Looks good to me, I was myself just testing a very simple CMM with the power lines and a short cable and had some interesting results (SQ-wise), and then found out that my DAC can still play with the GND disconnected (it has a battery), so that's fun stuff too (with more SQ results).
Haven't done the filtering on the data lines yet and have yet to try other types of filtering on the VBus line.
The thing I have read about the Adums is what they add that we like to avoid usually, namely jitter. Win some here, lose some more over there.
Good to know.
It will be interesting to see what he says.
The Uptone Audio USB Regen takes particular care of that sort of thing, IMO.
Some people go so far as to actually test these in their system and then either buy them or send them back. If it's affordable to them and provides them with a positive for their budget in their system, it's fine. It's up to other people who think they can do better to commercialise their own product for these end-users who aren't DIYers.
I have looked at the Uptone Regen, and rival products they are a different kettle of fish to the jitterbug. I have not looked closely at the Uptone Regen layout (I will if they want
for free).
I have quite clearly given my view on the jitterbug and the layout and my concerns regarding the layout. I would not entertain the product because the layout is wrong, if I checked one of our engineers boards and they had done that the layout would have had to have been re-done.
This is my area of expertise, laying out PCBs for low noise, best signal integrity and minimum EMC problems, mainly for life/mission critical products where noise margins for EMC are far more stringent than you will ever see in commercial designs. More importantly I know what I lay out works because they tend to go through some of the most stringent testing you can imagine.
for free).I have quite clearly given my view on the jitterbug and the layout and my concerns regarding the layout. I would not entertain the product because the layout is wrong, if I checked one of our engineers boards and they had done that the layout would have had to have been re-done.
This is my area of expertise, laying out PCBs for low noise, best signal integrity and minimum EMC problems, mainly for life/mission critical products where noise margins for EMC are far more stringent than you will ever see in commercial designs. More importantly I know what I lay out works because they tend to go through some of the most stringent testing you can imagine.
I wont its a sub-optimal design... a design that is suppose to reduce noise and EMC problems yet does not cater for this in the layout...
if you are going to play with USB yourself I would recommend reading some of the links I have posted and look at the USB specs , also searches for USB common mode chokes etc. will bring up acceptable solutions. You are playing with a standard bus, follow the recommended guidelines.
Hi Ultrafi,
Stack up details. Components and routes on Top elec, ground plane on W2 internal tracking. 0.035mm copper thickness, High Tg FR4 (Dk approx. 4.2 for frequencies of interest.) Traces are 0.5mm wide with a 0.2mm gap between the diff pairs. Approx differential impedance (+/- 10%) 90R
Attachments
I have found from another gushing thread on another forum regarding this device that John Atkinson has reviewed the Jitterbug in the Sept. issue of Stereophile, measurements with and without the jitterbug are the same.... yet it somehow improves the sound even though it does bugger all that it is claimed to do reduce noise.... I don't subscribe to Stereophile but look forward to reading this revue and looking at some measurements....
The power of suggestion is strong with this device... the other interesting theorem is
I would like to discuss this on the other forums but having looked at the way they treat "unbeliever's in audiophile magic" I don't think I would a. get a good reception and b. be able to have a sensible discussion because I don't believe, especially in a device that shows no measurable change yet magically changes the sound even though digital data is passing through it (the change claimed is down to its noise reduction, which as far as current data shows by measurement it does not do)
The power of suggestion is strong with this device... the other interesting theorem is
I have not yet found a link to this but I am interested as again it reeks of magic.
I would like to discuss this on the other forums but having looked at the way they treat "unbeliever's in audiophile magic" I don't think I would a. get a good reception and b. be able to have a sensible discussion because I don't believe, especially in a device that shows no measurable change yet magically changes the sound even though digital data is passing through it (the change claimed is down to its noise reduction, which as far as current data shows by measurement it does not do)
Very different and they don't have the same goal: the Regen is intended for the highest Signal Integrity and lowest noise nearest the DAC (uses frequency optimised PDN), the JB filters at the source computer.
That's very good, because EMC is a complex subject, but in practice, it is nearly impossible to validate the 3 rules of EMC for all devices in a typical audiophile system.
But knowing about it sure provides a good leg up.
I've been reading Ott, Morrison, Armstrong, Giddings, Whitlock, Muncy, Brown and trying to apply some of the material.
Well, after all this is diyaudio, build your own, but check that thing about jitter with the ADUMs.
