Big can of worms. .
For consumer audio less than 100 pS + No obvious deterministic jitter. Stop there and use the budget (almost zero anyway) to make better drivers.
For pro 20 pS is achievable even using standard SPDIF receivers and no magic.
For ego driving audio better is possible but published specs are misleading. Most published specs are for network performance and the LF limit is 10 KHz. Most of the audio related jitter is below that frequency and at higher levels. The conundrum is that below 100 Hz there is no evidence that it will generate any detectable or discernible artifacts unless the level is huge. if it were a problem no turntable or tape machine ever made would be usable. . .
Dividing down the clock reduces the jitter but thats built into most modern DAC's. Further higher frequency crystal oscillators tend to have higher phase noise at about the same rate as the decrease from dividing down.
I would content that most improvements heard from upgrading the oscillator are not from lower intrinsic jitter. More likely from better overall execution. However there are published reports of "good sound" from systems with essentially broken master clock solutions so who knows?
For anything above 24/96 (and really even for that) you need the 480 Mbps rate and the AD isolator chips do not support that rate. There is a box that does but you can get a laptop for less than the isolator: USB Extenders for Long Distances They do work quite well but are not an impulse buy ($450)
You are being hard on them
!Yes , me and Jwilhelm found that to read. This determined how "anal" we design
the clock and power supply for our DAC , considering what the article says
about bitperfect SPDIF/USB. Jitter won't be an issue with a PC.
But , for laughs ... someone might plug in a 20'th century CD transport (toslink/spdif). To avoid embarrassment , we decided jitter (might) be an
issue here - thus an "anal" overdesigned clock and digital PS WILL be used.
It's one thing to talk about it , to make the "real thing" requires a lot of
reading and the right choices.
OS
All UPS's use a variety of switching to create the output power. The cheaper ones use a "step" wave that has maybe 4 transients per cycle and some have an output transformer that further isolate the noise. The HF "sine wave" converters have a lot of HF hash in the output. 10A at 20 KHz will inevitably be noisy. Also most have pretty short battery life. They are supposed to keep things running long enough to have an orderly shutdown.
The full time versions won't have a transient when the primary power fails but will cost in excess power consumption. I have a big UPS in the lab but its to run the cable modem and the phone. its a standby type so normally it does nothing.
Isolation transformers are a better option. An ultra isolation transformer will get you the best practical isolation. They have become harder to find today. Here is a sample Topaz Electronics 91001-12 Ultra-Isolation Transformer Noise Suppressor ON SALE at RecycledGoods.com and more background Testing isolation transformers for attenuation. | content content from Electrical Construction & Maintenance (EC&M) Magazine
I don't understand why the guys who are concerned about jitter and many other aspects of digital don't make a stronger push to move to a more idealised system. Not a Pono, but a sensible Pono. Something that will accept cheap SS card memory, with a minimal operating system, high quality clock, simple display, TOTL DAC and maybe battery power. Something that can turn everything else off and just use the firmware to read from the card and write to the DAC during playback, as long as you don't wake up the OS. It's easy to create a small section in prog logic to handle this with a comparatively simple finite state machine.
This just takes you past pits and lands, data-related jitter... ever since SD cards got cheap enough to record HD video it's been a no-brainer. If you like minimal in a playback device.
A couple of threads here are about devices close to this spec.
You can certainly run asynch USB and reclock the data, that lets you use a networked HD for storage, but if you absolutely want to minimise the problems that are commonly raised such as digital noise getting onto analog circuitry, PSU noise, EMI, and, of course, jitter, then a 'super DAP' has a lot to offer. It's what I'd be seeking to build or buy if I didn't just use whatever DAC is in whatever device I'm using at the time.
Just why flow control on USB had to wait so long or be so obscure and expensive to implement is a mystery to me.
This just takes you past pits and lands, data-related jitter... ever since SD cards got cheap enough to record HD video it's been a no-brainer. If you like minimal in a playback device.
A couple of threads here are about devices close to this spec.
You can certainly run asynch USB and reclock the data, that lets you use a networked HD for storage, but if you absolutely want to minimise the problems that are commonly raised such as digital noise getting onto analog circuitry, PSU noise, EMI, and, of course, jitter, then a 'super DAP' has a lot to offer. It's what I'd be seeking to build or buy if I didn't just use whatever DAC is in whatever device I'm using at the time.
Just why flow control on USB had to wait so long or be so obscure and expensive to implement is a mystery to me.
Last edited:
Some interesting material here on noise, and USB galvanic isolation - Archimago's Musings: MEASUREMENTS: Corning USB 3 Optical Cable, Ground Loops, and Noise..
That is a good summary. You can get much better results than just balanced does by adding isolation transformer (which equi-Tech later added to their balanced config xfmr) and some filtering. I use the TOPAZ 1KVA combined with filtering for T&M.
https://www.gearslutz.com/board/high-end/682081-balanced-power-conditioners-bpt-vs-equitech.html
There are several inexpensive ways to clean up the noise, provide isolation and eliminate grounding issues...... not for here... this is JC's forum on analog design.
