Bigger is Better?

A writeup by John Alexander on DAC/Amp design and build. Originally Posted here

Dongle size does matter: Bigger is better

(Post 2 about VE Mega Dongle, Commissioned by: Wild Lee. This is a two part post, stay tuned for a discussion USB Drive size vs Terry size dongles)When “giant-killers” like the VE Odyssey HD, CX-PRO 31993 and JCally JM20 exists, why bother with the fat, bulky and heavy dongles? Well, think of it this way, all 6 of the 5-star rated dongles for Andy EFs dongle madness are “chonky” bois, dongles that are far larger than the usual apple dongle size like the Apogee Groove or the Ovidius B1, and all across the lineup of 4 – 4 ½ star dongles you’ll find other dongles that are about the same size.

You’ll also find that most of the smaller dongles don’t test as well as their bigger counterparts even with the same chip from the same brand (See JCALLY JM20 vs JM10, both with the same CS43131 chip, but different body size and implementation). So, anecdotally, bigger is better right? Well, categorically, yes but to say that without explanation is quite the claim, so if you’re interested, read on.

To understand why bigger is better it’s important to break down and analyze the chain we’re talking about here. At it’s simplest form, a dongle or portable DAC/AMP consists of 2 things, a DAC to handle the digital to analogue conversion, and amp to amplify that analogue signal before it’s fed to your power hungry cans or IEMs. Of course, there can be many more components in the chain, with larger devices having OP-AMPs, power filtering capacitors, etc. but for the sake of simplicity I won’t discuss those now.

The majority, if not all of the smaller dongles use something called a system on a chip or (SOC) one example of which is the ES9280 chip on the HI-DAC which is basically a single all-in-one chip that takes care of all the necessary processes, from the D-A conversion to the amplification stage and all the necessary in between actions. What this means is by using a good all-in-one chip, you can get rather good audio quality in a very very compact form factor, hence things like the “giant-killers”.

However, for getting into the legitimate high-tier audio space, these all-in-ones are simply inferior in performance and less viable than the discrete solution I’ll discuss below. Firstly, an all-in-one is difficult to tune, with most of them offering little to no software-based programming and will generally have their own “house sound” imprinted onto them by the manufactures that designed them. Other tuning trickery like using a specific cable to change the sound is also possible, but again, difficult. This means that it’s generally difficult for a manufacturer to get the exact sound they want out of an all-in-one, and thus they have to compromise and work around the chip.(PS:

If you’re still confused, here is an analogy. all-in-ones are like having a computer with an integrated GPU (i-GPU) or APU (for my Ryzen peeps). Now using these for high performance tasks like gaming or rendering you can get away with pretty decent results IF you’re using a good CPU (SOC) and if it’s implemented well (good cooling, enough ram etc.). However, the results will almost be never as good as a set-up utilizing a discrete graphics card (discrete DAC/AMP) owing to factors like customizability, upgradability and choice.)

Now, discrete solutions. These are what you see in a lot of the larger portable DAC/AMPs, and inside of some high-end phones as well like the LG Quad DAC phones and is usually the superior implementation when compared to the all-in-one chips. Essentially, what this means is that the components of the chain within your device are handled by dedicated components made to perform a specific function and only that specific function. This usually looks like XMOS Chip-AKM DAC-Custom Amp circuit (XMOS is actually a type of SOC, but in context it’s a discrete component because it basically only handles the receiving and conversion of the digital data into data that the DAC can process, although some companies will have a type of special proprietary equivalent like Chord’s USB implementation, or some variation of it.).PS: Some devices can also have certain parts be SOCs like the XMOS chip used as the USB interface of some “discrete” setups. Other dongles and plenty of phones (Samsung and Apple etc.) contain something called a CODEC, which refers to CODER-DECODER, basically a DAC with an amplification circuit built in and ALSO, an analogue to digital converter to allow for the device to also process mic inputs like with the OHD and CS43131 dongles.

Basically, what this allows is the freedom and headroom for a manufacturer to mix and match different components in the chain (USB receiver chip, DAC, AMP) to allow for different results in the final audio output. It also generally allows them to use higher quality components, or even custom programmed ones using CPLD or FPGA technology, which allows them to use a near infinite amount of tricks to change their sound, which means they can make it sound exactly how they want it to sound, without the limitations of an SOC. Now the actual implementations of all-in-ones and discrete setups look very different. Almost all of the smaller dongles use all-in-ones or a CODEC with a SOC for data interface, and discrete need to have a larger housing to accommodate the much more complex PCB and all the different components. So, if you want good sound quality in a small package, go with an all-in-one based dongle. If you want the best SQ a portable device can offer, go with something discrete. Food for thought.

Back to #donglemadness