Article: M2Tech hiFace Asynchronous USB To S/PDIF Converter Review

[The Computer Audiophile]

Hi sq225917 - Please start another topic in the forum if you want to discuss this further.

[digger945]

I thought you did a fine(beautiful?) job with the review. No doubt it was a tough one, with all the anticipation from others.<br />

<br />

I purchased a HiFace to compare with my Halide Bridge. I tried it for 2 weeks. Maybe I did not give it enough time, but I was left wanting more. My perspective will not be the same as another listener, who will have different tastes and equipment, as well as exposure to different levels of equipment(or lack thereof in my case).<br />

<br />

If it were my forum, and I was doing the reviewing, I would want to be known as the one to go to for an honest review. <br />

<br />

If I were a manufacturer who just created the latest audio product, I would want to submit it for review by a reviewer who is known for honesty.<br />

<br />

Of course, this is in a perfect world, and I am sure I live just down the street from Pat. :-p

[1audio]

I must agree with Gordon on issue of poorly understood specs and phenomena creating unwarranted firestorms. <br />

<br />

If the output of he HiFace is excessive it will cause problems with input receivers. However different receivers will react differently. in te receiver the time to propagate from the input to the output is dependent on the magnitude of overdrive. This is not spec'd for the Crystal receivers but a scan of general purpose comparators will show the influence of overdrive. Its probably not symmetrical either so alternate edges will pass through at different times. Whether this will have a significant influence on the output jitter or anything else is difficult to quantify. Each dac will respond differently based on a slew of factors including the input isolation (if any) the nature of the termination and even the layout of the traces from the connector to the input pins of the chip. The very high speed comparators degrade rapidly with overdrive but the more general purpose ones get better.<br />

<br />

If there is a reflection from the input circuit (and there will most likely be one) the next question is will it be absorbed by the attenuator (not 100%), will it affect the receiver (comparator) and how? If the cable is long enough the reflection will not get back to the receiver at a time where it can influence the switching. A very short link with a big mismatch COULD influence the transition but may not. Some systems have been designed to use the mismatch to boost the signal going into the receiver. <br />

<br />

I looked for the "kick back" you mentioned on a receiver and could not find evidence of it. One could speculate that the hysteresis could couple through the input transistor pair and kick a minute charge back through the input but I couldn't find it. Perhaps you can provide more details of how to see it?<br />

<br />

My experience with the HiFace was very similar to Chris's. Further no Linux support made it uninteresting for my efforts. But it may provide a good experience for others. <br />

<br />

[StateRadioFan]

I'm pretty sure Lexicon was "In it for the money" when they slapped a nice faceplate on the $500 Oppo BDP-83 and charged $3500.<br />

<br />

All manufacturers are in it for the money. If they weren't they would be producing the gear for personal use only. Just like the guy that loves to home brew beer or the artist that never sells his paintings. <br />

<br />

[jkeny]

I didn't want to be dragged back into this thread but feel I have to based on this post.<br />

<br />

Firstly should your sig not read "Auraliti, vendor of USB DAC" I've been told that is what all commercial bodies should do - like I had tagged to my sig? So now that we have established that you have a commercial interest in selling your USB DAC it will give us a reference point.<br />

<br />

Secondly can you say what your qualifications are? Chris believes that we need to evaluate your technical abilities in order that we can judge your statements or as Chris puts it "Can you share a bit of your technical background so we can put some of your comments into perspective?"<br />

<br />

<i>If the output of he HiFace is excessive it will cause problems with input receivers.</i><br />

<b>Can you specify what you consider excessive?</b><br />

<br />

<i>However different receivers will react differently. in te receiver<br />

the time to propagate from the input to the output is dependent on the magnitude of overdrive. This is not spec'd for the Crystal receivers but a scan of general purpose comparators will show the influence of overdrive. Its probably not symmetrical either so alternate edges will pass through at different times.</i><br />

<b>So your surmising here - the correct formulation of your statement should be " Its probably not symmetrical either so alternate edges <i></b>might</i><b> pass through at different times." </b><br />

