Yes, I've got that side of things taped.
Looking at the mute circuit for the LM49710 board on the JLSounds website it looks like I don't need R3.
Ray
So, I've been busy reworking the PCBs; hopefully there's something for everyone!
I changed my mind and kept the single boards, so no mute daughter board, but I've left the option to not use Fabio's timer (just leave out the parts and link J1).
I've dispensed with the TLP isolators by putting everything related to the mute control and DSD LED on the dirty ground.
I made some small positional adjustments based on my experience of the previous boards.
Here's the SE board;
and here's the flip-flop board;
Any observations/comments?
I'll order some of these boards later this week. In the meantime, if anyone wants the current boards I have a few I can send?
Ray
I changed my mind and kept the single boards, so no mute daughter board, but I've left the option to not use Fabio's timer (just leave out the parts and link J1).
I've dispensed with the TLP isolators by putting everything related to the mute control and DSD LED on the dirty ground.
I made some small positional adjustments based on my experience of the previous boards.
Here's the SE board;
An externally hosted image should be here but it was not working when we last tested it.
and here's the flip-flop board;
An externally hosted image should be here but it was not working when we last tested it.
Any observations/comments?
I'll order some of these boards later this week. In the meantime, if anyone wants the current boards I have a few I can send?
Ray
I'm amazed, a stupendous work. One thing, maybe that's why there are problems. I have the relay mounted on the output stage and put in the short the output stage when there is no music, in Your pcb, we put in short the dac outputs. rifletterei on this point, and possibly on the pcb I would put only a connector for powering the relay that will short-circuit the output stage that will follow the I2S or a connector to derive the pin of the relè for the shot circuit in the output stage. . Today, as I watched your wonderful pcb, I reflected on this point. when, in one of my early experiments I shorted the circuit after the low pass filter and before the output stage, the bump was present. Test, and define the schematics. Again, congratulations on your skills.
Fabio.
Fabio.
It is funny that you make that observation because I have been thinking that it would be useful to enable the mute operation to work at the output of your project, and not just after the LP filter. I'll think on it more.
Ray
Hi Ray & all,
Interesting to see this thread progress ;-) You ask
and in relation to the way the wires are run to and from the relay may I suggest that you download Saturn's PCB toolkit and take a look at the "crosstalk calculator" tap? It is free and available from here:
http://www.saturnpcb.com/pcb_toolkit.htm
At these frequencies even small parallel runs of wire will induce crosstalk in the adjacent wires thus leading to a reduced channel separation. A suggestion to this end may be to use separate relays for each channel, and - I know I've mentioned this before and don't mean to be unduly persistent - also separate FFs with very equal length wires to the clock pins on the FFs. Separate FFs may reduce the ground bounce phenomenon (for improved channel separation and individual channel "solidity") and equal lengths of wires to the FFs (one millimeter counts here) may give equal timing of the clock signal to each FF. To my knowledge this is especially important for sigma-delta conversion and thus possibly also here. Just some ideas ... ;-)
Cheers,
Jesper
Interesting to see this thread progress ;-) You ask
and in relation to the way the wires are run to and from the relay may I suggest that you download Saturn's PCB toolkit and take a look at the "crosstalk calculator" tap? It is free and available from here:
http://www.saturnpcb.com/pcb_toolkit.htm
At these frequencies even small parallel runs of wire will induce crosstalk in the adjacent wires thus leading to a reduced channel separation. A suggestion to this end may be to use separate relays for each channel, and - I know I've mentioned this before and don't mean to be unduly persistent - also separate FFs with very equal length wires to the clock pins on the FFs. Separate FFs may reduce the ground bounce phenomenon (for improved channel separation and individual channel "solidity") and equal lengths of wires to the FFs (one millimeter counts here) may give equal timing of the clock signal to each FF. To my knowledge this is especially important for sigma-delta conversion and thus possibly also here. Just some ideas ... ;-)
Cheers,
Jesper
Thanks for the feedback Jesper, I'll check out the link.
