What’s Inside the $13,000 dCS LINA DAC?

What’s Inside the $13,000 dCS LINA DAC?

Show Video

The circuit board's really sharp and if i  don't wear a glove I'll cut myself it's not   that it's like very delicate or I'm going to  do any magic or anything like that no no no   it's not for esd this doesn't have a power  inlet so we're we're we're all we're all   good okay ready you want to get bring come  a bit nearer so um I thought I'd just um do   a bit of a reveal and show you what's inside  a dCS Lina DAC so this is my constant travel   companion Lina uh she goes with me all over the  world and this is what she looks like inside and   there were a couple of things about the way  Lina is made that I wanted to show you guys So first thing I wanted to show you is that most  of the electronics that you'll see here at this   show the chassis are boxes and what I mean by a  box is something that's made out of flat plates   okay or or or maybe folded plates so the chassis  of the leaner is mil from solid billets so this   is the bottom plate this started life as a billet  of aluminium this thick and then it's been milled   out to make this shape you can see the the milling  marks here but you can't actually feel them if you   it's completely flat but but you can see where  the milling machine has has gone around okay   now one of the so this has a number of of benefits  of using this this uh way of constructing firstly   you can hear this has a very very low resonant  frequency if you make the chassis from flat   plates the resonant frequency is is much higher  so this is much more inert um the second thing   is we can make it exactly the shape we wanted to  so in this case you imagine this is upside down   so the circuit board is sort of folding around  the transformer here and the transformer is a   potential source of noise both mechanical and  electromagnetic noise so we want to move the   transformer as far from the circuit board as  we can so what we've done is instead of having   feet screwed to the outside we've actually  used the the milled construction to drop the   floor where we mount the transformer which  brings the transformer further away from the   circuitry that that can be affected by it um  this part of the circuit board is the output   uh sorry the uh power supply so this is all  the regulators and everything uh that take   the uh the voltage from the transformer and  smooth it out into really nice uh dc um that   can generate quite a lot of heat so what we've  done is we've mounted it directly against the   chassis which is the mil chassis so it's quite  thick to it's probably 3-4 mm thick and this whole this whole part of the chassis which  is again milled from one solid billet this   is all heat sink the entire chassis is a heat  sink so we don't need any discrete or external   heat sinks the whole of the chassis is is a  heat sink I was wondering why you didn't shut the contact another advantage of this construction  let's say uh we wanted something we had something   this side that was very sensitive to uh  electromagnetic interference and we wanted   to put a shield in between them well instead of  putting a a separate part in we can just mill a   wall into the case and put that the other side of  the of the wall so we can basically configure the   chassis around the needs of the of the circuit if  you look at this end plate here spacers which you   would normally need to mount a circuit board in  a in a box chassis these are milled into the end   piece not only milled in but they're pre-drilled  and they're pre-tapped so i can put a bolt in i   don't need a nut i just have the bolt and i put  it directly into the chassis so you can see all   the mounting points all the screw holes they're  all built into the chassis so the number of pieces   i need in this construction is much lower than  if I'm screwing plates together if i if i have   two plates like this and i need to put a bracket  to hold them together and then i need two screws   on each bracket and if i want to mount a circuit  board i need spacers which are like little turrets   that i put the circuit board on and a bolt goes  through and a nut on the other end here i can   just mill the spacers into the into the chassis so  that's about the chassis if we look at the circuit   board so sharp so looking at this you might think  that this is six separate circuit boards this is   one circuit board this is called a flex rigid  circuit board so basically this is a 12 layer   board so there are 12 layers of conductors inside  this and then there's this solid substrate what   this manufacturing technique does is it removes  the substrate around these edges and then the 12   layers of of copper traces are encapsulated in a  flexible resin which means that the whole board   can now bend so we can take a 30 cm by 50 cm board  and we can fold it up and put it inside a shoe box   the other benefit of this for from a packaging  point of view is we can take advantage of these   features like i can i can fold a a heat  generating component i can fold it up to   a thick bit of chassis and mount it directly  so that it can use it as a heat sink or if i   if this is very sensitive to noise i can fold  it down behind a wall that I've built into my   chassis so the combination of the flex rigid  board and the milled construction gives you   really really interesting packaging alternatives  the the second area where flex rigid gives us   enormous benefits is electrically so anytime  you have two um conductors meeting you get an   impedance mismatch between the the two the two  materials because this is one piece every one   of those 12 planes is completely uniform impedance  which means electrically this board behaves nearly   perfectly it's it behaves just like the spice  model there there are no impedance mismatches   because we have two wires connecting together  or we have a ribbon connector how many layers   12 layers 12 this ribbon cable is actually is 12  layers of conductor inside a piece of resin it's   already inside so it's not additionally solder on  or anything no no this is one board just one board   it was ribbon cable yeah exactly that's that's  why I'm explaining it because when i when i saw   this i thought these are ribbon cables