I can not believe that “new software” needs burn in. This borders on the ridiculous. Hardware exists. Software does not. IF a transistor now exists in a new state, 0 vs 1, it just does. If there is any “stabilization”, it would take milliseconds.
Peace
Bruce in Philly
I think the new release should be called Cloudnight, you know, after Redcloud?
Sort of and sort of not. FPGAs, like all digital, are really analog devices with conventions about how and when certain states are discrete, and try not to stay in other states for long. The Xilinx tools are well aware of the analog nature of the beasts and know details about how each resource in the FPGA acts over lifetime, temperature, power supply voltages, etc. with sub picosecond accuracy. They make sure that none of those variances change the digital output from a one to zero in a correctly designed program. However the user can easily ignore warnings (or turn them off) about some of the timing and have more metastability, that is, more places where the value isn’t known at a given point in time where some other place is reading it. Anytime you have more than one clock interacting you have the possibility of metastability (sort of like when you are driving and the light changes to Red: no matter what the light’s timing is, there is a place that a given human (running with it’s independent “clock”) will freeze for an unknown amount of time trying to figure out whether to brake or go for it.) A good design tries to keep metastability failures as rare as the product requires (e.g. perhaps about a 100 years between failures or for other products maybe 10 minutes between failures…) The fewer failures you want the more resources you need to consume…
So yes, the noise and jitter changes over time and use and hence each part of an FPGA ages differently for each release and there’s the clear possibility of separate “burn-in” for each piece based on how much it was used in each release…
Reflashing probably doesn’t make any difference at all for multiple reasons: The FPGA has no flash and is loaded new on each power up and reloading the FPGA doesn’t rejuvenate or change function/timing nearly as much as temperature and voltage during “normal” operation.
Needs burn-in? No, each release of software will work. But is there a change over time that depends on the software and running time? Certainly there is. Whether people can reliably hear a difference is another question, but in some past releases that used newer areas of the FPGA there were fairly consistent reports of a new “burn-in” going on from the members here. Those kind of reports didn’t happen in releases that used very similar resources to older releases. FWIW those burn-ins were not hundreds of hours, but also they weren’t just an hour or so.
I love my DirectStream DAC. This will be my first DAC update, which I’m looking forward to. I cannot imagine how it will improve but it appears it will. As far as i’m concerned, there is no rush; RedCloud sounds pretty amazing.
Interestingly, I went to a great dealership last weekend, HiFiLounge in Hertfordshire, where the hi-end Chord DACs were being demonstrated. There was a long talk, with time for questions, by Rob Watts, who has written all the code for the FPGA in all the current Chord models. Most of the talk was very technical and, quite frankly, well over my head. However, many of the design goals described, low noise floor etc appeared very similar to Ted’s for PS Audio. I asked Rob whether he was likely to update the code in the future and he responded that he had the code as ‘right’ as he could and there were no plans to issue further releases. Part of the reason for this was that he had used up nearly all the capacity for both the DAC and Scaler.
The very fact that PS Audio regularly issue updates seems to be pretty unique in this industry.
Obviously – Rob Watts is only a mere mortal…
I think we’re pretty set of Snowmass.
I think the nomenclature of the realeases should be more transparent.
Simply:
DirectStream 1.0
DirectStream 2.0
…etc.
This would help underscore which revision is at hand and its position relative to others; and since the sonic characteristics are so discernible, it reflects the reality that the releases are essentially new versions of the DAC.
But that’s the scientist in me talking.
The version screen has the information you may be asking for. For each component increasing numbers work the way I think you want.
Where did all this silly “cool” naming conventions start? - Mount X, Lollipop, Goldilocks, Beaver and Wally, Weeble, and Sam?
I want Version X, Black Faceplate, One Button, and hair to grow where it stopped. Is that too much to ask?
Peace
Bruce in Philly
Hi Paul, “Snowmass” sounds good to me, but if PS Audio has not already done so, it might be worth checking with the folks at Aspen Skiing Co. beforehand. See this September 10, 2018 article from The Aspen Times:
Oh, come on. Snowmass is the name of the mountain and an unincorporated community. There are no copyright issues here.
I believe the issue is TRADEMARK, not copyright, and yes, the word, SNOWMASS has been trademarked multiple times. Below are the live ones, according to the US Patent and Trademark Office. It seems highly unlikely that anyone would confuse a DAC operating system with a ski resort, but the Aspen Times article I posted indicates that the ski resort’s parent company is, shall we say, ‘diligent’ about protecting its trademarks. I imagine PS Audio has considered these issues already, but I was struck by the recentness of the Aspen Times article and thought I should mention it. The fact that Aspen/Snowmass and PS Audio are both in Colorado is another reason for potential concern, I think.
To be clear, I think that if PS Audio wants to call the next DAC OS “Snowmass,” it should be able to do so. But I would not want PS Audio to have to deal with a [threatened] trademark infringement case. (Strange things can and do happen in the trademark arena.)
Absolutely fascinating, Ted
Many thanks Ted!
If this keeps going over a name, the firmware is probably not going to get released by next Christmas let alone this one…
Nope. We’ve narrowed the field down to about 8. Darren’t taking them home tonight and will narrow it even further on a different set of speakers. Tomorrow we will make our final decision. Still to come is the marriage of the PIC code and then we have to recheck everything.
Yes, they are playing off the Snowmass Mountain resort bit - there’s no conflict because of what we do. Snowmass is an official name in the state of Colorado for the 14 teener. I think we’re fine.
I appreciate you watching our backs. The article was pretty explicit that it was the use of the Snowmass proper noun and Aspen leaf logo on a T-shirt sold in the town of Aspen that has them up in arms and rightfully so. I would hope they’re not going after an audio company using Snowmass for a product update. I’ll check with our attorney just in case.
