USB Regen from Uptone Audio

[Clive]

I've been trying out my Regen over the last couple of days. I have one from the latest August batch so it is an"amber". I run an untuned W10 laptop with Bug Head (Infinity Blade IQ) feeding a Metrum Octave mkII DAC with has an M2Tech USB card internally. .

Thinking of buying a Regen for 2012 M2Tech Young, Uptone doesn't know if it will work with M2Tech USB bulk data transfer technolgy used, is this tech used in the Metrum
Mike

Hi Mike, Regen should be fully transparent, I've not heard of any issues. All I know about my M2Tech card is that it's the original version and the 2nd gen version for some reason wasn't deemed to be as good. One more thing...I run the Young driver but if I swap between Regen or no Regen I have to reinstall the driver when going to running without the Regen.

[jkeny]

I've been following John Westlake's (johnW) measurements of the Regen on PinkFishMedia
His measurements & analysis, so far, points to low frequency crud riding on the incoming USB signal
As the USB signal is then regenerated by the Regen, most of this LF crud is eliminated
As the M2tech uses bulk USB communication, rather than asynch USB communication, the LF crud may well be of a different nature & less intrusive? Only measurements can truly show this - but I guess polling owners of devices that use this 1st gen M2tech USB board will help to sort this out.

[Mike19]

@Clive and @JKeny - Many thanks for the advice, I placed my order tonight and will record my views on sound quality with Young DAC. For those interested my Order No. was 2617.
Mike

[DaveF]

I've been following John Westlake's (johnW) measurements of the Regen on PinkFishMedia
His measurements & analysis, so far, points to low frequency crud riding on the incoming USB signal
As the USB signal is then regenerated by the Regen, most of this LF crud is eliminated
As the M2tech uses bulk USB communication, rather than asynch USB communication, the LF crud may well be of a different nature & less intrusive? Only measurements can truly show this - but I guess polling owners of devices that use this 1st gen M2tech USB board will help to sort this out.

All very interesting from a technical point of view John. I dont browse PFM forums at all or the Wam for that matter so I've missed out on any discussions there.
Whats the nature of this LF crud though? At what frequencies and what about SNR figures? I mean a before and after set of measurements. Sounds like some sort of filtering going on??
Does it sharpen up the edges of the signals though as what my idea of the regen unit is. I did work on a regen unit for HDMI a good few years ago. Different beast than for audio though.

[jkeny]

I've been following John Westlake's (johnW) measurements of the Regen on PinkFishMedia
His measurements & analysis, so far, points to low frequency crud riding on the incoming USB signal
As the USB signal is then regenerated by the Regen, most of this LF crud is eliminated
As the M2tech uses bulk USB communication, rather than asynch USB communication, the LF crud may well be of a different nature & less intrusive? Only measurements can truly show this - but I guess polling owners of devices that use this 1st gen M2tech USB board will help to sort this out.

All very interesting from a technical point of view John. I dont browse PFM forums at all or the Wam for that matter so I've missed out on any discussions there.

Clive gave me a heads up that JohnW was measuring the Regen with the intent of doing his own USB "Detox" device, as he calls it - that's why I went there

Whats the nature of this LF crud though? At what frequencies and what about SNR figures? I mean a before and after set of measurements. Sounds like some sort of filtering going on??
Does it sharpen up the edges of the signals though as what my idea of the regen unit is. I did work on a regen unit for HDMI a good few years ago. Different beast than for audio though.

His initial observations were that this LF crud he was seeing on the eye pattern plot was fluctuating in correlation with activity on the PC but he couldn't show this (taking a video of the oscope was suggested). So he used a digital scope instead which colours the statistical density of the signal (colour grading) - you know what I mean - traces that overlay the same path 99% of the time result in a white or yellow colouration, less visited paths show in red, green & finally light blue & dark blue as is seen on this plot which is the USB signal on the input to the Regen (i.e what would normally go directly to a USB audio device)


The Regen output looks like this:


I don't know, Dave, what the details are related to these LF events which he calls "runt events" :)

[DaveF]

That top graph looks pretty bad. How could it possibly work without occasional bit errors? The blue levels are very smeared all over the place. I wonder over how many seconds(uS/nS) was that plot taken over though: Just after playback was enabled or some random time afterwards like a few seconds or mins during playback?