<br />

<i>Whether this will have a significant influence on the output<br />

jitter or anything else is difficult to quantify. Each dac will respond differently based on a slew of factors including the input isolation (if any)the nature of the termination and even the layout of the traces from the connector to the input pins of the chip. The very high speed comparators degrade rapidly with overdrive but the more general purpose ones get better.<br />

<br />

<i>If there is a reflection from the input circuit (and there will most likely be one) the next question is will it be absorbed by the attenuator (not 100%), will it affect the receiver (comparator) and how?</i><br />

<b>The attenuator will absorb it somewhat close to twice the dB speced so if you are using a 10dB the reflections will be attenuated by 20dB (a power ratio of 100). This is significantly better than 0dB.</b><br />

<br />

<i>If the cable is long enough the reflection will not get back to the receiver at a time where it can influence the switching.</i><br />

<b>So what length do I need? Oh, yes this has to be calculated based on the various parameters. I can't do this myself so let's go to a cable vendor who knows how to do this. OK, Mr cable vendor "What price did you say again?" "Oh, OK, it's so expensive because I need 16 feet of cable, I see" "It has to be 75ohm without impedance variations, I see, so that's why it's so expensive per metre". "Jeez, it's beginning to cost more for the cable than for the devices connected to it!" "Can you tell me if I need a different length of cable for different data rates" "Why? Because I was told that if I have cables of certain lengths it will avoid the issue of reflections which I was told are on all digital lines" "Ok so I'll get a cable length of 20 feet, that way I can put my DAC where I want to" "What, I can't use 20 feet because now the reflections will hit the DAC at the next transition window & cause jitter?" "So it's only certain discrete cable lengths that work?" "My head has melted, I think I'll just buy an attenuator, I heard they work great & I'll save a bundle :)" "What's that you say, you want me to prove that they work?" "I think I'll risk it" </B> <br />

<br />

<i>A very short link with a big mismatch COULD influence the transition but may not.</i><br />

<b>"Huh, what's that you say? Hold on, a minute ago you told me I needed 16 feet of cable, so why are you trying to sell me a long cable, if the reflections are of no consequence, Mr cable vendor?" "My brain is truly melted now - I'm off before you do it any more damage"</b><br />

<br />

<i>Some systems have been designed to use the mismatch to boost the signal going into the receiver.</i><br />

<b>"Huh! what are you telling me this for? I'm already confused by the cable scenarios" <br />

<br />

<i>I looked for the "kick back" you mentioned on a receiver and could not find evidence of it. One could speculate that the hysteresis could couple through the input transistor pair and kick a minute charge back through the input but I couldn't find it. Perhaps you can provide more details of how to see it?</i><br />

<b>Don't believe me, Ask Jocko he knows all about it - he posts here as Pat from ART - look here http://www.diyhifi.org/forums/viewtopic.php?p=5879#p5879<br />

"In case you have not heard me before..........Schmitt trigger inputs are no bueno. Hysteresis couples energy back to the input. <br />

Yes, it screws up the timing by adding some anomalies that could occur at the decision point. " <br />

or here http://www.diyhifi.org/forums/viewtopic.php?p=5727#p5727<br />

"This is due to the presence of a schmitt trigger on its input, which according to Jocko Homo and others will tend to inject trash back into the transmission line."</b><br />

<br />

<i>My experience with the HiFace was very similar to Chris's. Further no Linux support made it uninteresting for my efforts. But it may provide a good experience for others.</i><br />

Great thanks for your opinion! <br />

<br />

[The Computer Audiophile]

Wow John, as my law school professors used to say, you shouldn't ask a question to which you don't have the answer. Many people in this hobby know Damien as one of the brightest and most talented engineers/designers around. He's been at this at an incredibly high level longer than most people. I won't defend him or list his accomplishments as I'm sure he can do this if he wants. <br />

<br />

That said, your combative tone and unwillingness to let things go or stop beating a dead horse have turned many a thread into a pissing match. Please take a break from Computer Audiophile in an effort to "sleep on it" and let things subside for a few weeks.

