Noise free 12v psu for Headphone amps/DACS

[Deleted]

Jan 31, 2011 0:23:30 GMT 405line said:

Hello, go on then enlighten me and everyone else, explain the CM and ground loop problems, don't just snipe explain the mechanism to me please, I am all ears mate, I don't pretend to know it all and if something is to be learned then I will learn it, if it is unsafe etc please advise.
Cheers

Frans is a highly experienced and busy Electronics Engineer. I feel sure that he will expand on this subject if he is able to find the time.

[405line]

Thanks.

[Deleted]

In essence you talk of differential mode 'noise' signals not being present with the use of a battery.
There won't be much of that except in the higher frequencies (top part of audioband and everything above).
This is due to wiring length and it's inductance.

As soon as you connect a charger to the battery there are CM signals on it.
IF and HOW they (might) influence the sound is completely dependent on WHAT is connected and HOW.

Stating a battery is a huge cap. is simply very... basic
The equivalent circuit you would have to mike to spice it will be rather complex and have many more parts then just a big capacitor.

Writing a tutorial about HOW CM enters the audiodomain and how screening and proper conditioning and filtering affects it AND explain it to those with basic knowledge and those with NO knowledge but like to have it explained to would take me several days to write up an advanced electronics course with lots of explanatory drawings and calculations.
Also I would have to give a crash course in proper decoupling and explain why it is important AND have to do some PCB making tutorials also as this is a major part of it.

You could ofcoarse follow the many courses I had professionally and be as dedicated as me in electronics in general with a specialism in analog and audio in particular and would have to have many years of field experience in troubleshooting EMC problems and PCB / schematic designing to fully grasp everything involved in this matter.

You will have to understand that explaining separate aspects won't 'enlighten' everyone as one would have to grasp the total concept of all these aspects that are 'connected' and together have a certain effect which may or may not be audible because of the conditions that have to be recognised for this.

I have opened a few curtains in the grotto.

Search for 'leakage current' and 'common mode' and you will find things that for those with at least basic electronics knowledge should be able to connect some of the dots but won't make you an expert in EMC field so to speak. Also many years of (learning) how to measure properly and spending time with measuring equipment and devicing many tests and procedures and even making test equipment and know HOW to test is called experience which cannot be tought in a 1 page crash course.

When using the search button notic the 'days' box below and set it to a year or so instead of 7 days.

I might write an article about it but do not know when and what single aspect I will explain.

Afterall a GP will not become a specialised heartsurgeon by reading a few books and looking over the shoulder of said surgeon.

[jkeny]

Jan 30, 2011 19:16:03 GMT @solderdude said:

Have a look at this study - it reports impedance dropping at higher freqs

Wow 100Hz that's a real high freq.
I was more talking about 1kHz and up (to compare it to 'normal' regulators and 10kHz and up for the better types. and 100kHz and up for comparison to super regs or add ons.
These values are of importance the DC to 100Hz region is more important for powering motors and actuators in high drain applications, yes also in audio.

Time issues prevent me from investigating it myself in the lab, but may do so in the future to shed some light if you are interested.

If you have already made up your mind I won't bother.
The high-face is probably locally HF decoupled in many critical places anyway and a stable PS like Alex suggested or the battery will NOT make things worse as it was already designed to be fed from extremely crappy PS (USB port)

Ahem, I might be reading this wrong & looking at the graphs incorrectly but I think it's you who is wrong here.

I quote from the paper I referenced on page 12:

The frequency range shall be large enough to encompass the anticipated
network response, and in this case corresponds to a range of 100 kHz to 10-4 Hz.
At least six different frequencies per decade shall be measured.

Now have a look at the AC spectral analysis graphs again but with this quote from the paper in mind

The expanded scale high frequency response is also shown in Figure 7. The high frequency impedance of the cell is essentially the same before and after PSOC pulsed cycling. The small decrease in impedance from 4.01 mΩ to 3.74 mΩ

I hardly think that the writers of the paper would say that 100Hz is a high frequency, do you?

