How can the "Analogue section" of a DAC decide or affect the total quality of that DAC?

[nyxnyxnyx]

The common argument I've heard about DACs from both objective people and full-blown audiophile folks are something like "well, it's not important what the dac chip is, it's more about how the company designs the analog output/section of it."

I've heard that line in similar variations from time to time now but yet I never crossed an article or some designer sharing his knowledge on what, why and how it affects everything. I assume the notion of using latest dac chips is because they are easier to work with and they have most if not more of the advantages of the older ones (think old PCM or TI chips for example). I don't know if that is correct? Would love to understand this subject better!

 

[nyxnyxnyx]

I would like to add a part of why I create this thread is I used to ask the same question to people who said that line but they just could not give me a clear answer. I don't mean to be rude but I think if one does not really understand it but say it like gospel then he's more like parroting it instead of knowing it.

 

[brandonhall]

Great question. Following

 

[voodooless]

The analog section has multiple tasks: in many cases it needs to convert the current output of the DAC to voltage. It also filters some remaining HF content out, and finally it should be able to drive a pre-amp or power amp well enough. All of these need a certain degree of engineering, so things can go wrong if done improperly. Think about the famous IMD hump in ESS DAC’s it took quite a few tries of various companies to eliminate that.

Next question would be how audible a badly designed output stage is? That will depend on what kind of problem it has and how these effect the output. There are probably very bad examples to be found out there, probably from some boutique companies as well, making the exact same claims and at the same time putting tubes in their products

 

[xaviescacs]

convert the current output of the DAC to voltage

Can someone explain to me what that means? In my book, without voltage there is no current, so I guess it's a kind of engineer jargon, that I fail to grasp. Thanks!

 

[voodooless]

Can someone explain to me what that means? In my book, without voltage there is no current, so I guess it's a kind of engineer jargon, that I fail to grasp. Thanks!

Yes, you always have both. But the question is which of the two will carry the signal. In case of many modern DAC’s, it is current. However, all other audio equipment needs voltage drive. Hence a conversation must be made. Simplest conversion from current to voltage can be made with a resistor. Obviously much more elaborate schemes are used in DAC’s.

 

[xaviescacs]

Yes, you always have both. But the question is which of the two will carry the signal. In case of many modern DAC’s, it is current. However, all other audio equipment needs voltage drive. Hence a conversation must be made. Simplest conversion from current to voltage can be made with a resistor. Obviously much more elaborate schemes are used in DAC’s.

That means the DAC chip produces very little voltage and it needs to be amplified to the nominal 2/4 V?

This is what they call the output stage?

 

[solderdude]

A DAC chip either provides a current (so no voltage on the output) or it provides an output voltage.
Cheaper chips (simple low parts count solutions) often provide a voltage out.
Easy to use but not top performance.
Internally the DAC section in the chip is probably current based but there is an internal I/V circuit in there (opamp + resistors) which saves costs and board space.

As soon as one simply connects a resistor to the output of a current-out DAC chip, hoping current is simply converted to voltage then non-linear behavior will be the result.
Such a current output must go into a '0ohm or 0 volts' inputs.
This is easily created using an opamp with the + referenced to ground (or a reference voltage from the DAC chip).
The - input will always want to remain at the same voltage as the + input and does so via the Rfeedback.
This means the input voltage (I in below) will never have a voltage on it but does have a current flowing into the Rf which 'sum' to '0' at the - input.
So current from the DAC chip - the exact opposite current (Vout/Rf) sum game.
The input current thus is converted to an output voltage (current x resistance) and this way current is converted to voltage I/V.

The higher quality DAC chips are current out and most of them even balanced (+ and - out).
As audio equipment is all voltage based the DAC chip has to be connected to an I/V converter.
This usually is an opamp where the + input is either biased by a ref voltage from the chip or referenced to audio ground.
The - input is connected to the current output of the chip.

When there are glitches (short spikes) often the right opamp has to be chosen or countermeasures must be taken to eliminate those. Gating or filtering.

There are several possible solutions for I/V conversion (filtered or not followed by summing or not) with op-amps or discrete parts, even using tube circuits which may even be including (partial) post filtering or not.
It is easy to just follow the directions laid out by the DAC chip manufacturer but requires the usage of specific parts and deviating may result in poorer performance.
A lot of audio manufacturers feel they know better or want to do it differently and believe in 'house sound' or use cheaper (less suited) parts.

