HP3577A: Does it include Cross-Spectrum Analysis?

[Angelo Campanella]

Just returned from ASA/Baltimore, where I was advised that the way to go to
200 kHz and higher in analysis is to use an HP(Agilent) 3577A analyzer.

My preliminary observations and questions:

1- From its specs (.05% full width bins) it looks like it can make 100 Hz
bins over 200 kHz for instance, which is OK for me.

2- It "has" 3 channels, A, B and R; I don't know what R means.

3- Does it do Cross-Spectrum Analysis as for Intensity and, say, impedance
tube/two-microphone techniques.

4- There is comment that it can be controlled by Basic commands, through an
HP interface... How about RS232?

4- It seems to be a blunderbuss; 67#, shipping 90#.

Any comments will be very welcome.

Thanks,

Angelo Campanella.

Post by Angelo Campanella
Just returned from ASA/Baltimore, where I was advised that the way to
go to 200 kHz and higher in analysis is to use an HP(Agilent) 3577A
analyzer.
1- From its specs (.05% full width bins) it looks like it can make 100
Hz bins over 200 kHz for instance, which is OK for me.
2- It "has" 3 channels, A, B and R; I don't know what R means.
3- Does it do Cross-Spectrum Analysis as for Intensity and, say,
impedance tube/two-microphone techniques.
4- There is comment that it can be controlled by Basic commands,
through an HP interface... How about RS232?
4- It seems to be a blunderbuss; 67#, shipping 90#.
Any comments will be very welcome.
Thanks,
Angelo Campanella.

I have one in my lab and another in my store room as a backup/spare.
It's a two-channel swept-frequency (hetrodyne) network analyzer. Since
it's not an FFT analyzer, it doesn't compute cross spectrum.
Nonetheless, being a network analyzer, it does measure and display the
phase difference between the signals in channels A and B. However, it
does not provide a means for compensating for the time delay (rapid
linear phase shift) that can occur in acoustic measurements. On the
other hand, the HP 3585A, another hetrodyne analyzer, does provide the
means to offset the frequency tracking filter and to perform a TDS
measurement. Unfortunately, the 3585A is a single channel instrument
that only computes and displays frequency response magnitude.

[Angelo Campanella]

Post by
I have one in my lab and another in my store room as a backup/spare.
It's a two-channel swept-frequency (hetrodyne) network analyzer. Since
it's not an FFT analyzer, it doesn't compute cross spectrum.

So much for that.

Post by
Nonetheless, being a network analyzer, it does measure and display the
phase difference between the signals in channels A and B. However, it
does not provide a means for compensating for the time delay (rapid
linear phase shift) that can occur in acoustic measurements. On the
other hand, the HP 3585A, another hetrodyne analyzer, does provide the
means to offset the frequency tracking filter and to perform a TDS
measurement. Unfortunately, the 3585A is a single channel instrument
that only computes and displays frequency response magnitude.

As Scott said, the only economical way is to use a National Instrument
Card..

Thanks for the manual... that settles many issues..

Sincerely,

Ange

[Scott Dorsey]

Post by Angelo Campanella
Just returned from ASA/Baltimore, where I was advised that the way to go to
200 kHz and higher in analysis is to use an HP(Agilent) 3577A analyzer.

I went looking for you there and I didn't see you! I was only up on
Wednesday so I missed all the interesting room acoustics papers on Thursday.
And the ASA, unlike the AES, doesn't require preprints for the proceedings,
so I can't even read them.

Post by Angelo Campanella
1- From its specs (.05% full width bins) it looks like it can make 100 Hz
bins over 200 kHz for instance, which is OK for me.

It's not really an FFT analyzer, but a swept-filter analyzer.

Post by Angelo Campanella
2- It "has" 3 channels, A, B and R; I don't know what R means.
3- Does it do Cross-Spectrum Analysis as for Intensity and, say, impedance
tube/two-microphone techniques.

No.

Post by Angelo Campanella
4- There is comment that it can be controlled by Basic commands, through an
HP interface... How about RS232?

I don't think it has RS-232, but it has HP-IB (IEEE-488) and you can buy
a conversion box.

Post by Angelo Campanella
4- It seems to be a blunderbuss; 67#, shipping 90#.
Any comments will be very welcome.

