How to deal with comm port assignments with FTDI chips

[Dwayne Reid]

Good day to all.

I'm having a buddy write a small utility for me using one of myriad
BASIC packages out there.  He gets to choose one of the following
flavors: MS VB dot net Express (free version), Real Basic, Power
Basic, Liberty Basic.  I don't care which of the four he uses - this
is (hopefully) a short term solution until I get around to writing
something that is stand-alone.

The utility needs to merge 2 serial comm streams into a single,
combined output.  It thus needs 3 serial ports which I plan to
implement with FTDI FT232BM or FT232BL USB to serial adapters.  I'm
stuck with using a small netbook for this project - no serial ports
but has 3- USB ports.  I'm using these particular adapters because I
have them handy - they are a tiny PCB with a USB jack on one end and
TTL terminals on the other end.  I need to interface to RS-422 serial
ports so the TTL levels are easiest for me.

Something that concerns me is that I don't know how to ensure which
USB port is tethered to which comm port.  Because I have 3 identical
USB to serial adapters, its entirely possible that the comm port
numbers will move around between the 3 USB ports everytime the
computer is rebooted.

Is there something that I can do to ensure that the comm port
assignment stays with a particular physical USB port?  Alternatively,
is there something that I can do that ensures that the comm port
assignment stays with a particular USB-serial adapter card?  I think
that I recall reading somewhere that some of the FTDI chips have
eeprom that can be used to lock a particular adapter to its comm
number but I'll be darned if I can find it now.

Many thanks for any ideas that you might have!

dwayne

--
Dwayne Reid   <dwayner@...>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
[phone] voice          [phone] fax
www.trinity-electronics.com
Custom Electronics Design and Manufacturing

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Xiaofan Chen]

On Fri, Apr 23, 2010 at 6:47 AM, Dwayne Reid <dwayner@...> wrote:

> Is there something that I can do to ensure that the comm port
> assignment stays with a particular physical USB port?  Alternatively,
> is there something that I can do that ensures that the comm port
> assignment stays with a particular USB-serial adapter card?  I think
> that I recall reading somewhere that some of the FTDI chips have
> eeprom that can be used to lock a particular adapter to its comm
> number but I'll be darned if I can find it now.
>

This one?
http://www.ftdichip.com/Documents/AppNotes/AN_132_Re-Assigning_COM_Port_Numbers_Using_Registry.pdf



--
Xiaofan http://mcuee.blogspot.com

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Bob Blick-4]

On Thu, 22 Apr 2010 16:47:11 -0600, "Dwayne Reid" said:

> Something that concerns me is that I don't know how to ensure which
> USB port is tethered to which comm port.  Because I have 3 identical
> USB to serial adapters, its entirely possible that the comm port
> numbers will move around between the 3 USB ports everytime the
> computer is rebooted.

Hi Dwayne,

In my limited knowledge of Windows and USB-serial adapters:

A USB-serial adapter should grab the same comport number each time it is
plugged into the same USB port. The port numbers will not overlap. If
you plug adapter A into USB port 1 and it is assigned com4, it will
always be com4 if it is plugged into USB port 1. If you plug it into
another USB port it will grab the next never-used-before comport number,
but plugging it back into USB port 1 it will be com4 again. A different
USB-serial adapter will grab different comport numbers. Comport numbers
stay reserved "forever" for their mated adapter/USB port combinations.

So if you label the USB-serial adapters and the USB ports nothing should
change.

All of this works fine until you get into two-digit com numbers, then
you mileage may vary and you might consider going to the control panel
or if they are FTDI adapters they have a utility that cleans out the
registry entries and you start fresh.

