December 2019

Stan Considers Microphony And Electronic Component Quality
Article By Stan Curtis Of HIFICRITIC

  We largely take it for granted that each generation of amplifiers will be audibly superior to those that came before even when occasionally they aren't. From my viewpoint as an erstwhile designer I do sometimes find myself doubting some of the claims when I can't find the evidence to substantiate a claimed improvement whilst applauding the efforts of designers who have spotted and solved a problem we old codgers missed or chose to ignore. Over the past few years I have given a lot of thought to the question 'is the design of today's amplifiers better than that of yesteryear?'

There is absolutely no question that the availability of better components has had an enormous effect upon the measured performance; reliability and sound quality of today's products. But has the circuit design itself advanced that much and if it has, has it had much influence on the sound quality? I say this because virtually every audio amplifier today can trace its circuit configuration to the differential amplifier popularised by H. C. Lin of Westinghouse whose custom operational amp for the Minuteman II missile was configured in 1963. This is a topic I shall return to in a future column because I am painstakingly reconstructing a transistor amplifier design from the early 1970s but using throughout the best top-quality components from today. It will be interesting to see how well it compares.

Misdirections And Goose-Chases
However, no significantly different amplifier configurations having found wide acceptance since there has been a tendency for designers to focus on working around the edges of the design; finding a hitherto overlooked problem and investigating how to minimise its effect. Personally I'm guilty of this over the years and it does have a couple of advantages. At its most cynical it is difficult for anyone else to argue that the audible improvement is due, say, to your unspoken use of super-fast transistors rather than your claimed use of nano-particles on all the switch contacts. This misdirection is great for sending your competitors onto a wild-goose chase. And second if you do hit on something that really does have some relevance it will lie in a very narrow field so you should be able to build up a goodly head of experience before your competitors start down that road. Put together you have all the commercial justifications needed to finance a bit of R&D to inject into next year's products. Which brings me to this issue's topic of conversation.

I recently enjoyed reading an interview with Steve Sells, Naim's Technical Director (HIFI CRITIC Vol 13 Number 3). In the course of this interview he explained the downsides of vibration and microphony and the steps being taken by Naim to minimise its audible effects. Those with long memories will recall that I gave Steve his first job in hi-fi as my assistant at Cambridge Audio in the mid-1980s. I can't recall if we ever discussed microphony as much of our time was spent considering how much better a design could be if we used his array of twelve transistors instead of my skinflint array of two.

Seated One Day At The Organ...
My own seminal experience of microphony came in the mid-1960s when I was playing the organ in bands and hooked up a Hammond organ to a pair of full size Marshall amplifier stacks (the sixteen 12 inch speakers then defining the term “LOUD”). I was initially plagued by uncontrolled feedback until a member of the band casually picked up the power cable to the Hammond as he rearranged the cables near the loudspeakers. The feedback stopped and I realised I'd discovered microphony coupled via the vibrating cable and so watched out for its influence thereafter. In those days it was a weakness in valve electronics and some designers went to great lengths to eradicate its evil influence.

I once acquired a very powerful public address amplifier made by Philips: its circuit design was fairly mundane but its construction was a marvel of mechanical engineering. Not only were all the valves spring mounted onto the chassis but many of the capacitors were housed in little boxes packed with grease. Most impressive of all was that every wire of any length was wrapped in cotton thread and housed individually and snugly in a steel tube; a tube that was periodically welded to the chassis.

I dread to think about the manufacturing cost but it was a joy to behold and furthermore it was audibly incredibly dynamic and noise free and putting the contemporary Leak and Radford amplifiers to shame. The Philips design was wasted by being used only for public address applications: I tried it in a mono system but with only the one example I couldn't try stereo. However, I did attribute its clean performance to its manner of construction and so ended my second lesson on the impact of microphony.

All electronic circuits create a mish-mash of noise which adds a layer of muddiness at the bottom of the sound stage. The noise generally comprises Thermal Noise; a kind of “hiss” generated by all electronic components; Shot Noise & Burst Noise which have an irregular sound and are mostly generated by semi-conductors; and a low frequency noise called 1/f or “Flicker” Noise. Added to that there will be a trace of ripple or hum from the power supply; audible as the “drone” of the constant mix of harmonics of the 50Hz supply, and then there are the microphonic effects, picking up vibrations generated or present inside the amplifier.

Tracing The Vibrations
Where do these vibrations come from? Well some from sources external to the amplifier, such the loudspeakers in the room, but in my opinion the most insidious vibrations come from wires and components carrying signal-related currents: not just those carrying the music signal but also to wires and components carrying the power supply feeds and the ground signals. Why? The power supply current will vary as it tracks the music signal: by way of an example the current in the wire carrying the supply power to the output stage of an amp will rise and fall with the music signal, as a severely distorted copy of that signal. Yet you can touch that wire (or indeed any other component) with your fingertip and you can't feel anything, so where's the problem? Well, it's easier to understand if you scale up the problem.

When I was a student at the City University we had a basement full of massive electrical machines: motors, generators and a mass of switchgear and cables as thick as your leg. And what was slightly terrifying was the way cables would jerk or jump when you suddenly switched 1000 amps or so through them. There are several reasons for this motion so I won't bore you with the technicalities, but bear in mind that if a wire moves visibly when carrying 1000 Amps, it will also do so when carrying, say, 5 Amps – just not as much.

Put a wire that might vibrate next to a high value resistor in a high gain stage and you've created a form of microphone next to a sound source, injecting that very distorted image of the music signal into the mix. A number of possible mechanisms can be at work – all are bad news if you are attempting “clean” low-noise amplification. True, you can seek to overcome some of these problems by injecting extra thermal noise as a mask but that's hardly the point, is it?

Can you do anything to minimise the pickup of these stray signal related vibrations? Here we come to the third of my seminal experiences of microphonics. In the 1960s I undertook my student apprenticeship with Smiths Industries, and one of my projects was the first thermal camera for early detection of breast cancer by picking up the “hot spot” locations of disease. A liquid nitrogen cooled cell picked up the tiny heat deviations but the microscopic signal was swamped by circuit noise. Over the months the signal-to-noise ratio was improved as I learned more about how to reduce the noise – experience that paid back well when I was called upon to design moving-coil phono stages – but eventually I came up against our old friend microphonics.

Cabled to avoid microphony: The comments by Steve Sells on the design of the Naim Supernait 3 (HIFICRITIC Vol 13 Number 3) set Stan thinking.

Tackling The Problem
With the aid of some Heath-Robinson creativity we eventually reduced that to tolerable levels and the equipment rolled into hospitals around the world. This was achieved through a whole host of small modifications some of which will now be familiar to the team at Naim. Cable routing; dressing and anchor points were important as was the choice of acoustically well damped components or the addition of isolating bases. The transistors (germanium in those days) were mounted on a dab of mastic and their wire legs were coiled to form a spring. The same was done with many resistors and capacitors; the wire legs being coiled into miniature springs before they were mounted on terminals called turret lugs. One thing I do remember is how long it took to build this thermal camera – there were identifiable shades of the Philips amplifier I mentioned earlier.

So what became of that laboriously acquired knowledge? Well except for paying attention to cable routing I never used it again, as I never needed to achieve such extremes of signal-to-noise ratio. But one thing I did learn is that it is a genuine effect and it can be frustratingly difficult to eradicate. But fortunately it isn't really a challenge until you've largely overcome the other problems that afflict amplifier operation. And that I guess is the challenge Steve and his guys found themselves facing at Naim.

Subscribe!

Click here to subscribe to HIFICRITIC