CA-42 cable -> Serial

Post new topic   Reply to topic    DD-WRT Forum Index -> Broadcom SoC based Hardware
Goto page Previous  1, 2, 3  Next
Author Message
Apokrif
DD-WRT User


Joined: 22 Aug 2007
Posts: 64

PostPosted: Fri Feb 25, 2011 4:02    Post subject: Reply with quote
Oops...
Found another reference to CA-42 with external power...
http://www.dd-wrt.com/phpBB2/viewtopic.php?t=62998&postdays=0&postorder=asc&start=123
He explained also why he didn’t use Vcc from router
Sponsor
LOM
DD-WRT Guru


Joined: 28 Dec 2008
Posts: 7647

PostPosted: Fri Feb 25, 2011 4:27    Post subject: Reply with quote
It is an annoyance when the primary side (USB side) of the converter is not powered from the USB port as CPhoenix pointed out.
It is also electrically wrong to power the primary side from 3.3V - it defeats the level conversion and gives low signal levels to the USB, not that I have had any problem with those low levels but it is still wrong Smile

The secondary side though (the TTL side) should have been made so it always takes the power from the unit it is connected to so that it could work with lower level TLL standards like 2.5V and 1.8V.
But the 3-wire cables either has an internal voltage regulator from USB side 5V to generate TTL side 3.3V or has sometimes the TTL side powered directly from the USB side which also defeats level conversion. Sad
3.3V TLL routers doesn't seem to care about 5V TTL signals from the CA-42 but it would prolly not be good to use the cable on a 2.5V TTL unit.

Many version of CA-42 there are..

_________________
Kernel panic: Aiee, killing interrupt handler!
Apokrif
DD-WRT User


Joined: 22 Aug 2007
Posts: 64

PostPosted: Fri Feb 25, 2011 4:53    Post subject: Reply with quote
LOM wrote:
It is an annoyance when the primary side (USB side) of the converter is not powered from the USB port as CPhoenix pointed out.
It is also electrically wrong to power the primary side from 3.3V - it defeats the level conversion and gives low signal levels to the USB, not that I have had any problem with those low levels but it is still wrong Smile

The secondary side though (the TTL side) should have been made so it always takes the power from the unit it is connected to so that it could work with lower level TLL standards like 2.5V and 1.8V.

Agreed 110%
Basically repeating what you wrote:
What we have so far:
3 wire is wrong – not taking Vcc from device for correct tx/rx levels
5 wire, which use external device Vcc to power whole cable chip – wrong also...
The only good should be 5/4 wire with power to cable chip from USB and taking tx/rx levels from external device (as you wrote above)
Does such a cable exist?

LOM wrote:
Many versions of CA-42 there are..

Yep, and they look exactly the same! (black cable/blue connectors)
Apokrif
DD-WRT User


Joined: 22 Aug 2007
Posts: 64

PostPosted: Fri Feb 25, 2011 4:57    Post subject: Reply with quote
IMO, the "right cable" supposed to have simplest (i.e. most straightforward) internal design too Wink
PetervdM
DD-WRT User


Joined: 11 Jun 2009
Posts: 282
Location: EU

PostPosted: Fri Feb 25, 2011 6:29    Post subject: Reply with quote
LOM wrote:
It is an annoyance when the primary side (USB side) of the converter is not powered from the USB port as CPhoenix pointed out.
It is also electrically wrong to power the primary side from 3.3V - it defeats the level conversion and gives low signal levels to the USB, not that I have had any problem with those low levels but it is still wrong Smile

the pl2303 chip ( http://www.prolific.com.tw/support/files//IO%20Cable/PL-2303HX/Documents/Datasheet/ds_pl2303HXD_v1.3.pdf ) used in this cable has two power in connections: VDD_5 for the usb part and VDD_325 for the rs232 level shift part. it also has a voltage regulator on board: VO_33, to convert the 5 volts vrom VDD_5 to 3v3. the quick solution is to use this VO_33 and connect it to VDD_325.
but you loose flexibility. VDD_325 ranges from 1v8 to 3v3, so if you let the equipment supply power to VDD_325, you have a "one size fits all" solution.
for whatever reason the designer ( NOKIA ?? ) chose to take this approach. electronically a clever decision: the usb side is 5v, the rs232 side can accept logic from 1v8 to 3v3
see also the reference designs: http://www.prolific.com.tw/eng/downloads.asp?ID=58

_________________
now running tomato by shibby
E4200v1 cfe 2010.09.20.0
Apokrif
DD-WRT User


Joined: 22 Aug 2007
Posts: 64

PostPosted: Fri Feb 25, 2011 22:06    Post subject: Reply with quote
PetervdM wrote:
electronically a clever decision: the usb side is 5v, the rs232 side can accept logic from 1v8 to 3v3
see also the reference designs: http://www.prolific.com.tw/eng/downloads.asp?ID=58

Cleverer will be to power up chip from USB and connect VDD_325 to device external power.

