Current recommendation for instrumentation grade audio interface

[Raqs]

Hi all,
I'm currently reconnecting with my old pursuit of a good audio interface to be used for FFT duty. I am currently relying on ARTA for THD measurements, and I used to have an EMU0404 which, as stellar as it was, I stopped being able to use it once I (accidentally) wiped my Win XP boot partition from shop computer. Win 10 w/ EMU0404 has been a no go for me as much as I've tried. Nor am I really interested in going back o trying.

So I switched to a Lynx on PCI, which is fine, but frustrating on two accounts:

• Noise is maddeningly dependent on exact position of the connecting cables' dongle. I thought an integrated card would make it easier with bulk and cables' noise, but it's quite the opposite.
• I can't get it to reliably perform an FFT with "external" generator option on. It needs forever before locking on the fundamental and performing a reliable, stable reading. A good part of my work is with FM tuners, and being able to perform alignment adjustments while looking for a THD dip is critical. I can get the Lynx + ARTA to do quick, reliable FFT if I claim the gen is "sine" (so internal to the interface), but I am really using this with my AP S1 over its monitor outputs. Therefore using "sine" although it's the AP generating the signal seems like trouble in a couple of ways: frequency (ARTA doesn't read the exact Hz of the fundamental and hence doesn't exactly calculate over that measurement) and phase (the generator and analyzer, at least as far as ARTA is concerned, and not phase-locked, which I feel doesn't guarantee realistic THD readings).

I guess I'm looking either for recommendations/input on ARTA tweaks (if anyone can see an error I may be doing, or a relevant setting, or maybe some hardware controls to be looked at), or on very clean USB interfaces. I've been spoiled with the EMU, and anything that may match it - if it exists - is an arm and a leg. But maybe there's a deal or surprisingly good interface I am not aware of as I haven't been exploring the topic in a pretty long time.

I've been looking at the Clarrett+ 2Pre and SSL 2+?....
Thank you in advance for all input and recommendations.

 

[Blumlein 88]

RME ADI-2 Pro FS ADC/DAC is the answer to your questions.

Now if you just want the ADC end of things, an E1DA Cosmos ADC is about as good as you can get.

EARLY LOOK: E1DA Cosmos ADC - affordable high performance measurements for the audio hobbyist!

A blog for audiophiles about more objective topics. Measurements of audio gear. Reasonable, realistic, no snakeoil assessment of sound, and equipment.

archimago.blogspot.com

E1DA Cosmos ADC

@IVX kindly sent me a sample of the new Cosmos ADC to try out. Here are my impressions of it so far. The Cosmos is housed in a compact metal case with rubber feet, USB Type C connector on one and dual XLR inputs on the other. Power is over USB only. Beside the USB connector is small LED that...

www.audiosciencereview.com

EARLY LOOK: E1DA Cosmos APU - high performance, low-noise pre-amp and 1/10kHz notch filter. [Quick peek - S.M.S.L. DO100 DAC.]

EARLY LOOK: E1DA Cosmos APU

archimago.blogspot.com

 

[AnalogSteph]

So I switched to a Lynx on PCI, which is fine, but frustrating on two accounts:

• Noise is maddeningly dependent on exact position of the connecting cables' dongle. I thought an integrated card would make it easier with bulk and cables' noise, but it's quite the opposite.

This part suggests that you're having to contend with ground loops in your measurement setup. Which does not surprise me given that both the average FM tuner and FM signal generator are going to have unbalanced audio connections.

The easiest way to get rid of those would be using a legit audio analyzer with galvanically isolated I/O, but those tend to be a bit hard on the wallet. Otherwise some creativity is going to be required. For starters, I would get a galvanic isolation thingy for the antenna connection and a line-level isolator like a Behringer HD400 between output and signal generator input.

Given that in the world of FM tuners, distortion of 0.01% (-80 dB) is considered very low, I don't even think you actually need pro-grade hardware for this. In fact, I would probably prefer driving a HD400 from a headphone output - the lower the source impedance, the better a transformer tends to perform, plus this particular model is like 540:600 ohms (presumably so it'll drive a modest length of cable at least) so some decent output driving would be advisable. Besides, consumer FM tuners may have fairly high output impedance and very limited output driving, so an input impedance of ~47 kOhms would be advisable... that rather is some Creative soundcard line-in than an audio interface mic input. A pro-level line input would work if it has plenty of dynamic range (I would suggest around 110 dB up), although you'll probably be wasting the top 20 dB entirely.

 

[Raqs]

Hi Blumlein 88 - thank you for your thoughts.

The RME is out of my budget on this (I did say I'm excluding "an arm and a leg" options). I'm maybe ready to go as high up as $500 or so.
The E1DA Cosmos is something new for me. Given all I'm looking for is capturing audio from the AP monitor out, it sounds like it may fit the bill - though I just very cursorily looked at this product. Is there a good source for it? Hopefully with complete specs? I need to figure how low can a load impedance go on the AP monitor outputs before it becomes a problem.