Already have been doing that.
b. be able to have a sensible discussion because I don't believe, especially in a device that shows no measurable change yet magically changes the sound even though digital data is passing through it (the change claimed is down to its noise reduction, which as far as current data shows by measurement it does not do)
Just a matter of knowing what to measure, where and how.
You are joking, I know this is the subjective get out of jail BS std. reply for measurements, but we are talking noise of a USB port.....
It amazes me how people will defend some sub-standard piece of kit, that in my world of electronics would have been rejected because it is not done properly.....
Jitter and packet data
Well USB adaptive audio generates master clock from the incoming USB stream so perhaps the incomming jitter may influence the clock quality. But I would assume the clock reconstruction is done reasonably not to take into account individual incoming level changes but slowly adjusting independent clock to keep some internal buffer optimally filled (eliminating thus the incoming jitter completely). But I may be wrong, there are probably many methods used.
Wot bugs me though Phofman is they have thrown together a device to sell to the gullible and haven't even considered proper EMC layout, that is a disgrace and does say a lot for the esoteric high end add-on side of audio...
Yeah, the stuff I and many others work on has to work without failure, in fact failure is not an option for us, but also there has to be pride in your work, and in the case of the jitter bug its some basic EMC filtering that is usually on the USB motherboard, that is being glammed up to be something more than it is... and the usual use two so they sell more when the second one is purely placebo and sticking ******** would have as much effect if not more. The rest is censored, but a crisp UK £1 note to who-ever guesses correctly
I would love to comment on the other threads, but they are a perfect example of the Emperors new clothes..... so all it would cause is the usual heretic against the true believers argument, there is too much of this in audio and it distracts from real problems and gives the hobby... a bad names.
Yeah, the stuff I and many others work on has to work without failure, in fact failure is not an option for us, but also there has to be pride in your work, and in the case of the jitter bug its some basic EMC filtering that is usually on the USB motherboard, that is being glammed up to be something more than it is... and the usual use two so they sell more when the second one is purely placebo and sticking ******** would have as much effect if not more. The rest is censored, but a crisp UK £1 note to who-ever guesses correctly
I would love to comment on the other threads, but they are a perfect example of the Emperors new clothes..... so all it would cause is the usual heretic against the true believers argument, there is too much of this in audio and it distracts from real problems and gives the hobby... a bad names.
Last edited:
It took me a while to learn not to visit those forums (computeraudiophile, audioasylum). I quit visiting hydrogenaudio too when I hit a few militant guys who plain refused to learn from source code (linux audio drivers showing detailed function of the sound devices, to be specific) and were offensive instead. It was all waste of time.
I have been trying to delve into this, but it is hard to form a true picture as generally most don't see it as a problem, but in audiophile circles even 1 femto second jitter is disastrous. I have been looking at some pro stuff and have talked to some in the pro world of design, they don't see it as a major problem, as long as a correctly engineered USB cable is used and things are set up correctly... Jitter is an emotive subject, but from most serious and sensible measurements I have seen for USB and SPDIF it does seem to be a problem that some audiophiles over exaggerate a fear fed by many who want to either make money or be an "audiophile guru" by providing the next great cure to a problem that isn't as bad as made out.
And now we have the USB packet noise problem.... that has always been there like many other switching noise problems. I have mentioned SSN (simultaneous switching noise) many times as a major cause of digital noise (not just SMPS) and part of the SSN spectrum will be packet noise from various interfaces used such as USB. On a previous project we studied the effect of SSN and how to minimise etc. for quite a while as some parts of the SSN spectrum where causing us to fail EMC tests (MIL EMC not commercial). We broke down the noise into its component parts as much as possible (turning of circuitry on a special test motherboard) to see what sub circuits contribute the most to this wide band noise spectrum. There is no way to completely get rid of SSN, but by careful consideration of layout (loop control), how the power delivery system is configured, shielding and isolation you can reduce it down to an acceptable level (for mil/aerospace), we achieve immunity up to 18GHz. The experience form this we now use on many designs
where noise immunity is critical (i.e. every design in my case), but I don't see this level of detail in many digital audio based products, just a lot of wonderful marketing and seductive prose......
An while I'm on one... the buy one and try it doesn't hold water, if I designed some sub standard boards for a medical monitoring device for when your under anaesthetic would you be willing to try it out, money back guarantee, or would you rather I take care and pride in what I design, the EE test it to the limits and we publish specs and measurements so you can have some confidence in the device.......
And almost finaly, on one of them sites they were on about digital cables with some extra capacitance or less (I cant remember) but the effect was to sharpen the digital signal rise time thus adding some brightness to the sound (as the leading edge was sharper!)... this when I am having to be teetotal.....
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.