On the balanced the theory is that the differential outputs cancel it; I believe that's why Hazard started down the flipflop route with his Llundahls.
Actually, posting that and cogitating on the noise issue with my TVC project got me thinking about this as up until now I had simply assumed... I now realise that I don't have a proper understanding of balanced/differential operation, so I have started to try and find out a bit more.
I found this to be a good overview;
Balanced and Differential
The first thing I realised is that we should perhaps be calling the board with the FlipFlop the 'Differential Board' because the FlipFlop delivers differential outputs, i.e they are of opposite polarity, which is not necessary for balanced operation. Differential signalling enables us to achieve a 6dB signal level increase, evidenced by Hazard's measurement of the output level of his project, and a beneficial improvement in the signal to noise ratio. I think I get that side now.
What I've not managed to find any information about is DC offset on the differential outputs or even if there is any offset on the outputs of the flipflops. If there is an offset then it feels as though it should be additive; later I will measure the DC voltage across the differential outputs - chores first though
Hazard is using the differential output from a flipflop straight into his Llundahl transformers without a problem so maybe I'm worrying over very little? Can anyone else help out with this?
Ray
Me and many others have had really good results with the DDDAC using the differential pos and neg outputs and ignoring the gnd. The DDDAC uses i/v resistors on the output and if those are accurately matched then there's almost no dc offset at all. Not sure how it would work in this instance but I imagine it will be a similar story.
Just as a reminder to anyone starting to put a board together;
http://www.diyaudio.com/forums/digital-line-level/273474-best-dac-no-dac-94.html#post4630381
Ray
http://www.diyaudio.com/forums/digital-line-level/273474-best-dac-no-dac-94.html#post4630381
Ray
Thanks Ray. I'll get some components ordered this week.
12mv dc offset doesn't sound like a lot to me, but I guess it depends on what your tvc provider recommends. My dddac has approx 4mv dc and that was considered negligible at the time. I would guess the dc offset between pos and neg out is directly linked to the component accuracy/matching of the rf resistors. So I guess you could improve on your 12mv if needed with closer matched components there.
12mv dc offset doesn't sound like a lot to me, but I guess it depends on what your tvc provider recommends. My dddac has approx 4mv dc and that was considered negligible at the time. I would guess the dc offset between pos and neg out is directly linked to the component accuracy/matching of the rf resistors. So I guess you could improve on your 12mv if needed with closer matched components there.
I would guess the dc offset between pos and neg out is directly linked to the component accuracy/matching of the rf resistors. So I guess you could improve on your 12mv if needed with closer matched components there.
The 12.7mV offset was measured directly off the output of the flipflop, I'm not currently using Rf and Cf because the TVC's frequency response rolls off like a first order filter after 35KHz.
Ray
The 12.7mV offset was measured directly off the output of the flipflop, I'm not currently using Rf and Cf because the TVC's frequency response rolls off like a first order filter after 35KHz.
Ray
I think 12mV is noisy enough.
Try a stright big bipolar in series with tvc's primaries, just to hear the differences - if any.
I think 12mV is noisy enough.
Try a stright big bipolar in series with tvc's primaries, just to hear the differences - if any.
I finally found some good information about the Silk TVC, it specifically says that any DC will compromise the performance.
I've decided the best thing to do is start from where I had good performance and work on from there; perhaps going straight to the balanced/TVC/no RC filter was a bit too much of a gamble. So, I'll put together another SE board with a 40KHz LP filter and a coupling cap and feed that into the TVCs; that way everything is the same as before except for the TVC replacing the potentiometer.
Ray
Can you run the TVC after your tube buffer instead of before?I finally found some good information about the Silk TVC, it specifically says that any DC will compromise the performance.
I've decided the best thing to do is start from where I had good performance and work on from there; perhaps going straight to the balanced/TVC/no RC filter was a bit too much of a gamble. So, I'll put together another SE board with a 40KHz LP filter and a coupling cap and feed that into the TVCs; that way everything is the same as before except for the TVC replacing the potentiometer.
Ray
A tube buffer should be ok with a small dc offset on the input.
LMS with SoX, I plan to use it.
Inviato dal mio D5803 utilizzando Tapatalk
Any link?
Can you run the TVC after your tube buffer instead of before?
A tube buffer should be ok with a small dc offset on the input.
I could, but I really wanted the buffers for current to drive the cables/amps. Although nominally an easy load my amps seem to thrive with some active drive.
I've had some suggestions off-forum for driving the TVC from the flipflop that I need to check out.
Ray
Here's the info about the Silk TVCs that I mentioned;
Chapter 5 – TVC theory and Application | DIY Hifi Supply
Ray
Chapter 5 – TVC theory and Application | DIY Hifi Supply
Ray
Ray received the boards, thanks.
Felipe
Boards received, look wonderful. Many thanks.
Thanks for the update. Enjoy!
- Home
- Source & Line
- Digital Line Level
- The Best DAC is no DAC