Valve DAC from Linear Audio volume 13

[agent.5]

I am using the TI series of LVDS drivers, because it has SOIC-8 and SOT-23 footprints with different pinouts. I am using both footprints to avoid using vias to route traces to the back side of the board.

Datasheet
https://www.ti.com/lit/ds/symlink/sn65lvds1.pdf

And the answer to the question of what is the rise or fall time is, according to Table 7.7, a maximum of 0.7ns.

I think the Carlisleit connector is okay to use. It is not a special order item, unlike those from Samtec and Molex, and is in stock for sale at Digi-Key for around $5.

I read that when two oscillators are placing close together, they may affect each other. So, it is a good idea to have some separation between the two. I am also wondering, instead of building a mezzaine, looking architecture and high-tech, the entire clock system may actually perform better by having two separate clock pcbs, each having an oscillator on top, and a power supply regulator underneath. They can connect to the DAC side of the board through existing u.fl connectors.

[nautibuoy]

I've made the refunds for the overpayments.

I've received some final payments - thank you.

Will publish an updated spreadsheet later.

[agent.5]

Originally Posted by agent.5

And the answer to the question of what is the rise or fall time is, according to Table 7.7, a maximum of 0.7ns.

correction: It should be typical 0.7ns and a maximum of 1ns.

[MarcelvdG]

Originally Posted by agent.5

And the answer to the question of what is the rise or fall time is, according to Table 7.7, a maximum of 0.7ns.

Assuming a speed of propagation of 200 000 000 m/s, the signal can travel 14 cm in 0.7 ns, so it can go back and forth four times over a distance of 1.75 cm.

Originally Posted by agent.5

I think the Carlisleit connector is okay to use. It is not a special order item, unlike those from Samtec and Molex, and is in stock for sale at Digi-Key for around $5.

I agree with you if the contacts are shorter than 1.75 cm or if it has the correct impedance or both.

Originally Posted by agent.5

I read that when two oscillators are placing close together, they may affect each other. So, it is a good idea to have some separation between the two. I am also wondering, instead of building a mezzaine, looking architecture and high-tech, the entire clock system may actually perform better by having two separate clock pcbs, each having an oscillator on top, and a power supply regulator underneath. They can connect to the DAC side of the board through existing u.fl connectors.

I have seen serious problems due to unintended coupling between oscillators at work, but always with LC oscillators working at very close frequencies that were placed within a few millimetres from each other. I would not expect any trouble with crystal oscillators with a frequency difference of a few percent, like you use.

[MarcelvdG]

Originally Posted by Mikky1975

Ciao,
I’m wondering if I’m underestimating the effect of having the BBB inside the same chassis as the dac. I’m interested in your thoughts and projects because I was thinking about just doing a box-in-a-box for the bbb and then isolate/reclock. Do you think there should be audible benefits in a two-chassis implementation? Is it worth the extra care one must give to the clock/signal across more than a couple of centimeters?
I’m an IT guy, not an electronic engineer so be gentle, I’m just asking ��

Cheers,
Mirko

I find it difficult to say anything useful about that, but maybe Ray can, as he has experience with putting a BBB close to a DAC.

[nautibuoy]

Originally Posted by Mikky1975

I’m wondering if I’m underestimating the effect of having the BBB inside the same chassis as the dac.

My Valve DAC, which sounds fantastic (and I think is why people are interested in building their own versions) has the BBB and reclocker in the same chassis as the DAC board so I wouldn't be overly concerned about it. Could it be improved by separating the BBB? - maybe, but it is unlikely to be more than a small difference, if it's discernable. Perhaps I'll find out with a second build, when it'll be handy to have the original build available as a baseline.

In the picture, the BBB/reclocker are under the main board in the vicinity of the yellow caps/large brown resistors.

[Mikky1975]

Thanks Marcel, Ray, agent for your kind replies.
@Ray: your realizations are amazing and incredibly your dac interior looks even better. I wish I'll reach a similar result. I came across the Valve DAC after reading a lot on the DSC2 and Pavel's work. Originally I was more keen on the DSC2 and then found this tube implementation... love at first sight

If I may ask: how you other diyers put the pieces together? I would try PPY reclocker but could not find an open GB so I found others went with Ian's fifo/clock and others with Cronus/Hermes. Any advice on this?

Best,
Mirko

[agent.5]

PPY posted gerber.

https://puredsd.ru/BBBreclk.zip

You can upload this to

PCB Prototype & PCB Fabrication Manufacturer - JLCPCB

and have it made inexpensively.

[MarcelvdG]

Originally Posted by nautibuoy

In the picture, the BBB/reclocker are under the main board in the vicinity of the yellow caps/large brown resistors.

So that's basically the same digital at the bottom, analogue and mixed-signal at the top strategy that I used for the FPGA board and other digital circuitry of the original valve DAC. It also worked well for me

[Mikky1975]

Originally Posted by agent.5

PPY posted gerber.

https://puredsd.ru/BBBreclk.zip

You can upload this to

PCB Prototype & PCB Fabrication Manufacturer - JLCPCB

and have it made inexpensively.

I didn’t know it was so easy nowadays, thanks!