I think the cheap way to do what you want will probably involve using a
general purpose Windows machine. I am pretty sure the Rockland FFT machines
don't go up that high, and I know the GenRad doesn't. This means you're left
buying something like the newer Agilent or Ono Sokki machines which are not
cheap. A PC with a National Instruments card and Matlab is much cheaper.
--scott

[Angelo Campanella]

Post by Scott Dorsey
I went looking for you there and I didn't see you! I was only up on
Wednesday

I was not always in the paper sessions on Wednesday, being sidetracked by a
committee meeting that lasted from 11 to 3, then in the plenary.

Post by Scott Dorsey
so I missed all the interesting room acoustics papers on Thursday.
And the ASA, unlike the AES, doesn't require preprints for the
proceedings,
so I can't even read them.

Look at the ASA http://asa.aip.org/ web sites, "Baltimore Meeting" link.
There may be access to some of them (can't promise much).

Post by Scott Dorsey
It's not really an FFT analyzer, but a swept-filter analyzer.

Gary was nice enough to sent me a PDF of it. After browsing through
that, I surmised that it was just a swept frequency RF receiver.... The
fellow who suggested it was right only in tha it will cover the desired
frequency range... But I would like something better than that. Found
several on E-bay... for a few thousand dollars, basically.

Post by Scott Dorsey
I think the cheap way to do what you want will probably involve using a
general purpose Windows machine. I am pretty sure the Rockland FFT machines
don't go up that high, and I know the GenRad doesn't. This means you're left
buying something like the newer Agilent or Ono Sokki machines which are not
cheap. A PC with a National Instruments card and Matlab is much cheaper.

I have recently checked the NI cards... It looked like seveal thousamnd
dollars is required, plus interfacing it to a PC. That makes me hesitate
still until I have a more firm grip on it all. The ultrasound work I did
ages and ages ago (JASA, Jan, 1980) was done with a hand tuned heterodyne
analyzer. The 3577A is an equivalent to that. Pulling the data out of it is
another challenge.

I'll keep looking...

Thanks for your interest,,

Sincerely,

Ange.

Post by Angelo Campanella
Just returned from ASA/Baltimore, where I was advised that the way to
go to 200 kHz and higher in analysis is to use an HP(Agilent) 3577A
analyzer.
1- From its specs (.05% full width bins) it looks like it can make 100
Hz bins over 200 kHz for instance, which is OK for me.
2- It "has" 3 channels, A, B and R; I don't know what R means.
3- Does it do Cross-Spectrum Analysis as for Intensity and, say,
impedance tube/two-microphone techniques.
4- There is comment that it can be controlled by Basic commands,
through an HP interface... How about RS232?
4- It seems to be a blunderbuss; 67#, shipping 90#.
Any comments will be very welcome.
Thanks,
Angelo Campanella.

If you are looking for an FFT-based analysis option, one possibility
might be an Analogic/DataPrecision 6100 waveform analyzer. It is
essentially a digitizing oscilloscope with numerous built-in signal-
processing capabilities such as FFT, correlation, convolution, etc.
However, its basic function is that of a digitizing oscilloscope. And,
if you've ever been there, you can appreciate the fact that forcing an
oscilloscope to behave like an FFT analyzer is about as
convenient/frustrating as forcing an FFT analyzer to behave like an
oscilloscope. With appropriate plug-ins for the 6100, you can get
anywhere between 50M samples/sec with 8 bit resolution to 1M samples/sec
with 16 bit resolution. The resolution and sample rate are determined
solely by the plug-in. All the mainframe does is crunch the resulting
number array and display the result. The downside is that the 6100
doesn't have a button that you can press to get either the cross spectrum
or the phase response. In order to get the latter, you need to program
the 6100 via front-panel push keys, a sequence of steps such as 1)
simultaneously capture signals in channels A&B, 2) perform FFT on signal
in the channel A buffer, 3) perform fft on signal in the channel B
buffer, 4) calculate phase difference, 5) repeat 1-4 and perform linear
average. Depending on the FFT size and available memory, it may be
necessary to make two measurements and program two calculations in order
to get both frequency response magnitude and frequency response phase.
Also, the 6100 is a very old instrument and is slow in performing the
calculations. It is, however, both GPIB and RS-232 programmable.