I just wanted to know if there was enough interest to start production, and that's definite the case
So there is no need to post more interest on this thread....Don't know exact date yet for production, my resistor sales guy is on vacation, like everyone else in danmark, but should know in maybe a week.
The problem now is that I'm only planning 240 pcs in the first run (Can't get that many parts quickly, will of course make more soon after), so we might run into a shortage situation instead....
Those posted so far, here and in private messages, will get priority. I would prefer this thread to stay technical, so everybody are still welcome to send a private message to get on the priority list.
Once in production sales will be though existing real webshops, one in the US and one in EU, they ship worldwide and take most credit cards. There also seems to be significant interest from asia, so I would actually be interested in locating an online ship in asia to handle sales there, mostly to cut down on shipping and custom cost. Anybody interested, PM me. But please note that it require that you already have the setup.
Soekris, can you inform us which version will be produced? From what I've read in this thread, the 0.02% version is the most likely, correct?I just wanted to know if there was enough interest to start production, and that's definite the case
Will the request for 0.02% components provide any possibility to negotiate extra discount on those parts (directly from manufacturer e.g.)?
I'm not sure what you're asking about with the resistors ?
On power, the DAC is designed to take input directly from one of those small toroid transformers, but you can do anything you want, either shield it, or feed regulated DC to the DAC., as long as you keep the peak voltage between 7V and 15V.
I assume you mean resistors here, the Vishay Z foil are probably THE best resistors, but also start at $5 each, my DAC use 218 precision resistors.... The leaded parts the TotalDac use is even more expensive, but that unit was designed some time ago, it's only recently that low cost precision SMT resistors has become available. Also note that the TotalDac is NOT a sign magnitude DAC and therefore need very precise resistors....
The filter caps use in my DAC are 0805 size, to have some more flexibility, but they're all NPO ceramics, which considering the 200 Khz cutoffs should be perfectly fine.
My plan was originally to start with the 0.05% version, due to longer leadtime on the 0.02% parts. But as you have observed, many want the 0.02% (or even the 0.01%) version, so I have decided to try doing 120 pcs 0.05% and 120 pcs 0.02% boards is the first run, but parts availability ultimately decides, the problem is the resistors in the qty's needed are not something stocked, all other parts has either short leadtime, or in stock at manufacturer or distributor.
And I'm already manufacturing other boards in my day job, purchasing a lot of parts though the official distributors, so I'm getting the right prices.
Yes, the osc is located next to the 3.3V LDO and will probably be affected a little.
But it doesn't matter as the uC continuously measure the incoming bit rate and adjust as needed, don't care about the absolute frequency.
About resistors : In fact I was not just asking about the ohms precision but also about temperature drift and choice of the substrate both about sound effects ? Rohpoint have a great choice of smd resitor temperature proof, xirewound in smd case size... if only the size is important with resistor than caps in such a design ! I don't know as I said the question can be naive as I'm not a tech !
Sorry my dyslexia about IPEDIA which are not résistors but glass caps: cracking proof and superstable at high temperature, Noisy proof....the price is higher also but as you need big quantity I'm not sure in the end it's more expensive than NPO's. Also IPEDIA has some values above 220 nf NPOs limit (iirc) . IIRC size begin at 01 or 02 case size : so super inductance ! 0805 are good for pcbs test but as you need (?) first the best inductance, then the best ESR, why not if factory assembled to go with the lower size case if smd caps are choosed to be cracking proof ? (NPOs are but my understanding is than with Ipedia you will not have any temperature drift noise !)
About the size I just believed resistor lowest size was important also for such a design. As import than caps are !
Rohpoint have also Z foil wirewound SMD resistor less expensive....than the one of the TotalDAC...
Did you choose the can smd caps on your photograph for the best ESR or highs frequencies ? which ref are tey ?
Sorry for my low technical understanding !
cheers
Last edited:
I'm an engineer and look for technical reason why things sound as they do, usually keep a good distance to religious beliefs....
So the resistors are selected based on lowest cost meeting the technical requirements, mostly size, tolerance and stability.
The whole project started out doing a low cost discrete DAC, after I noticed the lower cost and better availability of precision SMT resistors at Digikey....
Checked them out and noticed: "Capacitance variation versus RVDC
0.1 % /V (from 0 V to RVDC)". So their capacitance varies according to voltage, a big no-no in audio.
A important spec for audio use is also the Dissipation Factor, which they don't list....
All critical parts are carefully selected, the large electrolytic SMT caps are Nichicon CL series, a high performance low ESR part. But I might use other types depending on availability and/or price, as long as the specs are the same or better.
Please respect OP (and those who have a subscription notification) and post a PM instead
I just mind how can sound a dac maid by a religious ? I think you meaned " not empirically' or "sound empiricaly cheched in relation to the known tech spec" ? This is not specific to the engineer as to be engineer do not guarante a good sounding machine.
Thanks for answers and to have checked Ipedia caps. Well if the device is redesigned to cost and sounds good, this is a great and good performance and a kind attitude to share.
regards
Thanks for answers and to have checked Ipedia caps. Well if the device is redesigned to cost and sounds good, this is a great and good performance and a kind attitude to share.
regards
To get the benefit of the high resolution resistors doesn't every element in the R2R network have to have similar resolution? I doubt the internal resistance of the shift registers are matched better than 1%. The same goes for the lineal resistance of the PCB traces. What's worse, there doesn't appear any effort made to equalize all the trace lengths. Also, at the data rates envisioned, the total resistance of a PCB trace will vary depending on frequency due to skin effect.
OK, typically the absolute accuracy of the last few LSB's in an R2R DAC are not very important. Did you think of doing a basic analysis, maybe by Monte Carlo, and have a gradation of resistor tolerances with only a few at .01%. There must be a sweet spot in terms of cost performance. Number 2 have you thought of using some lower bits in an auto cal sequence to smooth some of the higher LSB major carries? The fact that your THD spectrum has a fairly flat amplitude would indicate impulsive errors as the boundaries are crossed.
At those low frequencies I'm not concerned about skin effect or 100% matching trace length, it's not like pS precision is needed. And trace resistance doesn't matter much when 0.02% tolerance on a 10K0 resistor is 2 ohm. Again, the FFT plots show how linear the result is.