[1audio]

I guess you need some more history on me. I do work for several companies including Monster Cable and for a few more accepted high end guys, NuForce (who do make USB DAC's) where I designed a preamp, Auraliti (who does NOT make a USB DAC) where I designed a Linux PC for playing high definition audio and a few other manufacturers for different projects. I am well equipped to both discuss the analog side of transmission and to measure it. However, I'm not here to sell stuff, that's a responsibility I aggressively avoid. I’m trying to shine a little light on some technical issues. <br />

<br />

If the transition takes 1.5 nS (rise time of the HiFace) then a reflection 2+ nS later will not affect the transition at a time when the comparator is near its threshold. A rule of thumb is 1 nS per foot of transmission line (air lines are quicker but totally impractical and many cables will be slower). However the connector and adapters are going to be the worst offenders and they are very close to the receiver chip.<br />

<br />

Here are some pictures of the input Z on a few of the available DAC's I have to illustrate this and what happens with a fast transition on the input. The source is a Tek s-52 pulse generator (25 pS rise time) and the monitor is a Tek s-3a probe (350 pS rise time). The signal comes through a GR 50-75 Ohm matching network and a GR 75 Ohm air line so the source is as close to a perfect 75 Ohms source as possible. <br />

Berkeley Audio:<br />

http://www.4shared.com/photo/TI9YqwKE/BADA.html<br />

Apogee MiniDAC:<br />

http://www.4shared.com/photo/NMxnM8ke/Apogee.html<br />

Anagram demo board:<br />

http://www.4shared.com/photo/HC9V4LJy/Anagram.html<br />

This is the output of the HiFace using a slower preamp than the above system. (Syncing a sampling scope on an SPDIF signal is a real challenge).<br />

HiFace output unterminated (5 nS/division):<br />

http://www.4shared.com/photo/TorUWoJM/HiFace_unterminated.html<br />

HiFace output terminated in 75 Ohms (5 nS/division):<br />

http://www.4shared.com/photo/RtRknr6n/HiFace_terminated.html<br />

<br />

<br />

The level of hysteresis on the Crystal data sheet is specified as 50 mV so there is evidence of its presence in the data sheet but I can't see it on the input of the DAC. Saying it must be there is not enough. If you don't have the actual circuit of the chip it is all speculation. It is quite possible to design a receiver where the hysteresis does not interact with the input. I waded through some of the thread you linked me to but it wasn't focused or illuminating to me. There were interesting but misguided ideas for low phase noise line receivers and lots of talk about compensation networks for transformers. I looked very carefully at the input pins on a crystal receiver fed with a fast edge and say nothing coming back from the input.<br />

Here is the image with the two conditions overlayed. There is a little difference but mostly due to the difference capacitance of the circuit with power off.<br />

http://www.4shared.com/photo/1kde2Ri4/Crystal_receiver.html<br />

<br />

However while the attenuator may not affect the line receivers ability to decode the audio signal perfectly it may well reduce the RFI rattling around in the DAC affecting something else. A lot of these discussions get lost in details like this without considering the other possible influences on the end sound. For example if we say adding an attenuator leads to better sound as evidenced by several listeners its legitimate to say it is doing something useful. However the extension that it sounds better because its reducing reflections or the reduced overdrive of the input receiver is not something that can be drawn from the first evidence. They are just possible causes. <br />

<br />

If the real underlying story of the technology is too hard for customers and other interested parties to comprehend then the task is education not avoidance.<br />

[chuck]

Hi Folks, Can someone please give me some help! I have been sitting in the back ground studying and trying to learn how to upgrade my audio computer. So I went out and bought M2 high face used for 100.00 for my laptop. After downloading new drivers I plugged it in And it hick- upps every 15 to 30 seconds. I tried it on four different computers one with xp 2 with vista 32 bit and 1 win 7 32bit. I used Foobar, Winamp, J Rivers and I Tunes.<br />

Please Help!!! Thanks, Chuck<br />

[blindjim]

<br />

Hi Chuck<br />

<br />

Knowing a bit more about your setups … drivers… plug ins and output formats might help. <br />

<br />

Assuming all is right on there, and you have the exact same or near same thing happening regardless the pc you use… well the common denominator then is the Hiface itself. …. Or the DAC the Hiface is feeding. Or the connections downstream from the Hiface…. Or, the pins on the Hiface USB side are out of wack. … ??<br />

