WTB: Spectrum Analyzer

[dennismiller55]

I am looking for a suitable spectrum analyzer for audio work.

What I have researched so far that has intrigued me is the HP 3582A, HP 3580A and a HP 3562A. I think all of these are suitable for audio and amplifier work. Right now my preference would be the HP 3562A.

Anyone have any advice?

Anyone have a spectrum analyzer for sale?

Thanks

 

[dennismiller55]

Another HP Spectrum Analyzer I just looked at on the "Auction Site" that looks like a good choice is the HP 3585A.

Anyone familiar with this unit?

Anyone have one they would like to sell?

 

[Pano]

I've seen these in use. Very good, but sloooow if you want narrow bandwidth.
How high do you need to go?

 

[dennismiller55]

Thanks Pano

For audio work I would think this 40 Khz spec on the HP 3585 would be sufficient. The HP 3561 goes up to 100khz. I have not used a spectrum analyzer in 30 plus years so this is a re-leaning experience. I am hoping someone here can comment on recommendations based on experience with their use.

I am also hoping someone has a good working unit that no longer needs or is closing up shop. It scare me to purchase something in this price range from the "auction site".

 

[PChi]

I have seen older equipment in use which I believe was an HP 3582 which I think is a conventional analogue spectrum analyzer architecture.
I have used a Hewlett Packard Dynamic Signal Analyzer a long time ago which I think was the 3561. Strictly speaking it isn't a Spectrum Analyzer in the conventional sense with a mixer and bandpass filter but uses A/D conversion and FFT to display level versus frequency. The dynamic range isn't that great because I believe it uses 16 bit A/D conversion. More than adequate for mobile telephones and acoustic testing but limited compared to a PC Sound Card for measuring low distortion, low noise electronics. Possibly OK for high distortion valve work though. The user interface can be confusing. It managed to fox a colleague who couldn't find the AC / DC coupling option. I also spent some time looking through the documentation to understand the equivalent measurement bandwidth of the frequency 'bins' until I found the nV / root Hz display option.
As Pano says they can take their time to calculate an answer if you have many frequency bins so narrow bandwidth.

My summary is good for low end electronics and for acoustics. Inadequate for low distortion electronics. I guess will struggle below 0.01 % distortion. I also vaguely remember that the input stage was noisy and I had to add a pre amplifier for some low level measurements.

 

[Pano]

If you can live with 40Khz, then get yourself an M-Audio Audiophile 192 card. You'll get at least 80Khz, it will be as fast as the PC you put it in, and it's super easy to save plots.
There is plenty of free software that does spectrum analysis, but it's probably worth investing in ARTA if you plan to use it a lot.

You'll have 2 channels of very good FFT and other measurements. Most guys who use this as a lab setup build a front end that allows higher voltages in, overvoltage protection, calibration and the like. Not hard to build.

It's amazing what $100 will buy you these days.

 

[dennismiller55]

If you can live with 40Khz, then get yourself an M-Audio Audiophile 192 card. You'll get at least 80Khz, it will be as fast as the PC you put it in, and it's super easy to save plots.
There is plenty of free software that does spectrum analysis, but it's probably worth investing in ARTA if you plan to use it a lot.

You'll have 2 channels of very good FFT and other measurements. Most guys who use this as a lab setup build a front end that allows higher voltages in, overvoltage protection, calibration and the like. Not hard to build.

I have considered this option. I guess I am just a analog guy. I like to push buttons.

Really, I have been looking at my options for using a software based system. I think using a "front end" for a safer and flexable input is smart. There is a member here on this site that has designed one. It looks like a lot of thought went into the design. And I do like the DIY aspect. The link is here: "Test & Measurement Interface for Soundcard".

I was also thinking of buying a really good PC soundcard. Unfortunetly, I do not have a desktop anymore. Just a laptop. But for the cost of a good used desktop PC, a $200 soundcard, and just under $200 to build the front end would be a good deal as compared to the cost of a guaranteed working HP Spectrum Analyzer. I could try one of the free FFT softwares available to start with.

Thanks for your input. It will help me make a wise choice.

 

[iko]

Another HP Spectrum Analyzer I just looked at on the "Auction Site" that looks like a good choice is the HP 3585A.

Anyone familiar with this unit?

Anyone have one they would like to sell?

I have a 3585A and it's a pleasure to work with. I really hate working with the sound card, even though it's capable of measuring lower distortion. The convenience of pressing a few buttons and to get a distortion plot is amazing. You have to accept that you don't get a thd number out of it, but you can measure H2 and H3 and so on. For this unit, in order to be able to measure very low distortion you'd have to build a very good notch filter. However, for tube amps...

Another unit to have a look at is the 3561A, which goes to 100kHz.

 

[iko]

If you go the sound card route, you may want to consider the TI PCM4222EVM ($150). For low thd measurement, see what it can do:

http://www.diyaudio.com/forums/equi...ject-audio-measurements-tool.html#post2072409

I got one of those myself; didn't get such good results as ghg, but not bad either. The truth is that I don't reach for it because unless I will build a nice interface, it's cumbersome to use.

http://www.diyaudio.com/forums/solid-state/1565-diy-distortion-analyzer-analysis-14.html#post2664662

 

[boywonder]

I was also thinking of buying a really good PC soundcard. Unfortunetly, I do not have a desktop anymore. Just a laptop.

I imagine that there are several excellent choices for USB based external sound cards out there.

I'm using an M-Audio Audiophile external card (it's Firewire though, and 24/96) and Rightmark Audio Analyzer (freeware) with a simple home-brew attenuator/load resistor for the amp under test.

The external cards have an added advantage of less noise/better shielding than an internal card, and some use external power supplies.

 

[Pano]

Thanks for your input. It will help me make a wise choice.

You're welcome! Both approaches have their advantages, so pick what works for you.
I was in Gary Pimm's lab and we were running his soundcard, interface and ARTA next to the big HP. For measurements above 80Khz, the HP was used. But under that we relied on the soundcard (with front end). It was soooooo much easier, faster and the results were easier to interpret. At least for me.

I also do not have a desktop computer in my lab, I use a laptop with a USB card, so am limited to 96 Khz sampling rate. I usually build whatever interface I need for the signal and device under test.

The link is here: "Test & Measurement Interface for Soundcard".

Yep, that's Pete Millett's interface. Very good indeed. Gary Pimm and SY have built their own based on Pete's. All three of these guys build very clean, low noise electronics.

 

[dennismiller55]

Yep, that's Pete Millett's interface. Very good indeed. Gary Pimm and SY have built their own based on Pete's. All three of these guys build very clean, low noise electronics.

I have read Pete very long thread on the his interface. I have not found anything on " Gary Pimm" or "SY". Can you point me to their post on their interface project if available?

Thanks,

Dennis

 

[dennismiller55]

Thanks everyone.

I read everyones response to my OP question and I have decided to try "Pete Millett's interface" with a good usb external audio box (yet to be determined) along with one of the audio analysis software packages.

I ordered the pcb to build the sound card interface today. I look forward to the build.

Again thanks.

Dennis