 

[retro]

@Raqs

Is it a Lynx Two u are using..? Never had the slightest noise with mine.
Are u using XLR out or RCA..?

 

[DVDdoug]

I thought an integrated card would make it easier with bulk and cables' noise, but it's quite the opposite.

Do you mean an internal sound card or data acquisition card? Computers have lots of internal electrical noise, and the power supply usually isn't "audio quality" either so performance can be unpredictable.

To some extent the same is true of USB powered audio interfaces because USB power from a computer also tends to be noisy. A little (or a lot) of noise is no problem for digital but it's a problem if it leaks-into the analog. That's usually only a problem with the microphone inputs since the high-gain amplifies any noise. (Most good USB powered interfaces are fine on most computers.)

So usually your best-bet is an external USB audio interface with its own separate power supply.

 

[retro]

Computers have lots of internal electrical noise, and the power supply usually isn't "audio quality" either so performance can be unpredictable.

Not with the Lynx.
A Soundblaster or the likes, absolutely possible.

 

[Rja4000]

RME ADI-2 Pro FS ADC/DAC is the answer to your questions.

Now if you just want the ADC end of things, an E1DA Cosmos ADC is about as good as you can get.

RME ADI-2 Pro fs R is the current version.

The DAC on this last version will pull an additional 2-3dB at the top (from 116dB for the Pro fs to above 119dB SINAD in mono mode for the Pro fs R).

This remains, by far, the best option, in my opinion: it's a workhorse, very reliable, highly protected against mistakes, it allows measurements that requires ADC and DAC to be perfectly synced, it is VERY powerful in terms of functionalities, and it's also sounding perfectly (of course).

The E1DA ADC is an interesting device. I own one (2, actually), and use it for SOME ultimate (SINAD or noise) measurements, for VERY high end devices to test (typically, only needed for a small number of recent, state of the art performance, DACs).

One of the benefits is that you may also connect and measure directly an amplifier (a dedicated input allows up to 43V - over 400W with 4ohm load). (You need an external load for that, of course.)

It has some drawbacks though, which makes it more difficult (or impossible) to use for some measurements.

My default measurement device remains the RME, which work as well or better for me 99% of the time.
(The main exception being... when I measure the RME DAC itself ! )
Price is in another league though.

E1DA is about to launch another interesting device: the analog APU. That's basically a notch filter with a low noise preamp (2 preset gains).
First measurements seem to demonstrate top grade performance (as becomes usual for E1DA).
That will, for sure, increase usecases and facilitate use of the E1DA ADC for measurements. It should bring benefits for ANY ADC, by the way.

Then, you'll have the recently spotted RME ADI-2/4 fs SE, which may even bring the competition to another level. But we don't know yet.

 

[Raqs]

Thank you all very much.

I think I did a pretty bad job at explaining my setup here. More detail below.

The interface I'm looking for would entirely only provide FFT duty from the monitor outputs of my Audio Precision System One w/ DSP. No direct interfacing with the DUT is to occur as far as this interface goes (this interface entirely sits "behind"the AP in my instrumentation setup). The AP S1 generates the signal and collects it after passing it through the DUT (tuner or not).

Not sure if this means grounds loops are not an issue entirely, but I think maybe at best a remote possibility. Where I see collecting of some noise is in the length of cables (very long due to the Lynx dongle etc, and there's nothing I can really do about that) between the AP outputs (BNC) and Lynx inputs (balanced). In any case, once I have the cables placed in certain ways, the noise floor is acceptable and somewhere around 105-110dB. What's bothers me far more is the inability of the Lynx (its FFT) to lock on an external signal when using ARTA.

Why do I use an interface for FFT when the AP S1 has DSP? Live THD readings. The AP DSP snapshots the FFT, which only has the advantage of reliable calibration over the Lynx/ARTA solution.

With the EMU, measuring tuners below -80dB THD was a breeze (those tuners that can go that low... not very many). The EMU would be easily able to "grab" an external fundamental and shoot out an FFT.

I wonder if this E1DA ADC can replace the Lynx in this duty.

 

[Rja4000]

Now it becomes more clear.
About E1DA ADC:
If the AP monitor output impedance is low enough, and if, added benefit, it has a fixed range (0dBFS remains same V), and if you can use it with balanced connection, than that should be perfect.
Just make sure to use a Y cable to feed both ADC inputs and use it in mono.

 

[Raqs]

Rja4000 - yes, the monitor outputs are 1V RMS nominal (scaling the input signal to this, agnostic of its level) and Rout is 560 ohms. The input impedance of the E1DA is really strange, and not sure it'd agree with the AP monitor outs given the above.

