AC Filtering, Grounding Boxes, Linear PSU and Balanced Power.

[YashN]

Here is a DIY Supercapacitor battery pack that might interest you. It includes a LM317 linear voltage regulator! 

Yes, saw it when I was researching ultracaps, it's a very cool little project. I don't like the DC-DC booster circuit for audio though, so it could be adapted.

 

The plan is also to build my own ultracap this holiday period and test it at the output of the LRP. It should blast the Rubycon performance to smithereens.

 

Fellow CA member, @hurka built his own 2-bank ultracap circuit for audio - it looks great.

 

For commercial solutions, I have no trouble strongly recommending either the Vinnie Rossi PureDC-4-ever or the Uptone Audio LPS-1.

[YashN]

Far from all devices is not even possible to power by unregulated batteries.

 

It is crucial to understand the actual input voltage requirement of the powered device before choosing the correct unregulated battery.

 

When you have found out you want a battery which voltage fully charged is a close as possible to the maximum input voltage of the powered device.

 

Clearly, you have misunderstood, so let me clarify:

 

We were talking about your recommendation about never using a regulator after a battery.

 

I said that you can make mistakes when using a battery followed by a regulator and this can give you the wrong impression about this configuration.

 

Here is a more detailed example:

 

Your DAC needs 5V. So you want to power it with a battery and a regulator.

 

Imagine you choose the LM317 as regulator, as I did for my DIY Linear Regulated PSU.

 

If you put a 5V battery in front of the regulator, and think you can get good, stable, regulated, in short bettter sound from the battery + regulator to the DAC, you won't.

 

What you need to know here is the that the regulator will drop out some voltage for it to be able to work on the regulated output in a stable manner.

 

For the LM317, it is said to work best when there's at least 2.5V - 3.5V to drop.

 

This means, that if you want the regulator to be able to supply a low-ripple, stable, you actually need to feed the regulator a rather stable (5V + 2.5V) or up to a (5V + 3.5V).

 

That is, the supply to the regulator, or in this case, the battery, must be from 7.5V - 8.5V for this configuration to work well and get better, regulated sound.

 

So, in case you have tried a battery with a regulator, and haven't paid attention to the drop-out and necessary buffer to consider powering the regulator, you may have been powering the device out-of-regulation.

 

What that means is that the device can still work (as it may be receiving around 5V but with ripple and noise), but the SQ will not be as good as expected, it can actually be bad if you compare it straight to the 5V battery.

 

So, if you made the mistake of voltage starving the regulator itself because the initial battery voltage was too low, then you could also make the mistake of thinking batteries can't be followed by regulators to get better SQ...

 

This should help make things clearer.

[YashN]

This is something that I have asked many others on this forum and you are actually the only one except zilch0md (whom I trust very much) that thinks that there is something to it as well!

 

Here again, no! There is not 'something to it' - not in the usual charging a battery with an SMPS vs charging a battery with an LRP. You shouldn't have any difference at all. I would be very surprised if you did.

 

If somehow there are still effects related to Leakage Currents in your setup and if you're keeping the chargers connected at all times (which you said you aren't doing), then there could be some differences, but it won't have to do with the battery power itself (assuming the same battery in both cases), but more with the Leakage Current interactions of each charger power supply and other power supplies in the chain.

 

Let's leave it at that.

[YashN]

Great project! 

 

Thanks. I finished my ultracapacitor prototype today. Was checking out its initial charge and discharge behaviour with a multi-meter next to me while watching Sherlock's Episode 2 for this year's season - great episode, far better than last week's.

 

It have been some discussions regarding improved SQ with short DC to DC hard adapters.This would also apply to USB to USB power cable. A hard USB to USB DC adapter might just improve that SQ further.

 

Yes, that works: the nearer the load (i.e. the device being powered) to the regulator's (or your other means of powering if you don't like regs) output, the better.

 

This can be mitigated to a large extent with the implementation of true remote or Kelvin sensing though, so that then you're not at the mercy of the length of the connector (hard or wire) that much.

 

However, I have come to the conclusion that the very USB connector design itself is damaging to SQ for async USB real-time audio streaming, this even more so as we use higher-frequencies of data transfer.

 

Ideally, I would have liked the DAC manufacturer not to include or allow the power lines connections or pins within the connector which takes the Data lines, and have a separate, ideally orthogonal power-only port. I don't know how much of that is possible while keeping impedance requirements standard, but that's globally how it should be IMO.

