Posted 25th January 2014 at 01:28 AM byabraxalito Updated 6th February 2014 at 03:49 AM byabraxalito
Last night I finished building the second channel of my dual mono approach to DAC building which I've called 'free radical'. Here's a picture of the second channel's build just prior to adding all the crapacitors (cheap shanzai 'Sanyos' which measure extremely well). While building the second channel I was listening to the first in mono and that was a spur to quick completion
From left to right there's the AD605 with its top hat array of MLCCs - outputs are isolated via ferrite bead chokes from the AD8017 under its own pile of SMT ceramics. In between the two active stages are the capacitors associated with power supply reference voltage filtering. I realized from the previous build that as the AD605's gain is controlled by DC voltages, these voltages need to be low noise to ensure gain stability. Hence lots of RC filtering with those Nichicon and Rubycon low ESR lytics. I'm using BC817 transistors as low drop-out regulators and the reference voltages (2.5 and 5.7V) come...
Last time, we'd got to a functional voltage regulator, with a pass transistor and op amp error amplifier but I cheated and used ideal voltages for the reference and op amp power supply.
This time I've sketched out a functional circuit using real parts found in the LTSpice library. I've chosen a rail-to-rail op amp to avoid problems with low voltage references. The LT1009 reference puts out 2.5 V, the op amp gain is 4, for an output of 10 V into 1 kohms.
Two versions of the circuit are included below. Voltageregulator5 has some additional RC filter stages to remove noise from the reference and op amp power supply Voltageregulator5b just takes everything straight from the input voltage. As you can see there's a fairly substantial advantage gained from judicial use of RC filtering.
So that's the end of Term 1. The basics have been covered, however briefly. I encourage you to download the LTSpice files and play...
At the end of Part 3 I promised to introduce feedback, and I will, but what we are really talking about here is the addition of the error amplifier, the heart of all modern series pass regulators. The error amplifier is a non-inverting DC signal amplifier, and its function is simple: amplify and buffer the reference voltage. The twist is that the amplifier output is connected to the base of the pass transistor, while the feedback connection is taken from it’s emitter. The pass element is thus placed inside the feedback loop of the error amplifier, improving the ripple rejection and output impedance of the regulator dramatically.
So here we are, the three building blocks of a voltage regulator are in place: the reference voltage, the error amplifier, and the pass device. In the LTSpice circuit I’ve cheated, deliberately, in order to make the operation easier to follow. Instead of building a practical voltage reference I’ve...
I’m going to have to make a detour to point out what we are doing here is learning how these circuits work, and get a very rough idea of their relative merits. We’re not trying to minimize the output impedance, or maximize the ripple rejection. Three reasons immediately come to mind for why it would be bad practice to try and do that:
1. Any such contest will be easily won by the largest capacitor placed on the regulator output.
2. There are clear limits on these parameters after which further “improvement” is unlikely to serve any useful purpose.
3. There are other considerations such as the output noise of the regulator and the stability under dynamic loads, which are equally if not more important.
Clear? No cookies for the “most bestest” circuit in LTSpice. The great utility of LTSpice is it allows you, the designer, to easily check if you’ve left performance...
Now how much of its improved SQ over rivals three or more times the price is going to be down to the active bass section offloading the most PSU-draining signals from the driving poweramp? The comments about the scale of the soundstage do reflect the kinds of improvements I've been getting by reducing LF noise in my DAC, so the reduced LF noise from the poweramp from having a more benign load to drive could indeed be key. 92dB efficiency certainly helps a lot in reducing poweramp PSU stress.
That article says this speaker is 'sure to shake up the industry'. Really? What do you guys think?
Instead of pulling the output current through R1, we add an npn pass transistor, Q1. The output current now "passes" through the transistor, while the Zener diode still regulates the output voltage by being connected to the transistor base.
The output impedance falls to a few ohms, but the ripple rejection improves only slightly.
This is a useful basic circuit block for audio, but the ripple rejection can easily improved by the addition of a couple of additional components, as we'll see shortly.
Posted 15th January 2014 at 11:34 PM byrjm Updated 16th January 2014 at 04:40 AM byrjm
This is the first of a series, where I will be investigating the output impedance and ripple rejection of various voltage regulator circuits using LTSpice.
Today, for the first "lesson" (I'm teaching myself, as much as anything) we will look at the very simple zener voltage regulator.
The load is 1 kohm, and the Zener breakdown voltage is 12 V. The load current is about 10 mA, and to avoid gross inefficiency we will limit the current flowing through the Zener to about 5 mA, by adjusting R1 accordingly. The input voltage is fixed at 18 V.
To measure the ripple rejection, we perform an AC analysis with the voltage source AC set to 1 and the current source AC set to zero. The ripple rejection is the negative value of the signal at Vout: so -20 dB means 20 dB ripple rejection (1 V ripple at Vin generates 0.1 V ripple at Vout at a given frequency.).
To measure the output impedance, again the AC analysis function is used but...
John Atkinson's 'WOW!' (more than once) to his measurement results from the Vivaldi DAC prompted me to have a closer look at how it does for noise modulation.
The AP's FFTs shown don't come with details about the number of points in the FFTs, so noise estimates are a tiny bit tricky. However there is enough detail to make some reasonable estimates.
I've attached the plot from which I'm making my estimates - if anyone notices I've made a slip-up, please do comment and correct me.
The red line shows white noise at peak level of -4dBfs. I generated white noise in Audacity at this level to compare - with the maximum 16k point FFT and BH windowing, I got the same level of noise as on this plot - -42dBfs in 22kHz bandwidth. That suggests to me that the FFT shown has 64k points.
Another way to estimate the noise is by looking at the difference between the blue plot (19.1kHz, 0dBfs) grass and the red. To my eye, the difference in level is 76dBr...
I've had the beta build Ozone (no input stage yet, fed from my QA550 via the I2S transcoder to down-convert to 32fs) playing out 24/7 for a few days now.
Overall I'm very happy with how it sounds, just a minor gripe about sibilance on some operatic vocals which I'd like to understand better. On the upside the jump-factor (read dynamics) and soundstage stability (holographic on the right disks) are about the best yet. I'm using a couple of Decca double CDs for this - 'La Traviata' and "La Boheme'. They're about the most transparent sounding and demanding disks I have. Demanding in the sense that they have lots of emotional drama which should positively engage my attention if the DAC's really up to snuff. I never much enjoyed opera until I got into building my own NOS DACs, but now I really enjoy my dramatic fixes and these two recordings are really top of the pile. They're about 50 years old but so far I've not found anything newer which touches them (not that I've looked...