I hear what you're saying, but isn't it difficult to really figure out who is using what equipment for which purpose? Maybe somebody has their equipment in the basement and wants to run a cable to their room on the 2nd/3rd floor...
Anyway, will there be provisions for at least a single-ended line-out?
It's subject to change, but right now the ODAC PCB will have a USB jack and a 0.1 inch header for the single-ended line output. That could easily be wired to panel mounted jacks of any kind if you're using it stand-alone. There's some discussion about a different version of the PCB that might have everything on board that would be exclusively for stand alone use. It may depend on demand.
And you're correct, if you have long cable runs to make, I would suggest true balanced outputs and inputs and the ODA/ODAC might be a poor choice. Long runs are really rare in home audio (except perhaps to powered speakers and subwoofers), and even more rare in headphone audio, so it doesn't make sense to add to the size and cost of the ODA/ODAC with balanced hardware. XLR connectors, the extra op amps, and all the trimmings take up a lot of space even if you don't populate the board locations. Just the added panel space would demand a bigger enclosure. I suppose someone could design a balanced add-on board if they wanted?
And you're correct, if you have long cable runs to make, I would suggest true balanced outputs and inputs and the ODA/ODAC might be a poor choice. Long runs are really rare in home audio (except perhaps to powered speakers and subwoofers), and even more rare in headphone audio, so it doesn't make sense to add to the size and cost of the ODA/ODAC with balanced hardware. XLR connectors, the extra op amps, and all the trimmings take up a lot of space even if you don't populate the board locations. Just the added panel space would demand a bigger enclosure. I suppose someone could design a balanced add-on board if they wanted?
Yes, I suppose so.
Still, PCBs for this (Figure 3A) are not cheap, especially for one-off orders.
Balanced Transmitter and Receiver II
If you really want balanced output, you can always use "The Wire" ( http://www.diyaudio.com/forums/head...igh-performance-headphone-amplifier-pcbs.html ) and couple it with a standalone ODAC. Unfortunately the GB for "The Wire" has ended, but maybe another one is on the way, I don't know.
Last edited:
ODAC digital volume control?
I'm wondering if you have any plans for an integrated digital volume control for the ODAC? Because i would prefer to have a volume control on the dac itself and if you have a 24bit dac this could be done in digital domain with a cheap rotary encoder. This should remove some issues regarding an analog pot but needs some more logic instead...
The main reason is that i don't like to only use the volume control in the computer because that one has a tenancy to change its settings some times (crash or you use the internal speakers or something) and that leads to uncomfortable listening levels...
I'm wondering if you have any plans for an integrated digital volume control for the ODAC? Because i would prefer to have a volume control on the dac itself and if you have a 24bit dac this could be done in digital domain with a cheap rotary encoder. This should remove some issues regarding an analog pot but needs some more logic instead...
The main reason is that i don't like to only use the volume control in the computer because that one has a tenancy to change its settings some times (crash or you use the internal speakers or something) and that leads to uncomfortable listening levels...
A Digital volume control would preclude the use of a separate line in connection for an analog input, which adds some flexibility to the ODA
My point is even stronger then.
Anyway, I suppose as long as there is a header I'm happy.
Digital volume options on a lot of USB chips and integrated USB DACs are HID controls for the PC. You press volume down on the DAC, and it's really lowering the operating system's volume control.
I'm not aware of any high quality DAC or USB chips that can read an encoder directly. So it would also require a microcontroller on board to read the encoder. Many DAC chips also cannot do digital volume adjustment from just hardware pins--a microcontroller has to write to registers in the DAC to change the volume.
Lots of computer keyboards have dedicated volume up/down buttons on them. You can also get volume knobs that sit on your desk and will adjust the volume directly (via USB). For a USB DAC (no S/PDIF) like the ODAC, these solutions make the most sense. I've not had many problems with Windows 7 surprising me with unexpectedly loud volume levels but I guess it can happen.
Some have suggested a digitally controlled analog volume control but the only decent chip I know of that won't totally destroy the DAC's performance is the flagship part from Cirrus which is really expensive and also would require a microcontroller.
