With out pictures, it's not real- though it still may be unreal!

[JonMarsh]

Some more Schiit...

A few more tests, which uncover some interesting things...

First, a 50 Hz FFT at +6 dB output- this shows a mix of harmonics of the test signal (50Hz) and power supply residual noise... (60Hz related)



A test I like which is fairly demanding but not unreasonably so, is the CCIF two tone HF IM test; here we have 18 and 19 kHz, though often 19 and 20kHz are used.



In fairness, components below -90 dB are likely inaudible for the most part, at least not readily identifiable by most listeners. Usually this CCIF IM test is done with a linear scale like this, to make it easier to identify frequency components. It also shows IM sidebands above the test signal range, too.

It can also be plotted on the log scale, which gives more low frequency information, and may reveal power supply issues in an amplifier.



Here the difference tone at 1 kHz is quite evident, but remember, it's -90 dB. The overall circuit does not appear to be a typical op amp though, the residual noise floor is very low, better than I would expect.

In fact, this last graph is just that- input signal turned off and volume turned down on the headphone amplifier.



This is a quiet little amplifier- you can see DACs with a noise floor 30 dB higher than this, for an SACD player, such as the Playback Designs model 210.

[JonMarsh]

One more thing...

OK guys, BTW, I have to confess, this little piece of Schiit is not really as quiet as the graphs I've shown you would lead you to believe- you see, I was using a trick that helps with bringing signals out of the noise background in order to see all the distortion harmonic frequencies- I used 10 averages of the FFT signal to create these plots, which reduces the impact of random noise significantly. Tell you what, for educational purposes, I'll test and plot this again tomorrow with just one FFT sweep for comparison. But as a diagnostic tool, this is quite useful, as anything which is not just thermal noise emerges and is easy to pick out.

Another comment- those power supply hum components are probably the major part of the noise floor limiting the THD+N performance below 2V output. Cleaning that up would make the low level stuff "look nicer", but may not affect the sound much. Also, remember, the human ear does have some capacity for extracting signals in the presence of noise- one reason why vinyl is appreciated as much as it is- though the noise floor may not be all that low, the information isn't truncated the way digital recordings can be.

to borrow from an old expression, watch out- there are/can be liars, damn liars, and test engineers. (Actually, they can't get away with all that much, but you do have to watch out for some things).

In closing for the evening, let me just say that this instrument and the AP500 software are real sweethearts- there are things I wish were implemented differently, but for the most part, this is one really cool package. I foresee (hopefully correctly) a long and fruitful association...

[JonMarsh]

OK, some updates and further comments about how one should be careful and look at the big picture...

First, I'd like to show you the scope monitor of the test signal output- may give you some appreciation for why mixing 18+19kHz or 19+20kHz at high level is something of a stone cold b*tch... you get a beat frequency output that has a peak level as much as 2x the individual signals, and 2x the dV/dt, too.



Now, we're going to go back and look at that log frequency FFT display, with only one sweep, so there's no noise averaging - this still is a pretty solid performer. This time we're displaying out to the FFT analyzer window, not just to 20kHz.



Now, we can see that in addition to the difference frequency IM down at 1 kHz, but at -90dB, there are also some high frequency clusters garbaging things up- but with this log x axis, its hard to tell exactly where they're located. So let's redo that with a linear x axis-



Now it's pretty easy to see that those secondary IM side band clusters are falling at nice integer multiples of the test frequencies- 2x, 3x, 4x, and 5x. The 2x one is of more concern to me- though the frequency itself is nominally inaudible, this is a lot like an image rejection issue with a CD player- that's falling in a band where a tweeter may have pronounced breakup resonance, and that can drive IM further down in the tweeter itself, due to the physical interaction. And in this case it's only at -60dB. That's actually the highest distortion product we've seen now, testing this little headphone amp. Ah, but maybe that's just it- it's a headphone amp, not a preamp. Will not likely see any kind of tweeter- though some headphones have them! :W

OTOH, this relatively high level high frequency IM shouldn't be a surprise, given the trend we saw with the swept frequency distortion measurement earlier.



The FFT is very, very useful, if employed with some flexibility and discrimination- while the level based distortion sweep does tell us something about overall trends, as it lumps thermal noise, power supply noise, and distortion all together, it's not very diagnostic in a detailed sense.

