Prolific PL2303 - possible to tap some power?

I haven't, but I'd recommend the FTDI 232R for the following reasons:

-- Datasheet readily available, no having to call up and possibly sign an NDA (OK, FTDI does make you sign an NDA if you want a detailed description of the EEPROM configuration, but I've used a lot of FTDI chips and so far never needed it -- I expect that you only do if you're having to write your own device driver for the part)

-- Easily available from DigiKey

-- It's a newer part, so it's a bit more integrated (no external Xtal needed)

-- Drivers are generally far less buggy: I've seen the PL2303 driver blue-screen within the past six months, whereas it's been a number of years since FTDI's drivers were that bad (not that they don't have bugs, but they're pretty mature and the occasional bugs seem rather benign)

-- Bit-bang mode is incredibly useful at times -- does the PL2303 have this?

I don't work for FTDI, but I've used their ICs on multiple occasions and generally been happier with the results than the one time I used Prolific.

---Joel

"Real time" and "PC" ceased going together as soon as it became clear that Windows was going to become the dominant OS on PCs back nearly two decades ago now. I'm always amazed at how much effort people will expend trying to coerce Windows into being something close to an RTOS when, these days, it's dirt cheap to stick in a card or USB device or similar with its own microcontroller and memory and possibly FPGA and build a far "harder" real time system than any PC allows (even back in the DOS days).

Stealing power from the serial port and using the parallel port for "general purpose" I/O was never a particularly great idea, IMO. Yeah, it saved money and was even clever at times, but it also introduced plenty of compatibility headaches.

Think of it as a cheap Linux box.

---Joel

The 50mA I mentioned is to be taken from the FT232R low drop regulator that is fed from the USB port. My idea was to also power a small circuit without the need for a 3.3V regulator. So not what you probably are referring to as powering a circuit from the -15 to +15 (or whatever) levels of a serial port. That is not directly available. The FT232R has cmos compatible i/o for TX and RX.

Joop

Thanks for the input. I can see what you mean if this was to become a commercial product. What I am pondering on right now is some hobby gadget that maybe could be based on a cheap USB/RS232 adapter. I wondered if it could be cannibalized and also deliver a little regulated power. I already ordered some amazingly cheap stuff from Hong Kong through ebay. It will be a little surprise what is inside, but I would gamble on the prolific.

they're

For a one off project, this has help me:

Might be a poor gamble. About a year ago, I ordered several inexpensive (and cheap) USB/RS-232 adapters from a U.S.-based eBayer, whose auction pages and web site made a point of stating that they were Prolific-based and had the best software compatibility. Same vendor offered another type or two which were less expensive and had an unidentified chipset.

The adapters arrived, matched the visual description in the auction, and didn't work. They had some no-name chipset in them (I could only find one or two references to the company name on the Web, in Chinese-language pages, and no technical documentation at all), their USB endpoint descriptions were very different from Prolific's, they didn't work with the Prolific driver even when I forced it to associate, they didn't work with the Linux "generic USB serial" driver, and one of the three was so dead it wouldn't even USB-enumerate.

I RMA'ed them to the seller per instructions. It took two months and a complaint to eBay to get my money back.

Caveat emptor... I hope you do better than I did!

"Joel Koltner" ...

Another 'slight' nuisance: I used a USB-RS232 converter with the PL2303 and found that, if you program it for 2 stopbits, it also expects to _receive_ two stopbits. Its the first UART I found (in 30 years) that did not accept anything above a half stopbit as useable.

I use that 2 stopbit setting to send data from a PC through a loopbacked embedded system and receive it back. When doing BERT (Bit Error Rate Testing) the embedded system buffers may overflow if its output side baudrate generator XTAL is slower than the PC's XTAL. I solve that by setting the PC to two and the embedded system to 1 stopbit. That works with all other (PC and USB-RS232) UARTs like the FTDI ones but fails with the PL2303.

Arie de Muynck.