sq, I had to manually edit your post because of your attempt to get around the obscenity filter. That makes work for me. When I have to work, I get very, very grumpy.yeah i suppose the spec is ok... Lol!
Thats pretty much exactly everything i need or want in a DAC preamp!
People have been waiting a long time for this to come out, im sure it will sound great and it will be a big success, if it comes in at under £500 i'll be very shocked, whatever the price is, i cant wait to hear it.
The other thing is, looking back through the thread it obvious to me that John W is not only a talented designer etc, but is a true enthusiast helpful and generous with his time and knowledge, an all round good egg!
Looking forward to the amps too (even if they take some time
)
best, John
Thats pretty much exactly everything i need or want in a DAC preamp!
People have been waiting a long time for this to come out, im sure it will sound great and it will be a big success, if it comes in at under £500 i'll be very shocked, whatever the price is, i cant wait to hear it.
The other thing is, looking back through the thread it obvious to me that John W is not only a talented designer etc, but is a true enthusiast helpful and generous with his time and knowledge, an all round good egg!
Looking forward to the amps too (even if they take some time
)best, John
I thought SQ was kind of clever with that all important "5".... maybe I'm behind the times... I can't even manage to send an SMS...
Anthony.
The MDAC is just going into Pre-production now – so unfortunately unlikely to reach the Sunny Shores of UK before Christmas - there’s little margin to Air Freight them...
John,
Well occasionally life can surprise - price has yet to be confirmed, however we are pushing for 5 Pence under GBP400.... if you get my drift... I'm just a little over BOM cost target… and asking for forgiveness (without having to reducing the products Specs).
Betto,
The amplifier will be released end of first quarter 2010 – however I expect demand will outstrip supply for a very long time – it’s not going to let the side down…
I didn’t mention the CD players with DAC / Pre-Amplifiers did I?
John
Anthony.
The MDAC is just going into Pre-production now – so unfortunately unlikely to reach the Sunny Shores of UK before Christmas - there’s little margin to Air Freight them...
John,
Well occasionally life can surprise - price has yet to be confirmed, however we are pushing for 5 Pence under GBP400.... if you get my drift... I'm just a little over BOM cost target… and asking for forgiveness (without having to reducing the products Specs).
Betto,
The amplifier will be released end of first quarter 2010 – however I expect demand will outstrip supply for a very long time – it’s not going to let the side down…
I didn’t mention the CD players with DAC / Pre-Amplifiers did I?
John
Thank you John, I am a happy man now
I am expecting the amp to be on par with the DAC...
Current dumping class A ... what a clever idea... finally done right!!
Now to put the cherry on the cake you really should work on a loudspeaker project...some grounbreaking design ... a full range panel speaker based on BMR concept for example. Why not?
All the best from rainy Brescia, northern Italy
Betto
I am expecting the amp to be on par with the DAC...
Current dumping class A ... what a clever idea... finally done right!!
Now to put the cherry on the cake you really should work on a loudspeaker project...some grounbreaking design ... a full range panel speaker based on BMR concept for example. Why not?
All the best from rainy Brescia, northern Italy
Betto
John
its surely not controversial to say that the (forseeable) future of recorded music for playback is the computer hard disk. Given that, I'd be very interested to hear your views on how the sonic performance of the new DAC differs when comparing spdif input from a CD transport to the USB input from hard drive - ?
Regards
Ian
its surely not controversial to say that the (forseeable) future of recorded music for playback is the computer hard disk. Given that, I'd be very interested to hear your views on how the sonic performance of the new DAC differs when comparing spdif input from a CD transport to the USB input from hard drive - ?
Regards
Ian
Betto,
I'm an ESL Man - nothing comes close.... I'm working on a pair of speakers for Peachtree Audio (based upon Ribbon Drivers), but that will be for later next year.
Ian,
Indeed - Computer playback / IPod / IPhone is the future... Erh... soon to be reviewed in a magazine in Dec or Jan issue of..... (I'm not sure how much I can say - so sorry for being so cryptic).
If the Data Stream is Bit Accurate - then ideally the sound quality should be the same whatever the replay device - if it where not for:-
1. JITTER of the Data source
2a. RF noise injection from source to DAC (Directly via RF Emissions, Mains and Signal cabling).
2b. Noise injected into DAC via Earth leakage currents and Ground loops.
Clock locking and Asynchronies USB go a long way to solving issue 1 (MDAC has both options), however the effects of issues 2a & 2b will also effect the Phase Noise of the DAC’s internal Master Clock (worsen the Clock Jitter levels via external noise injection onto the DAC’s Ground plane / PSU etc).
MDAC has a novel circuit block I’ve called a “Cascaded Asynchronies Time Domain Attenuator” to help reduce the detrimental effects of non synchronies clock sources (i.e. when using the MDAC without clock-lock or Asynchronies USB) – Still nothing beats a correctly synchronised Data Source.
The subject of Jitter, RF and Ground noise is endless….
Betto,
Yes the weather’s also been really lousy here in CZ, has not stopped raining – car’s covered in Mud as it’s churned up our once grass drive…
10 years ago or so, I spent 2 years developing and understanding “Analogue amplifiers”. After 2 years of development – and achieving results better than I could measure – I realised that really Analogue amplifiers where not the way to go – and then started R&D on switching amplifiers. However at the time, I fell in love with the elegance of Current Dumping – and could not understand why it had not been developed further. Current Dumping to this day is still very miss-understood – 10 years back I achieved breathtaking performance with the topology.
