Schiit Eitr Preview Thread

Guess it'll stack with my Vali 2 to boot haha. DOWN THE RABBIT HOLE I GO BOYS.

 

haha welcome to the rabbit hole. the good thing about the eitr given its price and performance is that you likely won't spin your wheels buying and selling and daisy chaining products like many have done in the past. one and done!

 

Not quite.

Clocks, potentially.

Here's @atomicbob on this:

And on a separate occasion:

 

I'd call the rabbit hole cozy and warm, but that might just be all the tube amps lying about.

But yeah. Waiting to see if headphonebar will eventually stock them. If not, can just buy direct from Schiit. Not bad for a one time problem-solving implement to my current and future audio chains.

Edit: Headphonebar won't be carrying the Eitr anytime soon, so I'll save up a bit of pocket change from odd-jobs and grab one when I can.

 

Thanks so much, Grahad2 and atomicbob. That's exactly the info I was looking for--I don't want to send in my Yggdrasil for the G5 upgrade, but I've already got my CD transport occupying the coaxial input now, so I'll need to move one of them to the AES input.

And it got me thinking: those of you on the fence about the Eitr vs the G5 upgrade should also consider whether you will want the upcoming "Manhattan Project" (DSP). In a previous post on his dedicated thread, Mike Moffat indicated that the MP would initially be released as a separate Magni/Modi-sized box. Given this earlier discussion, I'm not sure if USB output is certain. So if you want to play the odds, the Etir might be the safer bet for future compatibility with the initial flavor of the Manhattan Project box.

But, again, this is just a guess.

 

Newbie dumbass question - should I remove my XMOS based USB drivers from my PC when switching over to the Eitr from the SU-1?

Since the Eitr is not an XMOS device, it makes sense to me to do this, but is it necessary / recommended?

 

Their presence (or absence) will have no effect on Eitr.

They'll just sit on the disc, doing nothing, unless you connect an XMOS device to the machine. So removing them, or not, is purely a matter of whether you care about having unused files, driver entries in the registry, hanging around on your machine.

 

One factor that MAY lead to audible changes with wet non-polar electrolytic capacitors (conventional non-solid PS caps, not solid polymer OSCON), is they need an applied voltage across the plates to form an oxide barrier on the anode to act as the dielectric. This is aging is typically done during the intial manufacturing process to force out imperfections and "self heal" these production problems. Then they are subjected to shelf life, typically spec' ed for 1000 hours, after which the dielectric deteriorates and the capacitance decreases and ESR increases. The longer the shelf life encountered, the longer time an applied voltage is needed to self heal the capacitor to its stable values for its service life.

I see from your excellent articles on the Eitr that you were using engineering prototype units for your evaluations. Might they have had been power cycled many many times before you had a chance to audition them? So probably no need for you to reform the caps, as they had not sat for an extended period without an applied charge, and deteriorated.

Makes sense when you acknowledge "it takes a few minutes powered up" for the Eitr to be as good as it gets. Your units had been cycled repeatedly in the near term. Takes reforming the capacitors out of the mix.

However a fresh out of the box Eitr, might not have had enough time powered during final QA to fully reform the caps. Especially if they had been sitting well past their 1000 hour shelf life. That Eitr would need some time at the customer end for the capacitors to self heal and stabilize. A wrote a couple of days run-in, to be on the conservative side. I noticed my Eitr SQ relaxing most over 4 hours, and nothing noticeable in changes after 18 hours.

That is one mechanism for changes in Eitr SQ from power-up. You also mentioned others, even some with measured differences. All possibly valid, I just don't have the patience, tools or even want to investigate to root cause. Who really cares, just leave your Eitr on 24-7 just like your Schiit DAC and be done with it.

My original comments were made to Gidgate, to not judge his freshly powered up Eitr in direct relation to his established Rednet/D16/Mutec combo. My advise to him, lost in these posts were to give the Eitr some time to settle in. I experienced my Eitr sounding better with it powered on after a while. If you don't agree that is possible, I guess must be delusional. I'm too busy listening to music to care.

Thanks for the articles on the Eitr, Torq. Would not have known about it as soon as I did.

 

Keep me in mind whenever you order. I might grab an Eitr and something else, and we could combine orders to save on shipping.

 

And by what mechanism do you think that's going to result in audible changes to a digital signal? Particularly one where we can prove, without question, that the data that was sent is exactly what was received.

