For over 15 years now, Microsoft has been attempting to force the phase out of classic RS-232 serial ports on PCs. (They actually coined the term "Legacy-Free PC" to refer to new designs without serial or parallel ports.)
However, many , many amateur radio applications still require RS-232 serial "COM" ports. Some of the many uses are: packet TNCs, radio memory programming, connecting GPS receivers, sound-card-interface transmit keying (PTT), rotator controls for satellite tracking, and remote control & logging applications on HF transceivers.
Virtually no new PCs available at retail for the last 5 years have serial ports. Serial ports are still available on some expensive specialized machines intended for non-consumer use, such as Panasonic Toughbooks and certain vertical-market Dells. The majority of "civilian" users have been forced to resort to the notorious and frequently problematic serial<-->USB converter cables a.k.a. "dongles" if serial ports are required.
These cables are not just pieces of wire with a different kind of plug on each end. They actually contain an active microcontroller device, . One of a half-dozen or so specialized serial/USB conversion chips, provided by one of 4 or 5 manufacturers is inside the molded DB9 plug, powered from the 5 VDC available from the USB end. In turn, this chip is dependent on a software driver installed into the operating system to function. Normally, such drivers create a virtual COM port that serial-using applications can then use like an actual physical serial port.
The quality and stability of the drivers varies wildly from chip vendor to chip vendor. Some versions of these drivers emulate a physical serial port far more faithfully than others. Some drivers will work with certain versions of Windows but not others. Or only certain applications will recognize the virtual COM port created by the driver.
Further, users may be dealing with the headaches of serial<-->USB conversion without even realizing it. So-called USB GPSs, USB interfaces on transceivers, "USB" versions of TNCs and other devices are actually serial internally. Exactly the same kind of serial<-->USB chip as in the stand-alone cables, is used to offer a "USB" version of the device. The only difference is that the serial<-->USB conversion takes place inside the device rather than in a cable between the device and the computer. Many of these so-called USB devices use exactly the same chips and drivers as the stand-alone serial<-->USB cables.
The give-away is that the applications on the PC, that use these devices, still require you to choose a COM port rather than referring to the USB device natively by name.
For example, the "USB" interface on the Kenwood TH-D72 APRS handheld is actually the serial interface of it's predecessor TH-D7 with a serial<-->USB chip added internally. The "driver" provided for the PC reconstitutes the serial COM port of it's predecessor for the benefit of APRS and packet programs that only think in terms of classic RS-232 COM ports. The ultimate irony is that the radio-specific memory programming utility for this "USB" radio requires you to choose and access a virtual COM port, instead of natively accessing the USB port.
A second example. The USB-based "GPS Locator" bundled with some versions of Microsoft Streets & Trips, MapPoint & Autoroute, the Globalsat BU-353 low-cost puck GPS, the Pharos iGPS-180, many Holux GM-series GPSs, and countless other GPS devices are all based on the Prolific PL-2303 chip. They use exactly the same driver for the Prolific PL-2303 chip that you find bundled with countless low-cost serial<-->USB dongles. The only difference is that the device description (that shows in the Windows Device Manager) has been changed from the generic "USB-serial Bridge" to "Microsoft GPS Locator" or something similar.
One of the most widely-used chips in low-cost serial<-->USB dongles and "USB" GPS devices is the Prolific Technology PL-2303. It's drivers have been released in countless versions for every flavor of Windows beginning with Windows 95/98 up to and including the latest Windows 8-64. This device has also been supported on various flavors of Linux, the Mac OS and even on Android smart phones. Prolific's English-language home page in Taiwan is here:
This device and it's associated drivers worked reasonably well until manufacturers on the Chinese mainland began counterfeiting Prolific chips several years ago, and started pirating Prolific's drivers. Prolific responded by placing a kind of DRM (Digital Rights Management) into it's drivers. The driver somehow tests the chip for "authenticity" and refuses to run if the chip is deemed a fake. This yields the dreaded yellow triangle with exclamation mark in Windows, along with "Error Code 10 - Failed to install driver".
The fake 2303 chips have managed to infiltrate the world-wide electronics-manufacturing supply chain, and randomly appear in hundreds of products from dozens of legitimate manufacturers. Many manufacturers of GPS devices and serial<-->USB dongles have taken to bundling older versions of the Prolific driver (that lack the "genuine check" routine) with their products to work around this headache. However these older drivers often won't work with newer variants of the PL-2303 chip. Since many different devices use the PL-2303 chip, you experience some version of the following scenario:
You install a device with an older version of the Prolific
driver. Every thing works just fine.
You then acquire and install another device that happens to
have a newer version of the chip in it. Windows "Plug-N-Pray" then
discovers that a driver for this chip is already on your system the first time
you connect the new device, and "automagically" enables a second copy of the
existing driver for the new device. However, you get the yellow triangle and
the dreaded "Error 10".
The new device doesn't work with the old flavor of the driver.
The user downloads and installs the "latest and greatest" version of the
driver from the Prolific website.
Now the new device works, but the older device (with the counterfeit chip in it) stops working.