When Peter Walker (and the other Co-designer (sorry I currently forget his name)) developed the original Current-dumping circuit, transistors where a major part of a units cost, so the design was inherently compromised. Today, the cost of transistors constitutes very little to the total BOM cost of a design, allowing far greater design flexibility – a luxury that Peter Walker and Co. did not have.
However, one area which is still common today manufacturing in China (if not more so), and manufacturing 30 years ago in England is insuring stable and consistent performance over the products entire production life – unit to unit. Adjusting OPS bias Current is strictly a no – no. This is an area where current dumping really scores…
The limitations of Current dumping are not in its topology – but its cost constrained 30 year old implementation.
One of Class A/B’s greatest issue is accurately setting the bias current – which amazises me, as during my "R&D" 10 years back I developed a VERY simple circuit to perform the task that would added less then 10 cents to a standard design’s cost – in fact cheaper then the cost of the Bias POT… Its really poor refection that Class A/B designs still rely on Bias settings… I hope one day to release a commercial design with the simple “Bias circuit”… just Pls. don’t ask me when…
If you where about to ask, one of the best sounding amplifiers I’ve ever built was Class D, however there’s just to much public negative pre-judgment towards Hi-Fi Class D, the reason I chose to use a 10 year old development of Current Dumping for the coming amplifier.
John
I'm an ESL Man - nothing comes close.... I'm working on a pair of speakers for Peachtree Audio (based upon Ribbon Drivers), but that will be for later next year.
Ian,
Indeed - Computer playback / IPod / IPhone is the future... Erh... soon to be reviewed in a magazine in Dec or Jan issue of..... (I'm not sure how much I can say - so sorry for being so cryptic).
If the Data Stream is Bit Accurate - then ideally the sound quality should be the same whatever the replay device - if it where not for:-
1. JITTER of the Data source
2a. RF noise injection from source to DAC (Directly via RF Emissions, Mains and Signal cabling).
2b. Noise injected into DAC via Earth leakage currents and Ground loops.
Clock locking and Asynchronies USB go a long way to solving issue 1 (MDAC has both options), however the effects of issues 2a & 2b will also effect the Phase Noise of the DAC’s internal Master Clock (worsen the Clock Jitter levels via external noise injection onto the DAC’s Ground plane / PSU etc).
MDAC has a novel circuit block I’ve called a “Cascaded Asynchronies Time Domain Attenuator” to help reduce the detrimental effects of non synchronies clock sources (i.e. when using the MDAC without clock-lock or Asynchronies USB) – Still nothing beats a correctly synchronised Data Source.
The subject of Jitter, RF and Ground noise is endless….
Betto,
Yes the weather’s also been really lousy here in CZ, has not stopped raining – car’s covered in Mud as it’s churned up our once grass drive…
10 years ago or so, I spent 2 years developing and understanding “Analogue amplifiers”. After 2 years of development – and achieving results better than I could measure – I realised that really Analogue amplifiers where not the way to go – and then started R&D on switching amplifiers. However at the time, I fell in love with the elegance of Current Dumping – and could not understand why it had not been developed further. Current Dumping to this day is still very miss-understood – 10 years back I achieved breathtaking performance with the topology.
When Peter Walker (and the other Co-designer (sorry I currently forget his name)) developed the original Current-dumping circuit, transistors where a major part of a units cost, so the design was inherently compromised. Today, the cost of transistors constitutes very little to the total BOM cost of a design, allowing far greater design flexibility – a luxury that Peter Walker and Co. did not have.
However, one area which is still common today manufacturing in China (if not more so), and manufacturing 30 years ago in England is insuring stable and consistent performance over the products entire production life – unit to unit. Adjusting OPS bias Current is strictly a no – no. This is an area where current dumping really scores…
The limitations of Current dumping are not in its topology – but its cost constrained 30 year old implementation.
One of Class A/B’s greatest issue is accurately setting the bias current – which amazises me, as during my "R&D" 10 years back I developed a VERY simple circuit to perform the task that would added less then 10 cents to a standard design’s cost – in fact cheaper then the cost of the Bias POT… Its really poor refection that Class A/B designs still rely on Bias settings… I hope one day to release a commercial design with the simple “Bias circuit”… just Pls. don’t ask me when…
If you where about to ask, one of the best sounding amplifiers I’ve ever built was Class D, however there’s just to much public negative pre-judgment towards Hi-Fi Class D, the reason I chose to use a 10 year old development of Current Dumping for the coming amplifier.
John
Last edited:
It can be whatever you like (e.g. SD card reader) as long as it has clock- lock connection.
@JohnW
Now, since you have moved on from CS8414x -> PMD100 -> data splitter -> 2 x SAA7350 -> LPF with discrete opamps, could we have schematic in near future?
I'm asking this because I just love those SAA7350 and the "raw Bitstream" L/R output pins thay have so one can use 74AC175 per channel as a differential dac with completely passive LP filter at output (so, no active electronics in site) like in the virtual pcb image below.
Hi Loris,
As Aparatusonitus has already answered, any Digital source where the replay Master Clock can reside within the DAC - and be transmitted (sent) to the Data "Source".