 

Some digital designers may disagree with @Torq

from the perspective that even though it is digital packets, it is represented in the device as analog voltages and some designers do believe everything matters. perhaps it is not a cap, the disussion of which can lead to more pissing up a rope confrontations. Hard to prove one way or the other. Again there are designers who say leave a dac on 24/7 for best sound, they obviously have a reason or point of view.

 

I was curious if anyone could say for certain if the dirty side of the Eitr is powered by the USB +5V?

Not that it really matters.

Except to the very nervous.

 

And I'd agree with that.

At a minimum, it matters for clock stability (jitter) and it matters for things like resistor values (discrete or on an IC substrate).

The 24/7 thing is more about convenience, though. Saves having to plan ahead by an hour to have things at their optimum point.

Yes, it is.

And you either recover the correct values as they relate to the pulse-train, or you don't. In this case, we can prove we do.

..

There are really only three things of significance when it comes to DDCs and their potential to affect sound: jitter, data delivery and electrical noise. We've talked about the former, can prove the 2nd is a non-factor here, so that leaves noise - which half the point of the box is to isolate.

So my question remains ... by what mechanism does, what is effectively, "capacitor burn-in" affect these three variables as applied to Eitr?

 

I was positing on why you were not experiencing a lack of run-in on your end with your unit, i.e.: reforming capacitors from a long shelf life not a variable.

I for one would be very interested to see your stated proof via measured data. As I understand it, specifically directly comparing data in> out of a just powered up, received from common stock at Schiit Eitr, and one that has been continuously powered up on your workbench. Requires two specially selected DUTs, and it is doubtful it can be inferred from other less considered measurement results.

 

"Damn it Jim, I'm a doctor, not a physicist!"

 

And to complete that thought it's necessary to explain how the state of those capacitors might actually affect a relevant variable in the performance of the unit. Not just say "because capacitors change over time" and leave it hanging.

We can do this if you like. Plan on being present for the test, however, as I'm not jumping through hoops just to post the proofs and then have random arm-chair quarterback types attempt to nit-pick every aspect of the test.

Before that, you're still going to have to come up with a sensible explanation of how the ridiculously small deltas in performance of a capacitor undergoing "burn in" are making an audible difference to the behavior of Eitr, in the face of the comparatively gross deviances that result from the normal variation in performance due to normal component tolerances.

So, you address the first question without voodoo and handwaving, and then we'll get on with setting up the test.

 

Head-Fi anyone?

 

I merely was entertaining possible causes for the differences between our experiences. Not The Root Cause. Remember, my comments were in regards to the OP making a value judgement between his fresh Eitr and his reference too soon after first powering up upon arrival. I cautioned to let the Eitr run-in for a few days......

Would you advocate it is valid for everyone to make such an out of box comparison shoot out/judgement of the Eitr against a current TOTL USB implementation?

If not, and think some run-in (minutes on your well cycled Eitr) would allow a fresh Eitr to work to it's best, then we are only disagreeing on timeframe. I only sided on a more conservative estimate for the out of box unit.

Cart before the horse. Existing proof was offered then withdrawn with conditions.

I live in Portland, And can drive to Seattle whenever you wish to schedule the experimental proof in your home lab as outlined.

I can be there.

 

It was stated as existent, I didn't offer anything. But see my prior comment about the nonsense of posting "proof" online and then having the methodology critiqued ad-nauseum. Been down that road enough that I'm not going there again. Way too easy to post fake results, and everybody half-way competent knows this.

And the only "conditions" are on conducting a more specific test (known-fresh Eitr, which is your sole point of real conjecture) - which is your request per your second post.

If you can't posit a reasonable explanation for why the capacitor effects you cite would affect the performance of the Eitr, then it's a pretty laughable position to argue from, and the lame debate-school tactics are just that ...

..

I'll order up a pair of Eitr, and see what I can do to make sure they've been touched as little as possible (I doubt there's a way around the board-house power-on/verificatoin tests). Or if you want to be sure one hasn't been "pre-fiddled with", you can order one yourself and I'll re-imburse you after the test and we can use that as the "virgin" unit.

One thing you can be sure of, though, is that I'm not inviting some Rando into my home. It'll happen at the office lab. That's absolutely state of the art, even though it doesn't need to be for this silly little test.

 

That will work for me. Please forward the schedule and location when known.

Thanks.

BTW, by home lab, I meant in your measurement comfort zone, not literally your residence.