On the second plot there is a curious 'knee' on the rising edge. Phase noise looks a bit better on this one though.

Any spectral analysis done? I just wondering where these LF tones are and how dominant are they?

[tony]

Amir on Whatsbestforum has pronounced the regen of no benefit after his testing and graphs etc.
John has posted a bit on the thread. The jury seems to be out on this.

[DaveF]

Amir on Whatsbestforum has pronounced the regen of no benefit after his testing and graphs etc.

That's another forum I'm missing out on too! :-)

[jkeny]

That top graph looks pretty bad. How could it possibly work without occasional bit errors? The blue levels are very smeared all over the place. I wonder over how many seconds(uS/nS) was that plot taken over though: Just after playback was enabled or some random time afterwards like a few seconds or mins during playback?

I don't know the answers to your questions - I can just quote these words of JohnW - I can't actually see any grid on the plots so the 200ps per division doesn't really help much

The scope timebase was 200pS a Div, so very fast any LF modulation causes the location of the eye pattern to shift in time - so the location of the crossing points fell into other areas which results these hits being colour graded - Lower Frequency events in Blue, more frequent events in light blue etc.

The Green graded hits indicate "medium frequency events" - so these "odd states" happen frequently - Observing the Waveforms on a very fast Analogue scope "WITH IMAGE INTENSIFIER" to allow LF events to be visible on the CRT at such fast sweep speeds (500pS Div) you can see the "runt" events directly related to the processing tasks of the PC - the frequency of their occurrence is visibly modulated by the operating system workload / operation.

More will be found here http://www.pinkfishmedia.net/forum/show ... tcount=687

On the second plot there is a curious 'knee' on the rising edge. Phase noise looks a bit better on this one though.

Any spectral analysis done? I just wondering where these LF tones are and how dominant are they?

We are awaiting his spectrum analysis on the DAC's output but he states this:

I suspect I'll not manage to measure any difference on the analogue output - I cannot measure any difference between the Optimal transient and Optimal transient XD filters, but I can pick them out 100% accurately in blind test.

[jkeny]

Amir on Whatsbestforum has pronounced the regen of no benefit after his testing and graphs etc.

That's another forum I'm missing out on too! :-)

That's another forum you're not missing much on - it used to have a good balance between technical & subjective info but has become polarised by Amir's postings & the trolling that is now rampant.
Here's his Regen test thread if you want to read it http://www.whatsbestforum.com/showthrea ... post341173

Amir has gone directly to the DAC's (Meridien Explorer) output to measure the effect of the Regen (forget about the Jitterbug - it's of less interest) on it using a "jitter" signal & found nothing except what he interpreted as extra jitter spurs.
I contend that these are actually the 8Khz (& harmonics) noise spikes that result from the 125uS microframes that are part of the USB 2.0 high speed protocol.

Other than these extra spikes there doesn't appear to be much anything of note except the Regen replaces the USB 5V VBus with a cleaner 5V PS supply to the Explorer DAC. The Explorer uses Vbus to power itself so this cleaner 5V supply lowers its measured noise floor.

Funnily, Amir hears no audible improvement when using the Regen despite this reduction in noise floor - I'm sure he says that it is below the noise floor but I would like to find out what the real noise floor on his FFTs are. In fact, I've just asked him this but he will ignore me, I reckon.

I see my mistake - the jitter FFT is actually using FFT settings which bring the grass below the noise floor of the Regen & therefore no difference is seen between Jitterbug & Regen noisefloors. I'm sure you got this already, Dave?