View previous topic :: View next topic |
Author |
Message |
Will Reeve
Joined: 30 Oct 2003 Posts: 209 Location: Norfolk, England
|
FTDI 'v' USBN |
Posted: Tue Oct 26, 2004 2:28 pm |
|
|
I don’t want to induce a war (unlikely with the guys on here I know) it makes a better world when people have lots of different ways in doing things IMHO. My USBN9603 design draws to an end and I thought I would share some pros and cons with you (maybe a FTDI guru can do the same later in the thread). I’m not 100% set on the ‘9603 route and would consider the FTDI in future designs but in my last design it was ideal.
Advantages of USBN9603 solution:
1/ Good support and stack supplied to CCS and a no-brain plug and play development board to develop the USB side of your application
2/ Near complete control of descriptors and mode of operation (HID or bulk for instance). More get your hands dirty approach which is why we all use C instead of basic I guess! No need for additional eeprom hardware for configuration settings, or the hassle of two things to program on each unit.
3/ Cheap £4 and easily available from large distributors (you need a crystal and few R and Cs as well)
4/ From a large manufacture, good chance of being supported in 10years time.
5/ Drivers (HID) are supplied with 99% of (Windows) operating systems as standard and is 99.9% bug free.
6/ Access to driver (HID) is via well documented Windows calls no third party driver to mess with or install on the target PC.
7/ Very fast
8/ Nice big SOIC footprint for ‘coffee shake’ prone soldering iron users.
Disadvantages:
1/ PIC overhead, you need a largish (ROM space) PIC with a number digital I/O lines to implement the reference CCS design.
2/ A couple of days reading / asking question / learning is required to get your head around how the thing works.
3/ No VID or PID supplied for free.
4/ Additional power on reset chip required to sense PC cable connection (1 x SOT23 1 x resistor) and a PIC I/O pin |
|
|
Guest
|
Re: FTDI 'v' USBN |
Posted: Wed Oct 27, 2004 4:28 pm |
|
|
Advantages of USBN9603 solution:
1/ Good support and stack supplied to CCS and a no-brain plug and play development board to develop the USB side of your application
Plenty of boards with examples are also available to FTDI.
2/ Near complete control of descriptors and mode of operation (HID or bulk for instance). More get your hands dirty approach which is why we all use C instead of basic I guess! No need for additional eeprom hardware for configuration settings, or the hassle of two things to program on each unit.
I don't need the hassle. FTI gives me PID and that's all I need change in the INF
3/ Cheap £4 and easily available from large distributors (you need a crystal and few R and Cs as well)
£4 = about $6 US Last I checked FTDI 245 was selling for less that that in small quantities.
4/ From a large manufacture, good chance of being supported in 10years time.
No "big" does not give security. .... Cypress is a large man. and they have already dumped several chips that I started using 3-4 years ago !
5/ Drivers (HID) are supplied with 99% of (Windows) operating systems as standard and is 99.9% bug free.
May be... but I still run into crap on dual cpu systems. !
6/ Access to driver (HID) is via well documented Windows calls no third party driver to mess with or install on the target PC.
FTID drivers are WDM compatible and wroks veyr well including on dual cpu systems. Not sure about other FTID
drivers .
7/ Very fast
What does that mean... FTDI 245 can do 8 mega BYTES second.... USBN can do ... ?
8/ Nice big SOIC footprint for ‘coffee shake’ prone soldering iron users.
The FTDI chip is EASY to solder.
Disadvantages:
1/ PIC overhead, you need a largish (ROM space) PIC with a number digital I/O lines to implement the reference CCS design.
Yes it's a problem, I can setup the FTDI chip in about 3 lines of code.
2/ A couple of days reading / asking question / learning is required to get your head around how the thing works.
Never hurt anyone to understand the subject. Even for USBN or FDTI it's not a Disadvantages.
3/ No VID or PID supplied for free.
VID . I fyou are doing your own stuf from hardware to driver, you can "invent" a VID. This will only get your end product in trouble if you deliver it for certification ...
However no one can throw you in jail for inventing a VID or PID or TID.
The USB IF would love to take your money... for what... a number !
4/ Additional power on reset chip required to sense PC cable connection (1 x SOT23 1 x resistor) and a PIC I/O pin.
If your PIC has a free ADC pin no need to use the SOT23 simply sample the line .. It can be rough and ready... either the 3.3 or 5V is there or it's not.. |
|
|
treitmey
Joined: 23 Jan 2004 Posts: 1094 Location: Appleton,WI USA
|
|
Posted: Wed Oct 27, 2004 4:53 pm |
|
|
For my 2 cents.
I wouldn't use just any PID,VID
FTDI gave me a block of PID's when I asked.
That will make your USB unique,.. and your driver install won't step on anyones toes. |
|
|
alexz
Joined: 17 Sep 2004 Posts: 133 Location: UK
|
|
Posted: Thu Oct 28, 2004 5:33 am |
|
|
Can anybody tell me how to get around of working with the USBN9603?
Say, after initializing all the registers of the device, what next?
What is the structure?
the example of the CCS was not enough for me really.
Thanks _________________ Alex |
|
|
Darren Rook
Joined: 06 Sep 2003 Posts: 287 Location: Milwaukee, WI
|
Re: FTDI 'v' USBN |
Posted: Thu Oct 28, 2004 10:27 am |
|
|
Anonymous wrote: |
What does that mean... FTDI 245 can do 8 mega BYTES second.... USBN can do ... ?
|
Nice one, Hans. How can FTDI 245 do 8MB/s when full speed USB is only 12Mb/s? That's 12 mega BITS per second.
Quote: | Quote: | 5/ Drivers (HID) are supplied with 99% of (Windows) operating systems as standard and is 99.9% bug free. |
May be... but I still run into crap on dual cpu systems. ! |
So you are saying Microsoft would provide drivers that don't work on dual cpu systems????
Quote: | Quote: | 1/ PIC overhead, you need a largish (ROM space) PIC with a number digital I/O lines to implement the reference CCS design. |
Yes it's a problem, I can setup the FTDI chip in about 3 lines of code. |
Yes, and you lose all the intelligence of USB. |
|
|
Guest
|
Re: FTDI 'v' USBN |
Posted: Fri Oct 29, 2004 10:43 am |
|
|
Darren Rook wrote: | Anonymous wrote: |
What does that mean... FTDI 245 can do 8 mega BYTES second.... USBN can do ... ?
|
How can FTDI 245 do 8MB/s when full speed USB is only 12Mb/s? That's 12 mega BITS per second.
No 8 mega bits sec.
Quote: | Quote: | 5/ Drivers (HID) are supplied with 99% of (Windows) operating systems as standard and is 99.9% bug free. |
May be... but I still run into crap on dual cpu systems. ! |
So you are saying Microsoft would provide drivers that don't work on dual cpu systems????
Yes lots and lots...
Quote: | Quote: | 1/ PIC overhead, you need a largish (ROM space) PIC with a number digital I/O lines to implement the reference CCS design. |
Yes it's a problem, I can setup the FTDI chip in about 3 lines of code. |
Yes, and you lose all the intelligence of USB. |
Crap. |
|
|
Darren Rook
Joined: 06 Sep 2003 Posts: 287 Location: Milwaukee, WI
|
Re: FTDI 'v' USBN |
Posted: Fri Oct 29, 2004 9:06 pm |
|
|
You make me laugh.
Best. CCS. Thread. Ever. |
|
|
|