Anyway, read PL-2303HX pdf, one part is not clear for me:
Table 5-1 Pin Assignment & Description
VDD_325 P RS232 VDD. The power pin for the serial port signals. When the serial port is 3.3V, this should be 3.3V. When the serial port is 2.5V, this should be 2.5V. The range can be from 1.8V~3.3V.
VDD_33 P Primary Power (3.3V)
VO_33 P Regulator Power Output, 3.3V
VDD_5 P USB Port VBUS, 5V Power

I cannot download reference design. Keep getting "MS ODBS error".
I assume: VDD_5 is actually "Regulator Power Input (5v)" - Where it is otherwise?

So racheous cable need to:
Connect VDD_5 to USB power
VO_33 to VDD_33 to power up chip
Put switch between VO_33 and VDD_325 so we can power up VDD_325 either from VO_33 or external source.

On Peter and mine CA-42 disconnect VDD_33 from VDD_325

BTW: http://www.prolific.com.tw/support/files/IO%20Cable/PL-2303HX/Documents/Datasheet/ds_pl2303HXD_v1.3.pdf
mention driver for:
Provides drivers support for Windows, Mac OS, Linux, and WinCE
Windows 7, Vista, XP Certified Logo Drivers (x86 and x64)

All above is for PL-2303HX chip, I hope PL-2303 is not much different...
Apokrif
DD-WRT User


Joined: 22 Aug 2007
Posts: 64

PostPosted: Fri Feb 25, 2011 22:36    Post subject: Reply with quote
Update:
Got PL-2303 Driver Setup Installer, Prolific Edition from
http://www.prolific.com.tw/support/files/%5CIO%20Cable%5CPL-2303%5CDrivers%20-%20Generic%5CWindows%5Callinone%5CPL2303_Prolific_DriverInstaller_v130.zip

Supported device ID and product strings:
. VID_067B&PID_2303 for "Prolific USB-to-Serial Comm Port"

Driver Versions:
-------------------
- Windows 98/ME Driver: v2.0.0.19
- Windows 2000/XP/Server2003 (32 & 64-bit) WDM WHQL Driver: v2.0.13.130
- Windows Vista/7/Server2008 (32 & 64-bit) WDF WHQL Driver: v3.3.11.152


Anyway, it’s only for USB\VID_067B&PID_2303 device – only one device!
Mine CA-42 is USB\VID_0421&PID_0802
And nokia driver list 202 different devices starting with USB\VID_0421

So Prolific driver won’t work obviously…
I.e. it still could be same PL-2303 chip changed with "On-chip OTP (One Time Programming) ROM for startup device configurations"
No way to tell without breaking it - and it might be relabeled still.
Does anybody have reliable proof, it’s PL-2303 chip?
Or an idea how to get it?


Even if it’s same one, my knowledge is insufficient to tell how difficult to modify Prolific driver to pick up VID_0421&PID_0802 instead of VID_067B&PID_2303...
PetervdM
DD-WRT User


Joined: 11 Jun 2009
Posts: 282
Location: EU

PostPosted: Sat Feb 26, 2011 20:01    Post subject: Reply with quote
no, there is no way i can "proof" this to you, except that i have loaded the prolific driver on my windows 7 x86 system.
there are several ways to do this:
- reprogram the otprom with the original vid and pid. this is quite a hassle and you need knowledge and a programmer.
- alter the prolific inf file to include the nokia vid & pid. you need the knowledge to do so.
- force windows to use the driver. this is the easiest, but also the dirtiest way. this is the way the majority of people solve this problem, including me.

loading a wrong driver for a device doesn't break anything, generally the driver won't load at all.
in rare cases it can make a system crash, in which case a roll back, or uninstall by starting in safe mode is needed.

nokia altered the vid and pid to be able to make an own inf file, because they wanted to install a modem, not a serial port. the modem part is in the phone, and the original DKU cable was a real serial TTL level shifter. with the dimise of the serial port an usb solution was presented by the CA-42 cable.

