Some new Schitt in power amplifiers... the Vidar

[JonMarsh]

The Vidar was announced back in late 2016, but has just begun shipping- the same week that the first Tesla Model 3's go to owners. While I wouldn't claim that the importance or impact of these events on the world at large is equivalent, it just might be that it is that important to Schitt, especially after the sometimes mixed reception to their high end integrated, the Ragnorak. That unit appears to be a fine sounding component, but the behavior of it's micro controller bias system left John Atkinson at Stereophile scratching his head, because while it may have been well optimized for most music signals, it behaved very strangely with continuous test signal, with bias levels dropping and distortion rising.

The Vidar retains microprocessor control for many functions, but incorporates a more conventional bias control. What's more, it follows a system configuration that I rather like, with the ability to operate either as a small stereo amplifier (100W/ch at 8 Ohm, 200W/channel at 4) or as a balanced amplifier driven by out of phase XLR signals, delivering up to 400W at 8 ohms. The power ratings remind me of the Benchmark AHB2, but the size and weight are in a different class, as there isn't a switch mode power supply, just a standard class AB with linear diode bridge to capacitor supply. There are reputed to be some tricks in the circuitry, inlacing the use of a current feedback gain topology, and three sets of output devices are used in each channel.

The casework and build quality are pretty typical Schitt; quality execution of a very simple design that minimizes the number of parts and cost.







Steve Manning seemed to think it was somewhat reprehensible that I took one of them apart instead of listening to it, but we all get our kicks in different ways...






Test data has been taken and will be posted soon- gotta walk the dog for now, and get ready for a busy Monday... Stay tuned.

And no, I haven't had a chance to listen to them- I don't even have a system setup these days, because of all that has been going on.

BTW, the Freya is back from storage and will get tested soon, too!

[JonMarsh]

At this point I have some data, but overall still probably more questions than answers. There are some specific things Schitt did to in their view optimize the audible performance, including a separate regulated HV supply for the gain stages that ensures that clipping will occur in the output stage, not the gain stage- Stoddard says this sounds better. Now, there can be problems doing this, such as over drive and saturation of the output transistors- resulting in sticking in turn-off at high frequencies. Later I will look at these issues.

Also, my interest is mainly in the fully balanced mode, not as a stereo amplifier, so my measurements have all been done in that mode. Also, the test load I used is a little more demanding being about 7 ohms instead of 8 ohms, with 3x 22 ohm resistors in parallel.

Additionally, I don't have the variac at home so that I can adjust/correct for powerline drop; but then., when you're listening to music, you don't ether. OTOH. continuous power tests load down the power supply in a way that music with it's typical 5:1 crest to average power don't, something that NAD takes advantage of in most of their designs. It will take more investigation to see if anything is going on like that here, with higher rest supply rails. Storage caps are a bit on the small side, at 10,000 uF each, for 20,000 uF for each Channel. OTOH, there should be relatively little channel cross talk on the power supplies when running in stereo mode, what with separate windings and rectifiers and filter caps.

What I can say with any certainty follows...


First Check... 3V peak at 1kHz sine wave

This is sort of equivalent to using it like a preamp, as regards output signal level.

Output of the Vidar in balanced mode:



THD reading




FFT of 1kHz signal




Interpreting the FFT

The key thing is to understand the Y Axis, which is in dBV. Note that the 0dBV level is 1VRMS. So when you see the residual noise floor in the range of -90 to -110 dBV, that is below the 1VRMS level. This is a quiet amplifier. To get the line harmonics more clearly, I could use a linear scale, but then I would need to restrict the frequency range or there would be too much data to make sense of. 2nd harmonic is below -80dB, 3rd is at about -108dB, as is 4th. There are no higher harmonics in evidence. This is a nice result at this relatively low output level.


THD+N versus output level, 100mW to clipping at 1kHz



Don't forget we're loading this a bit harder, with a 7 ohm load. Below 5-10 watts, the THD+N is mainly noise and a bit of low order harmonics. The distortion level is pretty reasonably low from 10W to the clipping or limiting point- it will take more checks to see exactly what is going on- didn't have time yesterday. Remember that I'm not correcting the line voltage, and that we're loading the amp a bit harder- it's clipping/limiting here at a voltage that corresponds to 300W at 8 ohm in bridged. Again, what it might actually do with music signals is an interesting question. The thing I plan to do is take these down to my colleague who has my original Wavcor Ardent's and see how they do driving them. A pair of these at $1400 is clearly more powerful than a 651A at $999, which clips at 100W at 8 ohms. How it handles the 5 Ohm Wavecor Ardent load will be interesting to see. I have new load resistors purchased so I can do that test here, but haven't had the time to assemble the setup.


THD+N versus frequency, 100W output





This is a pretty typical characteristic for amplifiers that don't have some kind of exotic Wideband feedforward error correction, like my Halcro DM68's, with THD gradually rising with increasing frequency and fall off of loop feedback. Output level is 100W.



FFT of 50Hz at 20V Pk





This test is a little more demanding of the power supply, and shows a combination of distortion harmonics (50Hz multiples) and supply related harmonics (60Hz multiples) up to 1kHz. The levels are fairly low, typically -100dB to -120 dB. Certainly good performance at this price point. Yes, a Halcro measures better- but besides the cost difference, will you hear it in your system? That depends on a variety of factors, I think.


CCIF Two Tone HF Intermodulation distortion


Signal level here is 25V peak; if you're driving your tweeters this hard, they're probably blown by now...

Output from Amplifier


FFT analysis




The highest level sideband products are down about -90 dB. This is a decent result, but doesn't quite match the Cambridge 651 power amp at the same output level:



I'd say the 651 is a bit more "blameless" under this punishing test- if you know who the designer of this amplifier is, you know why I used that term... :W


And just to be Evil, here is my reference, the Halcro DM68.

FFT SpectrumCCIM by Jon Hancock, on Flickr

And to be realistic, you'd probably need some amazing speakers and ears to hear the difference...

[wkhanna]

Wow......not bad at all for its price range.....!

thank you so much, Meistro !!!!!!!

[JonMarsh]

I think it's quite good for it's price range. In one sense, Schitt has almost set themselves up to be a miracle worker, so some may be disappointed if they don't see world beating performance compared to really big buck amplifiers for this one at $699. It's really a question of perspective and expectations. If one goes the Charles Dickens route and has "Great Expectations" (i.e., unrealistic?) then you might be a skosh disappointed.

Another thing to point out which is relevant to the sonics, there is no roll off or modification of the frequency response in the audible range- among other things I should do a small signal frequency response, but looking at it briefly at high power, it's dead flat 20Hz to 20kHz, which isn't surprising given it's a DC coupled design. Not having an input cap can be a big benefit. So let's just say that the basic bones appear to be here at a very reasonable price- with a 5 ohm load like the Ardent, this amp should easily deliver 150W per channel, and it's likely the sonics are very, very good for the price point based on what I've seen on the bench. Particularly, it's well built at this price point, especially considering it's not made in China. That's only possible with their direct business model. Through dealers, these would cost between $1,000 and $1,200.

[JonMarsh]

A small update/warning:

Something I didn't really expect- I have been using the Vidar's in dual mono with a fairly benign load, some Wharfdale Reva's, connected with some 3rd party manufactured Canare 4S11 speaker cable. no problems. But I figured I should spiff up the little system, because I'm checking out a new relatively high end R2R ladder DAC (actually, a very high end R2R ladder DAC) so it was time to pull out some of the better cables, including either my Golden Reference Cards cables or the newer Cards Clear Reflection speaker cables.


Nuh uh.

:nonod:


Something about the Vidar's behavior (running bridged mono) and protection circuits doesn't like those Cardas speaker cables, and goes into a protection mode with the blinking front light; changing back the speaker cables and all is well.

40 minutes later, the speakers are repositioned so that the stereo NAD M22 can be connected with the 1.5 M Cardas Clear Reflection cables. No protection issues, and is that a bit more focus that I hear out of those little Reva's? Is Jonathan Brooke's new Album (Midnight Hallelujah) just a bit more "in the room" now? And the bass is even tighter and deeper?

Nah, must be in my head. Or the NAD M22.

Maybe I ought to find the time to put the M22 "On the Bench", but then I'd have to pull it out of the system...