EMI Shielding... sure... got anything for the drool that spills out when people first see that setup?? haha
Insane man, very, very jealous. Any mods done to the PC's power supply? I've seen a lot of debate on running the PC off a linear supply, since MBs have so many DC-DC converters anyways... I'm very curious to see if better shielding inside the PC is beneficial at all, especially when using an external DAC. Well... either way I guess.
EMI Shielding... sure... got anything for the drool that spills out when people first see that setup?? haha
Insane man, very, very jealous. Any mods done to the PC's power supply? I've seen a lot of debate on running the PC off a linear supply, since MBs have so many DC-DC converters anyways... I'm very curious to see if better shielding inside the PC is beneficial at all, especially when using an external DAC. Well... either way I guess.
theres some serious moneys worth of internal filters. Custom cpu/atx and molex filters.
they respond further to clean power but also do a fine job of cleaning up any power you may choose to employ.
the trouble at this point in time with whats available. A linear psu needs the pico to convert 12v to 5 & 3.3v
the pico is a big noise introducer so indeed how effective is clean power in the first instance. post pico its messy again.
I think maybe the effectiveness of decent linear is a question of how the linear design handles the noise its collecting on route.
im currently using a standard brick switch mode and its signature was clear to identify.
once I shielded up the custom filters and cables to lower the noise allowed to transmit and re-enter the power as it feeds the system internal. the results have become audibly more difficult to identify the switch modes signature.
and certain aspects in high end extension and sense of image placement have improved nicely. in particular after I added the polymer resin ferrite compound to all the Oscillators and Ic`s mounted on the motherboard.
I also placed some on the lid of the pc enclosure as this is a suggested use to break up the reflected energy that is still being emitted from the electronics.
all seems to suggest if you suppress the antenna effect and emi radiation then handle putting that energy back into the chassis ground a pc can then do a pretty fine job of processing audio and video with reduced noise and high energy being carried along and emitted to cables and signal ground.
it probably also makes it a more friendly beast to situate anywhere near your hi-fi (dnm are totally unshielded so testament to Mr Morecroft for even realising a design in that context)
another argument for linear is a switch mode creates noise, collects it and dumps it back into your ring main power for your audio components to have to clean up for there operation.
a cheap solution to eliminate that maybe to clean the power in and ground of the switching psu.
too many audio buffs debate this cable or that device. they all fail to look at the fact your end result is how your local sub station performs >>> to your ears and back again in grounding terms. in short.
e.g. buy one low field power cable. will you hear any subjective improvement on your hifi? not so much. but repeat the process until all your power cables have a low field of energy emitted and youve employed the same technique to a point where it effectively has massive results.
so I look at the issues and do all I can to rectify any potential problematic wiring and emitting electronics.
accumulative terms I could build two of these machines one with the mods and one without and given a decent listening window anybody can pick out the improvements.
I get some foot traffic in my home brew lab. my system was good enough 15 years back to make you sit up and listen. now they throw on there favourite material and enjoy often with me just off poking cats or cleaning pots elsewhere. they dont know or care what I just did a day or two back or even if I may have. they dont live with the system as I do.
but they always have an opinion. latest remarks from the blind panel is its sounding very analogue. (I think good because that suggests a larger scale improvement from these mods than - hair splitting terratory )
and the virtue of measuring everything scientific can often tell you very little about how that influences sound.
hearing detailing and tightening of possitional information, growing sense of depth (the illusive black background) and detailing of event trails. are much bigger clues something went right.
its a warts and all system that can show weakness in recordings and highlight anything less than ideal. and working with it seems to turn these warts into pimples if its heading in the right direction.
there aint no substitute for passion well other than an engineering background in digital design mayhap lol
Mr Morecroft had no engineering background and his out the box thinking gained him a knighthood after he went on to re-design the capacitor twice in his own lifetime. network T caps and slit foil.