_________________
now running tomato by shibby
E4200v1 cfe 2010.09.20.0
Apokrif
DD-WRT User


Joined: 22 Aug 2007
Posts: 64

PostPosted: Sat Feb 26, 2011 21:01    Post subject: Reply with quote
PetervdM wrote:
no, there is no way i can "proof" this to you, except that i have loaded the prolific driver on my windows 7 x86 system.

Wow - so it should work for W7/2008 R2 x64 as well! Shocked
I read before (in this forum) – people install XP mode for W7 x64 only to make CA-42 works! :?


PetervdM wrote:
there are several ways to do this:
- reprogram the otprom with the original vid and pid. this is quite a hassle and you need knowledge and a programmer.
- alter the prolific inf file to include the nokia vid & pid. you need the knowledge to do so.
- force windows to use the driver. this is the easiest, but also the dirtiest way. this is the way the majority of people solve this problem, including me.

I have checked inf file comes with latest prolific driver yesterday.
It’s really short – 2863 bytes (Nokia is 97970 bytes)
And it’s only 3 references to USB\VID_067B&PID_2303 in there
I guess, replacing them all will to ours will do it, or something else is needed as well?

Couple other thoughts:
If I write driver, I’ll include VID/PID check into it to make sure it communicate with right device…
AFAIR, cannot tell how easy to mod driver binary, but depend on win OS, either checksum or signature need to be corrected as well

But since prolific driver allow to change VID/PID – they don’t check them in driver code, most likely
That’s why "force load" works, I guess.

PetervdM wrote:

- reprogram the otprom with the original vid and pid. this is quite a hassle and you need knowledge and a programmer.

Wow, I thought, it is "write once" only!

PetervdM wrote:

loading a wrong driver for a device doesn't break anything, generally the driver won't load at all.
in rare cases it can make a system crash, in which case a roll back, or uninstall by starting in safe mode is needed.

Yep. That’s expected behavior :-)

PetervdM wrote:
nokia altered the vid and pid to be able to make an own inf file, because they wanted to install a modem, not a serial port. the modem part is in the phone, and the original DKU cable was a real serial TTL level shifter. with the dimise of the serial port an usb solution was presented by the CA-42 cable.

I can add, nokia software won’t work with prolific driver, they want Nokia modem only (At least that’s want I read)

I was trying to find some info about hacking CA-42 cables – that’s what I get so far:
Old cables:



New cables:


Buffers (I guess!) are gone and impossible to trace Confused
PetervdM
DD-WRT User


Joined: 11 Jun 2009
Posts: 282
Location: EU

PostPosted: Mon Feb 28, 2011 10:17    Post subject: Reply with quote
your second picture shows a pl2303hx chip ... Smile
about 64bit: never tried that. but yes, it could be possible because the prolific drivers are both 32 and 64 bit.
about altering the .inf file, never bothered to do that, but adding or replacing the ca-42 vid and pid should do the trick.
i don't think they scan for the vid & pid in the driver, because they can be altered by the customer.
reprogramming the otprom might not work if it is already done once by nokia.

if you don't want to mess with the driver, you can still use the original nokia driver. install it, go to devicemanager, modems, and choose properties of the CA-42 modem. don't know which tab, but look for advanced port settings. there you can find or change the comport # to use. afaik there are no 64bit nokia drivers.

_________________
now running tomato by shibby
E4200v1 cfe 2010.09.20.0
Apokrif
DD-WRT User


Joined: 22 Aug 2007
Posts: 64

PostPosted: Mon Feb 28, 2011 22:53    Post subject: Reply with quote
PetervdM wrote:
your second picture shows a pl2303hx chip ... Smile

Right, but it’s for old cables (3-4 year ago?)
Modern ones are last two pictures (including yours and mine)
Any ideas how to trace pins there?

PetervdM wrote:
about 64bit: never tried that. but yes, it could be possible because the prolific drivers are both 32 and 64 bit.
about altering the .inf file, never bothered to do that, but adding or replacing the ca-42 vid and pid should do the trick.

I did with XP – it doesn’t like something still.
Pointed to folder with updated .inf.
XP doesn’t like something – it went to internet and fetched new drives from MS site for Nokia modem

I wiped it out and did force install.
Got blue screen, rebooted fine, but there is exclamation sing next to driver still.

PetervdM wrote:
i don't think they scan for the vid & pid in the driver, because they can be altered by the customer.

That probably be true (sine it works for you)
Or they can charge every reseller for drives with its own vid & pid


PetervdM wrote:
reprogramming the otprom might not work if it is already done once by nokia.