Cheers,

Bob

--
http://www.fastmail.fm - Choose from over 50 domains or use your own

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Brendan Gillatt]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Dwayne Reid wrote:

> Good day to all.
>
> I'm having a buddy write a small utility for me using one of myriad
> BASIC packages out there.  He gets to choose one of the following
> flavors: MS VB dot net Express (free version), Real Basic, Power
> Basic, Liberty Basic.  I don't care which of the four he uses - this
> is (hopefully) a short term solution until I get around to writing
> something that is stand-alone.
>
> The utility needs to merge 2 serial comm streams into a single,
> combined output.  It thus needs 3 serial ports which I plan to
> implement with FTDI FT232BM or FT232BL USB to serial adapters.  I'm
> stuck with using a small netbook for this project - no serial ports
> but has 3- USB ports.  I'm using these particular adapters because I
> have them handy - they are a tiny PCB with a USB jack on one end and
> TTL terminals on the other end.  I need to interface to RS-422 serial
> ports so the TTL levels are easiest for me.
>
> Something that concerns me is that I don't know how to ensure which
> USB port is tethered to which comm port.  Because I have 3 identical
> USB to serial adapters, its entirely possible that the comm port
> numbers will move around between the 3 USB ports everytime the
> computer is rebooted.
>
> Is there something that I can do to ensure that the comm port
> assignment stays with a particular physical USB port?  Alternatively,
> is there something that I can do that ensures that the comm port
> assignment stays with a particular USB-serial adapter card?  I think
> that I recall reading somewhere that some of the FTDI chips have
> eeprom that can be used to lock a particular adapter to its comm
> number but I'll be darned if I can find it now.
>
> Many thanks for any ideas that you might have!
>
> dwayne
>

The FT232 chips have a small eeprom area (called the EEUA) which is
accessible from the d2xx device driver that FTDI provide freely.

I suggest you store an ID number into each chip's eeprom and then use the
d2xx driver's API to enumerate and identify each device. The d2xx driver
will then allow you to read from and write to the device in a more
straightforward way than using COM port emulation. The d2xx driver is
really simple to use so don't be too drawn to COM ports.

- --
Brendan Gillatt | GPG Key: 0xBF6A0D94
brendan {a} brendangillatt (dot) co (dot) uk
http://www.brendangillatt.co.uk
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (MingW32)

iD8DBQFL0NfjHEhZ5Ws5poERAsiVAJ4gZNO6YUWAfCTNZ6kBSF+6ElonvgCgjlI9
bXoNg3AicTTyQft3RWGTq3M=
=A1fJ
-----END PGP SIGNATURE-----
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[alan.b.pearce]

>> Something that concerns me is that I don't know how to ensure which
>> USB port is tethered to which comm port.  Because I have 3 identical
>> USB to serial adapters, its entirely possible that the comm port
>> numbers will move around between the 3 USB ports everytime the
>> computer is rebooted.
>
> Hi Dwayne,
>
> In my limited knowledge of Windows and USB-serial adapters:
>
> A USB-serial adapter should grab the same comport number each time it is
> plugged into the same USB port. The port numbers will not overlap. If
> you plug adapter A into USB port 1 and it is assigned com4, it will
> always be com4 if it is plugged into USB port 1. If you plug it into
> another USB port it will grab the next never-used-before comport number,
> but plugging it back into USB port 1 it will be com4 again. A different
> USB-serial adapter will grab different comport numbers. Comport numbers
> stay reserved "forever" for their mated adapter/USB port combinations.
>
> So if you label the USB-serial adapters and the USB ports nothing should
> change.

Only if the adapters have a unique serial number in them. Early FTDI chips
required an external SPI EEPROM into which the relevant info could be
programmed, using an FTDI utility. Later chips come with unique serial
numbers, but the hooks are still there to assign your own serial numbers and
other identifying info that USB uses when it enumerates.

> All of this works fine until you get into two-digit com numbers, then
> you mileage may vary and you might consider going to the control panel
> or if they are FTDI adapters they have a utility that cleans out the
> registry entries and you start fresh.

I understood it would work way up into 3 digit port numbers. IIRC the
maximum number of COM ports is 256!

 

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Michael Watterson]

Alan B Pearce wrote:
>
> I understood it would work way up into 3 digit port numbers. IIRC the
> maximum number of COM ports is 256!
>
>  
>
>  
On NT (XP, w2K, Vista, Win7 are really NT) since  1993,  you have com 1
to com 255

Visual basic's mscomm.dll  only allows  up  to 16 ports as it was
written for "should have been strangled at birth" Windows 95. AFAIK you
can edit a location in  the dll to have 255 ports.