but it it's  it's way clever so a ribbon cable if if any of you   have seen a ribbon cable close up a ribbon cable  connector uh consists of pins and like a spring   that that goes down on them so the spring can  fatigue and it grips less loosely it can there can   be gaps the pin can get oxidized or there can be  dirt so anytime you have a physical connector you   have a potential point of failure with the flex  rigid construction we have so many less potential   points of failure because this entire board can be  tested as one unit we don't have six boards that   we need to then assemble together and then retest  system test so this can be this can be tested by   the board manufacturer and when it's delivered  to us we know 100% that it's it's working   so the interesting the most interesting thing  from my perspective is the flex rigid construction   especially with this size of board there's there  are very very few board manufacturers who have   the machines that can stuff a flex rigid  board this this big it's very specialized   it's very new I think we're the only audio  company that's using flex rigid at this point   so this is a much more expensive way of of  building a board than if we use separate boards   the milled construction from solid is much more  expensive than using flat plates and and bolting   them together but with this construction we can  use one bolt size with one torque setting we   need no spacers we need no nuts cuz they're all  built in and the time that we save assembling   the product the the our previous generation of  products took 3 days to assemble and test this   takes 15 minutes to assemble and test and the  saving on the human costs more than offset the   additional costs of the components so by making it  in a much more sophisticated way that's uses much   more expensive components we've actually made the  product cheaper how much is this this Lina DAC is   20,500 Sing this is our entry level product if you  have any problem with one of the connector this is   not a connector it's just a trace yeah flexible  pcb right before yeah so if you have one problem   is cracked or something then you have to throw  it it can't be replaced no but but we the failure   rate with these boards on once they've been tested  with the with the the by the manufacturer i don't   we've been selling this product for i think 2  and a half years i don't think we've ever had a   failure they are incredibly reliable because there  is just nothing to go wrong is a new technology   yeah it's really exciting technology so happy  to answer questions that was what i wanted to   to show you the the milled construction the  flex rigid boards and what you get when you   combine them is really really powerful really  exciting these are resistors right yeah so this   is these this is actually the DAC oh so it's r2r  deck it's an r2r DAC so uh 48 uh resistors per   channel and then the the the the the uh dCS uh  kind of magic if you like our our proprietary   um kind of yeah uh advantage um or uniqueness  is the way we use these resistors we basically   so we we have 48 per channel but the signal that's  being passed through them is upsampled to 5bit 2.8   8 mhz so for maximum signal we're going to light  up five of these resistors the reason that we   have 48 per channel is because we want to move  the signal around more or less randomly so that   we're distributing the current statistically over  time distributing it evenly among these resistors   and so we have uh like a brain which is for every  sample is deciding which and how many resistors   to light up that's running in this fpga and the  algorithm is called the mapper and this is like   the brains of the of the ring deck every single  sample is being sent to a different combination   of these resistors which basically means that  signal correlated distortion is being turned into   non-signal correlated noise that's the job of the  mapper and rather like the the background noise in   this corridor your brain is able to filter it out  very easily you all you're focusing on my voice   my voice is the signal and the noise is easy to  ignore whereas if i if i introduce distortion into   my voice by having my nose that's very difficult  to ignore your brain can't filter that out but all   the background noise because it's not correlated  to my signal it's very easy to to filter out   so we're trying to and noise can be mathematically  noise shaped you can kind of move it around hide   it in the corner where nobody's going to notice  a shield or that is a shield around the it's a   toro so this is just a shield sd card is this sd  card is it yeah it's a micro sd yeah it's a micro   sd um so that's that's the um no the it could  be actually for the firmware though i mean the   firmware update yeah probably update yeah come  to the update 16 has been famous you know you   guys can always update your normal deck you can  send it for effects update and all the things   then this guy i think basically can't right yeah  since on the well because it's flex rigid yeah so   you can't upgrade the deck no exactly because  the the deck is this part here the apex board   is essentially replacing that part this is half  apex so this i i hope buying this because all   the other product was upgraded yeah yeah but so  waiting for this guy to go full effect yeah so   well it's it it's halfway there because the both  this Lina Dac and the Apex upgrade were results of   the r&d work done on the vess so vess is you know  the the project the project the project is moving   along and then the first branch eventually becomes  Lina the second branch becomes apex so the first   branch already has some of the r&d that went into  apex but not all of it cuz apex came afterwards   investment here so i was thinking wait for you  guys to go full apex or you know yeah maybe we   will you mean because it can be upgraded your the  other day all can it can be it can be we just need   to swap the whole board yeah yeah expensiveness  your i saw that your rosina and all the other deck   can be upgraded you know with at least partial not  no but as i say there are there are like 16 good   reasons why we went with flex rigid construction  and you know it it's like everything it has this   great