Selecting Audio Test Equipment
I've always been interested in test equipment and have built a fair amount of it myself. But these days with surplus equipment readily available on various forums and auction sites, it makes little (financial) sense to build your own gear. But what to buy...
Few people realize it, but an external USB or FireWire sound card is actually quite useful for measurements. However, it does have one significant limitation; its input voltage range is usually fixed at a pretty high value (> 1 Vrms). In addition, it's not very well protected against accidental over-voltage, etc. So if one is to use a sound card as test equipment, I would recommend building a sound card interface such as the one Pete Millett offers.
Personally, I prefer stand-alone test equipment. This for a number of reasons:
- Better input sensitivity
- Lower noise floor
- Better over-voltage protection
- Ease of use
- I don't have to fight with USB cables that are always too short and wall warts that are always in the way.
But those are my preferences. Feel free to disagree.
HP 3581A Wave Analyzer
Once equipment such as variable power supplies, multimeters, oscilloscopes have been acquired, the question is often, "what's next?" Well, if you're interested in audio, you'd probably want to look at the harmonic distortion of your amplifiers. For that you need a precision signal source -- one with very low THD. And you need some sort of distortion analyzer or spectrum analyzer. It doesn't have to be fancy. Something as simple as an HP 3581A Wave Analyzer can be used for distortion measurements. You would have to tune to each harmonic manually, measure their amplitudes, and calculate the THD manually. That's not hard, but it's not exactly convenient either.
The 3581A is a frequency selective voltmeter. Its input stage is actually really good with a nice, low noise floor. It's on par with the HP 8903 and 3562A described below. With an X-Y plotter, it can be used as an inexpensive spectrum analyzer as its source can be swept automatically across frequency. With an A/D converter connected on its X and Y axis outputs, it would even be possible to import the resulting output into a computer for further data analysis. The default sweep is linear, but circuits exist that can convert it to a logarithmic sweep if need be.
Another use for this instrument is to hunt down mains hum. The 50/60 Hz hum along with it's harmonics can be tough to trace. But with this instrument, it's possible to measure the magnitude of the hum, thus, determine if any modifications have made an improvement.
HP 8903 Audio Analyzer
The next significant step up is the HP 8903 Audio Analyzer. The 8903 comes in different 'flavours' (A, B, and E) with a couple of options. I describe the differences between them on the HP 8903 Page. The image below shows an HP 8903A.
The 8903 (A or B) provides a precision signal source and a distortion analyzer in one package. It also functions as a precision AC voltmeter, hence, can be used to measure amplitude response (gain vs frequency).
HP 3562A Dynamic Signal Analyzer
So now that you know the amplitude response and total THD of your amplifier, the next question -- at least if you're building with vacuum tubes -- is usually, "well, is it mostly even-order or odd-order harmonics?" For this you need a spectrum analyzer (or the aforementioned HP 3581A wave analyzer if you're on a budget). The HP 3562A Dynamic Signal Analyzer shown below is just about perfect for this.
The HP 3562A does a lot of stuff. It's made for use on mechanical systems and audio systems, hence, offers a measurement range of 64 µHz to 100 kHz. It has a DDS-based signal source with very low THD. Its two inputs can be plotted measured and plotted separately, or one can be uses for a reference input while the other is measured for frequency response measurements. Its −144 dBV/√Hz noise floor also makes it a very handy instrument for measuring the noise floor of an amplifier. Furthermore, with a couple of simple test fixtures, it can be used for measuring the line regulation, load regulation, and output impedance of power supplies. It's a very handy instrument.
Note that the HP3563A is the lesser known cousin of the HP3562A. The HP3563A is essentially an HP3562A with a digital output and a few other improvements. As it's lesser known, it tends to sell for less on the various auction sites.
Also note that the HP3561A is essentially one analyzer channel of an HP3562A. I.e. the HP3561A is analyzer-only and does not contain a signal source.
For amplitude response and THD+N measurements, I prefer using the HP 8903A along with some scripts to automate the frequency sweep. This method is faster and easier to set up. However, where the HP 3562A shines is for noise measurements, frequency response (amplitude and phase), and for spectrum analysis.
The best advice I can give for getting started with these pieces of gear is: READ THE MANUAL!
Yes, the manuals can be expensive. Yes, they're hard to understand and in most cases rather boring to read. But the HP manuals of the 1970'ies and 1980'ies are really quite well written. I recommend getting both the Operating Manual(s) and the Service Manual(s). The operating manual will tell you how to use the gear. The service manual will tell you (in detail) how the gear works. Understanding how the gear works will give you a better understanding of its limitations and how to use it to measure what you want.
Many of the manuals -- even for the older gear -- are available for download from Keysight's website. Free of charge... Use the search function on the Agilent site to search the Test and Measurement Section for the manuals. Use the HP product number ("3562A" for example) as the search term. The quality of the scanned manuals is usually decent but the schematics are in many cases very hard to read.
For good quality scans with readable schematics, I suggest contacting ARTEK Manuals. They offer quality products at very reasonable prices.
I used to recommend Manuals Plus, but sadly they have now closed. There does appear to have been an effort made to save the manuals, though: The Atlantic: Introducing the Archive Corps.
And, of course, there's always eBay. Just beware that some (less than honest) sellers sell the manuals that you can download for free from Agilent. Don't fall into that trap. Personally, I only use eBay for the printed manuals. The scans I get from ARTEK.
Expect to pay $40~$100 per volume for a printed manual in good quality. About $5 for a scanned version. The price depends a bit on supply/demand and on the complexity of the manual (number of pages, number of fold-out schematics, etc.)