This is where a potentially good DAC chip performance can be ruined by choice of not suitable alternatives, post filtering or silly designs based on 'theories' of designers.

It is why some DAC devices that use the exact same DAC chip (and parts before it) can have substantially different technical performance.
Design choices or component choices made on whatever motivation of a DAC device manufacturer determine the final result.

This is what is meant with: The analog output stage is important. It is easier to mess it up than it is to get it right. Sometimes even when following exact recommendations of the DAC chip application manuals.

The audibility of this all is another matter though. This could range from 'psychological' to 'real' issues.

 

[Atanasi]

Can someone explain to me what that means? In my book, without voltage there is no current, so I guess it's a kind of engineer jargon, that I fail to grasp. Thanks!

Voltage sources and current sources are two kinds of models in electronics, like mirror images of each other.
Voltage sources are more familiar, they are supposed to maintain a particular voltage between their terminals, regardless of the load in between. Power supplies are an example of voltage sources.
Currrent sources are supposed to maintain a particular current between their terminals, regardless of the load.
Both kinds of sources have limits: if a voltage source is short-circuited, the source cannot maintain its voltage anymore. Likewise, if a current source is connected to a very high resistance, like an open circuit, the current cannot be maintained.

Voltage and current sources can either be constant or driven by a signal. Analog audio outputs are voltage sources where the output voltage is driven by the signal.
In the case of DACs, a DAC chip with a current output acts as a current source where the output current is driven by the signal.
The output stage turns the current source into a voltage source.

 

[xaviescacs]

Voltage sources and current sources are two kinds of models in electronics, like mirror images of each other.
Voltage sources are more familiar, they are supposed to maintain a particular voltage between their terminals, regardless of the load in between. Power supplies are an example of voltage sources.
Currrent sources are supposed to maintain a particular current between their terminals, regardless of the load.

This is very neat. Thanks.

If I understand correctly, in the end, both apply a potential difference between the terminals to create a current, basically because it's the only way it works. So the difference, what makes them behave differently, is the impedance between the terminals or what? What makes something a current source or a voltage source?

 

[solderdude]

If I understand correctly, in the end, both apply a potential difference between the terminals to create a current,

A current source provides current, not a voltage potential on its output (what is needed for audio). The potential difference on the output of a current source depends on the resistor that loads a current source. There are practical limits and the resistance must be of a fixed resistance so can not be the RCA or XLR outs !
For this an I/V converter is essential when linearity is desired.
The output of the I/V converter is a voltage source (which all audio signals are) regardless of load resistance and is limited by the present power supply voltage rails and current limits.

A voltage source provides a voltage. Regardless of load resistance and is limited by the present power supply voltage rails and current limits.
Voltage to current conversion can be a simple resistor but only into a 0ohm load. In practice active components are used to create a linear current..

So very different circuits with very different properties and are not directly interchangeable and require conversion. Using only a resistor without any active components around it will result (in practice) in non-linear (= distortion) behavior.

 

[xaviescacs]

Thanks @solderdude.

 

[nyxnyxnyx]

Thank you guys for explaining the process behind making a DAC and how or why the "analog section" works. My knowledge on electrical components (or electricity in general) is shamefully poor so I have to say I will need to learn more to grasp everything we posted in this thread. @solderdude @voodooless @Atanasi

However after reading the main points I have another question related to the subject: Can a DAC (or other audio devices) gain significant improvements OR make up for said "not suitable alternatives" (both in audible or inaudible zones) by using finest/most expensive components possible? For example the volume pot, one can buy in bulks from Aliexpress for very cheap prices, but there also pots made in Europe and Japan that seem to be built in better quality and they are also way more expensive. Same to capacitors and resistors etc...

I see not only brands, especially Hi-End brands seem to advertise that their products use the finest components possible, but even avid DIY-ers like to opt for that option if they can too. One sighted example I can tell is the Kevin Gilmore KGSSHV on diy forums.

 

[solderdude]

Can a DAC (or other audio devices) gain significant improvements OR make up for said "not suitable alternatives" (both in audible or inaudible zones) by using finest/most expensive components possible?

Finest/most expensive does not mean best performance.
In some cases (one needs to know about the design and will need schematics or a LOT of experience) simply selecting better suited parts, which may even be cheaper than what's in there, or changing some values one could potentially increase measured performance.
An increased measured performance may not be audible.