Lovely down to earth guy. I met him once or twice... a true inspiration.
take a finger break... my keyboards melting with enthusiasm
As a note, digital signals are far more noise immune than analog are. Your precious digital signals are safe, if it isn't dropping packets you will notice absolutely no difference (because there will be absolutely no difference) with additional shielding.
at 192KHz with ripple in the carrier. suppose it is resolved by the receiving hardware which then has to work harder to correct the information and deal with bleeding off the high energy noise that comes with the bits.
at this level I feel rubbish in = rubbish out.
shielding may have some effect in lowering the unwanted energy with the signal and decent power should help define the square wave giving a much more accurate bit state to resolve.
pc`s are already crippled by timing issues. because there not built with exclusivity to deal with audio processing as critical.
anyone using a pc in a high end setup goes out there way to stop cpu cycles being taken away from the task.
I think theres much more to be understood about the situation than the fact digital is quite good at rejecting noise in general.
its the design question of current. companies making dacs using various processors and banks of in some instances to lower that noise floor and raise ability to process bits with ease.
the antenna effect of any cables from a motherboard is a known issue.
at its most basic bits intact or not the unwanted energy is still there to be dealt with.
if theres no benefit in a strict sense then why would they produce riser ribbon cables in two variants. shielded and unshielded.
and here is part of a digital design paper that in the top ten listing says shield and ground all high energy cables in a fully shielded config with drain wire.
also have some military and lab specification requirements I can point you to bob but not on here lol
hope lifes all well with you too bob. nice to see you again
also fair to speculate antenna energy from internal cabling unshielded will bombard the casework and charge the chassis in effect upping the caseworks field.
why did DNM use non ferrous casework? to eliminate eddy currents.
I will rest once I hear a digital system equal that of reference vinyl or a studio tape.
we fall short so theres more to be explained in this realm.
until we have zero latency and realtime processing of the audio file we will never get from a - b in a recording without trade offs in fidelity
edit: one of the fundamental problems with motherboards is they employ cost effective methods of grounding back into the chassis at various stages. its less than ideal to make an audio component from because given the brief from scratch they would not skimp on that element.
so asus & gigabyte wheres our true audio quality boards at?
If you edit your posts rather than double/triple posting, it makes things easier.
The first link confirms what I was saying, rather nicely. Thanks for that.
I've never run into a "timing skew" issue with regards to digital. Digital is rather nice in that either it works, and the data is correct, or it doesn't and the data is complete garbage. There really isn't much middle ground.
If the timing became skewed, your result wouldn't be noise, it would be utter garbage.
What signalling spec is it that you're using that has an analog carrier for a digital signal? That would be very rare. Active termination is common, but that's a very difference beast.
The square wave with noise in that first link is demonstrating the immunity to said noise that is granted with a digital signal and a schmidt trigger on the receiver. The receiver doesn't have to correct the rounded edges, because it never sees them.
If you'd like to be pedantic about things, there is no such thing as a square wave. They're all rounded. Shielding the wires makes them more rounded, as it increases the capacitance of the wire in question.
Moving along, shielded ribbon cables exist because at high frequencies (high frequencies, which anything measured in kHz is not. We're talking >40MHz here) there is no such thing as even an approximate square wave, and the sine wave is fragile enough that noise can be an issue over a long run. There are other ways to solve this of course, Intel came up with one that they used for FSB in order to close in on (and surpass) a 1GHz equivalent frequency. That, however, is not relevant here.
Your paper there is talking about reducing the EMI caused by digital designs. Not received by. Caused EMI has to be limited because the FCC says so. FCC testing is a major pain, even for unintentional radiator class B devices. If you want some real fun, read the FCC code on intentional radiators.
Note that it states that you should ground both ends of the shield. That is not something you generally do to protect a signal, it can to make the signal worse by introducing ground loop issues. The signal is better protected by shielding that is grounded only at one end.
Caused EMI however is reduced by grounding both ends.
Does the paper define "high energy"? That's rather important, I'd think.
I have no idea what you mean by the first sentence of your last post.
As for equaling vinyl, you're going to have to decide whether you want to equal/best the quality of reproduction, or whether you want to have the analog related distortion that you're used to and fond of. Much like the "warmer" sound of tube amps. Accuracy and sounding "right" are two very different things.