Right.

PetervdM wrote:
if you don't want to mess with the driver, you can still use the original nokia driver. install it, go to devicemanager, modems, and choose properties of the CA-42 modem. don't know which tab, but look for advanced port settings. there you can find or change the comport # to use. afaik there are no 64bit nokia drivers.

Yep, it’s the latter - no 64bit nokia drivers.
And the good thing about Prolific drives - they are signed, right?
Apokrif
DD-WRT User


Joined: 22 Aug 2007
Posts: 64

PostPosted: Tue Mar 01, 2011 5:14    Post subject: Reply with quote
Peter,
Could you tell how do you "force install driver", please?
I tried both XP/W7 x86 with same result:
W7x86 "This device cannot be start (Code 10)"

I can edit INF (and update CAT)
Or I can select driver for device manually (choose com port + have disk)

I tried MS GPS out of curiosity also:
I has Prolific chip + MS signed driver ser2plms.sys
Updated driver to latest ser2pl.sys from prolific site – got same result…

Tried ser2plms.sys for CA-42 - got same error + unsigned driver!
Does "signed driver" include all devices from its INF file and doesn’t cover anything else?

Am I missing something obvious?
Anyway, don’t know what else to try - looking forward to get your instructions!
PetervdM
DD-WRT User


Joined: 11 Jun 2009
Posts: 282
Location: EU

PostPosted: Fri Mar 04, 2011 12:03    Post subject: Reply with quote
forced install:

first remove all old stuff:
- make sure no drivers can automatically loaded from the net.
- uninstall ca-42 drivers from programs and features
- from devicemanager, tab view, by connection & show hidden devices
- insert powered up cable, and uninstall all remnants of the ca-42 drivers ( ca-42 modem and ca-42 parent device ?? ). in W7 uninstall and delete.
- remove and insert cable again.
- if drivers are automatically loaded, uninstall and remove them again, until no drivers are found anymore.

install driver:
- insert powered up cable, cancel driver request
- in devicemanager right click on the device, properties, driver, update driver, browse my computer for software, let me pick from a list, have disk, browse to the driver, select, ignore warnings.

but maybe i have an other cable??

if everything fails, look at ftdi. they have a webshop and the best serial cables, great driver support.

_________________
now running tomato by shibby
E4200v1 cfe 2010.09.20.0
Apokrif
DD-WRT User


Joined: 22 Aug 2007
Posts: 64

PostPosted: Fri Mar 04, 2011 18:44    Post subject: Reply with quote
Peter,
Thanks a lot for you replay!
I did exactly as you described, actually.
I.e. once I cleaned everything – Windows can find suitable driver by itself anymore.
Then I do "let me pick from a list, have disk, browse to the driver, select, ignore warnings "
Driver install succeeded, but have yellow exclamation mark AND the message "This device cannot be start (Code 10)" (And nothing changes after reboot)
I.e. windows tried to start driver, but driver initialization fails somehow. (the driver doesn’t like device)
And same result on both XP and W7

I guess, you right - even if it is same vid & pid it might be still different chip, since prolific is programmable.

Want to ask you one more thing:
I assume you don’t have any exclamation mark next to driver, right?
What the driver version windows shows?
And is the file version of ser2pl.sys?

Mine are (from prolific web site)
--------
XP
--
ser2pl.sys 51200 Nov, 19, 2009
2.0.13.130
MD5 b4664c1ee39a5b7fc112f4077f8d21a5


W7
--
ser2pl.sys 81920 March 12, 10
3.3.11.152
MD5 ac1f2a09b76b57356f906eeda43ccc2a

If yours are different – do you have a link where I can get them from?
PetervdM
DD-WRT User


Joined: 11 Jun 2009
Posts: 282
Location: EU

PostPosted: Tue Mar 08, 2011 6:12    Post subject: Reply with quote
i only have w7 at hand. the version is: ser2pl.sys 81920 bytes dated March 12, 2010 version 3.3.11.152.

it comes from an installer: PL2303_Prolific_DriverInstaller_v130.zip 2371549 bytes, download from http://www.prolific.com.tw/eng/Download-2.asp?ID=17

_________________
now running tomato by shibby
E4200v1 cfe 2010.09.20.0
Goto page Previous  1, 2, 3  Next Display posts from previous:    Page 2 of 3
Post new topic   Reply to topic    DD-WRT Forum Index -> Broadcom SoC based Hardware All times are GMT

Navigation

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You can attach files in this forum
You can download files in this forum