Bluetooth and 3G Modems use a lot of serial ports.
I also have virtual serial over TCP/IP and local virtual serial ports to
interconnect applications.

Linux has a different scheme. \dev\ttyNN and variants with different
device name for USB
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Wouter van Ooijen]

> I understood it would work way up into 3 digit port numbers. IIRC the
> maximum number of COM ports is 256!

I am up to COM152 at the moment :)

--

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: www.voti.nl/hvu

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Michael Watterson]

Wouter van Ooijen wrote:
>> I understood it would work way up into 3 digit port numbers. IIRC the
>> maximum number of COM ports is 256!
>>    
>
> I am up to COM152 at the moment :)
>
>  
Don't run my son's Java app then. He stupidly enumerates the ports and
then attaches to the lowest valid one by default. This worked fine on
his desktop (one real Serial port and one fictitious  com2, no header)
and  his  laptop  (one  USB  serial adaptor).

On my laptop it took 3 mins (real + USB serial + IR x 2 kinds on one
physical IR port + Bluetooth + TCP/IP serial + USB Modem "serials", and
virtual loopback serials to interconnect applications).  So I guess
maybe 10min to 25min on your system!

I told him there is a good reason everyone else just  has a  blind  list
and up to  user to  figure via  utilities or device manager  what  COMx
the Bluetooth,  USB,  or remote TCP/IP serial port server box is.

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Michael Watterson]

I need to drive a 3.3V  chip from 5V PIC (PIC18F4550). The PCB with the
3.3V chip already has 10K pullups to +3.3V on the chip inputs. (the 3.3V
chip is an off the shelf  iTrip  type module using A7303B with  5
parallel pins to set FM Radio TX channel, I'll only be using 3 or 4
pins, instead of the normal 4 way switch, giving me 7 channels minimum)

So I thought maybe drive with Open Drain. I thought  did this about 7
years ago in Assembler on 16F877. But I don't know how to do it
exclusively in JAL.

Also I can find NO mention of "open drain" output mode on any port on
18F4550 data sheet. maybe I imagined I did this before once.

The "weak pullups" mentioned in datasheet for port B and Port D would
appear to be approx 5K to 20K (probably FETs). That would give 1.7V
above 3.3V, and assuming clamp diodes to VDD (3.3V) then about 50uA
clamp diode current.

Can the Port B or Port D be open drain without pullups? Does the 18F4550
have open drain at all?


I could use 4 diodes, but if the V_low Max of the 3.3V chip is above
0.7V, that won't work (many 3.3v chips the max "0" is about 150mV). So
then option is 4 x BS170 fet (no resistor needed) or 4 x NPN transistors
with base resistor. Or an HC logic open drain driver (if I had one to
hand, I have a bunch of ULN200x chips, but more space than 4 x BS170s)

The datasheet of A7303 doesn't say what the levels are for its
programming pins. All is characterised for a 3.3V VDD (and off the shelf
PCB has 3.3 regulator). It claims Absolute Max is 5V VDD and 5V for "pins".

Thoughts?

I *COULD* see if diodes to ground "program" the channel, but without a
spec, how do I know that still works (if it works) when the unit is
hotter or colder?

A production version will likely use an 18FxxJxx or an ARM cortex M3
(I'm leaning to a higher end chip with Linux for production so MP3
playback via USB host from dongle or HDD driver using sw would work, an
8bit PIC core can't usefully do that).

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Dario Greggio (in giro)]

Michael Watterson ha scritto:
> I need to drive a 3.3V  chip from 5V PIC (PIC18F4550). The PCB with the
> 3.3V chip already has 10K pullups to +3.3V on the chip inputs. (the 3.3V
 >[...]

What if you use the "old" trick of setting LATx registers to 0 and then
set TRIS to output for dominant 0 and to input for recessive 1... ?