 

So, for all DACs with a single USB power jack, we're still putting patches on pimples without curing the inner disease though...

 

If you've got any pictures it would be very interesting to see them as well!

 

I had some and some more planned, my Macbook Pro's power supply has been misbehaving. I tried repairing it, and it would have been rather easy to do if only the +ve and GND/return wires were disconnected. But in my case, it's worse than that: the 5 small pins soldered to the tiny PCB in the computer-side plug broke off. I tried re-soldering them with various techniques, but this is much too small in terms of scale a job to do with my soldering iron. These are tiny connections, so soldering one wire can easily bridge to the adjacent one. The sense pin for the 1-wire protocol communication is really, really tiny and gets bridged all the time. So, I've had to order a whole new jack with wire. It's coming from Asia, but it supposedly going to be slow to reach here in Canada. After this I'll be able to post pics more easily.

[YashN]

I can only speak from my own experiance and according to it a unregulated battery is usually the best option since you remove one voltage regulator. Voltage regulators (even linear ones) makes noises and two of them makes even more noise.

 

Not true in all cases:

 

- If you build the circuit properly, two regulators can give you a new, much better circuit in terms of ripple rejection.

 

- Some batteries under load can have their noise profile raised a lot more than one or two regulators

 

However, this highly depends on choosing the correct voltage and mAh rating for the unregulated battery.

I'd say it doesn't depend on that: that is something you always do, no matter if battery or mains, or LRP - rather it depends on the noise profile under load and also whether you're using a proper cap with it, not a proper cap with it, or in a no-cap config.

It highly depends on how you choose the specs and in some cases it is absolutely necessary to keep the voltage constant by using linear voltage regulators which AFAIK should be high quality and close to the powered source (ie. external) to sound best.

 

Bears repeating here: proper use of the regulator necessitates you feed it with your final load's operating voltage but also, the additional drop-out buffer voltage to the regulator itself so that it remains in the optimal regulation mode.

 

If you're not doing that in a battery + Reg config, you're probably getting bad SQ and concluding the wrong thing.

As always it is best to try both. With my coming FMCs (MC100CM) I will try both unregulated (7,4-8,4v) and regulated (5v) battery supplies. I suspect that the FMCs will sound better with a low voltage than with a higher voltage...but I intend to find out the truth in my own setup.

 

Worth experimenting. I believe I gave some thoughts in another thread: it will depend a lot of how you do it, whether the FMC has internal regulators or switching circuits, etc...

 

Low-power to the FMC may create less noise/EMI/RFI, but too low a power may put it in out-of-regulation mode (we're considering the internal regulators here, not anything external), creating more non-linearities and damaging SQ. So, must be checked on a case by case basis, maybe the manufacturer can help with optimal values. You generally do not want a device operating out-of-regulation though despite it probably working even at the lower voltage.

 

The best thing would probably replace all internal regulators by battery or really good linear regulated PSU rails.

[YashN]

Ok, we leave it. I am not expecting a revolution...but I intend to experiment if a clean charging power could change the SQ of a battery disconnected from AC mains.

 

Indeed, never let anyone stop you from investigating yourself.

I don't expect any with our usual means of power though (we're assuming here that the charger itself is disconnected from mains after having charged, otherwise other noises through the AC mains and RFI/EMI would also come into play).

[YashN]

I have been experimenting a lot with the LM317 and concomitant circuits around it.

 

It has a variety of behaviours when it comes to output impedance in these circuits (simulated, not measured unless you count the first few listening session in between layout and sense line changes as valuable too), many of which show peaks and troughs and sometimes, depending on what elements you had in the circuit and where, more chaotic lines reminiscent of Picasso's sketches than what you would expect.

 

For a line-level audiophile PSU, it becomes quite convoluted to simultaneously reach great-sounding targets of PSSR, low noise floor, low output impedance for a wide-bandwidth, etc...

 

This internal schematic should give you an idea of what we're dealing with:

 

 

This, coupled with some sim + build experiments, should bring to mind a few things simultaneously:

 

- Complexity, Non-Linearity, Chaotic behaviour

 

- An off-the-shelf chip may have been designed with other uses in mind rather than our purpose (audiophilia), as such, some foibles may never be completely cleared out in practice

 

- Taking the datasheet recommendations for our purposes doesn't even begin to bring you close to solving the issues with some of these chips (not all fortunately).