<br />

On my laptop recently, following a reload of the vista OS, I was getting some intermittent drop outs. I found the issue was in my wireless setup in the Vista OS. I connected an Ethernet cable and all was well immediately. Consequently I reconfigured my wireless NIC.<br />

<br />

I also had an issue with using the Hiface and it’s latest driver with JRiver MC…. And WASAPI. No other app gave me problems, Fubar, Winamp, or itunes. Just JRiver MC 14.<br />

<br />

I rolled back the driver to a previous iteration with a few mouse clicks and a reinstall of the Hiface and all now works like a charm. Beforehand JRiver only worked in the Direct Sound mode… not with WASAPI.<br />

<br />

So having the right driver is important with the Hiface unit… BUT EXCHANGING ONE DRIVER FOR SOME OTHER ONE IS NOT a big deal, as some here would lead one to think.<br />

<br />

I’ve found out as well, configurations in Windows can add junk to the mix and streamlining the OS for or during playback sessions is a good thing. Eliminating background services, shutting down Antivirus apps, halting uneeded software from operating, etc.<br />

<br />

Setting Vista & Win 7 audio setup to exclusive mode operation. Helps too. <br />

<br />

If connectivity via wireless, or some downstream config issue isn’t it, and the software is setup well, you’ve enough pc resources to function adequately with, it must be what is in common… the converter, and/or DAC.<br />

<br />

If you have a receiver around try feeding it instead of the DAC from the Hiface and see what happens.<br />

<br />

Given your input so far, there’s myriad reasons on hand for your problem. It could be the fault too of a combination of things… please be more explicit, as to your particular configs, drivers, players output formats, settings, etc. <br />

<br />

I’ve used mine and one other on a few machines myself…. With no likewise issues. Either it worked and both mostly did, or as said, with JRiver MC 14 & WASAPI ONLY, an earlier driver was required to fix it. Maybe only a reinstall of the unit is needed.<br />

[chuck]

Thanks for the reply blindjim,<br />

I didn’t no I had to set up drivers… pluggins and output formats. I down loaded J Rivers mc,foobar,winamp and installed them with auto setup. My laptop that I want to use is a Dell studio 2 core T 9300 with 4 gigs mem vista 32. I went to High face web site and down loaded new drivers and installed them. My D/A is Musical Fidelity A3-24 I also go through a Genesis digital lens.<br />

I tried system without digital lens it made no difference, I don’t have another d/a to try.<br />

<br />

[1audio]

A few basic troubleshooting steps. First, can you play those files flawlessly using the native internal sound system? If you are getting glitches at this level it is an issue of optimizing your computer. You may need to remove or suspend some apps that are running ,in the background. Further things like mouse clicks will pop through the audio interface. That can be controlled but takes a bit of effort to figure out. Vista's sound system is quite different from XP but there should be info available on cleaning it for this use.<br />

<br />

You will get better results with one of the programs optimized for the high quality audio (Foobar or J River). They should also work as well with the internal sound system. Again I would verify they are working right.<br />

<br />

If the main audio is working right then add the HiFace. Stick to 44.1 content and make sure that works first. Other possible causes may be shared USB controllers in the PC, which means other activity could blast the HiFace off line or out of sync intermittently. A second possibility is that either the HiFace or the DAC are too far out of tolerance on their clock range and the DAC is having a lock problem. This seems unlikely but I have seen it happen with some hardware.<br />

<br />

[chuck]

I would like to thank Blind Jim for the several hours he spent on the phone with me Monday night. We did everything imaginable to my laptop and the HiFace still would not perform is it should, well I realized that my surround sound system in the house had coaxial inputs. Well much to my surprise the HiFace works fine and the problem was with my Musical Fidelity A3 24 DAC! It’s not compatible with HiFace M2TECH, so I either buy a new USB adapter or a new DAC. Thanks again Blind Jim

[The Computer Audiophile]

Great to hear Chuck! That's what this hobby is all about! <br />

<br />

Thanks for helping Jim!

