Does anyone have experience and/or recommendations
for analogue and digital I/O interfaces based on the Universal
Serial Bus (USB)?

Just an idea for the genesis of a homebrew spectrum analyser.

Does anyone have experience and/or recommendations
for analogue and digital I/O interfaces based on the Universal
Serial Bus (USB)?

Just an idea for the genesis of a homebrew spectrum analyser.



I can't give any specific advice about the USB interface, but I recall that
there was a PC based spectrum analyser published in QEX within the last
couple of years.

If you don't already, I would recommend subscribing to QEX, it's published
quarterly by the ARRL. It contains a lot of leading edge technology. A
Google search will give you more details.

On Jul 28, 3:59 am, "G \"Guglielmo\" Evans G4SDW"
wrote:
Does anyone have experience and/or recommendations
for analogue and digital I/O interfaces based on the Universal
Serial Bus (USB)?

Just an idea for the genesis of a homebrew spectrum analyser.

Ive done a little with USB, so I'll add a little background if not
provide complete answers.

The first hurdle is the interfacing. You need a board with a driver
thats easy to interface with your host PC software program.
Ive used a board from www.mirrorbow.com for general control...you can
access it every 1mS and get 12 x 12bit ADC samples for each access, or
you can alter 8 bit digital ports etc. However, the speed is limited
to how fast the operating system can access the board, which is 1mS.
This board pretends to be a com port so it makes programming easier
without needing to include libraries and dlls etc.
Theres a board from National Instruments which I think is faster,
though only around 2.8Msamples/sec http://sine.ni.com/nips/cds/view/p/lang/en/nid/202596
And of course there are some dedicated USB spectrum analyser
available, but thats not the same as doing a home version.
So, you could just get a DAC board and go for it, though the speed you
can make a spectrum analyser is limited.

Of course you could use the superhet principle with a narrow band
filter and a programmable LO. You could then use a simple IO board
like the one from Mirrorbow to setup the LO, then use a peak detector
and the boards AD to give an indication of level. You'd then have a
digitally controlled analogue spectrum analyser, and you woudlnt need
the large sample rate.

mirrorbow is not a bona-fide operation and does not advertise its
address, directors and other information as now required
by UK law.

"bigorangebus" wrote in message
ups.com...

The first hurdle is the interfacing. You need a board with a driver
thats easy to interface with your host PC software program.
Ive used a board from www.mirrorbow.com for general control.

The jobless bar steward wrote:

mirrorbow is not a bona-fide operation and does not advertise its
address, directors and other information as now required
by UK law.

Well, why don't you e-mail them and tell them what you read in your £350 law
book. There are two e-mail addresses on their website. I'm sure they'd be
delighted to hear from you as much as we are.

Better still, GET A JOB.

"John" wrote in message ...

Does anyone have experience and/or recommendations
for analogue and digital I/O interfaces based on the Universal
Serial Bus (USB)?

I can't give any specific advice about the USB interface, but I recall
that
there was a PC based spectrum analyser published in QEX within the last
couple of years.

The state of USB has improved pretty dramatically in the past couple of
years. The QEX article was a groundbreaker at the time, but the technology
has moved on.

Presumably you are aware that a USB interface implies computation on both
sides. The relatively recent USB PICs (18F2550, 4550, etc.) incorporate
everything needed for the USB interface except the connector. Only a couple
of years ago a USB controller chip was $26 and it needed a host of
additional parts to actually implement the interface. A smaller USB PIC
costs less than $5 and includes *everything*.

Go to http://ww1.microchip.com/downloads/e...Doc/39632D.pdf and take a
look at the capabilities. To see it in an actual circuit, try
http://www.microchip.com/stellent/id...ame =en529760
and download the users guide. The 2550 is the little TQFP to the right of
the board picture, although it is available in a variety of packages, some
more hobbyist-friendly then the tiny one on the board. The only additional
part is an inductor to filter the USB voltage when it is used to power the
target circuit.

This series of parts also includes A/D channels, tho they might not be the
speed and resolution you want. The particular board is a nice one for
experimenting because it includes one of the fast 16-bit parts that are
easily programmable in a high level language, along with the USB interface.

...