--

Ciao, Dario
--
Cyberdyne
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[peter green-2]

Michael Watterson wrote:

> I need to drive a 3.3V  chip from 5V PIC (PIC18F4550). The PCB with the
> 3.3V chip already has 10K pullups to +3.3V on the chip inputs. (the 3.3V
> chip is an off the shelf  iTrip  type module using A7303B with  5
> parallel pins to set FM Radio TX channel, I'll only be using 3 or 4
> pins, instead of the normal 4 way switch, giving me 7 channels minimum)
>
> So I thought maybe drive with Open Drain. I thought  did this about 7
> years ago in Assembler on 16F877. But I don't know how to do it
> exclusively in JAL.
>
> Also I can find NO mention of "open drain" output mode on any port on
> 18F4550 data sheet. maybe I imagined I did this before once.
>  

You can make any IO pin on a pic act in an effectively open drain mode
by leaving the data output (port or lat) bit permanently at zero then
controlling the output through the tris bit.
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Wouter van Ooijen]

> I told him there is a good reason everyone else just  has a  blind  list
> and up to  user to  figure via  utilities or device manager  what  COMx
> the Bluetooth,  USB,  or remote TCP/IP serial port server box is.

I think an 'auto search' is not that bad an idea, but its should be a
user-activated option. It is not up to an application to decide that it
can scan (and presumably try to communicate) on all available ports.
There might be something critical attached to one of the ports...

--

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: www.voti.nl/hvu

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Michael Watterson]

peter green wrote:

> Michael Watterson wrote:
>  
>> I need to drive a 3.3V  chip from 5V PIC (PIC18F4550). The PCB with the
>> 3.3V chip already has 10K pullups to +3.3V on the chip inputs. (the 3.3V
>> chip is an off the shelf  iTrip  type module using A7303B with  5
>> parallel pins to set FM Radio TX channel, I'll only be using 3 or 4
>> pins, instead of the normal 4 way switch, giving me 7 channels minimum)
>>
>> So I thought maybe drive with Open Drain. I thought  did this about 7
>> years ago in Assembler on 16F877. But I don't know how to do it
>> exclusively in JAL.
>>
>> Also I can find NO mention of "open drain" output mode on any port on
>> 18F4550 data sheet. maybe I imagined I did this before once.
>>  
>>    
> You can make any IO pin on a pic act in an effectively open drain mode
> by leaving the data output (port or lat) bit permanently at zero then
> controlling the output through the tris bit.
>  

Ah.. so it's only logic zero or input (= pullup). Good thinking 99.

But are the week pullup always enabled on port D and B for inputs?

though likely nn uA from 5V week pull up to a 3.3V input pulled up via
external  10K  to 3.3V isn't fatal.

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Dario Greggio (in giro)]

Michael Watterson ha scritto:
> But are the week pullup always enabled on port D and B for inputs?


no, and are there any for PORTD?
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Michael Watterson]

Wouter van Ooijen wrote:

>> I told him there is a good reason everyone else just  has a  blind  list
>> and up to  user to  figure via  utilities or device manager  what  COMx
>> the Bluetooth,  USB,  or remote TCP/IP serial port server box is.
>>    
>
> I think an 'auto search' is not that bad an idea, but its should be a
> user-activated option. It is not up to an application to decide that it
> can scan (and presumably try to communicate) on all available ports.
> There might be something critical attached to one of the ports...
>
>  

Indeed. Nothing apparently happening for 1min to 30min before splash
screen loads is a bad idea. The 1st time I thought I hadn't clicked
properly (the application also OUGHT to check if it's already running,
it didn't!).

I've no idea what the java enumerate ports function does, it may just
interrogate the registry or something. It doesn't appear to affect the
actual port(s).

On linux also this isn't simple as serial port naming conventions seem
to vary and if you look in /dev/ simply for existence of  a port name
you might see only ports that exist or all possible ports depending on
distribution,  The Java library in mention, is not part of Sun/Oracle as
they have dropped serial support(!) but does support Linux, Mac and
Windows. The enumerate function does find a USB based serial port on
Ubuntu (the Netbook has no physical serial) and also PCIe card (that
uses PCIe to PCI bridge internally) based dual serial port on XP laptop
also with no physical port. But on your PC, you might think the program
is broken :)

I'm fond of my 2002 Inspiron 8200 laptop as it has 1600x1200 screen @
15.6" approx (1200 line screen if you can even get it, is now WS, so 17"
1920x1200, or more likely 1920x1080). 1200 lines good for PDF
datasheets. Real serial, parallel, PS/2 and firewire ports. Sadly only
USB1.1, but I have a PCMCIA dual USB2.0 card. I dread trying to get a
replacement screen and port wise.