 

- Building a discrete, low-part circuit to reach our goals is probably the better option here - as simple as needed to get the job done (regulation, low-output impedance over the important bandwidth, rapidity of transients, recovery, noise floor, ...) but no simpler, and never at the detriment of overall SQ

 

 

I wouldn't recommend the LM317 for audio, but as mentioned earlier, as a learning platform, it's good, probably one of the best things you could do: here, the weaknesses are a path to learn some deeper things in Electronic Design and audiophilia.

 

My LM317 will probably be relegated in the end to Lab PSU duties where it belongs more than in an audiophile LRP...

[YashN]

Just finished one of my 100Base-TX (pin 1/2 & 3/6 on a distance) Ethernet cable with silver wires.

[ATTACH]32303[/ATTACH]

Cool, looking forward to your listening impressions and A/B comparison with a normal off-the-shelf Ethernet cable.

[YashN]

2. MC100CM powered by two regulated battery packs (uphill & downhill) vs downhill powered by unregulated battery & uphill powered by regulated battery.

 

If it's not too much, could you also please add:

 

2(b) Upstream MC100CM powered by battery, Downstream by Linear Regulated PSU?

 

What is the regulator in your 'regulated battery' device?

[YashN]

The end of Tinker, Tailor, Soldier, Spy has a great song from a young Julio:

 

[YashN]

This is the best regulator on the market now

great for DAC

 

0.8uV Ultralow noise DAC power supply regulator 3.3/5/7V 1.5A*x2 - DIYINHK

 

Quite good as a drop-in for a test, but without knowing the actual output impedance plot, there's a fair chance one could build a better circuit with the chip.

 

This is also not the optimal way for audiophile purposes:

 

4)The PCB follows the official demo circuit for ultimate performance.

 

The ultimate performance parameters in the datasheet rarely have to do with audiophile playback.

 

They do mention these cousins in the applications section of the datasheet though:

 

- VCOs

 

- Mixers

[YashN]

Best film I have seen for years.

put a 'smiley ' on my face 

 

Did you know that the real life events the story was based on included someone who blew "Le Carré's" cover, thereby forcing him into retirement prematurely?

[YashN]

I have three of these. They are very good, not great, but good. Voltage adjustment is a bit wobbly. They are low drop devices which is nice.

I see a couple of things to be aware of regarding these little boards, now that I have actually worked on researching, building and refining a regulator circuit:

 

First, as already mentioned, it is possible they go with the App Note or default schematics in the datasheet. As we know, this is often far from the best way to use the chip for our purposes.

 

Secondly, with the very low target noise and great performance (PSSR, stability) we're looking at in these applications, say Line-Level regulation for my case (the DAC), the actual circuitry (including Grounding, RFI/EMI) and PCB traces become an Engineering challenge, to such an extent that it can actually make or break a regulator board compared to its avowed purpose and best implementations.

 

Thus, there is a good chance these boards, which are readily available, aren't really that optimised.

 

NB: I'm not saying the chips themselves are bad, but that making a really good implementation is quite a complex affair.

 

NB2: We can also get lucky and get a really well-made and well-optimised affordable circuit if we look long enough, one day: there are a lot of talented individuals where these boards are being made, they have refined their facilities for board fabrications as well. Finding them might not be that easy though.

 

So, I'm not too surprised by your findings, @lmitche, thanks for sharing.

[YashN]

Just got word that my DIGITUS OM4 cable have been shipped...so not too far until I can evaluate it all!

Listening to Justice album "Woman" and enjoying myself!

Just heard that my birthday present was marked as 'delivered' but nothing was received here. Probably another mess up by Canada Post or people passing by on the street steal my packages...

 

I like Justice's "Cross". Haven't really heard their subsequent material.

[YashN]

Thanks guys.

 

Been working on enhancing the basic circuit design with a couple of new ideas which are quite promising, and perhaps a third iteration I am thinking of but which I yet have to model.

 

Potentially, this brings us is some extreme results in some areas, and we perhaps lose some performance in others.

 

I may also get some boards with newer regs that I may also do some comparisons with, which will certainly be very interesting.

[YashN]

I am always interested in new & promising ideas!

Please share more details when you feel that you're ready.

 

For now, just Wow!!

 

Overall impressions: the soundstage is just larger, much more micro-detail, especially the mid-range, bass expansion and dynamism although I do get the impression on some tracks that attack rapidity is lacking. Expansion in high regions.