[JonMarsh]

For fun and practice, I did a bench series for the AURILiC Taurus Pre- posted in Audio Hideout.

[dar47]

Wow that Aurilic pre is clean and it will be nice to see how your 851e measures for stereo pre not one optimized for head phones. That might be an option to upgrade the EMO. Wish I could give you my emo for testing to see if it's a week link. If the Ncore is in this range I'll feel like there is no more holding back sound for the Ardents. I think you could just play with your new gear for at least the next half year and forget the speaks for awhile. Jon I must admit you have great multi parallel processing capability, I would forget half this stuff with all your goings on, haha.

[JonMarsh]

Yes, the Taurus is very clean- I'm going to use it as an intermediate preamp between my condescending mic and the AP for acoustics testing. But I think the Cambridge will look very nice, too. I'm quite curious about that, but running out of time before two weeks of travel kick in.

Both the Minerva and upgrading the Isiris are going to get a lot of attention this year, I hope.

The Benchmark AHB2 tests are posted- pretty amazing for a power amp.

[TEK]

The measurement of the Benchmark is quite impressive.
I was a vit supriced by this:

Very compact for the amount of output power, and not being Class D

I thought that it was a class D design.
What design is it as it able to deliver that much power in such a small box. It can't be class A?

Have to say that I'm quite excited to see the measurements of the hypex modules with the standard inout board and with the "Jon Mash system board" design when we get to that stage...

[JonMarsh]

The Benchmark AHB2 is a class H design (switched multi-level power supply) with very low bias, relying on transistor selection and driver circuit and feedforward error correction to get very good results. (all audio power transistors are by no means created equal!) The entire chassis is a heatsink, the way it is put together, this also helps. Also, it is so small because it uses a patented (by THX) power supply concept based on dual resonant converters operating in variable phase quadrature to regulate the actual output power. So the power supply is quite small, quite efficient, and very clean as regards noise.

Last, it is only 100W for 8 ohms output, unless you bridge it. This keeps the voltage across the transistor relatively low.

[JonMarsh]

One of the few portable Windows 7 Professional PC's you can still buy runs 4K out

And it has a green racing stripe! (modern oscilloscopes are PC's at their heart- this is a big one- about twice as heavy as my 4 channel LeCroy, and can drive a 4K external monitor).

UPS delivered today- it's maddening that I'm so busy with other work things and getting ready for my trip to Munich that I don't have time to do anything with this beyond taking a few pictures.

[dar47]

Wow that's not light weight, even got some mean resisters hanging out the back. I think your planning to retire and start up a test / repair shop, haha.

[TEK]

Hmm, we have a couple of options here
- he is starting a repair shop
- he is going to be a professional reviwer (why not get paid for what he already does for free?)
- he is going to start his own speaker and amplifier brand

Or he is the most dedicated and engaged "amatur" ever ;-)
Or maybe it's all of the above ;-)

Joke aside - I really envy Jon for both his equipment as well as having the knowledge to actually use it - and I'm very greatful to be on the receiving end of his (and others here at the forum) effort.

[JonMarsh]

Repair shop? Nope. Too much aggravation and not enough money.

Reviewing? Well, certainly self reviewing, and also finally being in the situation where I DON'T have to depend on whether John Atkinson has tested something for Stereophile to know if it's Scheisse or not.

Own speaker and amplifier brand? Well, we sort of have the speaker brand, but not as a commercial endeavor. And in the past I have done consulting, would be possible to pick up that kind of work again...

But you know, I like being able to do in audio what I like to do, and not to have to make the kinds of decisions that might be necessary if one is trying to make a living from it. I've had the opportunity to quit my current job several times over the last several decades and join recognized audio firms, but somehow it has never charmed me, because then I would also have to buy in completely to their particularly philosophy and approach- and sometimes I find I there's an element of self delusion that arises out of the necessary effort to make sense of it all and have a corporate "philosophy" - whereas I'm more interested in discovering what really works... That's often not compatible. And finding it out, albeit at my own limited pace, based on my personal limitations of time and knowledge.

So the very last one is the most accurate or complete explanation- I like being a dedicated amateur, becuase then I am also the auteur in this effort, whereas joining another company I would be a minion. Not a bad role, in some ways, but limiting compared with what I want to learn.

I doubt at this stage I'm going to influence another future Charlie Hansen, but who knows? There's some years left to work at that!