A Lazy Day with Class D on the Brain
With nothing to do last Sunday I decided to do some research on class-D audio amplification development over the last five years. For those new to Parts-Ring.com or to electronics you’ll at some or other stage come across this classification and whilst it boasts all the bells and whistles of a good class AB amplifier, beneath the veneer lies a very sophisticated beast. I’m all for KISS (Keep it Straight and Simple) but you know what, there is merit in using a class D amplifier, lots of electronics inside. Let’s have a look at what cooks.
(I have rewritten this first paragraph to hopefully make more sense – Ed).
The Class D amplifier is essentially a high speed switch where an audio component decides the on to off pulse length (called modulation) where equal on off pulse lengths has no signal output. This is as simple as it can get… or
Well… an amplifier which takes it’s audio output from a variable duty cycle square wave… or…
Class D amplifiers are switching devices, switching at frequencies of usually 250kHz or higher where these same switched devices act as the output stage driven by a variable duty cycle square wave generated by an audio signal modulating a triangular shaped waveform…. or
Well, is that it? Not at all. As you can see there are many ways the operation of this amplifier can be described which hopefully makes sense but often words alone cannot describe the operation of this amplifier especially since… We have now a high frequency component and a low frequency component at the output. A cut off filter – in this case a low pass filter – removes the high frequency component from the loudspeaker connect leads. The loudspeaker is driven by the low frequency component. Ingenious? Yes. I left the filter until last because perhaps in many a designer’s eye it’s the least critical component. The problem is, it isn’t.
However, through this rather dubious means of biasing the output transistors to off only and driving them into full conduction at a very high speed and then off again comes a load of other complications (pun intended). The square wave switching creates magnetic or EMI/radio interference and of course how the hell does one think the audio component is going to look (theoretically speaking – says I with a huge scope connected to the speaker terminals). If there is high frequency switching taking place, directly coupled to a loudspeaker take a guess what will happen? Yes, the inductance of the voice coil should be sufficient to reduce the HF current but not always. A high frequency high energy pulse train is a transmitter of sorts, right? It makes the mind boggle but once broken down into stages and analysing each stage it becomes quite apparent that this amplifier can be used for high quality audio reproduction and of course, a carefully designed amplifier limits the radiation as well as reduce the high frequency component to a near zero amplitude superimposed on the audio frequency at the loudspeaker. Ingenious? Definitely. New technology? Definitely not – designs were already available in the 60s, maybe prior, but with modern technological advances this form of amplification has become cheap and of course, a major factor in following this trend – energy saving.
Curiosity Killed the Class D Cat
I’d like to set the record straight here first and foremost – I am always curious as to why certain research veers off in a direction which is either not feasible or more importantly, off the tried and trusted route and then again why use the Universal Brain’s Trust to go to the next level. To me the best audio must come from class-A where the entire sine wave is amplified, with as little outside influence made to the signal – in other words good, clean amplification. Why then would someone want to go this route which applies as many outside, disruptive influences as possible to the signal and on top of that create interference which can cause the authorities to sit up and take notice. The answer here would most likely be, because someone was bored, secondly, to make a more efficient amplifier and yes, something which only switches on and off should not create any heat – we relate then to a device that switches on rapidly, high current and low voltage and then off again, high voltage and low current. As power is equal to Voltage x Current the dissipation should be very low and indeed it is. So in a nutshell we won’t look at the dynamics of the amplifier or system design but rather at where to go when looking for amplifiers which will make your pockets empty in the build and bring you instant gratification when you buy the Lepai, maybe. But first a little bit of theory from those that know…
Ah yes, we may write in jest but the class D is with us to stay. To understand is simple, to design is great but to build to work properly is well – simply a miracle. Class D amplifiers are nature’s way of saying – hold it Tiger!
A neat little block diagram found below is from wiki: (sorry to steal their images but this is just too good to miss).
Simply put – Class D is not better than Class A, I beg to differ though.
The class D amplifier consists of (simply put) a high frequency sine wave oscillator in the region of a few hundred kilohertz, the audio signal then modulates this signal, generating pulse widths based on frequency and amplitude which in turn drives the output transistors hard on or off. The filter removes the high frequency component(s). An interesting fact here is that SSB radio uses a ring-bridge modulator which modulates the carrier wave which, depending on class, A3J being the most efficient, allows for only the audio and carrier to be passed on at any given time, never just the carrier wave – and it’s half the bandwidth of the older conventional design. What is conventional in any event? A great engineer never relies on conventional means in any case. One wonders how much our RF engineers of yesteryear played in this role of audio engineering. I digress, I apologise – there is a link somewhere however, I don’t know what and I’m not being sarcastic either. Certainly this concept of modulating an RF carrier is not new – tubes after all could do anything that modern transistors can also do albeit at the expense of efficiency. Blame it on the heater.
Better Information and Foresight makes Jack a Better Engineer
Moving on to a link to make things easier for the budding scientist – Worcester Polytechnic Institute Electrical Engineering students Briana Morey, Ravi Vasudevan and Ian Woloschin in May 2008 and under the guidance and advisory (AKA known as follow this or else) capacity of Professors John McNeill and Andrew Klein brought out this document on Class D amplifiers as part of the requirement for their B.Sc degree. Like most DIYers we often miss the small print when attempting to build something new – in this case I didn’t just fly off the handle and build the ultimate Class D amplifier with a couple of screws and a hammer. Good quality class D amplifiers are easy to build when you know how but beware of the pitfalls. I know of two companies building them locally, both use engineering graduates BUT you don’t need to be a university graduate. You do need the tools though. PC Board layout and design is very, very important. There is a solution though. But first….
As we rocket of into space comes design pitfall number two…. The Design of an Analogue Class-D Audio Amplifier Using Z-Domain Methods by M.Sc student (at the time) Pieter Stephanus Kemp, supervised by Professor H. du T. Mouton and B. Putzeys.
I have copied both thesis and do keep it on my system if anybody needs it.
Well OK, we now have a few obviously highly intelligent individuals building, testing and ranting about Class-D amplifiers. I personally was very impressed – I didn’t get proper supervision when doing a thesis on weather radar but who cares. In all seriousness, this is very somber reading and to put it bluntly, who in their right mind would care to build an amplifier based on the pitfalls of class-D?
Not so long ago there was a man building audio amplifiers called Bob Carver. If I was a manufacturer I would find his egocentric capacity to intimidate the audio industry totally off-putting. No, he wasn’t really ego-centric – a practical man? Yes. The thesis above is not to intimidate you but possibly to encourage you. You build at your own risk. But should you?
Tripath or rather let’s not plagiarise: Go to Wiki’s documentation on Tripath. Ineffectively known as Class T amplifiers they are of course still based on Class D and that’s where we will stay. Tripath went the liquidation way and was bought up by Cirrus. Read up about the Lepai Amplifier, a very highly rated design with minimum thrills and even less cost to own. Well, to date I haven’t purchased one and obviously then neither did the modifications enthrall me – i.e. turbo’d the unit but one thing you can be sure of, user contentment is very high. I find is peculiar that such a cheap amplifier should get resentment – you CANNOT build an amplifier this price in the USA. So you can gripe as much as you want – for the layman with low budget this may be the amplifier you want. In fact I believe some listeners with big budget think this amplifier is amazing. So what if you find the spindles of the tone controls or volume control are a bit loose – have you priced pots lately?
Lepai – thanks to our contributors
The Lepai is rated at 20W RMS per channel (or whatever rating they use these days). Some technogurus talk about RMS as being obsolete, some technogurus talk about continuous power being obsolete. You know what, shut the fuck up. Let the listener decide. I cannot hear that whistle through the fart of that old girl on ’10’, you heard it first so congratulations. At 21 U$ this sounds like a bargain.
Moving on to things better than the audio fundis out there – the sounds of success…
I will revert to Elliott Sound Products – I’m kinda addicted to their style and if I wanted another superpower it must be Australia – here’s their link to Class-D amplification, this time by Sergio Sánchez Moreno and edited by Rod Elliott, the maestro on all things audio. I sometimes draw a parallel to how Don Elliott reacts to Bob Carver’s own reaction on the audio superpowers that be. The whole world is a stage and most of the stage comprises of audio idiots. Used to be until the iPod.
Most of all this article is not about the design of Class D amplifiers but about class D amplifiers and what we should really know about them. There are some class A articles written about the same stuff but I can promise you, read the documentation ‘forward’ linked from this article, you will become more enthralled about the design, the total efficiency and you know what … build your own. Over the last ten years there has been a lot of conflict written about Class-D but they are here to stay. Read up on Texas Instruments / Digikey
Our own Smartness ruled by Hypocrisy – the Architecture that keeps us Obtuse
As we are normally, the architects of our own demise, there are some things we need to know about class D before embarking on a project: If you move off Class AB there will be crossover distortion. Class D or Class C (predominantly high efficiency switched carrier – Oh my gosh, Morse code) will carry quirks which is known to the electronics enthusiast, engineer – whatever. When writing a thesis on Class D the electronics student already knows about these quirks. I can key with Morse code a signal way more efficiently on a single sideband transmitter than a DSB unit – the problem exists in the receiving apparatus, your ear. Beauty is in the eye of the beholder. Efficiency is one thing, logic another. A 200W class A amplifier has to sound better than a Class D amplifier, right? Minimum fuss, minimum bother. When the band gets together that’s another problem entirely. Class-D amplification is not only going to get more popular – it will get better. Over engineered maybe, that’s the world we live in.
Thanks Guys, we needed you all along.
Thanks to the students and now hopefully leading engineers that facilitated this article, by adding it to the public domain, by not asking for funding and using the internet for the purpose it was originally intended. Thanks to Don Elliott for once again putting an article together in layman’s terms.
Class D may not be king, but it will win the war. Buy your parts thoughtfully!
P.S. I may have been wrong about the 5 years – things haven’t really changed that much really!
(Editor’s note – Ravin’ Turkey or Robin Trent as we know him wrote this article under the influence. We guided him on to this website where he could put his mark. The interesting thing is that he is a lover of Class D amplifiers and is an ICE man. Yes, he has designed and built his own amplifiers as well – he recommends you buy rather than build. If you need more information he may be able to assist but don’t bet on it).