Hacker News Reader

Comment by jjoonathan | Hacker News Reader

Does anyone know of good FTDI alternatives? Are there any clone makers that are relatively legit (i.e. they put their actual brand name on the chip, they support drivers, etc)? At $4.50 a pop for bog-standard bit-banging in a day and age where you can get ARM M4 SoCs for $2.75 a pop (n=1 prices) I would think FTDI would have more above-the-table competition than they do.

Is the subterfuge required for illegitimate cloning really that much easier than getting a website, writing docs, and supporting drivers?!

View on HN

Someone in the linked thread recommended the Cypress CY7C65213 and posted a link where Cypress claims it is a direct pin-for-pin compatible replacement that doesn't need proprietary drivers: http://www.cypress.com/?rID=83118

It's also cheaper; $2.44 for one and it goes down to $1.44 for 2500: http://www.digikey.com/product-detail/en/CY7C65213-32LTXI/CY...

As mentioned in the eevblog forum, Microchip make the MCP2200 which is about $2 in single quantities. Seems similar in features to FT232 chips and reasonable in price.

http://www.microchip.com/wwwproducts/Devices.aspx?product=MC...

The Microchip part needs an external crystal or ceramic resonator for a time reference. The FTDI part doesn't need any external parts beyond bypass caps.

If you're already soldering, a crystal isn't so much work or expense these days.

FTDI's major advantage is that their WHQL'd drivers are already installed with Windows. That's why people build chips to the same API.

Ah, that makes sense. The "right" answer would probably be to convince the USB-IF to standardize the interface. FT232 and FT245's popularity should be argument enough that it's worth doing. I have no idea how politically viable that would be.

EDIT: looks like they did standardize a USB-to-serial device class [1]. Maybe they just need to update it?

[1] http://www.atmel.com/Images/doc4322.pdf

Didn't see your post when I was posting mine, but anyway -- CP210x.

Do you know if it supports non-standard baud rates on OS X> (Maybe Linux too?)

I use FTDI because I can easily run comm ports at 3 Mbit/s. OSX (and I think Linux too) does support comm ports at non common baud rates. I'm looking for a way to add decently quick USB comms to a device without writing my own drivers. FTDI seems like the only option. Using HID is limited to 64KB/s.

Yes, Linux too. I'm running 2 MHz, the fastest that my jelly bean microcontroller du jour can do. I'm using FTDI on the remote end, and Python serial port drivers on the PC. I'm finding that with relatively little effort, I can make my gadgets and support software run on both Windows and Linux with no code changes, which probably sounds pretty mundane to experienced hackers, but is a convenient benefit nonetheless.

You could always just use libusb to talk to your device (e.g., an AVR w/ V-USB or another microcontroller with actual USB support) from user space, instead of fooling around with fake COM ports.

That's probably what I should do. I bought a book on usb and started learning a bit about the protocol. Are there any good Libusb projects that would be a good example? Most I was were for using libusb to talk to an existing device class. I'd like a sample that contains both device and host?

Take a look at how avrdude works over libusb. That's a pretty good example of bulk-upload and bulk-download and various types of things you're likely to want to do with a microcontroller.

I have not designed for FTDI and new in embedded desings but, for USB interfacing cypress PSoCs have served me well till now. They have dedicated USB chips as well.

Reply on news.ycombinator.com