I didn't just wiggle them about Marcel, I tried to find the 'sweet spot'. When I experimented it became clear that the artefacts only appeared with DSD256 (didn't try DSD512) which steered me towards looking at RV3/4.
It has been a while since I received my raw DSD board and finally got some spare time in the last few weeks. Due to the high number of stuffs inside (tube voltage regulator, 5V psu, 3 transformers + 1 choke, DAC board with tube output board stacked above, Amanero/BBB), I had to design the chassis from scratch and got it done at my friend’s factory. Quite large one (430mm x 360mm x 170mm) with transparent acrylic front panel so that I could enjoy all the tubes inside. Still haven’t decided yet on the color though I’m leaning to gold.
Very nice chassis. Are the screws temporary, they let it down a bit IMO. It would also look good if you illuminated the front panel, like this;
https://www.instructables.com/Edge-Lit-Acrylic-Sign/
https://www.instructables.com/Edge-Lit-Acrylic-Sign/
Just to update on the DAC I was building for my friend, featured in post 2013;
https://www.diyaudio.com/community/...-audio-volume-13.308860/page-101#post-7136410
It was finished and shipped across to the USA, arriving just in time for Christmas, and is now being very much enjoyed...
https://www.diyaudio.com/community/...-audio-volume-13.308860/page-101#post-7136410
It was finished and shipped across to the USA, arriving just in time for Christmas, and is now being very much enjoyed...
I got the chassis painted gold and quickly assembled the stuffs inside. Easily the largest and heaviest DAC I’ve ever built. The output stage is stacked above the ValveDAC board. The Sigma11 PSU is meant for the Amanero/BBB which I haven’t installed yet.
Hi Marcel,
I need a primer on the adjustment process of RV3 and RV4.
On my DSD-only board steered from the JLsound I2S board v.III I get artifacts when using anything above DSD64.
I use the bit clock.
Up to now, I thought that the trimmers 3 and 4 where like the valve trimmers for left and right channel. Obviously, this is not true, as there is only one clock for left and right.
Does this mean, that basically 3 and 4 should be set to the same value? Or do we need to find a good value for both independently?
Just how do I dial in the correct setting? I wish I had used the one-turn trimmers here as I did for the valve trimmers, but I have multi-turners here, so a lot of winding to do - grateful for a hint at how to do it wisely!
If you haven't already done so, a good first step would be to double-check the jumper settings and what 0 ohm resistors are and aren't placed against what is written in post #1 of this thread.
I would expect the optimal settings of RV3 and RV4 to indeed be about the same. RV3 has the same effect on one tap of the two-tap FIRDAC that RV4 has on the other tap.
With a scope, you could try to trim RV4 such that the rising edge of clockn5 is about a quarter clock period (or as close as you can get) after the rising edge of clockn at the highest DSD rate. The same holds for RV3, clock5 and clock.
I hope people who actually built the raw DSD version can share what RV3 and RV4 settings worked optimally for them.
I would expect the optimal settings of RV3 and RV4 to indeed be about the same. RV3 has the same effect on one tap of the two-tap FIRDAC that RV4 has on the other tap.
With a scope, you could try to trim RV4 such that the rising edge of clockn5 is about a quarter clock period (or as close as you can get) after the rising edge of clockn at the highest DSD rate. The same holds for RV3, clock5 and clock.
I hope people who actually built the raw DSD version can share what RV3 and RV4 settings worked optimally for them.
I've set up two Valve DACs now but didn't actually mesure the settings I arived at. I just used my ears and trial and error to get it sounding good - I didn't find the setting to be particularly critical in that there wasn't a narrow 'hot spot' but rater a reasonably wide band of adjustmnt across which it worked, especially with the second build, which used the 20turn potentiometers.
By the way, I don't know what artefacts Achim has, but some noise modulation at high signal levels is normal. I also have that with my original valve DAC.
Sorry about the horrible typos in my last post - shouldn't be respnding that late at night after a long day at work! 😵😴
Re: RV3/RV4
Thank you Marcel and Ray! I will be able to look into this next week and will report back.
Have a good weekend!
Thank you Marcel and Ray! I will be able to look into this next week and will report back.
Have a good weekend!
Re: RV3/RV4
Starting investigation. Marcel, I am not using the DSD input at the Amanero connector P13, using the U.FL connectors instead.
You stated: "P13 not used: R48, R57, R58 and R59 not mounted".
The 0-ohm-resistors are mounted in my build (Ray's friend did the SMT work) - is this detrimental or optional? Removing them would be a hassle I'd' like to avoid.
Which test points would be best for oscilloscope probing the clocks?
Starting investigation. Marcel, I am not using the DSD input at the Amanero connector P13, using the U.FL connectors instead.
You stated: "P13 not used: R48, R57, R58 and R59 not mounted".
The 0-ohm-resistors are mounted in my build (Ray's friend did the SMT work) - is this detrimental or optional? Removing them would be a hassle I'd' like to avoid.
Which test points would be best for oscilloscope probing the clocks?
Last edited:
The 0ohm resistors are basically a selector for whether you use master or bit clock and only one should be installed - take a look at the schematics.
Last edited:
When you use the U.FL connectors, having R48, R57, R58 and R59 mounted causes some crosstalk between data and clock that could have been avoided by not mounting those resistors. Its effect can't be too bad, because anyone who uses P13 (like Ray) has similar crosstalk.
The sides of R29, R30, R33 and R34 that are connected to relatively long traces, but I don't know if you can access those safely.