Now to be fair, I am not an audiophile, and I know little about such things.
What I do know about is digital signalling.
we have a confusing spread of subject matter going on now
ive shielded the power not any digital cable (there's none present to address)
with exception to the riser cable perhaps.
i`ll need some time to rediscover various articles on jitter to point to digital audio.
grounding and noise too.
its late in the uk bob. I will do my best to compose the information and post.
I need food and to find four kittens before its dark or they`ll be no rest tonight.
meanwhile I will try to take onboard all you say and ponder why I like these bits more than my laptops or other PC`s using the same output card. I have three identical cards and they all sound different feeding a supposedly immune dac.
you`d be hard pressed to find many audio types that think bits are bits. more to be understood.
Ahh, if you're shielding the digital bits and the power bits to keep EMI in so it doesn't wander off and infect the analog signals, that makes good sense.
not quite sure where you're going with this.. But if i understand right the goal is ripple/noise free signal.
And one of the problems being that the PSU is virtually unable to provide "clean" power at all required voltages.
Apparently you need both 12v & 5v and since the 5v is stepped down from the 12v that is one major "noise" source.
Now, i don't know what the power drain is, does this have to run 24/7? Or do you need a "clean" signal only at certain times a day?
Cuz, i was thinking that if you'ld use batteries -not the cheap throwaway ones, but the deep-cycle ones (eg Shorai Li-Fe)- of the required voltage/amps.
And then just re-charge when needed (which would introduce noise at THAT point, i know), would that no "clean-up" the noise at the source?
I might be taken a too simplistic approach here, i know
Batteries are low noise, right? That's why people want to use them in audio systems. The most common assumption is that because they are a pure DC source they must have no AC noise. Unfortunately we live in a world where nothing is pure, so let's look at why batteries are not so great a source of power as some may want you to believe.
How a battery produces power
Batteries produce power through a chemical reaction. Every type of battery will have a slightly different chemical mixture that is used to generate power.
Batteries have a voltage rating and a current/time rating called Amp-Hours. Amp hours basically say how many hours the battery will last if it's drawing one amp of current. Because many systems draw variable loads, a 10Ah battery will not necessarily last 10 hours. It may last 20 or it may last 2 hours depending on the system it is connected to.
As a battery drains, the amount of chemical compound available to produce electricity drains. That means there's an exponential decay on the amount of power a battery can produce. So a battery at 100% will have far more capability of producing fast transients than a battery at low charge.
Battery Noise Generation
Batteries have a substantial amount of high frequency ripple. This means there is a large AC component to their DC output voltage.
What causes this ripple?
The battery's chemical reaction will generate different types and frequencies of noise. The chemical reaction changes as the battery drains. That means the spectrum of noise changes as well. The change in load also impacts the way the battery is draining, which will alter the chemical reaction and thus the noise produced.
As a battery drains so does the output current and the output voltage. As these elements change the voltage will become less stable and the transient response slower.
So ultimately you are creating a non-constant source that has varying spectrums of noise, that is constantly changing, and produces the exact noise we want to get rid of.
Regulating and filtering battery power
The first thought is to regulate and filter the battery power. This can help the battery's performance considerably, but it does not compensate for the variations that occur as the battery drains. Batteries in particular do not have a pump, so as current demands stay the same and their ability to produce that current diminishes the regulators and filters become progressively less effective.
Here's the catch. Regulators are very sensitive to input impedance. They like a stable input or they must regulate both Line and Load variations. A battery's output impedance changes with drain.
Different batteries sound different
Every battery will have a different chemical composition and thus different chemical reaction. This translates to different levels of noise and different resulting sounds.
The noise generated is usually very high frequency noise up into the mhz or ghz ranges. This is not only more difficult to filter, but it's the exact noise that pollutes digital audio the most. The noise generated may be less than the noise filtered out compared to a SMPS, but it is still generating noise that is a challenge to compensate for.
Recharging the Battery
For one, recharging is a hassle. Many companies have circumvented this by having micrcontroller charging circuits to automate the process.
