Cronus - It's about time.

[francolargo]

I am using some really nice external clocks into the chronus. I need multiple master clock and I2s outputs to drive dual mono DAC’s synchronously and an extra word clock to source an ADC.

I'm sure Brian can suggest something, but buffering a clock typically adds some 'jitter'. Maybe not a lot, but possibly enough to negate (some of?) the benefits of your sweet external clocks. TI has a ton of information on line. The only question is where to start. Maybe here? Clock Buffers | Overview | Clocks & Timing | TI.com

Best,

Frank

 

[stephen1212b]

Thanks francolargo, that's a very helpful link. Knowing that proper impedance layout will go a long way toward getting the best possible performance I'm actually hoping to inspire a Cronus Plus. Just thinking out loud, couldn't clocks be paralleled on the input side to unused channels on the Cronus?

 

[francolargo]

I'm not 100% clear on all of your objectives. Perhaps a block diagram would help. But for the last few years I have been using the current Cronus to drive three Buffalo Pro stereo DACs, using 6 of the 8 Cronus channels. Perhaps part of the sauce that makes it work so well is the fact t that the DACs are enough remote from the Cronus that I used TPA teleporters from the Cronus to carry the synchronous I2S over to the Buffalos. Were I to use the u.fl. inputs to connect some external 'super-clocks' into Cronus, I don't think there would be any additional buffers needed.

So, a clear statement of objectives might help suggest work-arounds and/or time (or money) saving strategies.

Best,

Frank

 

[stephen1212b]

So the Cronus outputs can each drive at least 3 teleporters? I was intending 4 inch ufl fan out to two pro DACs and not sure what double termination or strange reflections might come from sourcing in essence from the middle of a coaxial wire. The ADC is remote so a teleporter pair could work but seems overkill as only one ch would be used to carry word clock which normally runs unbalanced for the ten or so feet that would be needed. A teleporter would also require additional power at the receiving end. Any idea how much jitter a teleporter pair would add compared to a simple line driving buffer?

 

[stephen1212b]

One other question to those who know, The Cronus aligns the various clock edges and we have been instructed to use the same length wires to connect them to the DAC. When driving the DAC synchronously the master clock needs to be longer to reach the connection point. Should we still use the same length wires for the I2S and the master clock?

 

[francolargo]

I was intending 4 inch ufl fan out to two pro DACs and not sure what double termination or strange reflections might come from sourcing in essence from the middle of a coaxial wire.

The advantage of the Hirose U.FL connectors is that they provide a 50 ohm termination, which helps with reflections. So I would find a way to use two separate u.fl. wires rather than messing with splicing. But there is another option that works equally well for me. Direct soldered silver wire - but very short since it is not shielded. Re: wire length, a couple cm difference is OK with soldered connections.

Any idea how much jitter a teleporter pair would add compared to a simple line driving buffer?

All I can say is that the sound seems faultless compared to other setups I've tried with similar active components but no teleporters. For example, the sonic difference between synch and asynchronous signals into the b3s is maintained, teleporter or no. The teleporter transmitter will handle up to 150MHz (IIRC), so the master clock traces look very good on the scope. Regarding distance, the teleporter transmitter creates a differential signal that enters the twisted pairs of cat 5-6 wire. So common mode junk after a 4 foot trip through cat5 is not a factor.

 

[Abra]

I have a BBB->Hermes-Cronus installation (i2s, sync mode).
I don't use it because the software hasn't been supported by Volumio for years, and it lags behind today's networking capabilities. RPi can offer more in this respect.

Nothing else to do hardware-wise other than swap BBB for RPI, and swap BBB Hermes for RPI Hermes?

What software images will work with?

 

[francolargo]

Here is the GitHub link for the Hermes RPi overlay. GitHub - miero/tpa-hermes-rpi-dev

I have not personally used it but it appears straightforward. I have used and like dietpi, so I would probably add the configuration file feature - a nice convenience!

 

[drone]

What are the right Jumper settings on Cronus v 1.4 (MCL+/MCL-)?
I would like to use it in async mode with 49/45 MHz clocks (J1:2)?

thx

 

[BrianDonegan]

If you are using it with our RPi board, select MCK-. For all other applications, MCK+ should work.

 

[drone]

If you are using it with our RPi board, select MCK-. For all other applications, MCK+ should work.

Hi,
thank you for helping.

Setup is Amanero-Hermes-Cronus --> B3pro9038

It is not working (Cronus v1.4) (also with setting MCK+ with BIII 9830pro).
While my Amanero/hermes/Cronus v1.1 is working stable in the same setting with out problem. When I swap it for the v1.4 cronus no lock or sound
Tried to recheck an solder the SMD-chips so far bur no success.
How to troubleshoot/measure to fin the problem?

thx

 

[pixelpusher]

After a frustrating few weeks with the BBB I throw in the towel. I went back to the Amenero/Hermes. This is working well but I had to change the clock jumper to 1:4. With the BBB it was at 1:2. My Cronus clocks are 49/45 and I have Cronus connected in sync with the Buffalo 9038.
Can anyone here explain how this works? I've read this entire thread and still do not understand the clock jumpers.
Thanks so much
Mike

 

[francolargo]

If you read Russ’s introductory post, it seems the design was intended to accommodate a broad range of options in clocks and outputs. In practice, slower clocks produce less phase noise, which is good, but they limit the max stream sample rate. So the dedicated tweaker can experiment with clocks and divider configurations.

I use a BBB with 45/49 clocks at 2x and it has always ‘just worked’, so I’m stumped. I haven’t heard a qualitative difference based on clock divider setting.

 

[pixelpusher]

Hi Frank,
What software do you use on your BBB? Sync or Async?

 

[francolargo]

Hey Mike, On the BBB I’m just running old Debian with old Botic - all of which creates a synchronous I2S stream, which is what the BBB/Cronus is designed to do. The fuss about sync/async has to do with the B3’s re-timing jitter correction, and whether the B3 uses the Cronus clock or it’s own 100MHz clock. …or do I not understand your question?

 

[pixelpusher]

Sorry for the confusion. You answered my question.
For several years I was blissfully listening to the BBB/Hermes/Cronus using GentooPlayer and Roon. Recently Roon came out with a mandatory update that broke Gentooplayer. The developer of GentooPlayer decided it was not worth his time to support the BBB/botic. I tried the PureDSD firmware but for me it has reliability issues. So now I am looking for alternatives.

 

[pixelpusher]

Wondering how many BBB/botic users out there. I think we are a dying breed.

 

[francolargo]

For my ‘big’ system I’m pretty well married to the multiple TI McASP cores because of running 6 channels. I would love having much more CPU for running FIR filters, but the beaglebone AI has thermal throttling issues and only 3x the BBB’s CPU. So, I’m just enjoying the music such as it is, hoping a path forward emerges. It might turn out to be HQPlayer, which I’ve been resisting because it doesn’t seem very “hands-on DIY”.