next topic »

[on:]

If you are using the Silicon Labs CP2102 USB-Serial bridge the behaviour of the RTS line changes.

Previously this line changed state when the port was opened by the host, now under the new OS it does nothing.

This is important because it can be used as an auto reset on the Arduino embedded board to synchronise the boot loader with the downloading software. No other pin changes on opening the port so while you can still communicate you have to manually reset the board for downloading.

I have contacted the Silicon Labs support team and they have promised an update in 3-4 weeks.

Note this does not affect the normal Arduino boards as they use an FTDI chip, this is only an issue if you are using the Silicon Labs chip.

[Reply #1 on:]

I don't like the SI Labs devices.  Their driver has been the only piece of software I've used that caused a kernel dump on Mac OS X.  The devices themselves do not have a unique identifier, so if you use more than one they are a huge pain in the neck.

If you haven't already bought some, choose FTDI instead.  (Sounds like that advice is too late for Mike.   )

-j

[Reply #2 on:]

Hi,

The devices themselves do not have a unique identifier, so if you use more than one they are a huge pain in the neck.

The CP2102 does support a Serial-Number ID just like the FDTI chips. I have a ATXMega-breakout board here with a unique serial. In their datasheet they also claim to have a tool comparable to MProg on their website.

Eberhard

[Reply #3 on:]

so if you use more than one they are a huge pain in the neck.

That's odd because I have used them for about four years and have had no problems both on Windows and Mac.

This is what I designed using it:-
http://www.proxcard.co.uk/acatalog/PAC_Desk_Top_Reader_Admin_Kits.html

[Reply #4 on:]

The CP2102 does support a Serial-Number ID just like the FDTI chips.

It's entirely possible that I'm mistaken, but I  have several of these device (or the 210x family, for sure) at work.  When using them with linux, udevinfo couldn't find any information that was unique to the device.  This is a serious problem, as I have a box that has 4 of these units installed, and I don't have a good way to differentiate between them.

Maybe they're simply blank and require use of their mprog-like utility to configure a serial #?

In any event, it was definitely an SI labs CP210x driver than kernel dumped my mac years ago.  This left me dead against the devices, and seriously disappointed in Apple (although in fairness I don't know if OX X could have prevented a bad driver from causing a dump, once I had installed and executed the code).

-j

[Reply #5 on:]

Anybody ever get this figured out?  I'm using a CP2102 USB->TTL converter with Windows 7 and Windows XP, using the Arduino host the reset pin is never asserted low so you have to manually trigger the reset when programming.

[Reply #6 on:]

I find that it works under XP but not on the Mac.

Silicon Labs promised to bring out a fixed driver for this, I am still waiting and periodically get emails from them saying they are going to do it.
It wouldn't do any harm if other people were to contact them as well. It could spur them into action.

[Reply #7 on:]

Waterppk,

The auto reset should work, at least if you use the schematics I had here: http://www.kerrywong.com/2010/07/16/an-arduino-compatible-using-cp2102/. It will automatically reset the chip just like in Arduino...

[Reply #8 on:]

It seems you are using the DTR line instead of the RTS line like the OP is.  Which I guess makes a little more sense, since you aren't requesting anything.

[Reply #9 on:]

The RST pin on CP2102 is used to reset the CP2102 not the ATmega, the DTR line is used for that.