 

Peter Gabriel's "Mercy Street" was moving.

 

Definitely the impression of a darker background. Here again, lower noise and output impedance has apparently increased perceived power.

All right, cake and champagne time for birthday boy, as he deserves it.

[YashN]

Cake, champagne & a successful hifi tweak. What a great way to spend the birthday! [emoji4]

 

Another gift for birthday boy! (by birthday boy himself too, who also played the part of Father Christmas!).

 

Fortunately, Amazon agreed to re-send the package, and today I managed to grab it from the mailbox outside before somebody else did (probably what happened to the last packet) and this is in:

 

 

[YashN]

@YashN

If you want another present you should wish for this

John Hopkins "Insides" is a really great album! Would be on my top 10 "to have list"! Go get!

 

Sounded a bit harsh and unmelodic for my liking.

 

I'm more inclined to be tripping to this:

 

[YashN]

When you lose 10 minutes searching but not finding anything about 'Volfcraft' because of Cornan's repeated typos instead of improving the performance of your Linear Regulated PSU...

 

[YashN]

*Haha* sorry YashN! VOLTCRAFT, VOLTCRAFT, VOLTCRAFT! 

 

Now I know, thanks to Alex, but Google kept showing me results for 'Wolfcraft'. <facepalm>

[YashN]

Audiophiles can have girlfriends? I thought we had to get permanently married to find a partner wiling to put up with our bizarre habits!

 

Yes they can, even if not married.

 

But I thought they liked fuchsia

[YashN]

I just made a expetiment that might give @YashN the shivers!

 

Frolicking around in the snow?

[YashN]

Isolation matters!

 

Greatly, and this is what we're doing here for power.

[YashN]

In position between my speakers and Auralex MoPads.

A good improvement over only MoPads. Better bass definition and clear treble!

 

To get excellent sound with bookshelf speakers:

 

- put them on dedicated stands (a desk will form a resonant body)

 

- replace the Mopads with a test of a ball-and-cup arrangement

 

- also try mass loading the stands themselves

 

The Vibration Isolation thread has a lot more info, including the use of air cushions in addition to the ball-and-cup and various other commercial offerings or DIY ideas.

[YashN]

A few thoughts on what we're doing here:

 

I had another opportunity at making yet another 3-level test, but this time the AC Filter Box is already hosting the iMac, the TV and the Blu-Ray Player, but not the SET Tube Amp, nor the SMPS or Linear PSU of the DAC.

 

So basically, these 3 when the AC Filter Box is already being used:

 

1. Default USB cable, with DAC powered by the Computer (NB: internally, the battery powers the clock and DAC chip apparently)

 

2. My DIY USB Connector with external power provided by an iPhone charger (you could use an iPad charger as well or a battery charger pack - and indeed I can also do this last test again)

 

3. My DIY USB Connector with my DIY Low-noise Linear Regulated PSU.

 

A friend listened and she already liked Level 1, but this is now unlistenable to me because so much details and power is lost, and the soundstage is collapsed.

 

At Level 2, the soundstage opens up considerably, micro-details become more apparent, like Adele taking her breath before singing Skyfall, finger articulations on guitar strings apparent with Coldplay, etc...

 

And of course, at Level 3, there's actually more body movements by the listener, sound separation is several levels above that of Level 2, clarity of the main vocalist is more apparent (you now hear distinctly what happens after a hard 'k' articulation), the sound stage is bigger, the bass more dynamic, perhaps a little roll-off in the highs and some rapidity lost.

 

What this shows is that whether I first use the custom USB connector only without AC Filter Box, or in conjunction with it, and if the latter, whether I use the connector first or second, doesn't really matter: the changes in SQ for both are readily apparent.

 

What I think that means is that the USB Connector is tackling the noise from the computer, now impeded towards the DAC (potentially in a couple or more ways), but doesn't perform the same function as the AC Filter box mains side.

 

Therefore, I tend towards the following hypothesis (I am not considering isolation transformers here in this example, but they can be very useful as well):

 

- The custom USB Connector, by its special configuration, not only is performing some ground isolation but is also reducing Leakage Current (of the many-device interaction type).

 

- In turn, the AC Filter Box, is reducing the noise re-injected by various SMPSes by virtue of their internal operation into the mains.

 

Conclusion: if you're not all on battery or all on Linear PSUs, you need both the interconnect work and the AC Filtering work.