Ladder Schumann R2R DAC Review

[Henry53]

You'd be heavily surprised with the age demographics outside of speakers (headphones and specially IEMs). I'm leaning into the older side of that spectrum at 27 and still can hear up to 17khz. Moreover, I personally know a lot of people into their late teens and early twenties (hell, I entered the "audiophile" world at 13 with my HD600...) that own a lot of the new chinese iems that broke the market as of late.

Given, almost no one of that demographic will be interested in a product like this though: too expensive and bulky.

You are 27 and can hear up to 17Khz and it appears prominently a headphone user, I think you prove my point. Concluding a DAC used in a home stereo system has 0.00% less distortion bears no relation whatsoever to what can be heard.

 

[antcollinet]

No it doesn't. Digital circuits only consume current when they transition from one logical state into another. This transition is so short that a simple ceramic 100 nF capacitor accross the IC takes care of that. In case of FPGAs, there are multiple power supply pins, each with their own 100 nF capacitor. That is all there is to it. These massive toroids are just for show and "paper weight".

Say what now?

So are you saying a digital circuit that is outputting say 5V half the time, and 0V the other half, into a load, is only consuming power when it transitions, and not when outputting 5V?

That would be a pretty miraculous circuit.

And the 90W or so used by a typical CPU - what is that then? Last time I looked, CPUs were also pretty damn close to 100% digital.

 

[mhsprang]

Say what now?

So are you saying a digital circuit that is outputting say 5V half the time, and 0V the other half, into a load, is only consuming power when it transitions, and not when outputting 5V?

That would be a pretty miraculous circuit.

And the 90W or so used by a typical CPU - what is that then? Last time I looked, CPUs were also pretty damn close to 100% digital.

I think you should do some reading on how CMOS circuits work. And the comparison with a PC CPU is daft because such a CPU runs on mich higher frequencies and never stops working.

I design digital circuits for a living and the IC's and CPUs I use consume in the mA range, even when running at 120 MHz internally, and when put to sleep they consume in the uA range. A CPLD clocked at 49 MHz and outputting samples at less than 500 kHz is asleep 90% of the time, so to speak. The average consumed power is what a power supply delivers, the peaks for swithing levels are delivered by the decoupling capacitors, not by the power supply. The inductance of the DC path from power supply to the CPLD and other IC's would be prohibitive of supplying the switching peaks due to their inductance. The large capacitor banks won't help either, their ESR is still far too high at 50 MHz.

Most of the current drawn by this DAC will be driven into the R2R network and how much that is, depends on the resistor values used. Look at the stated power consumption of that DAC: 20 Watts. That transformer inside has the size of an 80 Watt toroid. Total overkill, just paperwaight to give the unit a "premium" feel.

 

[antcollinet]

I think you should do some reading on how CMOS circuits work. And the comparison with a PC CPU is daft because such a CPU runs on mich higher frequencies and never stops working.

I design digital circuits for a living and the IC's and CPUs I use consume in the mA range, even when running at 120 MHz internally, and when put to sleep they consume in the uA range. A CPLD clocked at 49 MHz and outputting samples at less than 500 kHz is asleep 90% of the time, so to speak. The average consumed power is what a power supply delivers, the peaks for swithing levels are delivered by the decoupling capacitors, not by the power supply. The inductance of the DC path from power supply to the CPLD and other IC's would be prohibitive of supplying the switching peaks due to their inductance. The large capacitor banks won't help either, their ESR is still far too high at 50 MHz.

Most of the current drawn by this DAC will be driven into the R2R network and how much that is, depends on the resistor values used. Look at the stated power consumption of that DAC: 20 Watts. That transformer inside has the size of an 80 Watt toroid. Total overkill, just paperwaight to give the unit a "premium" feel.

I'll not dispute anything about the need for a giant power transformer.

I'm sort of disputing your claim that digital circuits "only" consume current during switching. It may be that in low power CMOS which is only driving other low power CMOS, the switching currents vastly outweigh on state currents. But even in CMOS ICS on state currents exist. Even ignoring that, your claim was about Digital circuits, not CMOS Circuits. I'm sure you'd agree that CMOS IC's are not the only form of digital circuit, and it is quite possible for digital circuits to include loads which have significant on state power.

 

[mhsprang]

I'll not dispute anything about the need for a giant power transformer.

I'm sort of disputing your claim that digital circuits "only" consume current during switching. It may be that in low power CMOS which is only driving other low power CMOS, the switching currents vastly outweigh on state currents. But even in CMOS ICS on state currents exist. Even ignoring that, your claim was about Digital circuits, not CMOS Circuits. I'm sure you'd agree that CMOS IC's are not the only form of digital circuit, and it is quite possible for digital circuits to include loads which have significant on state power.

All modern mainstream digital IC's are made in CMOS technology. CPLD, FPGA and CPU current consumption formulas all include a small constant current plus a part that is directly proportional to operating frequency.

These circuits could not even be made in biplolar technology because of the huge power consumption they would then exhibit.

 

[YSC]

You'd be heavily surprised with the age demographics outside of speakers (headphones and specially IEMs). I'm leaning into the older side of that spectrum at 27 and still can hear up to 17khz. Moreover, I personally know a lot of people into their late teens and early twenties (hell, I entered the "audiophile" world at 13 with my HD600...) that own a lot of the new chinese iems that broke the market as of late.

Given, almost no one of that demographic will be interested in a product like this though: too expensive and bulky.

Actually I am at 39, soon 40, and last time I check I can still hear 18.5khz... it isn't that uncommon tbf