[barrows]

I might know of a fix for you. The hiface is known to have an unusually high voltage SPDIF output. Most DACs can handle it, but it appears your cannot. There is a product mentioned elsewhere in this thread, an RF attenuator, that can be plugged in series with the SPDIF cable which will lower the voltage without creating any problems (these are impedance matched). I actually picked one up to try just to see if it would improive sound (they can, in some cases, reduce jitter caused by signal reflections). These are cheap at around $12.00 and worth a try. The company is Mini Circuits, I do not have their url, but google should work. I got a 6 dB, but you shoudl go for 10 dB with the Hiface. Make sure you get the ones for 75 ohm impedance, and rated to 2 GHz. <br />

Just found the url: minicircuits.com and the product is: RF Attenuators-I am 95% sure one of these will get the Hiface working with your DAC.

[1audio]

I'll be interested to hear if reducing the SPDIF level fixes the problem.<br />

<br />

This sounds more like a clock frequency problem than an input drive problem from the symptoms. Input overdrive would cause a continuous error rather than a periodic error. The fact that it works fine with the surround system suggests that the output is free of intermittent dropouts. <br />

<br />

Some DACs (ones that use voltage controlled crystal oscillators) have a limited lock range and may not be able to lock to another source if the two are too far apart. Another issue may be that it just isn't locking but they are close enough to work. I have had that experience with the Pacific Microsonics (its easy to forget to switch the clock source) and it works with an occasional click.<br />

<br />

Or its possible that the receiver is hiding the stream error in the spdif stream that causes the click. Figuring this stuff out is challenging to say the least. Good luck. <br />

<br />

[ar-t]

There is one more possibility, which is close to what Demian describes.<br />

<br />

The TI chip, used in that transport, always sends data, in the Channel Status block, that claims the TX rate is 48 kHz. Even when it isn't. We did experience this problem, not too long ago. A certain DAC had trouble deciding whether to believe its PLL stage, which provided a 44 kHz rate clock, whereas the status bits claimed it was 48 kHz. The secondary PLL, in the DAC, had trouble picking the right one. As a result, it just refused to work. Once we found a way to trick the TI chip into transmitting 44 kHz data, all was fine.<br />

<br />

(Our trick will not work with this unit, as I am sure most everyone knows ours only does 44 kHz. Turns out it is easy to trick it into sending 44 kHz data.)

[chuck]

Hey Barrows, <br />

Is this what you are talking about?<br />

[barrows]

That is it. After re-reading your post, and Demian and Pat's responses, I tend to believe that Demian may be right in that the attenuator may not help-your issues could be caused by another problem. It is a little hard for me to believe that the timing output of the Hiface is so far off that your DAC cannot stay locked on though... most measurements of the Hiface I am aware of just are not really that bad? Maybe Demian could attempt to make what he is saying a little clearer in regards to the measured performance of the Hiface's clocking.<br />

In any case, the attenuator is not too much of a financial risk to give it a try, as it will pad down the overly high output voltage of the Hiface (some have reported the attenuators can lower jitter...this will be dependent on the entire system though, and how the Hiface, SPDIF cable, and SPDIF receiver all interact).

[The Computer Audiophile]

Hi Barrows - I believe Demian is talking about the same thing mentioned in the measurements below my review. Check out the Status Bit info above and this screenshot that show 48k when it should be 44.1k<br />

<br />

<br />

http://images.computeraudiophile.com/graphics/2010/0822/HiFaceStatusBits.jpg

[ar-t]

Yep, that is what it does! Don't ask what we went through, trying to extract that info, without that nifty gizmo the Prism/Dscope people make.

[barrows]

Doesn't it just irk ya that they cannot get this stuff right? And Pat, you say this is a fault of the TI SPDIF chip. Why not just build it to the spec...really. I mean, the folks at TI are not incompetant.<br />

<br />

All these problems people are reporting with the Hiface and the Evo really are turning me off to these devices-although I give the Hiface a bit of a pass 'cause of the value it offers (if it works with one's DAC).