I take exception to your comment about me having made up my mind already!

I made no reference to the Hiface & wonder why you dragged it into this discussion - I was simply offering a suggestion based on my research & experience. But seeing as you did, please show me your evidence for these statements about the Hiface or are they pure guess woork?

[rowuk]

I guess that this will always be a question of what we are trying to accomplish. A battery has the beauty of avoiding ground loops and certain types of noise, but other than that will not "sound" better than a properly designed power supply with a bit of common sense.

If you are looking to clean things up, a proper uninteruptible power supply with mains regeneration (like the APS RT series) does a great job without breaking the bank. It in fact does isolate the mains and generates a clean 50 or 60Hz which is a big advantage for audio. It provides a common earth for everything connected too.

[Deleted]

@ JKeny,

What can I say.

Sorry you take offence but I cannot reach another conclusion reading your pioneering articles and your snap at me thinking I was snapping at you, which I wasn't.
I like your work and seems well thought out eventhough I CHOOSE to ignore the all important subjective parts of it as subjective has not been proven to exist, at least to me, in lab conditions and extensive listening tests by me and many other objectivists.

Disclaimer:
MY choice to choose so and mine only...
I know... I am 'unopen' and rigid EE hardliner with no decent hearing nor 'real' practical experience... to some.. but so be it.

If you read my post you would have noticed I sortof OFFERED to do some research for you regarding HF aspects of cells. (about which I cannot find any real data and thought you would like to know/add to your reserach)
I know you really like the battery with the hiface and simply offered to pitch in to do some research but I withdraw my offer as you are obviously not interested.

I also mentioned that I felt using a (unconnected to charger circuitery) battery will never harm but only improve things.

Sorry you saw it as an attack... it wasn't in fact if you read carefully it was the opposite.
Please read the mentioned article and my reply again carefully with the comments below in mind.

Perhaps reading a technical paper isn't as easy as it appears.

They talk about an 'anticipated network respons'.
THis doesn't mean they use it in a network !
What it means is in what application it will mostly be used.
In the case of these cells.. traction not feeding network gear !

In the article they said they shall measure from 100kHz to 10-4 Hz
All graphs speak of 100Hz to 10-4 Hz. the k is a 'goof' IMHO

Please check your calculus again and notice that 10-4Hz = 1000 seconds (16 mins)
So tests were conducted between 100Hz and 16 min intervals
So maximum 100Hz which is VERY high in the case of this 'anticipated network respons' and the other end of their 'spectrum' of 16 minutes is quite long for such a load.
all graphs are in HOURS not microseconds, giving 'hints' in this direction too.

So I look at a battery that during a 5A load drops 0.2V (5%)
at 6A load it drops 0.4V (11%)
At least that's what the graphs appear to tell me which they are also not because the righthand scale should read A instead of Ah which is confusing because timeaxle already has the hours.
Perhaps another 'goof' or I am simply missing something (not unlikely by the way)

That's EXCELLENT regulation b.t.w. ;D
I know even the cheapest regulated powersupply that can handle 10A for instance doesn't drop even a few mV opposite the 300mV of the battery.

This does NOT relate in anyway to the calculated internal resistances put puts them in the 40 mOhm to 66 mOhm range with a dynamic resistance of 26 mOhm around a 5.5A current.

Very well possible that the impedance falls to the displayed value when only a little current is drawn since the voltage drop of 1.5V at 10A suggests an increasing internal resistance at higher currents.
Fortunately in the hiface case this is the case as I expect the battery to last very long in this appl.

Everything above is not meant to offend you b.t.w....

The fact that their conclusion is the resistance at high freq (100 Hz in this case) did not change much BEFORE and AFTER the test says ABSOLUTELY totatally NOTHING about HOW it performs DURING the test nor is it mentioned in other things then graphs DURING the test.
It merely says the cell wasn't damaged or degraded during the rather high current tests.
Which is important for the writrers of the essay as they are not independent researchers but are the sellers or at least are employed by them so it seems.

