Review and Measurements of SMSL Sanskrit 10th DAC

[Sangram]

Usability:

The DX3 pro's muting circuits incessantly click depending on what is selected and what you're doing. The unit can be set to fixed gain mode with one button on the remote - which can result in serious injury or equipment damage, maybe both. The two outputs are not usable simultaneously or switchable by jack insertion, require manual switchover using remote control.

Circuitry:

The Topping design uses series capacitors in signal path that are not properly terminated and not needed, and you could probably get a small performance increase by eliminating them. The power supply for the SMSL uses a switching isolator, whereas the Topping has a hard connection between PS input and circuit output. That's sloppy design.

Value:

The SK10 is only a little behind in terms of sound (though both are much brighter than I like) but is overall not as far behind as the 2x increase in price might suggest. The move from a single 4490 to dual 4493s should have gotten a bit more than the DX3 is able to achieve.

These are my opinion, of course YMMV.

 

[Tatteredmidnight]

Usability:

The DX3 pro's muting circuits incessantly click depending on what is selected and what you're doing. The unit can be set to fixed gain mode with one button on the remote - which can result in serious injury or equipment damage, maybe both. The two outputs are not usable simultaneously or switchable by jack insertion, require manual switchover using remote control.

Circuitry:

The Topping design uses series capacitors in signal path that are not properly terminated and not needed, and you could probably get a small performance increase by eliminating them. The power supply for the SMSL uses a switching isolator, whereas the Topping has a hard connection between PS input and circuit output. That's sloppy design.

Value:

The SK10 is only a little behind in terms of sound (though both are much brighter than I like) but is overall not as far behind as the 2x increase in price might suggest. The move from a single 4490 to dual 4493s should have gotten a bit more than the DX3 is able to achieve.

These are my opinion, of course YMMV.

Serious injury? Really? I find it hard to believe that switching to full line level could cause major bodily harm resulting in death or permanent disability (if your system can literally deafen you under these conditions, there are other safety issues).

I agree its a design flaw, but maybe tone down the hyperbole. I think temporary/minor injury (at most), discomfort (more likely) is more reasonable, no?

 

[Sangram]

Serious injury? Really?

Depending on your overall gain staging, you may force extreme SPL levels from the system and depending on where you are in the listening chain, 0dB volume is capable of serious and permanent hearing damage.

Consider a system with maximum SPL level of 105dB before audible clipping. This can cause severe injury to children (who might be more likely to press a random button). I stick by my statement. It may not be true in your case, but that doesn't mean it can never be applied. In the right (wrong) set of conditions, severe hearing damage is a real possibility. This is why no system with switchable fixed level uses a single (and mutlifunctional) button to do this. It's basic ergonomics in electronics.

 

[GGroch]

Whereas I agree that in this case the manufacturer should have been explicitly clear about the need for a power source when using USB as well - or offer such a supply for sale.....All equipment using USB power will give different results depending on the test situation. It might be a good idea to test all USB-connected devices with an external supply when such a possibility exists.

Welcome new member!

While I appreciate the review of how USB works, in this case the Sanskrit User Guide is already explicitly clear about the use of the "Aux Power" connection. It is not for use with desktop connections, but is instead solely for low power phone/OTG connections. They follow up with a warning that using the wrong type of power supply may damage the DAC with over voltage...not a warning I often see.

Are you suggesting that test protocol should always involve disregarding the manual? Is it your practice to disregard user manuals from "day 1"? (or more probably like me, you do not read them

As Amir stated, he has never run into a test result before where insufficient current made a USB DAC sort of work..ie, work but with issues. Perhaps going forward should similar issues arise he will know to test them. In this case, I think the knowledge that the DAC works via USB only when disregarding the manual is useful information, and I appreciate SMSL so quickly stepping in with this suggestion.

 

[Tatteredmidnight]

Depending on your overall gain staging, you may force extreme SPL levels from the system and depending on where you are in the listening chain, 0dB volume is capable of serious and permanent hearing damage.

Consider a system with maximum SPL level of 105dB before audible clipping. This can cause severe injury to children (who might be more likely to press a random button). I stick by my statement. It may not be true in your case, but that doesn't mean it can never be applied. In the right (wrong) set of conditions, severe hearing damage is a real possibility. This is why no system with switchable fixed level uses a single (and mutlifunctional) button to do this. It's basic ergonomics in electronics.

I agree it's a design flaw, but this seems like serious blame shifting. If you are trusting a sub $200 budget device with no safety certifications to protect your children from a stereo you've turned into a literal weapon, the fault may not lie with the budget device.

I know you won't likely cede this point, but I think this is a situation where some common sense and dilligence is required. Many devices can fail and introduce high gain noise into a system (I've had a crossover fail and blast mains noise at very high levels before).

This, in my opinion, falls into the "other safety issues" caveat I made in my original statement.

But again, I think it's a poor ergonomic decision that could cause some amount of discomfort and/or temporary injury, so we are arguing semantics at this point.

Maybe if you are in that situation, you should use the DX3 as a dac/headphone amp, and use a separate preamp, considering the particular danger here.

 

[Sangram]

Are you suggesting that test protocol should always involve disregarding the manual? Is it your practice to disregard user manuals from "day 1"? (or more probably like me, you do not read them

Thanks for the welcome

As I said, the variability of power supply over USB ports is a significant issue with any DAC that power from such a port.

My point is not that the review is inaccurate or unfair - it is neither of those things. It is that the testing protocol does not test the maximum possible performance, but what is possible under a specific set of conditions.

I've been designing and using audio products for over 30 years now, power is always an issue whether you deal with fractions of, or thousands of watts. In fact the smaller units tend to be more problematic to get right.

I agree it's a design flaw, but this seems like serious blame shifting. If you are trusting a sub $200 budget device with no safety certifications to protect your children from a stereo you've turned into a literal weapon, the fault may not lie with the budget device.

I know you won't likely cede this point, but I think this is a situation where some common sense and dilligence is required. Many devices can fail and introduce high gain noise into a system (I've had a crossover fail and blast mains noise at very high levels before).

105dB is not a weapons-grade stereo, most consumer-grade systems will hit 103dB fairly easily and the 105dB level may come with lots of additional distortion but is still achievable. Budget is not the issue either. Whether a product costs a dollar or a thousand, it must achieve its maximum possible performance with minimal safety risk, with a good eye to all sorts of use cases. That is what separates great designs from average ones.

Yes, you're right, I can't cede the point of it being a serious risk, primarily because it is. I have this in the bedroom, where a -15 to -10 listening level is louder than needed and anything above -8.5dB is unusable. An accidental 0dB level would be quite the problem. I don't have children per se but if I was designing a product I would consider a scenario where a child had even a chance to get at the remote.

Failure mode effects are a different story and cannot be considered in the same breath as design flaws. Those should also be benign but rarely are, as you cannot account for every possible failure mode when designing, and cannot always force a failure mode when debugging the product.

 

[garbulky]

This sounds like a good time to see how much power your laptops USB ports are actually capable of delivering and how good a quality this power is. Also have you had a chance to do your own power line measurements?
The USB power would establish a baseline
1. That there's nothing weird/sub par going on with your laptops USB power.
2. If we know how much power is there, then we can definitively say the SMSL and other USB DACs that fail here would require a unit with more power than xx. This would also be useful due to some phones providing inadequate power for USB DACs.

 

[bennetng]

It seems that USB linearity remains the same even when using an additional power supply, who can explain this behaviour?
Also, could you check USB jitter again? Thanks!

 

[amirm]

One of the reasons the unit did not operate correctly when connected to USB 3.0 port was the lack of signalling. The PC assumes it is a USB 2.0 device, and does not turn the port into USB 3.0 mode. The additional power pins on the USB 3.0 A host port do not have corresponding contacts on the micro-USB cable, so you do not get the benefit of the additional power either.

My reference to using USB 3.0 port was not in regards to the DAC operating in 3.0 mode which it does not. But rather, the USB 3.0 port has been able to power some units that don't work well with my USB 2.0 port. The SMSL Sasnkrit did not work well with either over USB.

Most equipment using XMOS receivers will draw the power for the XMOS from the USB port, but power all other circuitry off other power sources (where available). There is also the issue of the USB cable, some are built with thinner wire and may have significant voltage drop. For example it seems highly unlikely that the unit will operate correctly when connected to phone using OTG cable, as the phone's port will never supply that much power.

As I noted, I also used the supplied cable and that actually made things worse in one channel. The cables I use have proven to work well powering countless DACs. Still, when in doubt I do switch them around just in case.

To be clear, my main system for the last 9 months or so has been a desktop, not my laptop and certainly not a phone.

 

[amirm]

It seems that USB linearity remains the same even when using an additional power supply, who can explain this behaviour?

We saw the same issue with Topping DX3 Pro (although not as severe) which didn't have power issues (has its own external power supply). Something is going on with budget implementations of AKM DAC I think. Note that S/PDIF performance is nearly as bad just the same so this is not USB related.

 

[amirm]

Get the Wyrd in front of this DAC, let's see if it helps. This seems a good candidate to test the Wyrd.

Here you go:

As you see, Schiit Wyrd does nothing to improve the situation.

 

[amirm]

Also, could you check USB jitter again? Thanks!

It is substantially cleaner with external power supply:

All spikes are below -120 dB (provably inaudible). There is a jitter source at 6.x kHz though as represented by the symmetrical spikes below and above our 12 kHz tone.

 

[rebbiputzmaker]

Whereas I agree that in this case the manufacturer should have been explicitly clear about the need for a power source when using USB as well - or offer such a supply for sale - the vagaries of depending on USB power are many and well known.

The USB ports of different computers/motherboards are powered differently. Not all downstream equipment will be able to operate correctly on all upstream ports. For example, when we order USB fuses you will have a variety of current and surge ratings, even if the overall spec is still 500mA. A fuse with zero surge rating, for example, will have a very high resistance and will sag above 50% of current draw. Overall power budgets dictate the choice of fuses, and the end-user has no control over the quality of USB port power.

One of the reasons the unit did not operate correctly when connected to USB 3.0 port was the lack of signalling. The PC assumes it is a USB 2.0 device, and does not turn the port into USB 3.0 mode. The additional power pins on the USB 3.0 A host port do not have corresponding contacts on the micro-USB cable, so you do not get the benefit of the additional power either.

Most equipment using XMOS receivers will draw the power for the XMOS from the USB port, but power all other circuitry off other power sources (where available). There is also the issue of the USB cable, some are built with thinner wire and may have significant voltage drop. For example it seems highly unlikely that the unit will operate correctly when connected to phone using OTG cable, as the phone's port will never supply that much power.

All equipment using USB power will give different results depending on the test situation. It might be a good idea to test all USB-connected devices with an external supply when such a possibility exists. Whereas I understand the need for test protocol, one of the founding principles is the need to eliminate variables. The USB port performance of a PC and cable system is a variable and cannot be assumed constant, which means the results may vary by user configuration, motherboard capability, and so on. You could consider eliminating this variable.

BTW I purchased this unit before seeing any reviews, and find it excellent at the price. Have been using with a simple linear supply from day 1, and I love it. Much better than the DX3 Pro which I also purchased together with this.

Curious how is it better that the dx3, how are you using it? TIA

 

[Aibo]

Thanks! I've been curious about this test ever since I bought it and wasn't impressed by the sound. This sheds more light on the matter.

 

[trl]

Here you go:
[...]
As you see, Schiit Wyrd does nothing to improve the situation.

Thank you!

It's odd, because of Wyrd's specs state that: "USB Power Current Capability: Over 500mA (full USB 2.0 spec)", so when using Wyrd with its own power supply, the voltage and current should suffice for all USB2 DACs, so I don't understand what's wrong in this setup?!

 

[GGroch]

Trl.
As per my previous post, you would probably need to use 2 of them (one for the Aux power) to solve this problem. Too Wyrd I guess
Amir is using a desktop and tried both USB 2 (assumed full mA spec) and USB 3 (significantly higher mA power spec) and that did not solve it, so I am not surprised this did not.

This further confirms that problem was not that Amir's setup did not meet spec or the result of a dirty crapified USB signal. It was that the Sanskrit requires more power than a single USB can provide. Nearly all desktop DACs with full function (like remote and LED display) do require more. SMSL should update the instructions and include a 5V power supply. With the aux supply, the numbers (and presumably sound) are great. No decrap needed.

It is interesting that in the description of the Wyrd on schiit.com they pretty much admit that in most cases it does nothing & that cheaper solutions may be available. Rare honesty from an audiophile mfg. They even write: We won't make any such claim to sonic nirvana—sorry, creating expectation bias and neuro-lingustic programming ain't something that we do. I am not sure they do not, but admitting the effect exists is the 1st step to recovery, and once again, rare, if tongue in cheek, honesty.

This test does not prove whether the Wyrd ever works or not (or under what circumstances). Just that it does not clean up the measurements when the reason is lack of USB power.

Follow ups: rebbiputzmaker, previously answered.
Aibo, Did you try it yet with the 5V power supply in aux (as per the 2nd test?). The specs are great that way. Does it sound better?

 

[trl]

[...]
This further confirms that problem was not that Amir's setup did not meet spec or a dirty crapified USB signal. It was that the Sanskrit requires more power than that on USB.

110% got it now, thank you! Now would be a good time to test the 5V/1A charging port from ThinkPad laptops.

[...]
It is interesting that in the description of the Wyrd on schiit.com they pretty much admit that in most cases it does nothing & that cheaper solutions may be available. Rare honesty from an audiophile mfg. [...]

I'm sure they're right, I also appreciate they stated such thing on their website, although I'm sure they were referring to self-powered DACs that don't care if the USB power is clean or not...or if it even exists any +5V there.

 

[folzag]

I'm going to repeat myself in case anyone missed it. Amir was not the only one to not realize a second power source was required with USB audio. This YouTuber made the same assumption...

OK, I find this super interesting... I wanted to see the back of the unit so I did a quick search and this YouTube review was the second or third hit. At 1:10 he describes the two USB ports as "so, if you're not going to be using the USB for signal input (ie. use S/PDIF), you're going to have to hook power to this upper USB port." Heh. Intuitive, rational, and, yet, exactly backwards.

And sure enough, his review of the sound, which starts at 1:23 matches about one would expect given what we now know. Always nice when the objective measurements can anticipate the subjective results, eh?

 

[amirm]

I'm going to repeat myself in case anyone missed it. Amir was not the only one to not realize a second power source was required with USB audio.

It is not an assumption. It is based on what the manual says:

Since I a not using a phone, it reasons that you don't hook up the aux power. A number of other DACs have such aux ports but I have not seen one necessary to use the USB. Even Schiit Modi 3 works this way even though it does come with the power supply. The use of the power supply is usually for other inputs other than USB (e.g. S/PDIF).

 

[skcheng]

It is substantially cleaner with external power supply:

View attachment 19002

All spikes are below -120 dB (provably inaudible). There is a jitter source at 6.x kHz though as represented by the symmetrical spikes below and above our 12 kHz tone.

What external power supply did you use for this test? And would this be a more universal improvement for other USB DACs? Or is this more specific to this DAC due to its design? Thanks.