So Clock Locked CD Transport, Async USB - or any device that can accept its Clock FROM the DAC. (DAC integrated within a CD player).
I believe you can even Sync PC sound cards by returning a SPDIF "BLACK" (no Data, just clock) to the sound card from the DAC (MDAC does not support this feature as we expect most people will be happy with Async. USB).
Aparatusonitus,
I can see you have spent a lot of time on your DAC’s Layout.
You’re basically developing a "Dacapo" Discrete DAC. A couple of points:-
The SAA7350 Bitstream outputs need to be re-formatted to "Return to Zero" Bitstream coding; otherwise you will have very poor THD+N due to "incorrect Pulse energy". Basically you need to logic "And" with the Master Clock - you also need to generate the Negative Data Pulse.
+ Pulse = Clock High & Data High
- Pulse = Clock High & Data Low
So you only can have a resultant Pulse (+ or -) if the clock is high - resulting in half pulse integration energy per clock cycle - but without "Edge Rise/ Fall time errors (If Balanced) & Missing Edge" distortion.
After Generating the 2 Phases from the Bitstream, you should then Latch to a Master clock 16.9344MHz, (in the case of the SAA7350 the Bitstream modulator is operating at half MCLK) - to resynchronise to MCLK and remove the Phase Noise (Jitter) from the SAA7350 outputs & subsequent Bitstream reformatting circuit. If you use a Dual D-FF with Q and Q-, then you can end up with 4 "Dac Elements" that can be finally summed.
The Bitstream outputs of the SAA 7350 are very heavily jittered - if you simply integrate its outputs directly you will hear all kind of "Birdies" due to its internal processing "Jittering" the outputs (typically heard as the Music ramps up and down in Level – such as the start of a track).
These 2 reclocked Output Phases need to be individually filtered then "differentially Summed" via an Opamp.
Remember that your are integrating the final Pulse energy - so anything that can affect the pulse area will cause Distortion and Noise - Clock Phase Noise, PSU quality (noise) on the Final D-Latch rails, Resistor Noise etc.
I noticed that you have used Ceramic Capacitors for the LPF - while these are best for RF (which there is a massive amount from the Bitstream), I strongly suggest PP Film Foil Type, with maybe a parallel 100pF (Parallel with the larger valve PP Cap) to Ground on the first stage of filtering.
The Latch requires very good PSU decoupling - its PSU PINS ARE the DAC Vref Node. Good Quality Ceramics directly across its VDD to Ground (Think about the decoupling Loop Path).
For latches I recommend using Fairchild’s UHS 74 (Single Latch) or VHC74 Dual Latch - with the Dual Latches you can be clever and arrange the Data Phases to "Null" the Latches internal "Data induced" PSU modulation.
John
As Aparatusonitus has already answered, any Digital source where the replay Master Clock can reside within the DAC - and be transmitted (sent) to the Data "Source".
So Clock Locked CD Transport, Async USB - or any device that can accept its Clock FROM the DAC. (DAC integrated within a CD player).
I believe you can even Sync PC sound cards by returning a SPDIF "BLACK" (no Data, just clock) to the sound card from the DAC (MDAC does not support this feature as we expect most people will be happy with Async. USB).
Aparatusonitus,
I can see you have spent a lot of time on your DAC’s Layout.
You’re basically developing a "Dacapo" Discrete DAC. A couple of points:-
The SAA7350 Bitstream outputs need to be re-formatted to "Return to Zero" Bitstream coding; otherwise you will have very poor THD+N due to "incorrect Pulse energy". Basically you need to logic "And" with the Master Clock - you also need to generate the Negative Data Pulse.
+ Pulse = Clock High & Data High
- Pulse = Clock High & Data Low
So you only can have a resultant Pulse (+ or -) if the clock is high - resulting in half pulse integration energy per clock cycle - but without "Edge Rise/ Fall time errors (If Balanced) & Missing Edge" distortion.
After Generating the 2 Phases from the Bitstream, you should then Latch to a Master clock 16.9344MHz, (in the case of the SAA7350 the Bitstream modulator is operating at half MCLK) - to resynchronise to MCLK and remove the Phase Noise (Jitter) from the SAA7350 outputs & subsequent Bitstream reformatting circuit. If you use a Dual D-FF with Q and Q-, then you can end up with 4 "Dac Elements" that can be finally summed.
The Bitstream outputs of the SAA 7350 are very heavily jittered - if you simply integrate its outputs directly you will hear all kind of "Birdies" due to its internal processing "Jittering" the outputs (typically heard as the Music ramps up and down in Level – such as the start of a track).
These 2 reclocked Output Phases need to be individually filtered then "differentially Summed" via an Opamp.
Remember that your are integrating the final Pulse energy - so anything that can affect the pulse area will cause Distortion and Noise - Clock Phase Noise, PSU quality (noise) on the Final D-Latch rails, Resistor Noise etc.
I noticed that you have used Ceramic Capacitors for the LPF - while these are best for RF (which there is a massive amount from the Bitstream), I strongly suggest PP Film Foil Type, with maybe a parallel 100pF (Parallel with the larger valve PP Cap) to Ground on the first stage of filtering.