In case of 'modders' the mods could even lower measured performance or knowing what's in there could alter the perception of the individual resulting in 'improved audio'.

Changing analog volpots could lower noise (when adjusting the volume) or have better L-R balance, particularly at lower levels. Better SQ/technical performance... no not really but again, knowing what has been replaced will alter the perception of the owner.

Of course the whole 'capacitor' story could potentially lead to better measured performance when very crappy or not suited parts were used to begin with.
In most cases changing parts for more expensive (audiophool) parts will lead to a guaranteed improved sound quality of the listener regardless if technical improvements have been made. That's how it seems to work for audio.

 

[DVDdoug]

Most "things" are voltage sources and most amplifiers (and pre-amps, etc.) are considered "constant voltage". The voltage isn't really constant because it depends on the volume control, but under normal conditions the voltage is independent of the load. It puts-out a voltage with no speaker connected (open circuit = infinite resistance) but no current flows out. Resistance (and impedance) resist current flow. Connect an 8-Ohm speaker and current flows. Connect a 4-Ohm speaker and as long as the voltage holds-up you get twice the current and twice the power. Some amplifiers aren't capable of double the current and in that case the voltage will drop and you'll get less power (and clipping unless you turn it down). If you connect a 2-Ohm speaker you might get excess current and the amp might shut down or burn-up.

...If you live in the U.S. there is always 120VAC at the power outlet (assuming it's not switched). When you plug something in, current flows. It you plug-in too many things you get excess current, the circuit breaker blows and voltage drops to zero. With a heavy load you can get a small voltage-drop caused by the resistance in the wires.

Most power supplies are also constant voltage. How well it can hold the voltage depends on how well it's regulated, or it's internal source resistance if it's unregulated.

Similarly, the source resistance of a headphone amplifier can affect how much voltage you can get out of it (depending on the headphone "load" impedance).

There are some constant-current supplies, most often used for LEDs which are "current driven". With no load (infinite resistance and zero current) the voltage jumps to the maximum the power supply can put-out. Unlike a regular power supply, a constant-current power supply can usually be safely shorted (zero load resistance). The voltage simply drops to (nearly) zero and the specified current flows.

Ohm's law defines the relationship between voltage, resistance and current: Current = Voltage / Resistance.

 

[giordy]

beautiful thread ....
I follow with interest

 

[solderdude]

This is why good lab powersupplies have a foldback feature.

 

[MAX232]

Bit late to the party, but i found this thread coming from another question: "how can we find out if a dac is constructed with a quality analogue output stage". As Solderdude explains, there is a obvious way to build this in the cheaper way (less parts, smaller board etc). How can we (non-electricians) find out if a DAC is build with a not-cheap output stage?

 

[notsodeadlizard]

Because a DAC is a gateway between the digital and analog worlds, the analog part of a DAC (including its powering) determines virtually all the resulting analog indicators.
It was obvious.
And everything else is too complex set of details for the professional analog designers.
If you are going to develop your DACs, this will be one story. If you use ready-made DACs, these are not very interesting details at all, because you will not fix anything in the DAC that you use (most often this will require a radical change in circuitry, PCB and powering).

It's good that the DAC is not a very complex device (that's why Chinese cooperatives compete in the assembly of DACs), 80+% of the functionality is implemented by a DAC IC and all that can be done in the best case is not to worsen the parameters of this IC.
Here is an example of a DAC developed long ago by a competent amateur based on AKM DAC IC, especially the analog output stage, and it's a good output stage by many reasons:

Arduino controlled Dual Mono AK4490 DAC (Part 3) | Dimdim's Blog

For the user, the main parameters declared by the manufacturer and the indicated nuances of circuitry are sufficient.
In any case, if the manufacturer claims honest (corresponding to measurements correctly performed for serial products) device indicators, this is an honest manufacturer.

 

[DSJR]

Up until several years back, I'd say that dac designers could get away with various tweaks to get a kind of sonic 'personality' from the dacs. Rega certainly did (or still do) and it was a nice listening experience while getting a then competent technical performance (i suspect it's way outclassed now in every way due to price - especially exported - and massive tech improvements in sub $/£500 dacs in recent years).

These days, dacs are appliances you can buy on facilities, looks and 'feel' I suspect and with that new Topping LA90D amp box reviewed today, smaller amps are becoming the same (takes all the fun out of it for us oldie hobbyists who liked to pick and choose ).