 

[Rja4000]

Rja4000 - yes, the monitor outputs are 1V RMS nominal (scaling the input signal to this, agnostic of its level) and Rout is 560 ohms. The input impedance of the E1DA is really strange, and not sure it'd agree with the AP monitor outs given the above.

That's not good.
Cosmos' input impedance in unbalanced mode is below your device's output impedance on the lowest range (1.7V). (640 ohm -30%)
You need a 0dB gain buffer in between.

 

[Raqs]

I am trying to understand why this device was designed with such hostile input impedance specs and am not coming up with much.

A buffer input stage can pretty easily be designed with pristine specs and very little would be lost - if anything - by adding that at its front end.

While a lot gained by having it there.

Still exploring this little and pretty remarkable device.

 

[Blumlein 88]

As far as $500 audio interfaces, the Clarrett is probably as good as you'll get in that price range. You'll probably be short of the other suggestions by 10-12 db in most parameters. That may still be enough for your purposes.

You could use the Topping L50 preamp as a buffer to the E1DA without loosing much, but it is a kludgy solution. Still both devices combined would fit inside your $500 budget.

Topping L50 Review (Headphone Amp)

This is a review and detailed measurements of the new Topping L50 balanced (input) headphone amplifier. It was sent to me by the company and costs US $229. It will be announced shortly. The industrial design of the L50 is different than older models with a plexiglass front: While I like...

www.audiosciencereview.com

 

[Raqs]

As far as $500 audio interfaces, the Clarrett is probably as good as you'll get in that price range. You'll probably be short of the other suggestions by 10-12 db in most parameters. That may still be enough for your purposes.

You could use the Topping L50 preamp as a buffer to the E1DA without loosing much, but it is a kludgy solution. Still both devices combined would fit inside your $500 budget.

Topping L50 Review (Headphone Amp)

This is a review and detailed measurements of the new Topping L50 balanced (input) headphone amplifier. It was sent to me by the company and costs US $229. It will be announced shortly. The industrial design of the L50 is different than older models with a plexiglass front: While I like...

www.audiosciencereview.com

I am not looking for the ultimate possible performance. There are limitations with the AP S1 itself, which is effectively the floor for my setup. I am trying to make sure, as much as possible, that my solution is both affordable, and hopefully right "below" the AP baseline.

The EMU0404 did spoil me, so I hope to be able to at least match its performance. So many years after that card was created, and almost as many since EMU itself went bust, you'd think it shouldn't be so hard (or expensive).

 

[Blumlein 88]

If you haven't seen this video review of the Clarrett Julian Krause does basic measurements to confirm what the Clarrett's performance envelope is.

Also one on the SSL+

 

[Raqs]

I grabbed a used RME ADI-2 for what i think it's a fair price from a large US retailer.

I am aware it's not at the level of the RMEs mentioned earlier, but by spec it should essentially "sit" just below the AP S1 specs, so it meets my expectations.

I will probably test it here and compare it with the Lynx in actual use (and ultimately keep one of the two), but at a minimum I expect I will be able to run very short BNC to XLR cables from AP Monitor outputs to RME inputs (probably as short as 6"?...). Some of what's going on at my bench - and probably most people's benches - is garbage splashed around by switching PSs, LED lights, etc etc. I probably had such easily clean signal into the EMU because I could keep hookups very short and clear of sources on noise. By itself, the Lynx does an uncanny job to stay as clean as it is while being inside a computer, but the unnecessarily long dongles are unfortunately not such a great idea.

 

[Raqs]

Well, shoot. I don't see a USB port anywhere on the ADI-2. Can this not be used as an interface?....

 

[staticV3]

I am trying to understand why this device was designed with such hostile input impedance specs and am not coming up with much.

The Cosmos ADC was designed for measuring DAC+Amp combo devices like the Tempotec Sonata HD Pro or E1DA 9038S. Those usually have a Zout of 0.1Ω.

Adding an input buffer to the Cosmos would've increased the price, complexity, and current draw while at the same time dragging down the SNR.

In its current form, those who don't need high Zin can enjoy the best SNR possible, while those who need it can slot in an external input buffer. The APU can serve as one btw, though it does not have unity gain. Specs are here.

 

[Rja4000]

The Cosmos ADC was designed for measuring DAC+Amp combo devices like the Tempotec Sonata HD Pro or E1DA 9038S. Those usually have a Zout of 0.1Ω.

Adding an input buffer to the Cosmos would've increased the price, complexity, and current draw while at the same time dragging down the SNR.

In its current form, those who don't need high Zin can enjoy the best SNR possible, while those who need it can slot in an external input buffer. The APU can serve as one btw, though it does not have unity gain. Specs are here.

I think indeed the APU is what you need to highly increase resolution and usability of the E1DA ADC.

And, indeed, measuring headphone amps is a perfect use case for the E1DA Cosmos ADC anyway (and Ivan's load board is genius).