[Deleted]

If you have already made up your mind I won't bother. The high-face is probably locally HF decoupled in many critical places anyway and a stable PS like Alex suggested or the battery will NOT make things worse as it was already designed to be fed from extremely crappy PS (USB port)

Frans
I am not surprised that John felt he was under attack. I took exception to the above as well.I would have joined in, but it's so damn hot here presently. (still 40C) that I haven't been spending as much time in front of the PC. What you are saying flies in the face of reports from myself, Will, Syd, John and numerous customers of John regarding crappy USB power supplies.You are saying that things like that will not make matters worse, whereas in the real world internal PSUs on many (most?) commercial PC related products are far from as well designed as you seem to be implying.
By implication, are you also saying that my JLH Linear +5V PSU for the Corsair Voyager is also a waste of time ?
Tell that to Will, and Erin (from DIYAudio),as well as some other Sydney based RG members, and see what their reactions are .
I apologise, if that is not what you meant to say in the quote above .

Alex

[Deleted]

I see my choice of words seems to cause confusion.
Must be a 'technical jargon' thing or me not being native English speaking causing confusion.

Feeding things from separate batteries or your PS suggestion is preferred over the crappy USB supply and the end result will NOT be worse when you do.

My point is the hiface (and memory devices, DAC's, converters and whatnot) are DESIGNED to do their work even on crappy power supplies like USB is, designers KNOW it's limitations in general) so already should have applied a fair amount of local decoupling inherently (or should have at least) I know some designs don't and are poorly made/though out.
Feeding it from a better power source (be it is battery or other PS) will always be better NEVER worse is what I implied.

The subject was about clean powersupply and my statement is simple:
Battery power is clean as long as you do NOT connect it to a charger of any sort at the same time it is in the audiochain as a groundloop is created regardless if one is able to recognise it or understands the mechanism of it or not.
This is even, as I explained, already difficult for professional trouble shooters in some cases let alone for noobs or those with basic knowledge.

In short feeding from batteries (stand alone) is beneficial by nature.

Feeding from batteries that are connected to charging circuits that is connected to mains defeats the purpose of the battery as a groundloop breaker and good regulators and the likes of JLH add-ons etc are provable BETTER then a battery as 'regulator' in this circumstance IMHO.

I hope that makes things clearer and hope John understands my comments above without being offended again.

The subjective differences perceived is another matter entirely and not the scope of my replies.
Clean power and HOW it is achieved is technical and can be explained/proven/measured/made visible.
Pure subjective differences seem only backed up by theories, ears, concurrance and have never been able to be explained/proven/measured/made visible.

The border where subjective and objective meets is fuzzy at least and causes confusion if certain mechanisms/aspects are partly understood and leads to 'funny and beside the truth' theories in some cases mixed with parts of truths in it.

I hope it is clear that this sometimes amuses me or pisses me of incase yet another 'theory' is launched based on partly right and wrongly interpreted input.

The subjective remarks have NOTHING to do with this subject or it's validity but are general observations from my (technical) point of view.

I hope the air has cleared and all noses point in the same directions as I only engaged in conversation because I saw part truth and wrong assumptions in one statement and thought I'd clarify some, not to bash people or make a mockery of certain convictions.

It was NOT my intention to bash, make people look foolish just to shed some light and do some research for free, which I won't do anymore unless it is beneficial to myself in this particular case.

[Spirit]

Feb 1, 2011 13:26:46 GMT @solderdude said:

I see my choice of words seems to cause confusion.
Must be a 'technical jargon' thing or me not being native English speaking causing confusion.

FWIW Frans, I understood perfectly what you meant in the bit that Alex quoted.

It seems to me that people have decided you have a certain point of view, and then manage to misinterpret what you're saying based on what they think you should be saying.
(Rather than what you actually wrote )

[Deleted]

YUP, that is the essence...

My 'disbelief' in subjectivity is causing most of this 'confusion' as some feel ridiculed because I have different boundaries then others.