Or everything worked great on your old Windows XP computer, but now nothing (or only some devices) work(s) on the new Windows 7-64 computer
Some manufacturers bundle new driver versions that have been hacked to disable the "genuine check". These work on 32-bit versions of Windows XP and Win 7. However, they fail miserably on Windows 7 & 8 64-bit editions. This is because the 64-bit editions of these OSes require drivers to be tested, certified and blessed by Microsoft with a "digital signature" before they will install. The "hacked" versions of the driver fails the digital signing validation and won't install on 64-bit systems.
Or you can go through a byzantine procedure involving registry hacks and reboots into "Safe mode" that defeats driver signing tests in Win 7-64. (There is a "Test" mode, intended for developers of new drivers, that will defeat the driver signing test and allow the installation of "unofficial" drivers for testing and debugging.)
After the "test mode" tweaks, the hacked driver works for a while.... until the ever-helpful Windows Update decides to automatically download and install the latest official Microsoft "signed" version of the PL2303 driver that DOES include the "genuine" check in it.
There are really only two ways to get around this mess.
Throw out all older Prolific-based devices and cables that
fail to run on the current DRMed verion of the driver, and replace them with
new ones that do work with the latest version.
Avoid PL-2303-based devices entirely. Other
chipsets from FTDI, Cypress Semiconductor and Silicon Labs don't have these problems.
(Silicon Labs chips are used by the built-in serial-to-USB conversions in the
Kenwood TH-D7 and some Yaesu products.)
Serial<-->USB interfaces based on the FTDI chip set do not suffer from the numerous problematic driver issues that plague the Prolific chipset. There are numerous serial<-->USB devices based on the FTDI chipset, but the one pictured below, a "Gearmo FTDI2X" is exceptionally useful for ham applications.
You get two DB9 serial ports from a single USB
connection. For APRS applications, this means one port for the TNC and one for
the GPS receiver. The two serial-side cables are 1 meter (approx 39") long,
while the USB-side cable is about .4 meter (approx 17") long.
The COM port number assigned to each DB9 "sticks" and remains
the same no matter what USB port you plug the cable into.
(Prolific-based devices randomly acquire a different COM every
time you plug the device into a different USB port, or via a hub instead of
The "goiter" in the middle of the cable is made of a smoky
translucent plastic that allows 3 LED indicators for TXD, RXD and POWER to
show through. These indicators can be quite useful for determining if
devices on the serial side(s) are actually putting out data.
The cable in the picture has two serial ports, but Gearmo offers similar cables with up to FOUR serial ports.
The installer for the FTDI driver is a single downloadable
.EXE that contains drivers for every flavor of Windows 32 or 64 bit from Win98
to Windows 8. The appropriate version is automatically extracted and installed
when you run the same setup on any flavor of Windows. I have installed this
device and it's associated drivers on two Win7-64 machines and 5 WinXP
machines of varying vintages. The longest time from starting SETUP.exe
to plugging in a functioning device was about 20 seconds. And it worked
on the FIRST TRY each time!
The FTDI driver produces a far more comprehensive screen of options when you drill down through the Device Manager "Port Settings, Advanced" dialog. Instead of the generic serial port dialog that only lets you set COM number and the size of the TX and RX buffers, you get a huge menu of options. One of these is being able to toggle "Serial Enumerator" ON or OFF.
Typical Serial Port Configuration Dialog
FTDI Expanded Serial Options
Serial enumeration means Window "Plug-N-Pray" tries to look through the serial
port to identify the device attached to the port. In the case of ports
connected to live NMEA GPS devices spewing out data once a second, this always
results in Windows falsely identifying the device as a "Microsoft Ballpoint
Mouse", an ancient clamp-on serial trackball for early laptops. The mouse
cursor then goes insane skipping all over the screen and randomly clicking on
things as it interprets the NMEA data stream as mouse movements.
See details on this headache here. Disabling serial enumeration completely stops this problem.
The FTDI chipset is capable of operating at serial speeds in
excess of 900K/second (!), and will operate with 5, 6,7 or 8 data
bits/character. The 5-bit mode, which is not supported by most dongles,
makes it usable for classic 5-bit/char Baudot RTTY operation.
The device and it's drivers are stable and rock-solid. I road-tested it for hundreds of miles on the way to and from the 2013 Dayton Hamvention, using both my 32-bit WinXP Panasonic Toughbook and my new Win7-64 Acer 756 netbook (review here), with absolutely no problems.
This product is made by "Gearmo". Their website is here:
Note that they offer a variety of FTDI-based dongles in versions
with 1, 2 or 4 serial ports from one USB port. They sell through the website.
I got mine for USD $10 less from this Amazon page:
Update 22 May 2013: Apparently the writeup on
this page has already caused a sellout of the limited supply of the Gearmo
cables at the cheap price this vendor had! Hope that they restock.
A version with a single serial port on a short (less than 1 foot) pigtail is here:
Googling for "ftdi usb to serial" yields quite a few other dongles based on this superior chipset, which should have about the same capabilities.
Another (single port) Gearmo from Amazon