The Latch requires very good PSU decoupling - its PSU PINS ARE the DAC Vref Node. Good Quality Ceramics directly across its VDD to Ground (Think about the decoupling Loop Path).
For latches I recommend using Fairchild’s UHS 74 (Single Latch) or VHC74 Dual Latch - with the Dual Latches you can be clever and arrange the Data Phases to "Null" the Latches internal "Data induced" PSU modulation.
John
Last edited:
Hi John,
as you may remember I have been waiting the digital amp for years, but for now I think I'll be very comfortable with the current dumping amp
BTW, how powerful the new amp is going to be?
Thinking about loudspeakers, as much as I am a fan of dipoles, I find very much intriguing the concept of Balanced Mode Radiator loudspeaker by Graham Bank: a wide band, constant directivity design which is about to be launched by Naim (not heard yet though).
Regards, Betto
as you may remember I have been waiting the digital amp for years, but for now I think I'll be very comfortable with the current dumping amp
BTW, how powerful the new amp is going to be?
Thinking about loudspeakers, as much as I am a fan of dipoles, I find very much intriguing the concept of Balanced Mode Radiator loudspeaker by Graham Bank: a wide band, constant directivity design which is about to be launched by Naim (not heard yet though).
Regards, Betto
Hi John,
as you may remember I have been waiting the digital amp for years, but for now I think I'll be very comfortable with the current dumping amp
BTW, how powerful the new amp is going to be?
Thinking about loudspeakers, as much as I am a fan of dipoles, I find very much intriguing the concept of Balanced Mode Radiator loudspeaker by Graham Bank: a wide band, constant directivity design which is about to be launched by Naim (not heard yet though).
Regards, Betto
as you may remember I have been waiting the digital amp for years, but for now I think I'll be very comfortable with the current dumping amp
BTW, how powerful the new amp is going to be?
Thinking about loudspeakers, as much as I am a fan of dipoles, I find very much intriguing the concept of Balanced Mode Radiator loudspeaker by Graham Bank: a wide band, constant directivity design which is about to be launched by Naim (not heard yet though).
Regards, Betto
How then does one use the clock locking feature of the MDAC, and with what sources?
Thanks.
John, just a quick question (hopefully!),
Does the MDAC have the ability to accept AES/EBU digital in? I ask as I would want to use it in a system with a DEQ2496 providing room correction in the digital domain, and then passing the bits on to the MDAC to convert and pre-amp.
Many thanks,
The Moog
Does the MDAC have the ability to accept AES/EBU digital in? I ask as I would want to use it in a system with a DEQ2496 providing room correction in the digital domain, and then passing the bits on to the MDAC to convert and pre-amp.
Many thanks,
The Moog
Hi John,
Some more quick questions if you don't mind. Excuse me sir, if some are bit dumb
What software can the MDAC remote control operate?
Is there going to be a "bit-transparency checker" feature for USB at launch or in the future? USP and genuinely useful I think!
How does selectable 3V unbalanced/6V balanced mode benefit driving power amps direct when 2V RMS is already over the input sensitivity of virtually all power amps?
The data sheet for the MAS6116 says SNR 100dB and THD+N -72dB @ 5V RMS. How does the MDAC analogue pre-amp have 117dB DR and THD -96dB @ 2V?
You say the DAC's sound will take no prisoners even versus anything out there at any price....what's your assessment of the sound of the analogue pre-amp under the same comparison?
Which do you prefer/recommend for best sound from the MDAC - digital or analogue volume control?
Can you give the real lowdown on the disadvantages of digital volume controls to the uninitiated - people say different things. Is it true that you can minus the source bitdepth from the processing bitdepth and have an attenuation range where it acts perfectly? Or that any attenuation will sacrifice something to the end sound?
And please mention some more the CD players with DAC / Pre-Amplifiers! Integrated CD player, DAC and pre? Separate CD player with DAC and separate preamp? Full width to match power amp? Same insides as MDAC? When
Many thank yous.
Some more quick questions if you don't mind. Excuse me sir, if some are bit dumb
What software can the MDAC remote control operate?
Is there going to be a "bit-transparency checker" feature for USB at launch or in the future? USP and genuinely useful I think!
How does selectable 3V unbalanced/6V balanced mode benefit driving power amps direct when 2V RMS is already over the input sensitivity of virtually all power amps?
The data sheet for the MAS6116 says SNR 100dB and THD+N -72dB @ 5V RMS. How does the MDAC analogue pre-amp have 117dB DR and THD -96dB @ 2V?
You say the DAC's sound will take no prisoners even versus anything out there at any price....what's your assessment of the sound of the analogue pre-amp under the same comparison?
Which do you prefer/recommend for best sound from the MDAC - digital or analogue volume control?
Can you give the real lowdown on the disadvantages of digital volume controls to the uninitiated - people say different things. Is it true that you can minus the source bitdepth from the processing bitdepth and have an attenuation range where it acts perfectly? Or that any attenuation will sacrifice something to the end sound?
And please mention some more the CD players with DAC / Pre-Amplifiers! Integrated CD player, DAC and pre? Separate CD player with DAC and separate preamp? Full width to match power amp? Same insides as MDAC? When
Many thank yous.
Sorry for my late reply, despite recently having “bravely” received the seasonal Flu jab, I’ve come down with full blown Flu – and of course while I’m meant to be relaxing back in CZ!