As much as some clearly want the ODAC to be a high-end, full featured stand-alone DAC with S/PDIF, balanced outputs, digital volume, displays, etc., please understand the main design goal for the ODAC is a relatively low cost add-on for the ODA headphone amp. The ODA already has a volume control. If the ODA has pre-outs it can also control the level for speakers.
If there's a future dedicated stand-alone version of the ODAC it's already been discussed to possibly put an Alps volume pot on it or at least have the option for one. That will outperform the digital solution in multiple ways except at extremely low volumes where tracking can be a minor issue. But one step at a time. First up is the ODAC for the ODA.
Think of the ODA/ODAC pair as a much better version of the NuForce uDAC-2.
i would say one thing, please dont use crappy pinheaders for the digital connections if any. one thing i dont quite get about the bal vs SE DAC argument, i'm assuming the ODAC will have balanced outs at least as an option, because in the vast amount of cases (all that i know of) it actually takes MORE parts to do a proper single ended output on a dac that is inherently balanced as they just about ALL are these days. So unless you are using something like es9023 or the Ti equivalent, which are the first SE audio dacs i've seen produced for a fair while, then you are doing it wrong if you get more noise on the balanced output =)
all of the es900X and ES901X series, some crystal semi, some new AKM all have internal volume control set in the dac registers, at least in the sabre this requires a very simple ADC, its not 40bit as such, but it largely leaves the audio data uneffected unless you have turned it down quite aggressively. and even then it fades in comparison to putting a carbon element in the circuit. only needs a single number to effect any 2 or 8 channels so channel matching is perfect all the way down.
the usb hardware i'm using however is the one with the 40bit signed volume thats independant of the dac and independent of the computer as far as i know, it controls the DSP, which is independently clocked
the usb hardware i'm using however is the one with the 40bit signed volume thats independant of the dac and independent of the computer as far as i know, it controls the DSP, which is independently clocked
Last edited:
@Qusp, I'm referring to balanced audio in general having more noise, not just DAC chip outputs. But I understand your point. Analog volume controls, and fully balanced signal paths, are problematic cost and design-wise. I wonder how many realize their $3000 pre-amp with balanced inputs and outputs may well convert everything to single-ended unbalanced at the volume control.
What's your concern about digital signals and pin headers?
What's your concern about digital signals and pin headers?
yeah but more and more hardly use their analogue volume control. i agree though, if you have to go that way a 4 deck stepper puts a dent in the bank balance and most 'balanced' designs are simply doubling up to serve the other phase and are still referenced to ground.
my problem with pin headers for high speed signals is the mess they make, impedance control is difficult and often ignored, termination often haphazard and thats when they are new. ive seen some newer type FPC/FFC connectors that i like the look of, but your average 2.54mm header just doesnt cut it for common mode coupling imo. its not limited to differential signals, its just as important for the 'wires' to be closely coupled for SE signals like i2s
for output from the dac its not quite as crucial, but i still wouldnt do it, thats probably more about me though =)
my problem with pin headers for high speed signals is the mess they make, impedance control is difficult and often ignored, termination often haphazard and thats when they are new. ive seen some newer type FPC/FFC connectors that i like the look of, but your average 2.54mm header just doesnt cut it for common mode coupling imo. its not limited to differential signals, its just as important for the 'wires' to be closely coupled for SE signals like i2s
for output from the dac its not quite as crucial, but i still wouldnt do it, thats probably more about me though =)
Last edited:
Thanks. I generally agree. I commented several pages back about my reservations in "going vertical" (relative to adding S/PDIF on another board) and I have even bigger reservations about using cables/wiring to interconnect boards. It's much easier (and cheaper) to maintain signal integrity, analog or digital, on a single board where possible.
But sometimes it's not practical. In this case one board is DIY through-hole analog, and the other requires fine pitch SMT, licensed parts, programming, etc. I hope to keep the design to just two boards and keep signals close to their ground return paths to minimize loop area, noise pickup/radiation, impedance problems, etc.