Many chargers only begin trickle charging when a battery is below a certain charge level. Introducing AC noise into the battery's path means the system has to filter out both AC noise and battery noise. You also loose the benefit of being off the grid. If the AC noise is somehow isolated from the battery's DC output using dual batteries with a switching circuit or other method then the charging will still alter the chemical reaction of the battery.
When a battery charges it creates the opposite chemical reaction to return the battery to it's initial chemical components. In this event you have opposite chemical reactions occurring at the same time.
GROUNDING
This could be one of the single biggest issues with battery power. By using a battery you are no longer connected to ground in the same way. That means that shields become a noise "cage" rather than pulling noise away from the system.
In addition, you end up having ground loops along cables and your system must find an alternative path for ground noise.
Enter the Linear Power Supply
Linear power supplies are going to be larger than batteries. But unlike batteries they can be built to a much higher bandwidth that functions consistently. A well-designed linear power supply can be considerably lower noise, faster, and higher bandwidth than any battery. Linear power supplies can filter the incoming AC noise, eliminate ripple to exceptionally high frequencies, and have much higher slew rate and current output. This means their sound will always be consistent and their potential much greater.
Comparing a battery to a SMPS or entry-level power supply will put a battery ahead as the clear winner. It will be much lower noise than a SMPS. But a well-crafted Linear power supply has several advantages and should be the only consideration when looking for the ultimate power solution.
-------------
has its advantages but still less than ideal.
i've lived with this for a couple of years. its a step up from switching. but in working practice you can hear a slight variation in performance as the battery loses charge and the regulation adjusts.
also in its particular application lower frequencies had a little false body and bloom to the sound and the highs were a little too delicate and lacking in authority.
its two batteries ones on the go while the other charges and automatic seamless switching operation.
nice unit but ultimately too slow and restrained a bandwidth.
in short it sounds like battery power when there should be no attention drawn to the fact
at its price point still good value for the improvement powering the final stage of the usb card
note. my hydra x+ is in the post (currently lost in a depot in Germany I think) this usb dac uses a 3.7v battery but can also take a clean 5V 2amp supply. so I look forward to comparing the two power options. and I can also remove the 5V from the usb termination as it can pull just the data. which may deliver less noise to the dac and benefit from data signal free from power cabling within the usb cable
I intend to grade the various techniques and results then set there performance perspective in relation to cost.
so far a well designed linear holds its mark but a full linear ATX would need to come on wheels, cost the earth and would be a fair overhead to power.
I`m buzzing... me that is. not the design in progress.
I cant disclose the letter as its confidential but a company that supplies for the aerospace industry said yes I have the right idea and also will help in design of a Faraday cage that will allow airflow to the pcie card to keep it within operational spec.
so the leading experts say yes!
so looking forward to working with these people
they state that it doesn't matter in the cards situation which direction I drain the shielded energy to ground.
they also expect the modifications will be expensive to implement but will benefit
after a weeks pondering I have to say that bobnova is probably quite right in terms of digital data in block transfer mode when digital is a voltage tolerance above of below the base line.
In final stages when the digital signal has to be constructed to PCM and a waveform has to be reconstructed then the latter information becomes relevant.
its such a complex broad subject this digital malarkey. (DSD streaming and 1bit makes me feel quite ill knowing I have a totally different set of pain lined up to understand and digest... humph! lol )
the application regarding my efforts is valid according to a few companies I have put it to.
but some of the lesser studied areas like EMC and eddy currents is something to be worked through between companies, myself and some additional outside experts to get to the bottom of.
so some steady and careful co working we will get through theories, experimentation and application methods.
its an ever forward project but chasing ideals makes it hard to see the finish line. lol
still immense fun, insight and excitement.
My passion to mod PC's for Audio and visual is the hobby of choice for me.
reflexes are too slow to be a gaming fanatic... computer wont let me win then im not playing
Frankenputer three years messing to get it audio worthy... its a bit delicate. wouldn't fancy moving it until I make that jumble of chaos a little safer
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
)
Okidoki, i'll shut up about things i obviously do not know enough about. Thanks for explaining 