Betto,
I’ve not heard a BMR – I basically have always used ESL’s – which apart for the very lowest of frequencies – perform without equal.
The amplifiers final output power has yet to be determined – electronically it’s well in excess of over 150W PCH – however the final power ratings will be dictated and limited by the thermal performance of the heatsink. The PSU is rated for 850W continues – 1.3KW short term (10 minutes).
SQ,
Forgive me if I don’t go into to much detail at this time (as soon as the MDAC’s are on the high sea’s, I can divulge greater in-depth design details). However, yes there are x256 1Bit DAC elements per channel, arranged into 2x 128 1Bit DAC arrays for each phase of the balanced DAC. Resulting in 512x DAC elements for a stereo channel DAC.
Moog,
By AES/EBU I presume your referring to 110R XLR digital inputs? Due to space limitations* on the rear panel of the MDAC, I had to make the design decision to forgo Balanced Digital XLR inputs (* There’s so little spare panel area that finding space for the silkscreen proved challenging). However, the RCA 75 Ohms inputs are truly balanced (via input Transformer coupling), so it’s a simple matter of converting the 110 Ohm environment into 75 Ohms (while still retaining the advantages of Balanced interconnect) – which can be done with a couple of resistors fitted within the XLR to RCA cable… there’s no real requirement to insure 500mV pk-pk signal levels upon conversion – in fact, within reason the higher the Digital signal amplitude the better.
Fauntleroy,
I’ve left your reply for last as I wonder who you really are – and your true intentions. Your new to this forum (apparently you’ve only recently joined our forum to post solely in this thread) – and the direction (and Tone) of your questing is indicative of possible ulterior motives. I’m happy to support true forum members on honest and open grounds – however possibly in my Flu impaired judgment you seem to have some hidden agenda – perhaps maybe a worried competitor? Or someone with an axe to grind with me for whatever reason?
I guess it’s to be expected that the price / performance and feature set of MDAC is going to have a few competitors worried – with the less scrupulous resorting to “Whispering campaign’s” to protect there own interests.
Anyway, I’ve nothing to hide (have you?), and on the chance that in my current Flu induced state I’ve popped a few too many pills and have lost my sense of reality (more so then usual), I will answer your question this instance. And if I have truly misread you, then please forgive my bluntness’, rudeness etc etc – and I’ll blame it squarely on the excessive pill popping…
The Clock-Lock interface is optical based (the clock is outputted via Toslink). Via the front panel the user can select within a “range” any output clock frequency, permitting fixed synchronization between the MDAC’s internal modulator and external data source. The Clock-Lock frequency is generated via an internal PLL / Digital divider circuit. By insuring the Digital source is synchronized to the modulators clock, allows the internal DPLL to be set to an ultra low BW for greatest performance (Greatest Jitter Attenuation).
I primarily intended the Clock Lock interface to be used with possible future transports from AudioLab – but can be used with older CD transports from Arcam, DPA/ Deltec or the Cambridge Audio CDT I designed (if there are many left working) and possibly other brands I’m not currently aware of.
Any CD transport / Digital Data source that facilitates an external Toslink clock input within the frequency range of the PLL can be used.
I’m happy to publish circuits and possibly design and ship small PCB’s that Forum users can retrofitted to there CD players to allow clock-locking – resulting in “Perfect” digital transports from “any old CD player” that has a Bit accurate SPDIF Data output.
The MDAC not only enumerates over the USB as an Audio Device, but as an HID (Human Interface Device) – via the HID, it sends the standard user playback control commands (Play Pause etc). The operating system sees the MDAC as a Human Interface Device like a keyboard, Mouse etc. Any Media Player (software) that recognizes HID playback control via the OS will work with the MDAC.
USB "bit-transparency checker" was a feature reserved for the original Dacapo replacement – It maybe possible to incorporate a simple USB Bit Accuracy checker into MDAC – I fully agree it would be very useful, however it will solely depend on how much free memory is available within the MDAC MCU after the “Core” MDAC functionality has been finalised – this is in Dominik's domain (the Co-designer of MDAC). If we find methods to compact the code at a later date, it can always be offered as an update – the MDAC allows easy firmware update via its USB Port.
3V / 6V balanced DAC output mode. I’m not sure how to answer your question – as I fail to understand the issue.
1. The MDAC’s Dynamic range is limited by the laws of nature – Thermal noise, and not an “Inherent” system non-linearity as such. By increasing the absolute output voltage level allows us to gain the benefit of its “inherent linearity” – increasing the headroom between thermal noise and Maxim output. This “high dynamic range” signal can be transmitted over a significant distance – to say remote power amplifier or active speakers which have input attenuators to reduce the signal to a level they are comfortable with. Any external interference would have a lesser effect on the 3Vrms signal / then say the 2Vrms signal.
2. As the MDAC can function as both a Digital an Analogue preamplifier, the “Digital Pre Mode” needs some Gain – at 3V rms it’s a little over 3.5dB (Granted – not much) – so in reality the 2Vrms Mode has a small digital scaling factor of -3.5dB or so. With over 132dB dynamic range (Thermal noise, not DAC linearity limited per say), we can afford this small amount of scaling without loosing sleep.