I agree, say 480 Mbps USB 2.0, is not the sort of thing you want to send through 2.54mm headers. The good news is it's easy to verify digital signal integrity with a suitably fast scope and good probes. The dScope also does full digital signal eye pattern tests including margins and allows intentional degradation its digital outputs in a variety of ways including levels, timebase deviation, adding jitter of various kinds and levels, adding noise, etc. It can also verify bit absolute bit transparency over hours or even days.
Any problems on the analog side will show up in the dScope measurements well before they're bad enough to be audible. But it's best to avoid as many problems as possible in the first place--especially when the solution involves getting yet another set of proto PC boards made.
But sometimes it's not practical. In this case one board is DIY through-hole analog, and the other requires fine pitch SMT, licensed parts, programming, etc. I hope to keep the design to just two boards and keep signals close to their ground return paths to minimize loop area, noise pickup/radiation, impedance problems, etc.
I agree, say 480 Mbps USB 2.0, is not the sort of thing you want to send through 2.54mm headers. The good news is it's easy to verify digital signal integrity with a suitably fast scope and good probes. The dScope also does full digital signal eye pattern tests including margins and allows intentional degradation its digital outputs in a variety of ways including levels, timebase deviation, adding jitter of various kinds and levels, adding noise, etc. It can also verify bit absolute bit transparency over hours or even days.
Any problems on the analog side will show up in the dScope measurements well before they're bad enough to be audible. But it's best to avoid as many problems as possible in the first place--especially when the solution involves getting yet another set of proto PC boards made.
You mean you don't have a PCB or you just don't want to read the blog articles? There are several small things that are important to watch out for and really hard to fix after the fact. So I'd suggest reading at least:
Circuit Board Construction
And:
O2 Cautions
And make sure you download (or view) the latest PDF and BOM from the Google Docs links here:
O2 Resources
And if you're mean building it from the schematic without a PCB, I strongly advise against that. It could be unstable, and isn't likely to perform anywhere near as well with out the ground planes, ground routing, etc.
Circuit Board Construction
And:
O2 Cautions
And make sure you download (or view) the latest PDF and BOM from the Google Docs links here:
O2 Resources
And if you're mean building it from the schematic without a PCB, I strongly advise against that. It could be unstable, and isn't likely to perform anywhere near as well with out the ground planes, ground routing, etc.
If that's the CS3310, then the TI PGA2310 has a better paper performance (0.0004% THD+N, 120dB S/N). PGA2310/11/20. The PGA2311 is actually the better part for a DAC (+/-5V) @ ~$8.50. There's an A-grade (0.0002% THD+N) part too.
Here's one in an application you might find interesting:-
wakibaki's world - technology
well if you must do it (connect offboard), i can highly recommend the amphenol and hirose 6GHz u.fl and w.fl smd bnc, or UMCC, not too expensive if bought in bulk, cheaper than a decent coax cable and neat as hell. if only 3 connections are needed for MCK, BCK, LRCK (ground handled by the shield) then if bought in bulk you might be looking at adding 10-15 to the cost for u.fl headers both sides and 3 short cables, considerably less if you use the TE connectivity branded UMCC cables
the problem with DIY in this aspect is its difficult to keep the ideal of the single board, unless you offer it in various states with different parts populated or not, because we all need different things and dont like paying for things we dont need, people like to try different power supplies etc. though it should be pretty easy to do that with spdif, which should take up bugger all room. i can highly recommend the spdif transformers from sceintific conversions
shutup about you magic bullet will you dammit?!! haha, most scopes wont even measure phase noise in a meaningful way, let alone injecting specific quantities of it as an aside, just so it can measure something else
the problem with DIY in this aspect is its difficult to keep the ideal of the single board, unless you offer it in various states with different parts populated or not, because we all need different things and dont like paying for things we dont need, people like to try different power supplies etc. though it should be pretty easy to do that with spdif, which should take up bugger all room. i can highly recommend the spdif transformers from sceintific conversions
shutup about you magic bullet will you dammit?!! haha, most scopes wont even measure phase noise in a meaningful way, let alone injecting specific quantities of it as an aside, just so it can measure something else
Last edited:
