Links

PEARS - Easy to use cross platform compatible file encryption application using AES encryption and RSA keys.

Startpage - A safer way to use Google.

Decibel Dungeon - A guide to Hi-Fi DIY.

G. Randy Slone - High-Power Audio Amplifier Construction Manual. - A great manual for understanding and building solid state amplifiers (PCB artwork include). Randy Slone passed away on April 2, 2010.

Julian Assange - Cypherpunks; Freedom and the Future of the Internet - An interesting discussion on internet privacy.

Speed Bag Stuff - Build a Free Standing Speed Bag Platform, Make a Super Fast Speed Bag Swivel, Steve Ciaramello (boxing style training), Speed Bag Central (Bible style training).

Gumbo Recipe - Make a pot of gumbo!

Black Forest Colorado Fire Map - The most destructive wildfire in Colorado history. El Paso County posted this map of the burn area on June 15, 2013 and then quickly removed it. Repeated requests to make it public again on the El Paso County website have failed.

Handy Pressure Cooker Time Tables


FAQ

Q: Why are the heat sinks on the front of the unit? Isn't that dangerous?

A: During the planning stages of this project I drew each major component to scale and moved them around exploring different layouts within the constraint that the amp chassis had to be the conventional width (about 17"). I quickly realized that the PC mainboard must face to the rear due to the I/O ports. That left just enough room across the rear for the ATX power supply and amplifier transformer. It turned out that the amp heat sinks and PC drives would fit across the front and leave space on the side of the unit for the amp power supply capacitors. This is the original concept sketch.

Pi-Amp Chassis Conceptual Layout

I thought about the aesthetics of having the heat sinks on the front and concluded that since tube amps leave glass vacuum tubes on display, why not have power transistors on display. It is easy to protect the transistor cases by installing a Plexiglass cover in the heat sink as shown below.

Pi-Amp

Q: What about noise? Do the PC circuits cause interference with the amplifier circuits?

A: No. The amp is dead quite with PC booted, line in connected, and volume control fully open. Any noise would come from improperly grounding the line in signal to the amp and/or active cross over (for the bi-amp build).


Q: Have you considered a fan-less option?

A: Yes. I think it would be worth trying out. The power supply and CPU fans are audible and may bother some people (I listen to music pretty loud!). There are commercial ATX12V fan-less 500 Watt supplies though I have not tested them. When it comes around to replacing a power supply I will check them out. The CPU is fan cooled. An option here would be a larger and quieter exchanger-fan unit or liquid cooling. I'm not sure liquid cooling would fit the chassis. I used the stock CPU fans and quickly upgraded to quieter after market fans.


Q: Why class B amplifier? Have you considered D or T class?

A: I did not consider the other classes of amplifiers though I'm sure they would work fine if the fidelity and power meets your specifications. I used the Leach design because I understand the topology and I found the build within my clumsy capability. I don't think I will attempt surface mount components.


Q: Did you consider using a DAC?

A: No. I can't convince my self that the incremental sonic performance of a stand alone DAC vs. a quality sound card is worth the cost.


Q: Why did you change the toroid transformer location in Pi-Amp Bi-Amp?

A: Better weight distribution and easier access. The bi-amp build is very heavy (about 50 lbs) and the chassis is taller. Placing the toroid opposite the PC power supply gives better weight distribution and easier access. If I were to build Pi-Amp again I would change the position of the toroid transformer to the same position as used in the bi-amp build.


Q: Have you had any thermal problems with the recessed heat sinks?

A: Initially, the Pi-Amp Bi-Amp heat sinks ran very hot. I thought it might be an air flow problem but traced the problem to my use of op-amp sockets in the ESP active cross over. The ESP site recommends not using op-amp sockets because of potential parasitic oscillation which will drive the amp to overheating. I pulled the sockets out and soldered the op-amps directly to the PCB and now power transistor heat sinks run cozy warm as they should.


Q: Can you use a nano-ITX main board?

A: Yes. In fact I agonized over this in the design phase. But ... it did not change the component lay out. The PC would be in the rear with the heavy gear (power supplies) and the amps and drives would be in the front, and these set the chassis width. Given that, I saw no reason to go with a main board with less CPU power. This was several years ago and we all know that CPU performance is always improving so today's nano-ITX boards may out perform the micro-ATX boards that I am using.


Q: Is there a significant difference between CD and MP3 audio quality?

A: This question has been tossed around audio and PC communities for quite awhile. The typical answer is given in a Maximum PC article entitled "Do Higher MP3 Bit Rates Pay Off?" which is a purely subjective taste test.

I will answer the question by letting you see and hear the difference between loss-less (WAV) and compressed (MP3) digital audio. Then you can decide if the difference is significant and we can stay friends. By difference I mean subtraction, not "oh, that one sounds better, I think". Subtraction takes out the subjective by showing exactly what is lost. You may be surprised at the results.

Let's take a look and listen at a very short piece of Haydn's Symphony 52 from the original WAV file. The signal graph below is a plot of signal strength (vertical axis) vs. time (left channel on top, right channel on the bottom). Listen to the WAV audio file here.

Pi-Amp Track WAV Rip Subtract 128 Kbps Rip

Now let's "rip" the WAV file using the LAME MP3 encoder at 128 Kb/sec and 320 Kb/sec, constant bit rate. The resulting signal graphs are shown below (the top two graphs are 128 Kb/sec, left and right channel, and the bottom two graphs are 320 Kb/sec). Listen to the MP3 128 Kb/sec audio file here and the 320 Kb/sec audio file here.

Pi-Amp Track 128 Kbps Rip and 320 Kbps Rip

The signal graphs shown below are a very short time segment of the graphs shown above. Can you tell them apart?

Pi-Amp Track 128 Kbps Rip and 320 Kbps Rip Close

Neither can I unless I invert the MP3 file and add it to the WAV file. If the WAV and MP3 files are identical, the result is zero signal strength for all values of time. (Note: The MP3 is ripped with no gain and time shifted to exactly align with the WAV file). The three signal graphs below show the process up close. The top WAV track is duplicated and inverted to form the middle track. Note that the signal values are equal in magnitude but opposite in sign. The top and middle signal tracks are added together resulting in the bottom signal track. Since the top and middle are identical, the result is zero signal strength for all time values.

Pi-Amp Identical WAV plus Inverted WAV

The signal graphs shown below are the sum of the WAV and the inverted MP3 files for both bit rates.

WAV + (-128MP3)

Pi-Amp Track WAV Minus Inverted 128 Kbps MP3

Listen to what got lost ripping WAV to 128 Kb/sec MP3 here.


WAV + (-320MP3).

Pi-Amp Track WAV Minus Inverted 320 Kbps MP3

Listen to what got lost ripping WAV to 320 Kb/sec MP3 here.


The portion of the frequency spectrum lost ripping to 128 and 320 Kb/sec from WAV is shown in the spectrogram below.

Pi-Amp WAV to MP3 Spectrogam

The compression ratio from this WAV file to 128 Kb/sec MP3 was 10 to 1. Going from WAV to 320 Kb/sec was 4 to 1.