I try my best to ONLY react to factual things and stay away from the subjectivity but these 'fields' have overlap.
For some more in the subjective realm and for me more in the objective realm as in general are able to look further in the technical part as most non-technical or a bit technical members/visitors.

I understand that sometimes it looks like I discredit part of their work/ beliefs as it is felt not to coincide with their points of view in certain matters.

complicated... but never intended to 'dis' people but meant explanatory as some do like things explained if possible and get some of audio's mysteries into the open.

Common mode garbage and power supply impedance being very mysterious to most people and often misunderstood.

As I sometimes say... I don't belong here as most audiophiles are only interested in the subjective realm with theories that accompany it.
But it amuses me most of the time and feel I can contribute to make audio more 'down to earth' and less magic which makes me stay in the Grotto (exclusively).

[jkeny]

Hold on, hold on - this has nothing to do with subjectivity Vs objectivity

Firstly, I objected to you dragging the Hiface into this thread when there was no need for it & I still don't understand your reason for doing so. I suggest to the O/P to try LiFePO4 batteries - simple.

Secondly you, like a lot of other e'ees, seem to make the assumption that equipment is designed adequately & any modifications are the whim/belief of the modder based on purely subjective results. Did you ever ask me for any evidence of before/after PS scope shots?

What I also objected to in your original statement was your carefully worded "a stable PS like Alex suggested or the battery will NOT make things worse as it was already designed to be fed from extremely crappy PS (USB port)" This never mentions the fact that it may improve matters just that it won't make matters worse. I accept that what you intended is now stated differently but as a native English speaker I'm sure you can understand the implied slight in this comment. I see you have amended this to "Feeding it from a better power source (be it is battery or other PS) will always be better NEVER worse is what I implied."

Thirdly, you are making general statements about a device, Hiface, without having looked at the PS de-coupling or indeed any aspect of it's operations. Again, I don't understand where you are coming from on this?

Finally, I'm not sure I fully understand the Nyquist graph in the linked paper, do you & could you explain it, please? It plots impedance Vs imaginary impedance. I have some other papers I can link you to, if you are interested, showing other Nyquist plots which might throw more light on the high frequency impedance behaviour of this technology - search for EIS LiFePO4.

Edit: I see you have substantially edited your above post since I wrote my reply !

Here's an EIS plot of impednace - note that the freqs are given (up to 2mHz) but may be a bit confusing - they are to me. Ignoring this the impedance never gets above 0.08ohm

[Deleted]

> Hold on, hold on - this has nothing to do with subjectivity Vs objectivity <

IMO it does to an extend as already explained above.

> Firstly, I objected to you dragging the Hiface into this thread when there was no need for it & I still don't understand your reason for doing so. I suggest to the O/P to try LiFePO4 batteries – simple. <

Google JKeny and LiFePO4 …
Hiface mods with LiFePO4 sites.google.com/site/hifacemods/
MOT: Modified Hiface
The 3V Lithium battery man.
Also I found it a GOOD example of battery use to bypass crappy USB (SMPS) fed devices.
See it as a positive remark, kuddo's...
Not all things are meant in a negative way but it appears you do see it as such, perhaps you have encountered some skeptics along the way ?

> Secondly you, like a lot of other EE's, seem to make the assumption that equipment is designed adequately & any modifications are the whim/belief of the modder based on purely subjective results.

It's your assumption that I seem to make such assumptions BUT is NOT my opinion although you seem to assume it is.
In some cases your assumption that I assume the assumed is correct though but wouldn't say this about a lot or most mods, it certainly would undermine validity of my own modifications if I did.
I assume most designers do their best but some fail or are simply tied by lack of space or budget or knowledge.

> Did you ever ask me for any evidence of before/after PS scope shots? <

Do I need to ?
Why do you assume I have not seen it and have not read your articles and need more.. errrm .. proof
It's all right there on the web..?

> What I also objected to in your original statement was your carefully worded bit "a stable PS like Alex suggested or the battery will NOT make things worse as it was already designed to be fed from extremely crappy PS (USB port)" <

Please read my previous posts, the answers are already there and don't need to elaborate yet again..

