Hi,
i am currently designing a USB DAC (I would consider it as entry level high end
).
Here you can find the project description.
The basic concept is quite simple:
- Using PCM2707 as USB receiver
- Reclocking and resampling using an AD1896
- WM8740 in mono mode as DACs
- separating analogue (DAC) and digital circuit (receiver, resampler) to reduce noise
I have now some questions which lay way beyond my knowledge and experience:
- Using transformers for output coupling seems quite usefull - the output can be easily configured balanced or unbalanced. But is it sonically wise? which transformers con be used (regarding sound quality and size footprint). Or is it better to use capacitor DC coupling?
- Are there any experiences regarding separating digital and (semi-) analogue circuits? My idea to separate it came from suggestions regarding high precision ADCs (I simply copied the principle of having separated VCC and GND with a buffer for transmitting data from the digital part to the DAC (which are the "analogue" part). Does it make more harm than good?
Thanks for your input.
_atari_
i am currently designing a USB DAC (I would consider it as entry level high end
).Here you can find the project description.
The basic concept is quite simple:
- Using PCM2707 as USB receiver
- Reclocking and resampling using an AD1896
- WM8740 in mono mode as DACs
- separating analogue (DAC) and digital circuit (receiver, resampler) to reduce noise
I have now some questions which lay way beyond my knowledge and experience:
- Using transformers for output coupling seems quite usefull - the output can be easily configured balanced or unbalanced. But is it sonically wise? which transformers con be used (regarding sound quality and size footprint). Or is it better to use capacitor DC coupling?
- Are there any experiences regarding separating digital and (semi-) analogue circuits? My idea to separate it came from suggestions regarding high precision ADCs (I simply copied the principle of having separated VCC and GND with a buffer for transmitting data from the digital part to the DAC (which are the "analogue" part). Does it make more harm than good?
Thanks for your input.
_atari_
DougL said:
I know a lot of studio gear uses transformers for input/output coupling and especially the lundahl transformers have a very good reputation.
But as a übernoob i have two problems:
a) what transformer can be recommended
b) is this simple approach I took really feasable?
Here is a link to he schematics:
An externally hosted image should be here but it was not working when we last tested it.
The capacitors at the output are optional alternatives - therefore t looks a bit odd.
analog_sa said:
That is what I try to find out
The purpose of the transformer is twofold:
1. By using some configuration as sketched it should give DC protection - the same you can also achieve with some capacitors. My question is: Gives the us of a transformer some advantage?
2. The use of a transformer makes it very easy to configure balanced or unbalanced output - no need for opamps or stuff like this. I have a new versio of the schematics - I will post later which has jumpers to configure balanced vs unbalanced output.
The main aim was to get rid of DC which is always a problem with DACs.
analog_sa said:
oh sh*!
That is something I completely overlooked. You mean that I forgot to think about that both differential outputs can have different DC offset (e.g. positive on the positive and negative output) which can lead to a lot of problems if I just connect them with a wire (which goes through the transformer). So I would need a split transformer and connect OUTP and OUTN to ground?
Or would something like this help:
Code:
OUTP ---------
)
)
)
GND ----------
)
)
)
OUTN ---------Are these kind of problems reasons that not very much people use transformers and stick to capacitors????
Please don't get so excited. It's a Sunday
The offset may be mentioned in your dac's datasheet. For the AD1852 which i use it's about 50mV. I think this rules out most of the better transformers. Your dac may be more suitable. In any case look for a transformer capable of taking a few mV without collapsing.
The offset may be mentioned in your dac's datasheet. For the AD1852 which i use it's about 50mV. I think this rules out most of the better transformers. Your dac may be more suitable. In any case look for a transformer capable of taking a few mV without collapsing.
analog_sa said:Please don't get so excited. It's a Sunday
The offset may be mentioned in your dac's datasheet. For the AD1852 which i use it's about 50mV. I think this rules out most of the better transformers. Your dac may be more suitable. In any case look for a transformer capable of taking a few mV without collapsing.
I had a week of holliday - so excuse my excitement
And thanks for your patience.
The DAC is a WM8740. It is a voltage output DAC. The datasheet tells nothing about DC offset. The output Voltage is said to be max 2.5V (VDD/2).
Anbody got experience with the WM8740?
_atari_ said:
Nothing to do with patience. I am genuinely interested in your project.
At present i use a dac with a 2707 usb/i2s which is pleasant but certainly not spectacular. After a week of mulling over the 1896 datasheet i'm still not convinced that ASRC is the best way to improve upon the 2707 jitter-wise but it's certainly worth a shot.
The TAS1020 is probably a better solution if one manages to get eeprom contents and drivers off a commercial product. I haven't been successful.
With regards to the analogue stage, as already mentioned i use a DAC with voltage output which is obviously an easy cheap-skate solution. Once satisfied with the USB part i'll probably switch to a current output chip.
The transformer solution for your dac may very well sound great. Sadly, i am not in a good position to try it out. One of the transformers i have is a Lundahl LL1660/18mA and although it will easily swallow the offset it doesn't seem very suitable in other parameters. I also have an LL1544 which can take absoliutely no offset but will work as a parafeed. Don't know if it's worth trying it out.
analog_sa said:
Yes, to amplify the small signal of a passive I/V and to convert from balanced to single ended, connected parallel to the I/V resistor(s).
But then the primary impedance of the transformer becomes part of ( lowers ) the I/V resistance ? ? ?
Doesn't the transformer impedance depend a little on the frequency ?
I tried a microphone transformer and it worked so far.
What about MC transformers ? A part of the 1mA will flow through the transformer primary and maybe kill it
Bernhard said:
But then the primary impedance of the transformer becomes part of ( lowers ) the I/V resistance ? ? ?
Doesn't the transformer impedance depend a little on the frequency ?
What about MC transformers ? A part of the 1mA will flow through the transformer primary and maybe kill it
The primary impedance is only determined by the transformed secondary load within the transformer bandwidth. Only at very low frequencies it may deviate due to limited primary inductance or DC current.
At the low levels of IV resistors it's really a non-issue.
1mA won't kill an MC step-up but may saturate the core. Not recommended anyway.
If you don't mind paying the price (close to $100 if memory serves), you could use something like the CineMag CMOL-2x600T2.
Connect VOUTP and VOUTN to each end of the primary. Tie the primary center tap to ground through a decent size capacitor (say 470uF). Tie the primary Faraday shield and can to ground.
For balanced output on the secondary, tie the secondary outputs to pins 2 and 3. Tie the center tap and secondary Faraday shield to pin 1.
For unbalanced output on the secondary, tie YEL to the RCA hot, ignore the secondary center tap and tie BLU and the secondary Faraday shield to RCA ground.
se
Connect VOUTP and VOUTN to each end of the primary. Tie the primary center tap to ground through a decent size capacitor (say 470uF). Tie the primary Faraday shield and can to ground.
For balanced output on the secondary, tie the secondary outputs to pins 2 and 3. Tie the center tap and secondary Faraday shield to pin 1.
For unbalanced output on the secondary, tie YEL to the RCA hot, ignore the secondary center tap and tie BLU and the secondary Faraday shield to RCA ground.
se
Steve Eddy said:
Thats sound somehow like the idea I had in mind. On the other hand the DAC wants to see a high impendance - perhaps the transformer is still no good idea ...
By the way: How much does the capacitor has effect on the sound? Should it be boutique or is a Panasonic FM be sufficient?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.