Hey all,

Anyone know of a portable USB/Firewire/PCMCIA-based signal generator
that is supported under Mac OS X? I'm looksing for something that has a
frequency response down to at least 5 Hz.

Herb

--
Herb Singleton
usenet[at] cross-spectrum.com
Sound & Vibration Measurements
http://www.cross-spectrum.com

In article <oster-EE002C.18040528082003[at]news.sf.sbcglobal.net>,
David Phillip Oster <oster[at]ieee.org> wrote:

> Some ideas:
>
> 1.) It is a shame you can't just use the audio out, but that probably
> filters out anything below 20 hz or so (although, you might try it and
> see.)

In my experience, you can generally get down to 10 Hz or so (usually -5
dB), but since the internal audio is usually rated at 20 Hz, it's not
practical for production work.

> 2.) National Instruments makes a PCMCIA card, the 6024e that may do what
> you want, and they also have it in a PCI card version.

I saw that - it's a nice card, but unfortunately NI doesn't support it
under Mac OS X.

> 3.) A USB -> Parallel Port converter will give you a byte of data and a
> strobe. You could use that as the input of a D to A chip. Open the USB
> as a serial interface port "/dev/tty???" and just write a byte stream to
> it using a Time Manager timer interrupt.

That's interesting... I may look into doing something like that if all
else fails.

Also it seems that the Nuendo AudioLink 96 Mobile & AudioLink 96
Multiset combo may fit the bill.

Herb


--
Herb Singleton
usenet[at] cross-spectrum.com
Sound & Vibration Measurements
http://www.cross-spectrum.com

In article <usenet-BC6161.00483229082003[at]news.comcast.giganews.com>,
Herb Singleton <usenet[at]cross-spectrum.com> wrote:

> In article <oster-EE002C.18040528082003[at]news.sf.sbcglobal.net>,
> David Phillip Oster <oster[at]ieee.org> wrote:
> > 2.) National Instruments makes a PCMCIA card, the 6024e that may do what
> > you want, and they also have it in a PCI card version.
>
> I saw that - it's a nice card, but unfortunately NI doesn't support it
> under Mac OS X.

<http://www.ni.com/mac/mac_daq.htm> says that NI supports its DDK under
OS X, and <http://www.ni.com/pdf/products/us/mhw_ddk.pdf> says the DDK
works with E-series boards. That looks like support to me.

In article <oster-9D4263.22193428082003[at]news.sf.sbcglobal.net>,
David Phillip Oster <oster[at]ieee.org> wrote:

> In article <usenet-BC6161.00483229082003[at]news.comcast.giganews.com>,
> Herb Singleton <usenet[at]cross-spectrum.com> wrote:
>
> > In article <oster-EE002C.18040528082003[at]news.sf.sbcglobal.net>,
> > David Phillip Oster <oster[at]ieee.org> wrote:
> > > 2.) National Instruments makes a PCMCIA card, the 6024e that may do what
> > > you want, and they also have it in a PCI card version.
> >
> > I saw that - it's a nice card, but unfortunately NI doesn't support it
> > under Mac OS X.
>
> <http://www.ni.com/mac/mac_daq.htm> says that NI supports its DDK under
> OS X, and <http://www.ni.com/pdf/products/us/mhw_ddk.pdf> says the DDK
> works with E-series boards. That looks like support to me.

DDK = Driver Development Kit.

NI does not currently offer a Mac OS X driver package for their cards.
The DDK provides documentation for customers who want to try writing
their own drivers. I looked into this and discovered that it is a
non-trivial project, even if you have some experience in driver level
programming.

Also note that the support provided by NI for the DDK is very limited.
From the above web pages:

"Support is available only through online newsgroups, downloads, and
register-level programming manuals. Standard phone and email support is
not available for the Measurement Hardware DDK."

By "online newsgroups" they mean usenet groups such as this one, and the
user forum on the NI website.

I called NI a few months ago to ask about the current state of their
support for Macs. Their website seems to imply that the cards and
software are supported for Macs. But when you read the fine print, you
discover that most of the support is still limited to OS 9 and earlier.
LabView for OS X only talks to a small subset of their cards (the GPIB
and ethernet stuff, I think). The DDK is probably intended for
enterprise customers who have the manpower and expertise to do their own
driver development (and maybe the cash to buy partnering support from
NI). The NI guy I spoke with was pretty candid. He said they are still
taking a "wait-and-see" approach on Mac OS X, watching the market.

For the project I am currently doing, we decided to go with a data
acquisition interface from InstruTECH, their model ITC-1600. This is a
nice piece of hardware, with a standalone rackmount box for the
converters, a DSP card that installs on a desktop Mac's PCI bus, and a
pair of optical cables to link the card and box. Very low noise, with
sampling rates up to 200k. Analog outputs are DC coupled, so no problem
with low frequencies. But it's not cheap, and not very portable, so it
might not be useful for Herb's application. InstruTECH does provide a
driver package for OS X, although this is still something of a work in
progress, with spotty documentation.

--
Julian Vrieslander

In article
<iintbeumtnbfkcpntohw__NOSP-6B8A0C.12000829082003[at]news1.news.xs4all.nl>,
Maarten Sneep <iintbeumtnbfkcpntohw__NOSP[at]M__xs4all.nl> wrote:

>Worse still: Only GPIB _over_ Ethernet, normal GPIB-PCI cards are not
>even supported. (Which is what most people in a lab are using I think).

I checked this recently too. NI are only "supporting" the Ethernet product
on OS X (at double the cost) precisely because they do not want to debug
hardware cards on a system with no proper real time debugging tools.

NI also claim that the PCI-GPIB card does not work past OS 9.1 although I
have one running under 9.2 - which shows the depth of their Mac knowledge.

Which begs the question of why use GPIB at all, it is an extremely
expensive solution. What type of signal generator is required? I could
build an audio one of similar spec to an HP 8903 with RS232/422 or MIDI
control. Give me a spec and I'll give you a price.

Graham Hinton wrote:
> Maarten Sneep wrote:
>> Worse still: Only GPIB _over_ Ethernet, normal GPIB-PCI cards are not
>> even supported. (Which is what most people in a lab are using I
>> think).
>
> I checked this recently too. NI are only "supporting" the Ethernet product
> on OS X (at double the cost) precisely because they do not want to debug
> hardware cards on a system with no proper real time debugging tools.
>
> NI also claim that the PCI-GPIB card does not work past OS 9.1 although I
> have one running under 9.2 - which shows the depth of their Mac knowledge.

Which is a shame really, given the history of labview, and the fact that
LabVIEW is already running on three unix systems (GNU/Linux, Solaris and
HP-UX) all with drivers for PGIB-PCI hardware.

Of course the IOKit is a different beast, but other companies are doing
fine in this respect.

> Which begs the question of why use GPIB at all, it is an extremely
> expensive solution. What type of signal generator is required? I could
> build an audio one of similar spec to an HP 8903 with RS232/422 or MIDI
> control. Give me a spec and I'll give you a price.

GPIB is expensive, and the fact that NI has cornered the market
certainly doesn't help. Agilent (or the company formerly known as HP)
has some GPIB cards, but they don't support the mac either.

OTOH, every time I have to work with a RS232/422 based instrument, I
remember why I put up with GPIB.

Maarten

In article <BB763470966830EFB[at]0.0.0.0>,
hitman[at]spammer.deathsquad.com (Graham Hinton) wrote:

> Which begs the question of why use GPIB at all, it is an extremely
> expensive solution. What type of signal generator is required? I could
> build an audio one of similar spec to an HP 8903 with RS232/422 or MIDI
> control. Give me a spec and I'll give you a price.


In my case, I suppose I don't really need a full-blown signal generator,
I just need an audio output with reliable output down to at least 5 Hz.

Herb

--
Herb Singleton
usenet[at] cross-spectrum.com
Sound & Vibration Measurements
http://www.cross-spectrum.com

In article
<iintbeumtnbfkcpntohw__NOSP-253591.13315130082003[at]news1.news.xs4all.nl>,
Maarten Sneep <iintbeumtnbfkcpntohw__NOSP[at]M__xs4all.nl> wrote:

> Graham Hinton wrote:
> > Maarten Sneep wrote:
> >> Worse still: Only GPIB _over_ Ethernet, normal GPIB-PCI cards are not
> >> even supported. (Which is what most people in a lab are using I
> >> think).
> >
> > I checked this recently too. NI are only "supporting" the Ethernet product
> > on OS X (at double the cost) precisely because they do not want to debug
> > hardware cards on a system with no proper real time debugging tools.
> >
> > NI also claim that the PCI-GPIB card does not work past OS 9.1 although I
> > have one running under 9.2 - which shows the depth of their Mac knowledge.
>
> Which is a shame really, given the history of labview, and the fact that
> LabVIEW is already running on three unix systems (GNU/Linux, Solaris and
> HP-UX) all with drivers for PGIB-PCI hardware.
>
> Of course the IOKit is a different beast, but other companies are doing
> fine in this respect.
>
> > Which begs the question of why use GPIB at all, it is an extremely
> > expensive solution. What type of signal generator is required? I could
> > build an audio one of similar spec to an HP 8903 with RS232/422 or MIDI
> > control. Give me a spec and I'll give you a price.
>
> GPIB is expensive, and the fact that NI has cornered the market
> certainly doesn't help. Agilent (or the company formerly known as HP)
> has some GPIB cards, but they don't support the mac either.

Which is ironic when you think about it, since GPIB arose from HPIB (at
least, as far as I understand it.) So, do Agilent now refer to their
IEEE 488(.2) products as "AIB"s now?

In article <usenet-CB856E.11172930082003[at]news.comcast.giganews.com>,
Herb Singleton <usenet[at]cross-spectrum.com> wrote:

>In my case, I suppose I don't really need a full-blown signal generator,
>I just need an audio output with reliable output down to at least 5 Hz.

I was not proposing a full 8903, which is more analayser than generator,
but if you don't say the frequency range & resolution, output range,
resolution & drive capability or waveforms & distortion required it makes
it difficult to help you.

You might find that a cheap MIDI sound module could generate what you
require.

In article
<iintbeumtnbfkcpntohw__NOSP-253591.13315130082003[at]news1.news.xs4all.nl>,
Maarten Sneep <iintbeumtnbfkcpntohw__NOSP[at]M__xs4all.nl> wrote:

>OTOH, every time I have to work with a RS232/422 based instrument, I
>remember why I put up with GPIB.

When HPIB first came out you were lucky to get a serial port on a computer
that worked reliably at more than 1200Bd unless you made all the hardware
and software yourself. Now 115kBd is not uncommon and more than sufficient
for most tasks.

I've been making RS232 controlled equipment for more than 20 years with
fast protocols for computer control and VT100 paged support for terminal
control. GPIB always strikes me as pretty clumsy. I used HPIB to prototype
a LAN in 1979 whilst developing a coax solution, but I've never seen any
instruments that push it to those limits. There may be some, but most bench
test gear is slow and cluggy.