Here is the best explanation I’ve seen of a resistor ladder.
One other thing worth mentioning is the line between Delta Sigma designs and Mutlibit designs are blurring. None of the new Delta Sigma DAC’s use 1 bit output exclusively they are now mostly 3-5 bits, so technically “multibit”.
That’s why Schiit adopted the whole “true multibit” nomenclature to signify that their is direct decode of the bulk of the signal.
FPGA dacs are getting interesting and more prominent as well, so that’s cool
I’m some what skeptical in that space, I’ve done some FPGA work.
All they really allow you to do is make low volume custom logic, so I have to wonder if the FPGA’s are just allowing manufacturers to tune the sound, or if they are really implementing something different.
You should go and see if you can try a ps audio directstream dac or a nice chord dac as they can do some pretty great stuff with them. They also commonly send updates that can improve the dac in sound and measurement
wow! those are expensive. is there a more affordable FPGA dac?
In theory they shouldn’t be too expensive but end up being so because of the skill needed to make them sound good compared to a more traditional sigma delta.
Software updates is certainly a huge win.
It has no way to model passive components, so on it’s own the best it can do is model a DeltaSigma DAC, albeit with potentially much higher base clocks.
They could use it in conjunction with Passive components though to do a custom multibit DAC.
I guess there is probably some advantage to just having it on the same silicon as the USB interface and decoder.
I’d just be interested in what they are doing with the fpga.
Big FPGA’s are expensive chips depending on how big they are, they’re generally used for small production run applications where cost is less of an issue, or where massive parallelism for a single task is useful. In this application, the cost is probably for the software given the tiny markets for high end audio gear.
Yeah, perhaps you may want to look into a fpga dac lol, you could have some fun with them 
The main cost comes from the design and r&d to get to sound good. It’s not as simple as writing some code that works and calling a day
Yes R&D cost is everything in low production run designs, software or hardware it doesn’t really matter.
I’m tempted to try and build a couple of the obvious FPGA DAC designs on the dev boards I have here, but without a really good way to measure them it’s not a very interesting excercise. My scopes aren’t accurate enough for a significant bit depth.
Also check out something like the AudioGD R2R 7 which uses 5 fpgas lol
Well there are plenty of obviously good uses for FPGA’s in these things, programmable digital filters, anything a DSP would be used for…
I’m just intrigued by the idea of a DSP based DAC and what differentiates it’s implementation from hard wired designs.
Couple days ago I posted this. Found it pretty interesting especially as to where technology is heading. In the end, DACs are like cell phones. The technology and capabilities are moving so fast, that waiting a year is always going to get you a better product. We still have to decode stuff today though.
I didn’t rank a DAC as high in the scale of importance, now though I think after speakers or your HP they’re probably the 2nd most impactful link in the audio chain. Excluding the media of course.
1 Like
Thanks that was worth the watch for what Chord was doing with the FPGA.
It’s basically still a Sigma Delta DAC with some specific signal manipulation.
It was also interesting to see how much control the DAC manufacturers have over implementation.