When you say "At these frequencies...", we are talking about audio frequencies and aren't the parallel tracks only likely to induce feedback when they are in circuit, which is when the mute is on? Crosstalk did cross my mind when I was laying out the board but I rationalised it as I've just commented so perhaps I have more to learn.
I'll stick with the single FF for this project as it seems to work well in my build but appreciate the benefits of dedicating one per channel; perhaps in board Mk. 186
Ray
Hi Ray,
Smiling here ;-)
I've measured my own DSD converter and there's quite some HF noise left even if the filter is active (will try using a ferrite bead with a thick wire here). This noise will be removed by subsequent components (if not earlier then by the loudspeakers) but I would expect the noise (and its intermodulation components) to have a say in how such a circuitry sounds in various setups (due to HF sensitivity, capacities, inductances etc. in electronics, cables etc.).
Cheers,
Jesper
Smiling here ;-)
... Yes, you're right - I missed this, sorry. However, since this is a 6 dB/octave filter some of the HF noise will pass through the filter and still cause cross-talk.
I've measured my own DSD converter and there's quite some HF noise left even if the filter is active (will try using a ferrite bead with a thick wire here). This noise will be removed by subsequent components (if not earlier then by the loudspeakers) but I would expect the noise (and its intermodulation components) to have a say in how such a circuitry sounds in various setups (due to HF sensitivity, capacities, inductances etc. in electronics, cables etc.).
No, not that I know of. It somewhat depends on the impedances present, i.e. a very low impedance power supply track may be less prone to noise pick-up, but in higher impedance contexts I would say it's there (yet frequency dependant). May I suggest you try the Saturn software - IMHO it's quite accessible to use and also quite useful in illustrating the phenomena happening.
Also smiling here and thinking about myself and life as a whole ;-)
Cheers,
Jesper
Chaged 1 SE to flip flip
Hazard500: flip-flop x1
Supra: single-ended x1
Stijn001: flip-flop x1
dwjames: single-ended x1
ichiban: single-ended x1, flip-flopx1
PJotr25: flip-flop x1
Pegasus21: flip-flop x1
skrstic: flip-flop x1
Snicklefritzz: balanced x 1
flowerpot: flip-flop x 2
Umarcus: balanced x 2
Giulio: flip-flop x 1
Caad: balanced x 2
randytsuch:balanced x 1
deanoUK balanced x 1 SE x 1
Vdi_nenna: single-ended X 2
merlin el mago: single-ended x 2
Nikola Krivorov: single ended Х1
pgour : Flip_Flop x 1 + SE x 1
ligascon: balanced x 1 SE x 1
matejS: single-ended x 1
analog_sa: balanced x 1 SE x 1
lukaluka: flip-flop x 1
xaled: flip-flop x 1, single-ended x 1
Anjump123: Single-ended X 2
asanden: balanced x 1 + SE x 1
whmok: single-ended x 2
JimS: balanced X1, single-ended X2
Guglielmope: balanced X1, single-ended X2
iestynrw: balanced x1
torb: balanced x2
Dimdim: balanced x 1
Leon77 : Balanced x 2, Single-Ended x 2
jborden: Balanced x 2
ravid: single-ended x 2
kumori: Balanced x 2
RushBattle: Balanced x 1
francolargo: flip-flop x 1
Barryblue7: Balanced x 1, Single-Ended x 1
Hazard500: flip-flop x1
Supra: single-ended x1
Stijn001: flip-flop x1
dwjames: single-ended x1
ichiban: single-ended x1, flip-flopx1
PJotr25: flip-flop x1
Pegasus21: flip-flop x1
skrstic: flip-flop x1
Snicklefritzz: balanced x 1
flowerpot: flip-flop x 2
Umarcus: balanced x 2
Giulio: flip-flop x 1
Caad: balanced x 2
randytsuch:balanced x 1
deanoUK balanced x 1 SE x 1
Vdi_nenna: single-ended X 2
merlin el mago: single-ended x 2
Nikola Krivorov: single ended Х1
pgour : Flip_Flop x 1 + SE x 1
ligascon: balanced x 1 SE x 1
matejS: single-ended x 1
analog_sa: balanced x 1 SE x 1
lukaluka: flip-flop x 1
xaled: flip-flop x 1, single-ended x 1
Anjump123: Single-ended X 2
asanden: balanced x 1 + SE x 1
whmok: single-ended x 2
JimS: balanced X1, single-ended X2
Guglielmope: balanced X1, single-ended X2
iestynrw: balanced x1
torb: balanced x2
Dimdim: balanced x 1
Leon77 : Balanced x 2, Single-Ended x 2
jborden: Balanced x 2
ravid: single-ended x 2
kumori: Balanced x 2
RushBattle: Balanced x 1
francolargo: flip-flop x 1
Barryblue7: Balanced x 1, Single-Ended x 1
A bit more refinement based on the feedback from Fabio and Jesper.
