hello Thanks for the hints, however, I am still lacking some information in order to get the MSP-FET430UIF to operate under Mac OS X. I have performed the two steps : $ make boot3410 $ ./boot3410 but there are no new devices in /dev. Were you able, with your code, to connect a target for debugging, for example using msp430-gdbproxy? Laurent von Allmen KIMURA Shigenobu wrote: > > > http://homepage.mac.com/skimu/fetuif-0.0.tgz > > BOOT FET430UIF's TIUSB3410. > > 0. Files > > boot3410.c : boot TIUSB3410 from USB port > boot430.c : boot MSP430 > boot430.h > firm.flat : flat binary for TIUSB3410's firmware > hex.c : Intel hex utility > hex.h > titxt.c : TITEXT->Intel hex > libusbcom/ : Library for USB communication. > tusb3410/ : Sources for firm.flat > > libfet430/ : Empty because this is using TI's IP. > main.c : > > 1. How to play with. > > When you plug FET430UIF into your PC (w/o TI's VCP driver) > or Mac, it has only one bulk output port. Then it is the time for > you to load USB3410's firmware with this program. > This program will put firmware into the port according > to Section 11.7 of ``TUSB3410/I Data Manual'' (SLL519D). > > If you successfully load firmare to USB3410, FET430UIF will > have both Bulk Input and Bulk output port. Basically you can > assume that these I/O port is MSP430's serial port. > (In another words, USB3410 is now transparent) > > On MacOSX, just do, > > $ make boot3410 > $ ./boot3410 > > On other system (BSD or Linux) > > 1. Write code as needed in libusbcom/usbcom_libusb.c > > > 2. What I can do after loading firmware? > > You can either, > > 1. Get FET430UIF's source code from TI, > and write host program to deal with FETUIF. > > 2. Develop your own FET430UIF's firmware. > As far as circuit diagram from TI goes, > FET430UIF newer than Rev1.4 has ability to bootstrap > by USB3410 (i.e, host computer through USB). > > Good thing for 1. is that you do not have to > modify anything on FET430UIF and let TI work > to improve firmware. > > Bad thing for 1. is that your can not distribute > your program as free software (like me), even > binary form (violation of GPL). > > Good thing for 2. is that you know everything > what's going on inside. Possibly you get much > better gdb interface. You can distribute in binary > form without violating GPL. > (if you link GPLed library such as BFD.) > > Bad thing for 2. is your FET430UIF is no longer > work with TI's official environment. (you have to > write back TI's firmware.) There is a good chance > that you have to deal with each device/family's > quirky behavior. > > References > > SLAU138C > SLAA276 > SLAA149A > SLLS519D > SLLC139.zip > > --Awk script to split SLLC139.zip > BEGIN { file = "README.txt" } > /^File/ { file = $3; print file } > !/^File/ {print >> file} > -- > > 7/1/2006 > Shigenobu Kimura <sk...@ma...> > > > > Using Tomcat but need to do more? Need to support web services, security? > Get stuff done quickly with pre-integrated technology to make your job > easier > Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo > http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 > _______________________________________________ > Mspgcc-users mailing list > Msp...@li... > https://lists.sourceforge.net/lists/listinfo/mspgcc-users > > -- View this message in context: http://www.nabble.com/USB-FET-for-Linux-tf1854102.html#a6502034 Sent from the MSP430 gcc - Users mailing list archive at Nabble.com. |
From: ezmez <db...@ya...> - 2006-08-03 08:18:04
|
Hey I have an application where I need to use the TUSB3410 with a linux box but I wanted a little bit of info from you guys before I went and built the whole thing. I read in one of your replies that linux only supports the 3410 with the default VID/PID. I want to use the TUSB3410 as an alternative to the FTDI ( I already have a few TUSB but it costs me a heep to get FTDI chips because of postage and handling etc) as a USB to serial bridge but I want to do this on a linux box. When I make my board for the USB to serial bridge should I include a blank EEPROM or should I leave the EEPROM disconnected? Also could you clafiry weather I need to program the TUSB3410 before use? Or can I just build a board around them and connect it to a linux system and have myself a USB to Serial bridge??? what do I need to do in linux from a software/ driver perspective to use the TUSB and send serial out? If anyone has an app/ ckt/ any more info could you please fwd it to me @ db...@ya.... Many thanks, Ezmez -- View this message in context: http://www.nabble.com/USB-FET-for-Linux-tf1854102.html#a5628354 Sent from the MSP430 gcc - Users forum at Nabble.com. |
From: Chris L. <cli...@gm...> - 2006-08-06 18:07:58
|
ezmez schrieb: > I read in one of your replies that linux only supports the 3410 with the > default VID/PID. yes. but i think the bug is found. i used the chip with the default ID and it worked flawless (unlike others... FTDI is fine too) > When I make my board for the USB to serial bridge should I > include a blank EEPROM or should I leave the EEPROM disconnected? both is OK. only when you want to reprogram an alredy used EEPROM, you need a way to disconnect during startup > Also could you clafiry weather I need to program the TUSB3410 before use? Or > can I just build a board around them and connect it to a linux system and > have myself a USB to Serial bridge??? Yes. no programming is needed, it works with the default ID without EEPROM (or an empty one). the driver downloads the code when the adapter is plugged in (reuqires that you have set up hotplug) as alternative you can also download the firmware to the EEPROM, which you can do in curcuit (at least with the Windows tools). i never tried this variant myself. > what do I need to do in linux from a software/ driver perspective to use the > TUSB and send serial out? while the driver is included in the kernel it seems that the hotplug scripts are not installed by default (at least not on my distro) there is a separate driver download (http://gate.brimson.com/downloads/) which includes the hotplug script ("make install"). chris |