To work on audio electronics, you need some test equipment. In this article I discuss what you need and how much you need to spend on it.
Old electronic test equipment can now be bought at a very sharp price. Gear which would have cost thousands 15 years ago, is now available for often less than 100€. This is a real bargain – the quality of equipment made by companies like Hewlett Packard and Tektronix back in the 70s and 80s was very high indeed, and it can be more robust than almost anything made today. When you buy it in good condition, it often works as well as it did when new.
There are also PC based tools available, some of which are very good. But to buy professional quality equipment – well, for a test set like this one, by Prism Sound, we are up at more than 5000 UK pounds. The brand leader in audio test is Audio Precision, but even an old second hand set typically sells for thousands on eBay. (Update – since this page was written I did manage to find an Audio Precision Portable One Plus at a fair price.)
At the very minimum, you need these basic things to start doing basic audio test:
- some way of injecting a (generally sine-wave) signal into the device you are working on.
- some way to measure the level of a sine wave level across the audio frequency band (20Hz-20kHz) and preferably to somewhat higher frequencies.
- a multimeter to measure voltage, current and resistance.
- an oscilloscope.
- DC power supplies to power circuits you are working on.
Let’s start out with the basics:
This is probably the first thing you should buy, for pretty much any type of electronics work. A basic one will measure AC and DC voltage and current, and resistance. These days you can buy inexpensive ones which also measure temperature, capacitance and so on. I have two – a very venerable Fluke 77 (pictured) and a newer cheap one made by Tenma, which is actually very usable (although a bit inaccurate on capacitance measurements).
Fluke are the big brand name here, but perhaps a bit pricey if you are just starting. A less expensive brand will serve you just as well for most purposes.
OSCILLOSCOPES – DIGITAL VERSUS ANALOGUE
This is the next item on your shopping list. A decent scope is something you will use very often. These days digital models are common, and can be bought from about 150€ upwards (for the very cheap models). I own one but to be honest I don’t like them much for analogue work. Because of the way that they work, what you see depends a lot on the way they sample the input, and it is quite easy to be misled, especially on cheaper models.
What’s the difference? Well, an analogue scope is generally used to look at “periodic” signals – i.e. signals that repeat regularly. Mostly we test and diagnose audio circuits with sine waves, which are just such a signal, so an analogue scope is perfect. Digital scopes can be used to capture a signal “event” (such as a narrow pulse of only a few microseconds which happens only once) into memory, and you can then examine the event (for instance the pulse edges) in detail, which would be pretty much impossible on most analogue scopes. Of course you can also look at sine waves and so on on a digital scope – but they are not quite as good at it – more about that below.
Older analogue scopes made by Tektronix and HP can now be bought at very reasonable prices second hand, and I think that they are an essential. For most of my repair and dev work I go to my Tektronix 2225. Perhaps it’s because I learned my trade working with these things, but I love the immediacy of having everything I need available on the front panel controls.
I do have a a digital model though – a Siglent SDS1102 which has a very nice large screen and was also great value for money. If I am working on digital circuits, especially when the signals I am trying to see are irregular
So where is a digital scope not quite as clear as an analogue one? Well look at the pictures below:
In all of these pictures, the input to the scope is a 50kHz sine wave of constant amplitude (the period of the wave is therefore 20us). In the first two pictures the horizontal timebase is 500us/division, so there are 25 cycles in one horizontal div, which is a bout 1cm on screen.
Obviously that means that individual cycles cannot be seen, but the Tek still shows the envelope of the signal (which is flat) faithfully. It will continue to do so even at slower time bases.
The Siglent DSO however, because of the interaction between the signal and the sampling rate, starts to introduce odd artefacts into the display. When the timebase is slowed down to 5ms (the third picture) the image is even more misleading.
When you work with a digital scope for a while, you do get used to this kind of thing a bit and recognise it. But even so it takes effort to do so, and can be distracting. Also, when testing audio gear, we are often watching the screen for oddities or anything unexpected in the output of the circuit. Having to adjust for the oddities of the test gear makes that harder.
So, for audio work I do prefer a good quality second hand analogue scope to a new digital one. It’s just more intuitive.
We need some way to generate a signal (at least a sine wave) into the circuit we are testing.
The most basic solution for this could be almost any sine wave oscillator in the audio band. If you have really nothing else, you could use a laptop with a soundcard, of the type that is used for audio recording and playback. You can find free software online which will generate sine waves.
This will get you going if you just need to generate a waveform to trace through a circuit to find a fault (with your trusty analogue scope – see above!) But there are some good reasons why you might want to get something a bit better than this fairly soon.
