When I started thinking about a MKII version of the Caiman DAC, the first thing that I decided upon from the outset was that it would have to be just as ground breaking in terms of price versus performance as the MKI. It is quite easy to fall into the trap of capitalising on the name of a successful predecessor, without any technical justification. But pushing the boundary even further is easier said than done. I often wondered if I had bitten off far more than I could chew.
In order to get an idea of what would make technical and audible differences versus what were just suggestions from couch potato experts, I decided to build a few prototypes and spend some time listening to them. Each build sounded good enough to sell in its own right, but the sound quality difference was in most cases merely a step up. But for the Caiman, I was only interested in improvements that were a substantial leap.
A selection of the prototypes that I designed and built to evaluate the performance and technical complexity of some parts that might be suitable for the final build.
Anyone who knows about the cost of designing and building electronic products would also know that the total cost of those prototypes would have been astronomical. Looking back, I reckon that I got possessed by an obsession for perfection, without keeping an eye on the R&D costs.
Things only got more complicated when some new components came on the market, and when some manufacturers gave me permission to use their parts in a consumer product, or use them in a manner that was not listed in their official datasheets. The ultra low noise regulators from the satellite communication industry made a significant difference especially. So much so, that I decided to try them in a 32 bit D to A configuration and compare that prototype against an otherwise similar 24 bit design and layout. The test results ended up costing me a year delay, partly due to a revised layout change that was optimized for 32 bit D to A number crunching.
Running in tandem to that were my tests on different types of USB audio conversion methods, and the added complication of software drivers. In the end I decided to settle with a 24/96 USB chip operating in adaptive mode that would not require any special drivers when used with different operating systems. Combined with my own async firmware routines, the outcome was actually better sounding than the alternatives I tried. Running the USB chip from the Caiman internal ultra low noise power supply, instead of using the BUS power, made even more of a difference.
The addition of an I2C BUS will allow the Caiman to be expanded even further. But I am not looking that far ahead yet.
My next goal is to get it all up on my website for a June launch.