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Michael Watterson]

Dario Greggio wrote:
> Michael Watterson ha scritto:
>  
>> But are the week pullup always enabled on port D and B for inputs?
>>    
>
>
> no, and are there any for PORTD?
>  
Yes,
Port D has them as well as Port B
But re-reading the datasheet reveals that by default the pullups are off.
/RBPU turns on all port B pullups
/RDPU turns on all port D pullups

my tact switches are on port B and the LCD and FM TX are using Port D,
so I can leave port D as it is and turn on pullups on port B



--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Dario Greggio (in giro)]

Michael Watterson ha scritto:
>>
>> no, and are there any for PORTD?
>>  
> Yes,
> Port D has them as well as Port B
> But re-reading the datasheet reveals that by default the pullups are off.
> /RBPU turns on all port B pullups
> /RDPU turns on all port D pullups


thank you Michael, must've missed them then :)
I usually use PORTB for inputs (must be legacy) and PORTD for output.
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Vitaliy-14]

Dwayne Reid wrote:
> Is there something that I can do to ensure that the comm port
> assignment stays with a particular physical USB port?  Alternatively,
> is there something that I can do that ensures that the comm port
> assignment stays with a particular USB-serial adapter card?  I think
> that I recall reading somewhere that some of the FTDI chips have
> eeprom that can be used to lock a particular adapter to its comm
> number but I'll be darned if I can find it now.

If it's a one-off application, as Bob has suggested once you plug in the
adapters, they are each assigned a COM port number, and it sticks as long as
you plug them into the same USB ports.

A better solution is to use FTDI's D2XX API. It allows you to sidestep the
issue completely, as it allows you to communicate with each FTDI device
directly, eliminating the hassle of having to deal with virtual COM ports.

http://www.ftdichip.com/Documents/ProgramGuides/D2XX_Programmer%27s_Guide(FT_000071).pdf

You will need to use FT_Prog to assign each adapter a unique name.

http://www.ftdichip.com/Resources/Utilities.htm#FT_Prog

Vitaliy

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Vitaliy-14]

Brendan Gillatt wrote:
> The FT232 chips have a small eeprom area (called the EEUA) which is
> accessible from the d2xx device driver that FTDI provide freely.
>
> I suggest you store an ID number into each chip's eeprom and then use the
> d2xx driver's API to enumerate and identify each device. The d2xx driver
> will then allow you to read from and write to the device in a more
> straightforward way than using COM port emulation. The d2xx driver is
> really simple to use so don't be too drawn to COM ports.
>

D'oh. Didn't see your post before sending mine.




--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Sean Breheny]

Some of my colleagues at work have told me that com port assignments
to FTDI devices can change even if they are not unplugged. Apparently
the enumeration order on boot is not guaranteed and sometimes the
assignments can change due to this.

Sean


On Fri, Apr 23, 2010 at 7:52 PM, Vitaliy <piclist@...> wrote:

> Dwayne Reid wrote:
>> Is there something that I can do to ensure that the comm port
>> assignment stays with a particular physical USB port?  Alternatively,
>> is there something that I can do that ensures that the comm port
>> assignment stays with a particular USB-serial adapter card?  I think
>> that I recall reading somewhere that some of the FTDI chips have
>> eeprom that can be used to lock a particular adapter to its comm
>> number but I'll be darned if I can find it now.
>
> If it's a one-off application, as Bob has suggested once you plug in the
> adapters, they are each assigned a COM port number, and it sticks as long as
> you plug them into the same USB ports.
>
> A better solution is to use FTDI's D2XX API. It allows you to sidestep the
> issue completely, as it allows you to communicate with each FTDI device
> directly, eliminating the hassle of having to deal with virtual COM ports.
>
> http://www.ftdichip.com/Documents/ProgramGuides/D2XX_Programmer%27s_Guide(FT_000071).pdf
>
> You will need to use FT_Prog to assign each adapter a unique name.
>
> http://www.ftdichip.com/Resources/Utilities.htm#FT_Prog
>
> Vitaliy
>
> --
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist