Back in 1999,
relatively unknown manufacturer 47-Labs released the "Gaincard"
to rave reviews. It immediately caused controversy because it was based around
a $5 power amplifier IC (Integrated Circuit), yet a typical setup would cost
you around $3300.
This wasn't the first so-called "chip-amp" that I'd come across - some years ago I read a Hi-Fi World review of a product by Final Laboratory that also commanded a four-figure price tag for a handful of cheap components. But it was the Gaincard that caught the imagination of audio diy-ers around the world and before long, copies of the Gaincard were appearing on internet sites and discussion forums. The Gainclone was born!
Chip-amps?
Generally, audio power amplifiers for "serious" hi-fi applications are constructed from discrete components - transistors, resistors, capacitors, diodes, etc. While integrated circuits (ICs) have revolutionalised much of the electronics industry, the necessary compromises have always been unacceptible for high quality audio power amplification. Therefore, power amplifier ICs have always been used in less critical applications, such as portable and mobile audio, televisions, computer audio, etc., making only occasion appearances in the hi-fi scene.
But recent offerings from semiconductor manufacturers are much improved. And using ICs is an attractive proposition because a complete amplifier can be built from maybe a dozen components, compared to perhaps 100+ for a discrete design. Designing audio power amplifiers from scratch is a non-trivial task fraught with difficulties - an errant amplifier can catch fire or destroy your loudspeakers - speaking from experience!
This is good news if you're new to DIY audio. The IC manufacturer has taken care of the difficult issues - you just need to provide a few surrounding components and a suitable case to build a complete amplifier that stands a high chance of working, and actually sounding good.
Gainclone philosophy
In addition to the use of a power IC, there are a number of details that differentiate the Gaincard from normal amplifiers.
Minimalism
As mentioned, the IC enables a very low component-count, which in turn leads to a very compact layout. A Gainclone can be tiny!
Very short signal path
This is something that 47-Labs boast about on their website - the length of the signal path is just a few centimetres, and the feedback resistor is mounted so close to the IC that the length of this path is less than 1cm! Some people feel that this is very important to acheiving good sound quality - they would have a shock if they saw the insides of a typical mixing desk!
Tiny PSU capacitors
The original Gaincard and early clones used extremely small smoothing capacitors - typically around 1000uF per channel. To the unitiated, you'd expect more like 3 to 5 times this value. Subjectively, this adds a certain character to the sound, which you may or may not like.
Physical construction
According to 47-Labs, construction is very important, and their website contains references to this, with mention of "stored energy", ect. I'd advise you to draw your own conclusions about this - it's easy enough to experiment with these aspects.
Point-to-point wiring
The amplifier is so simple that printed circuit boards are often considered unnecessary and the components are soldered directly to the pins of the IC. This can work well, but can be difficult for beginners. The final assembly can be fragile and potentially unreliable if the connections rely on just solder for strength and support. While working on the inside of the amplifier, it's easy to disturb the connections, causing faults to occur. But having said all of this, point-to-point wiring can be very good, and there are some beautiful examples around.
Exotic components
A number of forum participents are on record as stating that certain resistors and capacitors sound better than others. The point here is that the basic amplifier is so simple that there are not that many "engineering" variables to experiment with - so perhaps it was inevitable that these claims would emerge. Of course, there is nothing "simple" about a power amplifier IC containing hundreds of components!
I'm not saying that components can't be different; after all, any electronic component will contain "parisitic" componets in addition to the wanted one - for example, both resistors and capacitors have a small amount of series inductance caused by their physical construction. But you have to ask yourself some questions here - such as can these parisitic components alter the circuit behaviour, and by how much? In some applications, such as power supplies, parasitics can be significant (and readily measured). But what about a 1kΩ input resistor where no current flows (and therefore no voltage drop develops)?
Still, I wouldn't want to discourage anyone from squeezing the last 0.1% of performance from their creations - rather, I would just urge you to keep everything in perspective and don't panic because someone has written that a Caddock sounds better than a Riken. Well, not until someone has performed controlled A-B testing and published the results in full for peer-review. Don't hold your breath!
Inverting mode
The original Gaincard uses a standard non-inverting design, but many DIY designs use inverting mode. The differences between the two are reasonably subtle and the arguments for either are fairly nebulous in all honesty.
The biggest practical issue is caused by the low input impedance of the inverting version, which some preamps and source components object to. Perhaps ironically, this loading effect probably causes more difference in the percieved sound quality than the configuration. To read some more background about these differences, check out op-amps for beginners and my Gainclone experiments articles.
Input buffers
Something that came about because of the above issue - inverting Gainclones present a low-impedance load to the preamp, so people found that adding buffers improved the situation. This rather conveniently backs up my theory about load-effect - after all, a piece of copper wire will be better than the best buffer in the world! Personally, I can't see how adding a lot of extra electronics to solve this issue can be a good thing when the problem is so easily avoided in the first place. Still, it's worth experimenting with this, and plenty of information can be found on the internet about it (although nothing much on this site currently, but I do plan to experiment with this later).
Power supplies
Chip-amps are rather more fussy about the quality of the power supply than conventional discrete amplifiers that are able to use RC decoupling for the input stages. People have experimented with different capacitors, rectifying diodes,regulation, "snubbing" and even batteries. You'll find lots of opinions regarding these options - my advice is to experiment (while retaining a sense of perspective!) and make your own decisions.
In summary
The simplicity of the design makes it a good project for beginners. The parts are relatively cheap, there's little to get wrong, and the sound quality is far better than you might think. And, most importantly, there's lots of support and information on the internet, meaning you actually stand a good chance of ending up with a working amplifier rather than a pile of smoldering components!
GainClone Resources
There's lots of information on the internet that will help you if you are building a GainClone - please check out my Links page (click on the "Gainclones" category on the right of the page).