3. It’s the same for Pre-amplifiers that allow higher output levels then 2Vrms to facilitate “Positive Gain” without clipping – you don’t use the Pre-amplifier at 3Vrms, but you need to be able to increase the gain of lower amplitude signals…
Your knowledge and concern for the MAS 6116 while converting my THD % result into dB only heightens my suspicion that you’re not a purely interested forum member.
The DA6116 has an interesting history. Its roots stem from the original MAS9116 (WM8816) which I was first introduced too while consulting for Wolfson on advance DAC architectures. The MAS9116 (WM8816) was fabricated on a small Fab line in a Nordic country – which unfortunately went out of business. The original Fab process had the ability to produce highly linear on-chip resistors – resulting in the original MAS9116 (WM8816) stellar performance. To support there customers and with great credit to VLSI Solutions OY, they where forced at great expense to redesigned the device, this time however on a more standard CMOS process (due to a lack of foundry options) – with CMOS’s inherent non-linear polysilicon resistors. However despite the hit on performance by being forced to a CMOS foundry process, the performance of the DA6116 in balance operation is still very good. The DA6116’s greatest advantage over all other Gain control IC’s is the ability to use external off-chip Hi-performance OPAMP’s – and in the case of MDAC these are operated in pure class A.
The MAS performance results you quoted are measurements with MAS in a single-ended application and at 5Vrms.
This in no way reflects the way the DA6116 is used within the MDAC. Despite the cost implications, the MDAC uses a FULLY BALANCE TRUE DIFFERENTIAL signal path - not one, but two DA6116 are used per design, one “Dual Channel” stepped attenuator per channel (resulting in a fully balanced dual mono attenuator).
During the development of the DA6116 various combinations of resistor switching configurations where tried to try and improve upon CMOS’s inherent limitations – and after much evaluation and listening, it was finally decided upon released into full-scale production the MAS6116AA1 variant.
Pls. see the graph below which is from VLSI Solutions OY internal development and evaluation documentation of the DA6116 in Balanced operation – note that the MAS6116AA1 variant is the current production release standard (lower THD at lower levels – which is more important when listening to real music, not measurement test tones). You can see that at 0dB 2VRMS, the MAS6116AA1 THD hovers around -106dB (and nowhere near the misleading THD+N of -72dB you are trying to suggest).
When the DA6116 is used in true balanced mode, Even-order harmonic products are “Cancelled – read Nulled” to a significant degree. At lower signal amplitudes in single-ended use, the MAS’s dominant distortion components are Even-order (Thus are cancelled via MDAC’s fully balanced signal path) – with Odd-order harmonic distortion components (due to the MAS Poly-Silicon resistors) only becoming dominate at higher signal amplitudes. In fact at 2Vrms, it’s not surprising that it’s the Odd order components that dominate the 0.0015% THD result – as is to be expected for a correctly implemented “Balanced” signal path.
I cannot comment on MAS Single-Ended SNR results (I’ve never tried to use the MAS in SE Mode) – however at 0dB (2Vrms output from the “Analogue” Pre-Amplifier), I can say that the MDAC achieves an 117dB Awtd Dynamic range in Balanced mode.
See the attached FFT plot of the MDAC -60dB Balanced output showing the 132dB Awtd performance, as you can see the performance in (Digital Direct Mode) is limited by the noise floor at below -160dB Ref. 6Vrms
I’ve often been very surprised how poor most pre-amplifiers sound, despite most “Sources” requiring little in the way of Gain – rather attenuation in most applications. The Pre-amplifier is really the most simplest of units in a modern Hi-Fi system, and has no excuse to significantly degrade the audio quality.
To that end – the MDAC’s Line level pre amplifier relies on High Quality Relays to select between the Single Ended Line level Sources, which is then converted into a balanced signal path via a pure Class A Biased stage – then onto the MAS Gain control stage (operating in pure Class A) – Finally to the fully Balanced Folded-Cascode Hi-Current Class A Output Stages.
Each MAS's is fed from its own, ultra low noise and Low impedance “Discrete” regulator – with each channel of the (Dual Mono) Balanced Analogue sections fed from there own “Cascaded” regulated PSU.
So within reason, the preamplifier is simple, all active stages are biased into pure Class A, while operating in true Balanced mode (once past the input selector) – the pre-amplifier does not let the side down – limited ultimately by the Balanced MAS stepped attenuator performance – however at 117dB Awtd… its performance can hardly be described as disappointing
Basically, a pre-amplifier should do nothing other than select the inputs, allow adjustment of the system Gain and ultimately drive the power amplifiers – it should never be heard…
To answer your question (or not) concerning the Digital / verses Analogue gain control, well we have made it user selectable for the very reason that there will always be two opposing schools of thought in the subject. People love to have there own opinions and beliefs – as Peter Walker would always say “Give the people what they want”…
I will say however that the Audio signal path is much simpler (less Complex) in Digital Direct mode.
With 16 Bit Audio there’s plenty of internal Data path width and dynamic range headroom within the DAC that I believe if there is any disadvantage's to Digital Gain control – these are heavily outweighed by the greater complexity of the Analogue pre-amplifier circuit. I would lean towards choosing the Digital Pre Amplifier mode… But it also depends on the system and at what part of the Gain control range you use while listening…
That’s why it’s user selectable…
Once again, if I have misread you, then Pls. forgive me
John
Betto,
I’ve not heard a BMR – I basically have always used ESL’s – which apart for the very lowest of frequencies – perform without equal.
