I’ve found this nice article on op amps for audio applications: Op Amps: Myths and facts. It not only gives some information, why there should be no audible differences, but also explains, why there may be audible differences if you simply replace an op amp by a “better” one. That does not mean, there is no need for high-quality op amps, but you should think which characteristics are necessary in which part of your circuit. If you don’t understand how a circuit works, try not to improve it by replacing op amps by more expensive types.
To debug digital circuits (mostly I2C, I2S and SPI communications between chips), I still use my old LogicPort 34-channel logic analyzer. When I bought it several years ago, it was a real bargain. There where no comparable logic analyzer available in this price range. However today, there are alternatives. One that I found today is LabTool from Embedded artists. It is not only a logic analyzer, but also features an oscilloscope, a logic level signal generator, and an analog signal generator.
The specs are not very impressive: 100MHz samples rate only with 2 channels used, only 20MSamples/s per channel with all 11 channels used. Also the integrated oscilloscope has a bandwidth of 6MHz and a max. sample rate of 60Ms/s – even cheap DSOs have much more impressive specs. However, I like this board. First it is inexpensive: €99 for a device with so many used is a very attractive price point. Also the software is open source. If I look at the software of my LogicPort, there are so many things I would like to change. However, the vendor does not seem to be interested in further development. With an open-source software there is at least a chance, that people from the community will expand the software.
This is not a top-notch device for professional use, but I is interesting for hackers and makers that do not want to spend too much money and want to use it to design digital circuits. I’m not sure about the analog part. 6MHz bandwidth should be enough to debug audio circuits. I’m interested if somebody is using it for audio – leave a comment about your experiences if you do.
The TempoAutomation team that want’s to build a DIY minifab for electronic circuits has a prototype pick-and-place machine running. It looks like they’re using a Makerbot as the base. There is one problem with this approach: the Makerbot mechanic is not designed for high positioning speed. Using a larger CNC mill could improve speed a lot. But this is only a development prototype, we might see big improvements in the future. Having a machine that could place solder paste and components on the board would be cool.
One of our readers sent me a link to a filter design site: T-Filter. It specialized in the calculation of FIR filters. I had a quick look at it and it looks nice. The filter calculation was very fast. But then I noticed that I was only calculating a filter for 2000 Hz sampling frequency. This is not really a real-world audio calculation. Then I tried 48 kHz sample rate. This also work, but the default passband ripple of 5 db is way too much for HiFi audio applications. Reducing the ripple to 0.5 db often led to empty filter. But the software is still in beta, therefore it may work in the future.
Here is another interesting eLearning for DIY audio enthusiasts. Digital signal processing from the École Polytechnique Fédérale de Lausanne. With this course you will understand the basics of digital signal processing. The course does not focus explicitly on audio applications. But it might still help you to understand, what’s going on in an audio DSP and also what their restrictions are.
There is another goodie: the textbook for this course is available online for free. You can buy a printed version or the iBook version, but if you’re ok with a PDF file, you don’t have to pay for it. I just had a look at the textbook and it seems to be well-written.
However there is one thing, that some people may not like: There is a lot of complex algebra in digital signal processing. You will not just learn ready-to-use algorithms, but you have to do some algebra by yourself. If you hate math, you will not like this course.
Wow, how did this happen? Intel developed an Arduino compatible microprocessor board! The CPU is a Quark X-1000 processor which is a lot slower than current desktop CPUs, but a lot faster than simple ATMega microcontrollers or low-end ARM chips. It used the Pentium command set and runs at 400 MHz. I don’t know if the Arduino IDE will create low-level machine code for the hardware or there is an operating system running on this board (Linux?). I would suppose, that there is an OS running, otherwise it would be quiet complex to use all the nice onboard interfaces like USB and Ethernet.
I had a quick look in the datasheet and it looks like there are no I2S connection that can be used to connect audio DACs or ADCs. There is an onboard sound connector, but I would expect, that the DAC/ADC on the board are low-end chips. However, USB ports are available to connect USB sound cards. I’m looking forward to the availability and pricing of this module. It could be an interesting alternative to the Raspberry Pi for audio applications.
While still many audio enthusiasts favor linear power supplies for their circuit, switching mode power supplies (SMPS) are used more and more. They have several advantages to linear power supplies. Not only are they usually more efficient, but also the switching frequency can be far outside of the audible frequency range. Even though you can use the circuit diagrams in the data sheets of available ICs, you might want to learn a bit more about this kind of circuits. In this case, the Coursera course “Introduction to Power electronics” is the right course for you. The course started today and you can attend for free.
Our HiFiBerry project is still ongoing. The PCB design is almost finalized and we expect the device to be ready in about 4-6 weeks. With HiFiBerry you will get an inexpensive and high-quality sound card for the Raspberry Pi. However, there is one thing it cannot provide: high-resolution sound, that means sampling frequencies above 48kHz are not supported. There are a lot of Pros and Cons for or against higher samples rates than 48kHz. At least when it comes to post processing like equalizing or digital crossovers, higher samples are a good idea.
You could also add an external USB sound card. But we are looking for a real DIY solution ;-) The 2nd revision of the Raspberry Pi provides access to the I2S pins of the processor. You can add an I2S capable ADC or DAC on these pins.
Unfortunately, the Linux kernel of the standard Raspberry Pi does not support devices connected to the I2S pins. Therefore you need to compile your own kernel. Check out the “Noise is good” blog for more information. Hmm, looks like an interesting project for another Raspberry add-on board. We will have a look into this.