Each board now has two separate sections, the filter section and the mute section. There are no connections between the sections and the mute section, which includes the DSD indicator LED circuit, is completely on the 'dirty' ground. This approach gives flexibility and removes the parallel PCB tracks.
If you want to use the relay to mute the output from the filter section just run some wire links from mute1 and mute2 to the spare pads I've included at the outputs, otherwise, use mute 1 and mute 2 to ground the output from your project (i.e. hook up your phono output sockets to mute 1 abd mute 2.
I've retained J1, which allows you to not use Fabio's delay circuit
Here are images of the boards, first the SE board;
and the flipflop board;
I'll be placing an order for the boards shortly.
Ray
Each board now has two separate sections, the filter section and the mute section. There are no connections between the sections and the mute section, which includes the DSD indicator LED circuit, is completely on the 'dirty' ground. This approach gives flexibility and removes the parallel PCB tracks.
If you want to use the relay to mute the output from the filter section just run some wire links from mute1 and mute2 to the spare pads I've included at the outputs, otherwise, use mute 1 and mute 2 to ground the output from your project (i.e. hook up your phono output sockets to mute 1 abd mute 2.
I've retained J1, which allows you to not use Fabio's delay circuit
Here are images of the boards, first the SE board;
An externally hosted image should be here but it was not working when we last tested it.
and the flipflop board;
An externally hosted image should be here but it was not working when we last tested it.
I'll be placing an order for the boards shortly.
Ray
My next step is to hack my current flip-flop board, removing the TLP185s, to ensure the switching arrangement for the LED and relay work correctly, then, assuming a positive outcome, I'll order two batches to meet the requirements in the GB spreadsheet.
Ray
The decoupling capacitor on the 555 device is fastened to a GND star point, not a good idea, the added inductance makes the cap pointless, it should go straight across the pins, smallest loop area.
Unless I'm misunderstanding things, the NE555 hasn't got a decoupling cap? What do I need to change?
http://www.diyaudio.com/forums/digital-line-level/273474-best-dac-no-dac-82.html#post4615802
Thanks
Ray
Sorry its the 100nF across the CONT pin and GND, I would still route it direct to the chips GND pin. its loop is pertinent to the 555 keep it small.
You should add a decoupling cap across the power pins as per data sheet....
You should add a decoupling cap across the power pins as per data sheet....
I would have used a solid GND for all the boards GND connections, and ditched the star wiring for a timer device and a board this small, it just adds inductance and extra loop areas.
Thanks Marce.
OK, will do, it's simple.
I'll see if I can squeeze one in...
I have used solid grounds overall, but the mute function operates by switching on/off the ground connection of the circuit so essentially all of the ground connections star at the transistor anyway?
I used the star arrangement so that people could easily omit the relay delay function (NE555) if they wished and a single jumper link to the transistor seemed to be the simplest way?
OK, will do, it's simple.
I'll see if I can squeeze one in...
I have used solid grounds overall, but the mute function operates by switching on/off the ground connection of the circuit so essentially all of the ground connections star at the transistor anyway?
I used the star arrangement so that people could easily omit the relay delay function (NE555) if they wished and a single jumper link to the transistor seemed to be the simplest way?