THx-RNMarsh
Last edited:
Who's Keith Richard? I tried Googling but it kept wanting to return results for Keith Richards.
se
Here are two rooms out of 6 that house diagnostics and controls for 6 experimental chambers (HE weapons). The big black thing hanging from the ceiling is a 10KVA isolation transformer to power the equipment in the racks. It was my responsibility as Sr EE Technical coordinator to coordinate the design/development and construction for the electronics and power for the experimental areas in the HEAF facility. That included system grounding.
Same for other projects --- three phase? Three-phase isolation transformers.
THx-RNMarsh
Last edited:
USB2.0 specifies three speeds:
Low speed 1.5 Mbit/s (Mouse, Keyboard, Joysticks).
Full speed 12 Mbit/s (POTS, Audio, Broadband)
High speed 480 Mbit/s (Video, Storage, Imaging, Broadband).
Then AduM4160 is specified as
"USB 2.0 compatible
Low and full speed data rate: 1.5 Mbps and 12 Mbps"
Then some channel bandwidth calculations for our needs
16bit/44.1kHz 16 (bits) x 44100 (samples per second) x 2 (channels) = 1.41Mbps
24 x 96000 x 2 = 4.61Mbps
24 x 192000 x 2 =9.216Mbps
What is that I miss Demian?
Thanks for that
Pavel
Yesterday night I listened to your drum files through the confirmed 24bit card. Results the same as I already reported.
So the system’s resolution bottleneck for me is my ears (I already knew that but let hope be the last one to die
)Regarding isolation transformers, UPSs, noise, grounding myths and facts, recommended reading:
FIPS PUB 94, GUIDELINE ON ELECTRICAL POWER FOR ADP INSTALLATIONS (21 SEP 1983)., This recommended guideline for Federal agencies identifies and describes the electrical environment for safe, reliable operation of automatic data processing systems (ADP). [withdrawn July 29, 1997]
FIPS PUB 94 GUIDELINE ON ELECTRICAL POWER ADP
George
Last edited:
Good move! 1 will get you 10 they are more robust than the nicely packed, shiny ones.
We are numerous in this case.
Does-this mean that good music (or what our ages tell as it ;-) is more important than distortion numbers ?
I haven't tried it with all FETs yet, but doing this with a cross-coupling (i.e., use the sources of the bottom pair to drive the gates of the opposite top pair) using pFETs and tubes got me better than -100dB IM (19/20kHz) at 1.3V out and CMR that was high and nearly constant across the audio bandwidth. There's a lot of power in balanced 4Q indeed.
Any time, anywhere, Christophe. Thankfully, music is so varied that there is something for everyone, and if you care for it (music) enough, distortion figures will mean little if anything to you, unless the recording is truly a premeditated crime. And such are not too many, here and there.
I read somewhere that the Beach Boys' best selling ever LP, Pets, was actually recorded on a portable machine in a hotel room, and that they were stoned and/or drunk half the time. Does anybody really care to measure it? I don't, such were the times, I just tap my foot to some numbers on it.
When you listen to the music of the period from Johnny Holliday, Adamo, Nana Mouskuri or Demis Roussos, do you listen to them or the distortion, which is probably bad by modern standards?
Same with me, but no, it is not so much a matter of age but rather of statistiscs.
It is improbable that the past 20 years more listenable music would have been recorded than in all those decades before.
It's a different system effectively.
For a while I have been running a wireless digital audio link, or Kenwood CDP feeding a Behringer Mic2000 preamp feeding balanced signal to the B2031 active speakers.
For now it's an Edirol UA-25ex directly feeding balanced signal to the B2031s.
Ok, overall the system is very nicely lower perceived distortion, cleaner and lower lows and clearer/more extended highs, perfectly good centralising, perfectly good side and depth information, sharper/bigger transient sounds, and able to run louder......that's with the USB isolator in circuit.
With the isolator out of circuit, all of the above reduces....unacceptably so when compared to all of the above.
First order effects is fine noise over everything, reducing clarity and frequency extension more especially in the highs.
Transient attack and max usable SPL suffer also.
The A/B isolator comparison is with the netbook running on batteries and disconnected from the charger.
For now I am not sure where the benefit of the isolator lies.
It could be that the Edirol is powered by cleaner DC, or maybe partly the Data regeneration....the power and data connections to the Edirol using the Type A to B adaptor is as short as can be.....ie, no long USB cable to the device.
I can bridge the host/device power and earth connections leaving only the data regeneration as the variable as a future experiment.
So, the result for now is a really clean, quick, powerful sounding cheap system.
It goes loud and big, but with no real colourations.....not falsely clinical, not falsely grooving, just matter of fact clarity that is renovating quite a few tracks I thought I knew.
Adding a Quantum Purifier formerly known as Bybee, detracts from the sound from quick trial...transients and highs suffered and a wrong signature over the sound.
Dan.
Last edited:
- Status
- Not open for further replies.