> Thirdly, you are making general statements about a device, Hiface, without having looked at the PS de-coupling or indeed any aspect of it's operations. Again, I don't understand where you are coming from on this?

I can make pretty 'educated' guesses from looking at boardlayouts and used parts if local decoupling is present (picture search on google)
I don't have to see it in real life or measure it to make those statements.
Call it 'occupational designers experience' if you will.

> Finally, I'm not sure I fully understand the Nyquist graph in the linked paper, do you & could you explain it, please? It plots impedance Vs imaginary impedance. I have a paper from Stanford I can link you to, if you are interested, showing another Nyquist plot which might throw more light on the high frequency impedance behaviour of this technology.

The graphs lack info on which freq. which impedance was measured (between 100Hz and several mins) so is unrevealing and says little without the extra information that they did not print.

The impedance above 100Hz is not of much importance for traction batteries so there is likely to be no/little info on impedance behaviour within the audioband and above to be found.
Chopper HF effects (spikes from speed controls and collector sparks) are heavily filtered in the electronics so is of no concern and thus will not be measured.
Graph 3 already shows a factor 10 increase in dynamic resistance at a few Hz already.
Physics dictates this will not become lower with higher frequencies.
It has to be tested and a testdevice will have to be constructed to find out HOW it performs.
It is quite safe to say the impedance will be higher as even the simplest regulator for audio frequencies.
Let alone the peaks generated in the RF band in digital circuitery where the battery will have a very high resistance (physics dictates, not me).
This has to be taken into account when designing power-management electronics.
For this reason I mentioned the internal resistance values say absolutely nothing about their behaviour in the audioband let alone in the Mhz region.
It doesn't have to either it is not it's function that role is for other parts.

These cells in general only need to power motors, actuators and regulators for said devices.
Small Lithium cells in portable electronics are NOT used for their low impedance but for size, convenience, weight, selfdischarge rate, shelflife and life predictability.
LOW impedance comes from the regulators and local decoupling so it is NOT the function of the battery itself ... supplying power is.
I would prefer it over plain lead-acid.
That is why Robin and me commented on that part which I assume you assume is very low and matters, which it doesn't.

When not connected to a charger the power is 'clean' and the HF impedance in a connected circuit will not come from the battery itself but the present decoupling caps (10-100nF for RF, 10uF-100uF for lower freq.)
For DC and the powerstages of HP amps in the bass dept the battery, or regulator are responsible.

I hope we can put this at rest because it appears to me I am repeating powersupply impedance lessons over and over again.
I firmly believe all aspects have been covered.

[jkeny]

Feb 1, 2011 19:43:54 GMT @solderdude said:

..............
It's all right there on the web..?

....

No, it's not - a wrong assumption again


I appreciate your answer but how do you square your impedance guess with the graph I posted above in my last edit - no impedance greater than 0.08 ohms - I'll give you the article link if you are interested?

BTW, I don't use regulators with these batteries specifically because of their low impedance & low noise characteristics

[Deleted]

BTW, I don't use regulators with these batteries specifically because of their low impedance

The above refers to the MK1 modified Hiface. Since then a MK2 has been introduced which replaces the on-board 3.3V & 1.8V regulators with ultra low-noise regulators

From your website by any chance ? ;D

You are free to do and think whatever you want John.
by the way 2mHz = 500 seconds, not 2MHz.

Only a graph is shown to 65kHz NOT the all important MHz region for digital circuitery, but that's where regulators do nothing only the smaller decoupling caps.

I'll leave it at that.
have a nice evening.

[jkeny]

Feb 1, 2011 20:26:28 GMT @solderdude said:

BTW, I don't use regulators with these batteries specifically because of their low impedance

The above refers to the MK1 modified Hiface. Since then a MK2 has been introduced which replaces the on-board 3.3V & 1.8V regulators with ultra low-noise regulators

From your website by any chance ? ;D

You are free to do and think whatever you want John.