The amplifiers final output power has yet to be determined – electronically it’s well in excess of over 150W PCH – however the final power ratings will be dictated and limited by the thermal performance of the heatsink. The PSU is rated for 850W continues – 1.3KW short term (10 minutes).
SQ,
Forgive me if I don’t go into to much detail at this time (as soon as the MDAC’s are on the high sea’s, I can divulge greater in-depth design details). However, yes there are x256 1Bit DAC elements per channel, arranged into 2x 128 1Bit DAC arrays for each phase of the balanced DAC. Resulting in 512x DAC elements for a stereo channel DAC.
Moog,
By AES/EBU I presume your referring to 110R XLR digital inputs? Due to space limitations* on the rear panel of the MDAC, I had to make the design decision to forgo Balanced Digital XLR inputs (* There’s so little spare panel area that finding space for the silkscreen proved challenging). However, the RCA 75 Ohms inputs are truly balanced (via input Transformer coupling), so it’s a simple matter of converting the 110 Ohm environment into 75 Ohms (while still retaining the advantages of Balanced interconnect) – which can be done with a couple of resistors fitted within the XLR to RCA cable… there’s no real requirement to insure 500mV pk-pk signal levels upon conversion – in fact, within reason the higher the Digital signal amplitude the better.
Fauntleroy,
I’ve left your reply for last as I wonder who you really are – and your true intentions. Your new to this forum (apparently you’ve only recently joined our forum to post solely in this thread) – and the direction (and Tone) of your questing is indicative of possible ulterior motives. I’m happy to support true forum members on honest and open grounds – however possibly in my Flu impaired judgment you seem to have some hidden agenda – perhaps maybe a worried competitor? Or someone with an axe to grind with me for whatever reason?
I guess it’s to be expected that the price / performance and feature set of MDAC is going to have a few competitors worried – with the less scrupulous resorting to “Whispering campaign’s” to protect there own interests.
Anyway, I’ve nothing to hide (have you?), and on the chance that in my current Flu induced state I’ve popped a few too many pills and have lost my sense of reality (more so then usual), I will answer your question this instance. And if I have truly misread you, then please forgive my bluntness’, rudeness etc etc – and I’ll blame it squarely on the excessive pill popping…
The Clock-Lock interface is optical based (the clock is outputted via Toslink). Via the front panel the user can select within a “range” any output clock frequency, permitting fixed synchronization between the MDAC’s internal modulator and external data source. The Clock-Lock frequency is generated via an internal PLL / Digital divider circuit. By insuring the Digital source is synchronized to the modulators clock, allows the internal DPLL to be set to an ultra low BW for greatest performance (Greatest Jitter Attenuation).
I primarily intended the Clock Lock interface to be used with possible future transports from AudioLab – but can be used with older CD transports from Arcam, DPA/ Deltec or the Cambridge Audio CDT I designed (if there are many left working) and possibly other brands I’m not currently aware of.
Any CD transport / Digital Data source that facilitates an external Toslink clock input within the frequency range of the PLL can be used.
I’m happy to publish circuits and possibly design and ship small PCB’s that Forum users can retrofitted to there CD players to allow clock-locking – resulting in “Perfect” digital transports from “any old CD player” that has a Bit accurate SPDIF Data output.
The MDAC not only enumerates over the USB as an Audio Device, but as an HID (Human Interface Device) – via the HID, it sends the standard user playback control commands (Play Pause etc). The operating system sees the MDAC as a Human Interface Device like a keyboard, Mouse etc. Any Media Player (software) that recognizes HID playback control via the OS will work with the MDAC.
USB "bit-transparency checker" was a feature reserved for the original Dacapo replacement – It maybe possible to incorporate a simple USB Bit Accuracy checker into MDAC – I fully agree it would be very useful, however it will solely depend on how much free memory is available within the MDAC MCU after the “Core” MDAC functionality has been finalised – this is in Dominik's domain (the Co-designer of MDAC). If we find methods to compact the code at a later date, it can always be offered as an update – the MDAC allows easy firmware update via its USB Port.
3V / 6V balanced DAC output mode. I’m not sure how to answer your question – as I fail to understand the issue.
1. The MDAC’s Dynamic range is limited by the laws of nature – Thermal noise, and not an “Inherent” system non-linearity as such. By increasing the absolute output voltage level allows us to gain the benefit of its “inherent linearity” – increasing the headroom between thermal noise and Maxim output. This “high dynamic range” signal can be transmitted over a significant distance – to say remote power amplifier or active speakers which have input attenuators to reduce the signal to a level they are comfortable with. Any external interference would have a lesser effect on the 3Vrms signal / then say the 2Vrms signal.
2. As the MDAC can function as both a Digital an Analogue preamplifier, the “Digital Pre Mode” needs some Gain – at 3V rms it’s a little over 3.5dB (Granted – not much) – so in reality the 2Vrms Mode has a small digital scaling factor of -3.5dB or so. With over 132dB dynamic range (Thermal noise, not DAC linearity limited per say), we can afford this small amount of scaling without loosing sleep.
