I wonderd the same thing. In Ian's FIFO thread it was said that it was impossible to use two or more of Ian's reclockers for multichannel setups since (in layman's terms) the output of the FIFO buffers of multiple devices could not be synced. Søren (how) is that possible with your dac?
Wonderful project BTW!!
My DAC's can use same clock frequency, and use shorter FIFO's. Although not done yet, there are FPGA pins that can be used to sync things, I don't see any problems in multichannel or crossover setups that can't be solved....
Accuracy
This DAC is very interesting. Ladder DAC's have a cleaner sound to me.
One question though. Since this design uses discreet resistors, even at 0.01% tolerance, what is the equivalent error in Bits?
My math might be wrong, but 0.01% is a maximum error of 1 in 10,000. Does that not give the DAC a potential effective resolution of less than 14 bits? (16,384).
Nice looking design.
Jeremy
This DAC is very interesting. Ladder DAC's have a cleaner sound to me.
One question though. Since this design uses discreet resistors, even at 0.01% tolerance, what is the equivalent error in Bits?
My math might be wrong, but 0.01% is a maximum error of 1 in 10,000. Does that not give the DAC a potential effective resolution of less than 14 bits? (16,384).
Nice looking design.
Jeremy
If you read the thread from post 306 or thereabouts you will find an answer...
http://www.diyaudio.com/forums/vend...n-magnitude-24-bit-384-khz-7.html#post4040870
Would not...
the best way to do a discrete R2R DAC be to commision Texas Components to do each of the 4 (assuming a balanced design) resitor arrays on a single piece of metal foil, laser trimming the entire thing to as high a tolerance as is possible? I look at a DAC, like the "Total DAC" (commercial DAC using individual Texas Components high precision metal foil resistors soldered into a PCB) and it looks to me like solder joints and PCB traces would dominate the innacuracy of the resistor array, so that even .0001% resistors might not be taken full advantage of.
if we look at the measured performance of MSB (another discrete R2R commercial DAC) as published in Stereophile, and we can see that they are getting very good resolution for an R2R design (better than 1704 chip). Perhaps their resistor arrays are made on a single piece of metal foil... I asked one of their engineers about the resistors they use, but they would not reveal the info.
Perhaps DIYaudio members could fund such a project through crowd funding?
the best way to do a discrete R2R DAC be to commision Texas Components to do each of the 4 (assuming a balanced design) resitor arrays on a single piece of metal foil, laser trimming the entire thing to as high a tolerance as is possible? I look at a DAC, like the "Total DAC" (commercial DAC using individual Texas Components high precision metal foil resistors soldered into a PCB) and it looks to me like solder joints and PCB traces would dominate the innacuracy of the resistor array, so that even .0001% resistors might not be taken full advantage of.
if we look at the measured performance of MSB (another discrete R2R commercial DAC) as published in Stereophile, and we can see that they are getting very good resolution for an R2R design (better than 1704 chip). Perhaps their resistor arrays are made on a single piece of metal foil... I asked one of their engineers about the resistors they use, but they would not reveal the info.
Perhaps DIYaudio members could fund such a project through crowd funding?
IIRC a resistor network from TC is almost $40 a piece.
How many we need for a R2R DAC?
You would need four, but I don't think it would be as simple to cost as 4 * $40 as the cost will depend on the production run quantity for what would be a bespoke product; I'm sure most of the costs will come from the initial setup and that will be relatively fixed.
Anyway, having said that, its maybe worth a reminder that this thread relates to Soren's R2R DAC and he has elected to use discrete resistors.
Ray
Those have 4 resistors, so one resistor network per bit when doing a sign magnitude R-2R DAC.... You do the math.
I choose to use much cheaper precision thin film resistors, those are good enough.... And then manufacturing using cheap SMD technology.
Btw, Texas Components seems to be just a value added distributor for Vishay Precision Group, you can also buy them at Digikey.
The presentation linked in the first post of the following thread is worth reading:
http://www.diyaudio.com/forums/parts/237237-low-frequency-modulation-distortion-foil-resistors.html
It was written by Bruce Hofer (co-founder of Audio Precision) and looks at selecting components for ultra-low THD+N. Hofer states that "All factors considered, the best resistor technology for audio applications is a low TCR thin film". He recommends using precision thin foil smd resistors in 1206 size and better than +/-25ppm tempco. The presentation also suggests that bulk metal foil resistors have "low frequency modulation distortion that is much worse than expected".
Last edited:
You are confused. TEXAS COMPONENTS makes these things for themselves and for VISHAY.
If you contact TC they will make you what you want. If DIGIKEY has what you need, that is nice but they have a very limited selection.
You can specify any resistance you want and they offer tolerance to +/- .02% - all you need t do is pay. Some values are available at even closer tolerance. Two weeks later they will arrive.
Just what kind of value do you think they could add? Best to check before making such pronouncements.
Maybe because they say "As a licensed and certified Precision Center, Texas Components is an authorized part of the Vishay Precision Group global supply chain for manufacturing and distributing resistive products made using Bulk Metal® Foil, including standard Vishay Foil Resistors as well as our own unique design applications and derivative products."
And maybe because Texas Components is listed under "Vishay Foil Resistors' (VFR) authorized precision centers, with several global locations, provide fast delivery of small quantities of Bulk Metal® Foil resistors with any value/tolerance combination. Shipping small quantities in five days or less, our precision centers trim each resistor to the precise value ordered, giving the designers the exact resistor they need in the shortest time possible. For your local Precision Center please click here. "
Sounds like they customize things, that what you normally call "value added". Afaik, Vishay Precision Group is the manufacturer of Metal Foil resistors.
Hi Søren,
I am not sure if you have posted already somewhere but could you please provide a block diagram of your DAC implementation especially logic side of it covering isolation, re-clocking and external clock options. I am assuming your new version covers the above options.
TIA