"xpyttl" wrote

Presumably you are aware that a USB interface implies computation on both
sides. The relatively recent USB PICs (18F2550, 4550, etc.) incorporate
everything needed for the USB interface except the connector. Only a
couple
of years ago a USB controller chip was $26 and it needed a host of
additional parts to actually implement the interface. A smaller USB PIC
costs less than $5 and includes *everything*.

Go to http://ww1.microchip.com/downloads/e...Doc/39632D.pdf and take
a
look at the capabilities. To see it in an actual circuit, try
http://www.microchip.com/stellent/id...ame =en529760
and download the users guide. The 2550 is the little TQFP to the right of
the board picture, although it is available in a variety of packages, some
more hobbyist-friendly then the tiny one on the board. The only
additional part is an inductor to filter the USB voltage when it is used
to power the target circuit.

This series of parts also includes A/D channels, tho they might not be the
speed and resolution you want. The particular board is a nice one for
experimenting because it includes one of the fast 16-bit parts that are
easily programmable in a high level language, along with the USB
interface.



Many thnaks for the heads-up on the PICs. I've always tried to avoid USB
interfaces for control applications, but I have a project on the horizon
where I will almost certainly go down that route - simply because laptops no
longer have RS-232 ports..

"John" wrote in message ...

Many thnaks for the heads-up on the PICs. I've always tried to avoid USB
interfaces for control applications, but I have a project on the horizon
where I will almost certainly go down that route - simply because laptops
no longer have RS-232 ports..

If this is a one-off project, you might even consider that particular board,
depending on your requirements. The USB is used as a peripheral to the 28
pin part, so that from the bigger part you can simply do a printf() to
communicate to your PC over serial. It relies on a Microcoft provided
serial emulation driver. If your requirements are more demaiding, then you
will need to really hit the books. Doing the USB software interface isn't
real trivial.

...

"xpyttl" wrote in message
news
pin part, so that from the bigger part you can simply do a printf() to
communicate to your PC over serial. It relies on a Microcoft provided
serial emulation driver. If your requirements are more demaiding, then
you will need to really hit the books. Doing the USB software interface
isn't real trivial.

mmmm ... I can see I should have been more explicit there, and maybe caught
a couple of typos.

There are a number of USB classes with different capabilities. There is a
serial emulation class, which Microsoft provides a driver for. It appears
on Windows as a new serial port, and applications access it as if it were a
serial port.

All USB devices must offer up a manufacturer code and a device code whenever
they are first connected. The OS reads a .inf file to connect the
particular codes to a driver. MicroCHIP provides an .inf file with the
board that points to a MicroSOFT driver which does serial emulation.

When the 16-bit part wiggles it's USART, the PIC18F2550 catches that and
sends it out over USB to the PC in serial emulation mode. The application
on the 16 bit part must set up the baud rate for the USART, but beyond that,
normal C I/O can be used, although it is pretty fat for many applications.

The documentation for the board says in comes with a dsPIC33FJ12GP202 but
mine came with a PIC24FJ64GA002 -- more memory but no DSP engine. The board
has jumpers to accept any 28 pin, 16 bit part. I've been doing most of my
experimentation with a dsPIC30F012. The 30F parts are 5 volts, the others
3.3, and the 30Fs have a slightly different pinout. Jumpers deal with both,
as well as deciding whether the board wants to be powered from the USB or a
separate supply.

For a lower cost implementation, the 18F2550 and its cousins have plenty of
gas to do more work, and Microchip does provide sample code for the USB
serial interface, but I've always been a bit suspicious of Microchip sample
code ... often it isn't quite as solid as one would hope.

...

"John" wrote in message ...
Many thnaks for the heads-up on the PICs. I've always tried to avoid USB
interfaces for control applications, but I have a project on the horizon
where I will almost certainly go down that route - simply because laptops no
longer have RS-232 ports..

Many products still use RS-232 (albeit at TTL voltage levels) internally and
just use, e.g., an FTDI USB--RS-232 converter chip (such as the FT2232R).
This is actually easier than using a USB microcontroller, although it does
tend to end up being a little more expensive as well (i.e., $5 for the FTDI
IC, $5 for a microcontroller vs. $5 for a USB controller).