3. It’s the same for Pre-amplifiers that allow higher output levels then 2Vrms to facilitate “Positive Gain” without clipping – you don’t use the Pre-amplifier at 3Vrms, but you need to be able to increase the gain of lower amplitude signals…
Your knowledge and concern for the MAS 6116 while converting my THD % result into dB only heightens my suspicion that you’re not a purely interested forum member.
The DA6116 has an interesting history. Its roots stem from the original MAS9116 (WM8816) which I was first introduced too while consulting for Wolfson on advance DAC architectures. The MAS9116 (WM8816) was fabricated on a small Fab line in a Nordic country – which unfortunately went out of business. The original Fab process had the ability to produce highly linear on-chip resistors – resulting in the original MAS9116 (WM8816) stellar performance. To support there customers and with great credit to VLSI Solutions OY, they where forced at great expense to redesigned the device, this time however on a more standard CMOS process (due to a lack of foundry options) – with CMOS’s inherent non-linear polysilicon resistors. However despite the hit on performance by being forced to a CMOS foundry process, the performance of the DA6116 in balance operation is still very good. The DA6116’s greatest advantage over all other Gain control IC’s is the ability to use external off-chip Hi-performance OPAMP’s – and in the case of MDAC these are operated in pure class A.
The MAS performance results you quoted are measurements with MAS in a single-ended application and at 5Vrms.
This in no way reflects the way the DA6116 is used within the MDAC. Despite the cost implications, the MDAC uses a FULLY BALANCE TRUE DIFFERENTIAL signal path - not one, but two DA6116 are used per design, one “Dual Channel” stepped attenuator per channel (resulting in a fully balanced dual mono attenuator).
During the development of the DA6116 various combinations of resistor switching configurations where tried to try and improve upon CMOS’s inherent limitations – and after much evaluation and listening, it was finally decided upon released into full-scale production the MAS6116AA1 variant.
Pls. see the graph below which is from VLSI Solutions OY internal development and evaluation documentation of the DA6116 in Balanced operation – note that the MAS6116AA1 variant is the current production release standard (lower THD at lower levels – which is more important when listening to real music, not measurement test tones). You can see that at 0dB 2VRMS, the MAS6116AA1 THD hovers around -106dB (and nowhere near the misleading THD+N of -72dB you are trying to suggest).
When the DA6116 is used in true balanced mode, Even-order harmonic products are “Cancelled – read Nulled” to a significant degree. At lower signal amplitudes in single-ended use, the MAS’s dominant distortion components are Even-order (Thus are cancelled via MDAC’s fully balanced signal path) – with Odd-order harmonic distortion components (due to the MAS Poly-Silicon resistors) only becoming dominate at higher signal amplitudes. In fact at 2Vrms, it’s not surprising that it’s the Odd order components that dominate the 0.0015% THD result – as is to be expected for a correctly implemented “Balanced” signal path.
I cannot comment on MAS Single-Ended SNR results (I’ve never tried to use the MAS in SE Mode) – however at 0dB (2Vrms output from the “Analogue” Pre-Amplifier), I can say that the MDAC achieves an 117dB Awtd Dynamic range in Balanced mode.
See the attached FFT plot of the MDAC -60dB Balanced output showing the 132dB Awtd performance, as you can see the performance in (Digital Direct Mode) is limited by the noise floor at below -160dB Ref. 6Vrms
I’ve often been very surprised how poor most pre-amplifiers sound, despite most “Sources” requiring little in the way of Gain – rather attenuation in most applications. The Pre-amplifier is really the most simplest of units in a modern Hi-Fi system, and has no excuse to significantly degrade the audio quality.
To that end – the MDAC’s Line level pre amplifier relies on High Quality Relays to select between the Single Ended Line level Sources, which is then converted into a balanced signal path via a pure Class A Biased stage – then onto the MAS Gain control stage (operating in pure Class A) – Finally to the fully Balanced Folded-Cascode Hi-Current Class A Output Stages.
Each MAS's is fed from its own, ultra low noise and Low impedance “Discrete” regulator – with each channel of the (Dual Mono) Balanced Analogue sections fed from there own “Cascaded” regulated PSU.
So within reason, the preamplifier is simple, all active stages are biased into pure Class A, while operating in true Balanced mode (once past the input selector) – the pre-amplifier does not let the side down – limited ultimately by the Balanced MAS stepped attenuator performance – however at 117dB Awtd… its performance can hardly be described as disappointing
Basically, a pre-amplifier should do nothing other than select the inputs, allow adjustment of the system Gain and ultimately drive the power amplifiers – it should never be heard…
To answer your question (or not) concerning the Digital / verses Analogue gain control, well we have made it user selectable for the very reason that there will always be two opposing schools of thought in the subject. People love to have there own opinions and beliefs – as Peter Walker would always say “Give the people what they want”…
I will say however that the Audio signal path is much simpler (less Complex) in Digital Direct mode.
With 16 Bit Audio there’s plenty of internal Data path width and dynamic range headroom within the DAC that I believe if there is any disadvantage's to Digital Gain control – these are heavily outweighed by the greater complexity of the Analogue pre-amplifier circuit. I would lean towards choosing the Digital Pre Amplifier mode… But it also depends on the system and at what part of the Gain control range you use while listening…
That’s why it’s user selectable…
Once again, if I have misread you, then Pls. forgive me
John
Attachments
Last edited: