Conversion of a Chord DAVE to DC power using two Uptone Audio JS-2s

[ecwl]

5 hours ago, DAVE JS2 said:

I'm now turning my thoughts to the Blu Mk II, and I'm curious to know if you use one (or a Hugo MScalar) with your DAVE, and if so whether you've (or anyone else reading this) tried similarly improving the power supply and with what result?

I have gone kind of the diametric opposite direction as you did in order to improve the DAVE sound.

First, I traded in my Blu Mk II to an M-Scaler because I don’t actually use the CD transport as I’m 100% computer audio. Then like you, I have some UpTone products that do improve the sound a lot but having read @JohnSwenson’s comments about ground loop leakage current noise being a major component of noise affecting performance and that solution is proper grounding and not using SMPS and using linear power supplies instead, I just wondered instead of constantly trying to upgrade the power supply, what if I just remove the ground noise issue all together. So my UltraRendu is actually connected to a USB-Toslink converter (unfortunately max 96kHz) and the M-Scaler is fed Toslink. So my M-Scaler, DAVE and Etude are still powered by the original SMPS and are the only devices connected to each other. In fact, I discussed this with the Chord DAC designer Rob Watts and that was what he recommended since I don’t even have high-res music beyond 96kHz. (Actually he said I can try a USB to 192kHz Toslink converter). But on Head-Fi forum, there are definitely some people who have replaced their M-Scaler SMPS with linear power supplies instead and report a sonic improvement.

 

I also noticed your comment about Dirac. I recently bought Acourate and spent over a month learning how to optimize the convolution filter. My experience seems to suggest that it is very difficult for an automated program to find the optimal convolution filter so even if you have programmed to your preferred frequency 

response, it is very easy to end up with a phase correction that loses PRaT. I suspect that’s why you weren’t happy with Dirac. 

[ecwl]

7 hours ago, DAVE JS2 said:

That's so interesting @ecwl, sounds like good thinking to go into the device optically. I also spoke to Rob, a couple of years ago now... Although I do beg to differ with him about those SMPS he uses, given my recent experiences...

 

Where in Canda are you? My wife is Canadian, from Nova Scotia.

 

Your room convolution experiences sound fascinating.

I think Rob Watts (and maybe even John Franks) have this really odd philosophy about SMPS. They fully acknowledge that if you have a grounded computer audio device connected via USB to their SMPS Chord components, you’re probably going to get a little bit of ground loop leakage current noise that’ll affect the performance of their products. But Rob Watts definitely thinks that’s not his problem. It’s the problem of the computer audio device. Because when he tests his DACs, he just runs them with his laptop running on battery. And truth be told, I cannot tell the difference in sound when I connect my iPad via USB to M-Scaler/DAVE vs Toslink. But ideally, I would much prefer to be able to connect my UltraRendu via USB into M-Scaler. And despite using LPS1 and grounding my router, NAS, cable modem, etc. I keep hearing progressive improvements but it’s still slightly different than playing straight from the iPad or Toslink. So I finally caved and just went Toslink. But it’s definitely possible that replacing Blu2/M-Scaler/DAVE with high-end linear power supply would improve the sound. I just never bothered trying it. Felt like too much effort. With all that said, I think Chord has definitely lost some sales because people plug all sorts of things into their USB and the “incompatibility”? with the SMPS limits the performance. I think I’ve even heard this problem at Chord dealers.

 

I live in Winnipeg now but grew up in Toronto as a teenager. Originally grew up in Hong Kong. I actually stayed in Nova Scotia for a month. Very nice people. But yeah, I know what you mean... Different parts of Canada definitely has slightly different culture even if people generally are very nice.

 

I was talking about my convolution experience in another thread. I think the consensus is that you’re better off asking @mitchco to help you create a convolution filter in Roon. I strongly suspect you’ll be shocked by the improvement, even with the B&W 800 D3. I don’t know what @mitchco charges but it’ll probably be more worth it than changing to M-Scaler + a linear power supply.  It may even convince you to abandon plans to do separate digital XO for each driver. It took me a month to fully understand how to optimize the convolution filter in Acourate. It was a good learning experience but I think it’s time I could have spent doing something else.

[ecwl]

6 hours ago, asdf1000 said:

 

Maybe I misunderstood here but I don't think @JohnSwenson ever generalised that the solution for proper system grounding necessarily means using linear PSUs instead of SMPSs.

 

In fact I'm pretty sure I've read both John S and Rob Watts mention there are SMPSs better than linear PSUs and vice versa..

 

The solution for proper system grounding isn't always straight forward because it will vary between each unique total system.

 

You’re absolutely right. I can’t remember John Swenson mentioning SMPS can be better than LPS but he definitely never thought they are inferior, just different. Rob Watts definitely thinks SMPS when done properly is superior to LPS. 
 

I think the real challenge is that for someone like me with limited understanding of electronics and grounding, I can see it’s so much easier to say let me switch the SMPS to LPS so that I don’t have to worry about the high impedance leakage current noise. Rather than to analyze where the noise and grounding issues are. And I did try but in the end, since the Chord DAC architecture is very jitter immune, it was just easier to go Toslink so I don’t have to think about grounding. 

[ecwl]

@DAVE JS2 I forgot to mention that even if you’re not totally ready to add a convolution filter to your system from Mitch Barrett, just measuring your system (through DAVE) with a calibrated microphone through Room EQ Wizard and taking out the bass peaks (if there are any) using the parametric EQ function of Roon would usually yield quite a significant upgrade in sound. 
 

of course for now, enjoy the JS2 upgrade. But if I were in your position and debating what to do next, I’d probably look to see if there are any bass peaks to tune out using parametric EQ in Roon. It’s free to do if you already have a calibrated microphone. 
 

by the way, all this talk about the great products from Uptone audio makes me want to upgrade from my LPS-1 to LPS-1.2. It is a great company

[ecwl]

1 minute ago, DAVE JS2 said:

Thanks for the thought @ecwl; thats actually exactly what I did after I got my room as good as I could the 'old fashioned' way, and had only minor uplifts corresponding to the room L, W, & H.

 

I found that I couldn't really tell any difference with the parametric EQ in or out, so decided to leave it off.

 

I'm pretty certain now that I'm going to swap Blu Mk II for Hugo MScaler and use a JS-2 for power, I think there were RFI improvements to it compared to Blu Mk II.

Sounds like you just don’t have any major problematic bass peaks. Usually if you have peaks more than 5dB, an exact parametric EQ to tune it out would make a pretty obvious difference. Correcting troughs usually don’t work.

 

And yes, I suspect in your system, JS-2 + HMS probably would sound better than Blu Mk II. Even without the JS-2, the USB input of the Blu Mk II is not galvanically isolated whereas the HMS one is. So that alone would probably gives a bit of improvement. Although maybe not because you already have lots of ferrites on your dual BNC cables. I don’t use ferrites. I think Rob Watts thinks adding ferrites between HMS and DAVE actually makes things worse? Whereas ferrites between Blu2 and DAVE are better.

[ecwl]

4 minutes ago, DAVE JS2 said:

 

Ah thanks for the info on HMS - sounds like he learned some lessons from Blu Mk II then! I did not know that about the USB.

 

At least ferrites on vs. off the BNC cables is a pretty easy A/B to do!

As i said in my first response in this thread, Rob Watts has this funny idea that if your USB streamer needs galvanic isolation, that’s your streamer’s problem, not the Blu2/DAVE problem. Now that’s an exaggeration because DAVE USB input does have galvanic isolation.

 

When Blu2 came out, I think Rob Watts was expecting most people to just play CDs. And if they’re going to use the USB, they’ll use it like he does which is to play it off a laptop off batteries (which would not require galvanic isolation). So the only thing he did was to have galvanic isolation with the dual BNC output of the Blu2 into the DAVE. Which when you’re playing CDs, is at least IMHO just fine (even without the ferrites on your BNC cables).

 

Except Chord at the time did not have HMS so people like me ended up buying Blu2 and not use the CD transport just to get the million taps. And a lot of users on Head-Fi gave feedback that better dual BNC cables sound better and how different USB streamer sound totally different through the Blu2. Some people also noticed, like me, that going through Blu2 sounds better even without the upsampling, probably because my old USB streamer leaks so much noise that it used to overwhelm the galvanic isolation of the DAVE so going through the better galvanic isolation of Blu2 dual BNC made the sound better. So Rob Watts looked into it some more, thought about what the next generation product should be like. So he ended up keeping the dual BNC galvanic isolation in HMS and added USB galvanic isolation AND added additional ferrites within HMS to remove the extra RF noise that can leak into the HMS. 

 

With all these changes, some users found that they no longer have to do any modifications to DAVE/HMS or their USB streamers. But there are still some people who do. As @asdf1000 said, proper grounding of the system to minimize leakage current noise is not always straight forward. At least that’s why I think people still do all the different modifications.

 

I think it’s also important to keep in mind that with other DACs that I’ve listened to, they’re often more sensitive to these issues. Part of the problem is that other DACs tend to have other distortions that would overwhelm the more subtle differences the leakage current noise make. I would argue DAVE is too revealing so these subtle changes are more easily audible. But I would listen to a stock DAVE hooked up to a noisy USB streamer any day over another DAC hooked up to a better streamer and with better power supply. Of course, I’m sure lots of people on this forum would disagree with me. But to each their own.

[ecwl]

3 minutes ago, MarkusBarkus said:

...actually, I think I recall reading in his posts that he personally used Wave dual BNC cables, which are heavily "ferrited" between HMS and DAVE due to RF "issues." 

Oh yeah... I got confused. You’re right. He initially didn’t think it matters but lately has noticed that Wave dual BNC cables does make the sound slightly better. Haha... Maybe I should just go back to USB into HMS and then switch the dual BNC cables... And get the LPS-1.2... The more I participate in the forums, the more ways I can come up with to spend money

[ecwl]

38 minutes ago, skatbelt said:

 

Thanks! Hmmm, there goes my assumption of purposely built or at least custom built to spec for Chord. This is an off the shelf supply. For medical, among other applications.

My understanding based on what was discussed in Head-Fi by Rob Watts is that he picked the “best” SMPS he thinks is appropriate for his DAC design and then design the regulators and RF filters subsequent to the SMPS around the specific SMPS. So to him, the PSU of the unit is not the $50 SMPS or whatever LPS is in use but the entire system from the SMPS + RF filters and regulators, etc. So it didn’t surprise me that the SMPS was only $50. In fact, I thought it was going to be cheaper than that.