- you don’t really have a way to know what the amplitude of the test signal is. You could connect it to one channel of your scope, and that will tel you the level in volts. But for audio, we really would like to know the level in dBu (if you don’t know what that is, I’ll explain in a future article – for now it’s just the way that we almost always talk about signal levels in decibels) at least.
- the soundcard may not go much above 20kHz, and the amplitude of the output may even vary up to 20kHz.
- sometimes we just don’t want a laptop or simlar on the bench. Why? They can be a source of electrical noise (even when not connected to the circuit we are looking at).
Again, second hand test gear can come to the rescue. Old signal generators exist and can be picked up for about 50€ or so. The type you want is a “sine wave oscillator”. The one above was made by Trio, and it goes from 20Hz right up to about 200kHz. The output level drops off a bit at teh top end, but it is very flat right the way across the audio band. You can alter the output level with the combination of a switch (which gives you 0dB, -20dB, and -40dB positions) and a variable knob. The signal quality of teh output is very clean and free of noise.
SIGNAL LEVEL METER
We also need some way to measure our signal in dBs. For this, we want a level meter.
The one above was made by Leader, and it has a dual meter which is nice for monitoring stereo signals. The scale is calibrated in volts, dBV and dBm and this is fine (dBm are the same as dBu for most purposes). It also has a very wide range of measurement – from -60 to +50dBu. This was again quite inexpensive – about 70€ (it’s a long time ago).
AUDIO TEST SETS
So, for not too much money, we get a way to inject a signal into our circuit and measure what comes out. This is enough to do quite a lot of fixing, repairing and even design work.
If we want to spend a bit more – we could invest in an “audio test set”. What do we get for that? Some or all of the following:
- a signal generator and meter in one unit
- ways to make typical audio measurements (frequency response, noise, distortion) built in.
- often, a way to upload pretty graphs and so on to a computer (or print them out).
- very useful – a variety of options for input and output (balanced, unbalanced, ways to change input and output impedance, and even digital outs on some expensive models).
Whether or not you want to go this far depends of course on your budget and intentions. Again, the second hand market might come to the rescue. Let’s take a look at some common options:
Hewlett Packard 334A
These things were pretty common in the audio world a few decades ago. They’re built like a tank and last forever so you can still find them second hand. Prices are usually about 150-300€. They tend to sell fast though. Here’s what you get:
- a good quality oscillator and level meter
- possibility to do audio distortion measurements – useful.
- a cool old piece of gear on your bench
- no balanced inputs or outputs outputs. Damn. That’s a drag, HP.
If you’re on a budget and find one at a good price, this is really not a bad option.
AP was started by some engineers that used to work for Tektronix, and they pretty much took over the audio measurement world in the 1980’s, with the AP System One (which I remember fondly from my days at Soundcraft Electronics around that time). Now they are probably the brand leader.
A new AP will cost you quite some thousands, but you can pick up a used one for about a thousand give or take. The features available vary from model to model, but you will get at least:
- top quality signal generator and auto-ranging metering (in whatever units you care to use)
- all common audio measurements such as frequency response, noise, distortion, IMD – at the touch of a button.
- balanced ins and outs
- an even cooler piece of gear on your test bench.
The Portable One Plus shown above has become my essential workhorse. It automates a lot of the measurements that are part of my test and service workflow (noise, frequency response etc). I couldn’t do without it.
The System One (if you can find one) needs a PC or laptop to work, and then you get the niceness of seeing all your test results on the computer. For bench work that is not something I really have space for, but I can see that for some people it’s a good option.
A new AP is fairly serious money, and beyond the budget of most private individuals. There is however a less well known company that makes very good audio test sets at a quite reasonable price.
Lindos are a small British company who have made a few excellent quality test sets over the years. The original, the LA1, was pure analogue and none the worse for it (if you can find one it’ll probably still work great after a service). The next, the LA100, was a staple of the BBC amongst others and is still being made but it’s not cheap.
These days they also make the Minisonic MS10, which is probably about the cheapest decent audio test set you will find. It starts at 450UKP, but of course you pay VAT. Note that the PC software to get the nice graphs and so on increases the price a fair bit, but still a very interesting option.
Lindos also do a PC package, the DS10 which “turns your soundcard into a test set”. However – not just any soundcard – they recommend the YellowTEC PUC which goes for about 450€ by itself. With the software at around 500UKP, it’s not a cheap option, though doubtless very good.
No decent electronic workshop is complete without some way of applying clean DC power to whatever you are working on. Again I like second hand. Old linear power supplies beat the noisy switch mode beasts which you can find all over eBay.
An essential feature that you need (and will find on any decent model) is current limiting. If you work on a prototype piece of gear where things can get accidentally shorted out, the PSU can be set to limit its output current. This saves sparks and “magic smoke”.