DIYHiFi.org

In the event this forum ever does close, I want to say thanks and give you guys credit for taking the time to answer what must have been to most of you, banal and elementary questions. In this tribute, I will be listing some of the things I believe I have come to learn about good audio design:

1. It all starts with the power supply. Amplifier supplies need to be as quiet and stiff as possible while meeting the required voltage and current demands of the speaker amplifier system. Loudspeakers without relatively flat impedance curves and easy phase angles will be more demanding on the power supply. Large power supply caps can introduce electronic reflections that are picked up and amplified. This effect can be minimized by using a 'snubber' but it's a better bang for the buck to choose just enough capacitance to keep the caps off the flattening part of their discharge curve while they are being drained. Amplifiers are micro-phonic so take steps to reduce mechanical noise and vibration. If you are mindful of stray magnetic fields (inductance), larger transformers will always sound better than smaller ones, 'ceteris parabus'. I still don't know how to choose or use bypass caps or why they are so critical to the sound of an amp. I also don't know the best methods to mitigate RF, nor do I understand how or why it effects audio output. Much of what I understood above about power supplies also applies to line stages and pre-amps (line stages with gain) but now the current and voltage requirement is dictated by the amp input and the selection of the transformer requires compromises to manage and maintain the integrity of the original signal. Larger transformers are more likely to introduce stray inductance and smaller ones may introduce mechanical noise and transient spikes as they are stressed. The electrical signals moving about here are very small and easily altered so these problems are difficult to measure but can probably be heard when amplified.

2. Amplifier circuits can be created by all kinds of active devices and successful designs have been created with everything in the history of electronics from tubes to op-amp chips. Each has pros and cons but the most preferred active device for the ultimate amp seems to be the lateral mosfet for it's output bandwidth, durability, linearity and speed. The main problem is circuits with it like to oscillate. A design philosophy question that anyone attempting an audio project should ask themselves is whether you can achieve better results with devices you understand very well instead of learning about a new device while working with it. No device is perfect so if you can manage well the imperfections you know about, why create a new headache? It seems to me that the key to creating a good amp is understanding the data specs of the devices selected and measuring their performance. In many devices, what you would need to know is not published or so far off spec as to be unusable. This is the ugly truth for the casual hobbyist like myself, there is no way you can build to this kind of precision and matching without expensive instruments and the knowledge of how to use them. Some of the things that need to be matched on the active devices are; trans-conductance (beta), gate capacitance, slew rate, etc., even the passive devices used with them need to be 'specked' for optimum results. It is not surprising that many hobbyists prefer their chip amp implementations to their other diy attempts. This unfortunate reality is even more evident with the other components in the audio system. My knowledge is so poor in this section that I don't even know what I don't know so please fill-in what I missed.

3. Volume control is a sub-category of the line stage component and even here, there are multiple paths to success. Everyone on this board has stated a preference but the consensus seems to be that a digital implementation with signal isolation is the best way to go if you know how to implement it. I could probably write the microcode for the display functions but big effin' deal. I may never acquire the knowledge to do this properly so the next safest bet is a stepped attenuator with good quality, closely matched resistors. When the audio signal is not isolated, the distance and path it travels becomes critical and the physical layout becomes more important. Also important are the resistor values selected because they will create the signal impedance for the sources and amp. Source selection is yet another function of the line stage and I don't know if it is possible to isolate the switching from the signal path or if there is indeed any benefit over a plain metal switch because I never asked the question and never looked it up. If you guys discussed it here then please point me to the appropriate thread.

4. I never formed an opinion on whether a buffer with no gain (or negative gain) was superior to a slightly amplified line stage. I guess a lot has to do with the design choices of the amplifier and the interconnection scheme. I seem to remember a passive line stage built in an acrylic box with bare silver wire suspended in the air within that box connecting the input jacks to the source switch. Volume was controlled with a simple pot connected to the source switch and it sounded as good or better than any of the commercial line-stages I auditioned with a Pass Aleph-30. There is a crap load of stuff on interconnects but I will try and condense it here in a few brief sentences. Low inductance, low capacitance and good RF shielding are desirable in an interconnect but you will usually be trading off one for the other depending on your amp and environment. No one should worry about interconnects until everything else is in order. The only thing I do know about all interconnects is that they work better suspended in air. I suppose this is to avoid the micro-phonic effects mentioned earlier but it could also be to reduce the eddy current effects surrounding the wires. Help?

5. Loudspeakers are probably the most critical component in the audio system and I know less about them than the others. Once again, the path to success seems to be through matching and synergistic component selection. The devil is in the crossover and some opt to avoid it altogether by going with a single driver system on a narrow baffle or transmission line. To complicate matters, designing speakers requires a mostly different set of expensive tools. I just buy good used speakers and avoid this headache altogether.

Well, this is about all I believe I got from the eight years I have dedicated to this hobby. If anyone notes anything wrong, incomplete or misleading in what I've written above then please point it out as a goodbye present. Thanks again for allowing me into your collective brains, I don't know that I am happier for it but I do have a greater respect for what it takes to be an audio great.

Thanks but we hope to keep it open

Elso Kwak
Okay, so I guess I can go back to watching some more Squidbillies
Since I have your attention, are you familiar with the audio qualities of digital media players like the Zune, iPod and Gigabeat? I heard a hi-res FLAC of Huey Lewis on a Gigabeat straight into an Aleph-30 singing to single speakers mounted on narrow baffles that sounded like I could reach out and shake his hand. The drivers had unspectacular specs, 52-14,800Hz +/-6dB but did have some contour filtering. I am planning to revisit this audition with a high end denon DVD player as soon as I find the original recording in plastic but if you have any thoughts on these digital players or know anything about their D/A conversion, I'd be interested.

Hi,
I am sorry I am absolutely not familiar with computer audio.

I play CDs with a 16-bit DAC non oversampling.