everything has has you know advantages and  disadvantages this is less modular but in terms   of the improved yield we we get no test failures  uh we get very very you know so far i don't think   we've had a single one that needed to come back to  the the factory so it's it you know it's it's very   resistant to failures uh gives us the packaging  advantages um the ability to to utilize the mil   construction much more flexibly to it and and  you know and the electrical advantages of common   impedance the uniform impedance so it's half  effect that some of the technology more or less   half those or is it just uh no the so the clocks  this is the dac so the clocks in the dac are both   vcxo because we have vcxo voltage control so in  the master clock we have one ocxo and one vcxo i i   always curious you know isn't a better clock in a  deck better than external clock i i just sometimes   i don't get it no so the excellent question um and  the answer is the the so the the clocks are like a   conductor in an orchestra yeah they're beating  the time he's the last guy yeah he's he's he's   yeah he he is governing him this is the mapper  he's making the decisions which resistors and   he is telling him now now now now these clocks  always manage the digital to analog conversion   always even if you have a master clock that's why  i say i i somehow i this is a three-piece design   you know i was considering this set for a long  time so it's three piece design i always don't   get it if the tech clock is good enough why  do we need a master so the DAC so so the the   let me answer so the the clock in the DAC is as  absolutely as good as we can make it but it has   an fpga next to it running at six mhz okay it has  a transformer it has all this circuitry around it   in the master clock we we have been able to scale  down the environment to only what's needed for the   clock crystals so even if we use the same crystals  in the master clock we don't we use one ocxo and   then one vcxo but it's the same as as this vcxo  it's exactly the same crystal but the environment   that it's running in is so much better yeah  there's an envir advantage here but then the   cable loss no i mean clock it's it's a ttl signal  we can we can we we can recover it perfectly with   the plll the fact is it's a it's a less jittery  clock we're not sending the clock signal over an   audio interface it's not embedded in in in audio  we're just sending us ttl square one of these rj45   connector right i mean you no it's it goes through  the bnc's oh the bnc's yeah just a 75 ohm bm bnc   so the the the fact the the i mean you can hear  the improvement i can i can i can play upstairs   they have the master clock in the in the in the  i hear it i hear it so start playing a track yes   with with the with the with the master clock  and then just pull the cable out in the middle   of the track and then put it back in i do this at  shows and you can immediately hear the difference   it's making the amazing part that technically is  very hard to understand yeah but the main thing to   understand is these clocks are always the ones  in charge the master clock just puts a hand on   the shoulder and just steadies these clocks these  clocks your design is different because the rest   of people design they just cut off these two  clock and then they use the external clock i   i don't think you can do that i i i can't speak  for other people's designs what they do is that   they also using the master cloth they do a pl  or whatever they do it and then after that they   it's 10 10 mega then they come in and then they  just you know just drive the thing yeah people   who do that we've been doing clocking for 30  years you guys there is a chance that we ask   you this question because you are here so now  i learned something you do something different   but trust me these clocks are always in charge  yeah you guys do something different than than   than your clockwork and these clocks are very  very accurate clocks but there you know there   is so much going on in this chassis that there  will be a little little bit of emi which will   be affecting them which we can eliminate in the  in the master clock just because it's in its own   chassis you know clock crystals are incredibly  sensitive to mechanical vibration temperature   fluctuations emi they're sensitive to everything  they're like they're like the most sensitive woman   you've ever met they break out crying at the least  least thing yeah well and temperature i mean the   the resonant frequency of the crystal changes for  every you know tenth of a degree uh dcs is always   so expensive it's also a mystery to every one  of us that we never got a chance to try it you   know so this is the most basic entrance level  thing that at least i dare to ask question so   you're saying that you your the master clock and  this clock you have some secretary to take in the   master clock and with this as a reference still  as a reference and do some compensation to correct   exactly that's why so that's why interesting so  that's why we use vcxos vcxo stands for voltage   controlled crystal oscillator so you can change  the resonant frequency you can tune the resonant   frequency of this crystal by changing the voltage  if we used ocxos's here we would we wouldn't be   able to use a master clock because they're always  at the same frequency because they're tuned by   temperature and they have a they have a resistor  inside them which which has something there you go   around exactly so the voltage that is driving  these is what's tuned by the master plan that   is different from other people other people just  don't do it i thought it was obvious no they don't   yeah they just i don't know whether they yeah most  of them they just chop off the and then just take   the external clock and then they go through some  repro and then just generate back the the the the   master clock mclk or or the the eck something  you just general back yeah it's different

2025-06-01 21:40

Show Video

Other news

World's Greatest Military Inventions and Technologies 2025-06-03 18:26
Closed captions on DVDs are getting left behind 2025-05-28 19:46
JRE: Google's Quantum Computer is Communicating with Multiple Universes? 2025-05-28 12:26