Sorry but wrong - the regs are not run from the batteries. Why would I try to run a 3.3V reg from a 3.3V battery

What about the graph shown - any comment on impedance of LiFePo4 battery at 2MHz & below?

[Deleted]

> which replaces the on-board 3.3V & 1.8V regulators with ultra low-noise regulators <

Your text... not mine !

> What about the graph shown - any comment on impedance of LiFePo4 battery at 2MHz & below? <

See my reply above.

[jkeny]

Feb 1, 2011 20:35:04 GMT @solderdude said:

> which replaces the on-board 3.3V & 1.8V regulators with ultra low-noise regulators <

Your text... not mine !

Yes but batteries don't power these regs, a wrong assumption I'm afraid. Can we leave the Hiface out of this - it has nothing to do with the batteries particularly

No comment on the impedance shown in the graph?

[Deleted]

Not playing anymore John,

READ my comment on the graph reply #38.

Have a nice evening,

[jkeny]

OK, Have a nice evening. Your comments were not on the graph I posted above which show the impedance at up to 2Mhz & is at variance with your stated estimation of 44 to 60 mOhm. The graph shows a max of 8mOhm at 2MHz for these batteries.

[Deleted]

John,

Get a life and learn to read the graph YOU posted yourself..
2mHz is NOT 2MHz but 500 seconds.
and my comments WERE on the graph you mentioned.

Still the audioband is well behaved (upto 65kHz) and is certainly better then a lead-acid above 65kHz there is NO info.
This 123 type battery is much smaller then the bigger types and because of this has totally different properties.
Apples and pears.
But is irrelevant anyway as the decoupling caps take care of it.

and... have a nice evening... again.

[Deleted]

Feb 1, 2011 13:54:25 GMT Spirit said:

Feb 1, 2011 13:26:46 GMT @solderdude said:

I see my choice of words seems to cause confusion.
Must be a 'technical jargon' thing or me not being native English speaking causing confusion.

FWIW Frans, I understood perfectly what you meant in the bit that Alex quoted.

It seems to me that people have decided you have a certain point of view, and then manage to misinterpret what you're saying based on what they think you should be saying.
(Rather than what you actually wrote )

Phil
The statement that it will not make things worse,combined with the statement that it was already designed to be fed from extremely crappy PS (USB port) implies that there is no further advantage to be gained over the existing supply when using either batteries or a linear PSU, as the internal circuitry is already more than adequate.
IIRC, you have experience of listening to the HiFace in a modified form,. and should be in a position to either refute or confirm
John's claims based on your own SUBJECTIVE impressions.
Alex

[jkeny]

OK, sorry, mis-read. Ah yes I see you edited post 38 & now include some comment on the posted graph after I posted. Yes, let's leave this now!

[clausdk]

I protest
I protest about the words objektive and subjektive, objektivity is an illusion, it is just not possible to be objektive, no matter how hard you try, you will ALWAYS be infuenced by what has come before and also what you know and expect, you simply can not throw that away..

It is part of human nature and is not something that can be put aside..

It is that simple..

You guys can disagree and discuss all you like, till the end of time if so be, but none of you are objektive and none of you will ever be..
Unless you throw out your brain and replace it with a computer and even then you will be influenced by the guy who installs the thing, but if he is deaf you got a good shot., at least in audio


Sorry but it is a fact of life..

[leo]

I don't give a toss what anybody thinks anymore, I post an opinion, people either like it or don't
Still good to discuss any technical issues though

[rowuk]

How about just using the decoupling caps that Frans recommends regardless of the battery technology....... I think that is a real life solution for lower impedance, noise and higher SQ.

I fly battery powered model airplanes and through the year have tried all sorts of types. When we fly competitions, they measure the impedance of the battery to even the playing fiels. The lower impedance batteries (albeit without the decoupling caps) really sounded better to me. I tried motorcycle batteries and never was able to get the sould that the higher performance types offered. The newest generation of batteries really do seem to have an audible edge when using my Focusrite 8 channel AD/DA mic->USB converter. The same applies to the NAudio fast track pro.