[MR]

This is not a comment directed at any one in particular, just a general observation.<br />

<br />

It's not reasonable to expect every USB interface to work out of the box with every DAC out there (and there are a lot out there). I hold on to my Theta Generation Va which was hot stuff in 1995. Because it's Red Book, it still makes great music and benefits in a stunning way when used with an asynch USB->SPDIF converter.<br />

<br />

However, if I were to assume that every USB interface out there will work or sound good with it, I'd be a moron. A lot has happened in electronic interfacing since 1993. A lot. There have been a lot of changes in standards, copy protection, etc, etc.<br />

<br />

That suggests there are going to be differences in what DACs expect -- most of them were engineered to go with the manufacturer's transports. Back before computer audio, mixing a transport from one mfgr with a DAC from another was asking for trouble. Makes me wonder if the all-in-one players didn't sound the best after all.<br />

<br />

The bottom line is the more that technology changes, the more the need to use our ears remains. Nothing is a silver bullet and almost anything can be made to sound lousy under the right conditions.<br />

<br />

It took me years of experience before I could trust that what I hear from my system is 'good enough.' It happened when I began to trust my own sense of enjoyment. If I enjoy the music, then I think my system sounds great.<br />

<br />

Despite the goals of the "Absolute Sound," almost anything can mess up the best sound out there. Just set it all up in the garage and have a listen.<br />

<br />

[barrows]

players when properly designed did/do sound best (because they do not have to rely on SPDIF, and good ones slave the transport to the DAC clock internally). That's why high end companies just about stopped making DACs for awhile. The resurgance in popularity of stand alone DACs owes alot to the popularity of using computers as a source.<br />

There is a standard for SPDIF, it includes the parameters required for one device to communicate with another. It is not difficult to engineer products that work within the standard. I really do not find it acceptable that there are some products that are so far outside the parameters of the standard that problems occur. Occasional problems, with very unusual components I can handle, but the volume of reported problems with both the Hiface and the Evo seems a little high to me. Like I said, I'll give the Hiface a bit of a pass on this, due to its very modest cost, but the amount of reports of problems with Evo as well really bothers me. It is hard not to come to the conclusion that these products need more refinement, and perhaps were rushed into the market with insufficient testing.

[ar-t]

You don't think TI is incompetent? Well, maybe they are not. But, here are the facts:<br />

<br />

The Hiface unit has an output level, that is too high. Why? Because the designer (obviously <i>not</i> an RF guy) copied the circuit, in the data sheet. Companies frequently give the "menial" jobs, of coming up with "typical application circuits" to junior engineers. Who, are only junior engineers, for a reason. (Inexperience, and time on the job, being key factors. Some will get it together, but most will find some way to hang on, just out of sheer luck.) As a result, we have generated a cottage industry, of learning how attenuators can reduce reflections. (This is old hat, to RF guys. Older even than we are.)<br />

<br />

Second, the "It's not a bug, it is a feature!" is mentioned <b>nowhere</b> in the data sheet. Indeed, calling TI tech support, almost none of them are aware of it. I ended up teaching quite a few of them, about this problem. After I found the right person, who knew all about it.<br />

<br />

"If you want the data in the Channel Status Block to be correct, you will have to serially enter it in. There is no way we can have the chip transmit the right data, for all sampling rates."<br />

<br />

"Well, if I had read that in the data sheet, I would have known about it. And if I wanted to go to all the trouble to use a chip that required a micro interface, I would have gone with the Wolfson, instead."<br />

<br />

So, you decide: incompetent, lazy, or too unconcerned to mention it, up front. The only good thing to come out of all of that is their marketing weenies (in Atlanta) have finally gotten the message, that we don't want to have them calling us all the time. (Or ever, from now on.) They didn't get the hint, when we (and lots of their other customers) pointed out that telling us about what great products they have is a waste of time, if they have lead-times >18 weeks, on damn near every part in the catalog.<br />

<br />

Moral of the story: I have lived and worked in the Dallas area, for almost 40 years. I should have known better than to go with a TI chip, but on the surface, it looked like the fastest way to get a product to market.<br />

<br />

Answer your question?

[barrows]

It sounds like they suffer from the problem of being a big company, where the details of how things work often get lost in all the BS.<br />

When I worked for PS Audio I spent about a half hour chatting with a TI senior design engineer (during CES) who was all hot about their new receiver/ASRC chip. <br />

Later I tried contacting him as I had some questions about what happens to the data, exactly, when it gets received by one of these chips (I also asked about the 4192 ASRC only chip) but you tell the chip to output the same data rate as is received. It appeared that no one could get me the answer-was the data re-sampled, did it just pass through with "no" processing... I never did get the answer.