Will this supercapacitor bank work for my Ifi Zen Can?

[Nkam]

Except that is not the CEA-2006/490A standard...*

The original EIA from IHF A-201 ('66) standard, which became the EIA 490a/CEA 2006 is 1kHz 20mS at full power (0dB), 480mS at -20dB, (500mS total) repeated. So an 8.8% duty cycle.

The test you referenced and pasted is the EIAJ (Japan) burst test. It's also constantly repeated, not just a single 32mS period. It's basically a 25% duty cycle.

Amir already performs the CEA-2006 burst testing on amplifiers he tests.

* Copy and paste will bite you.
You'd think a manufacturer (Marani) quoting a 'standard', showing an AP's generator screen, and measuring their gear (supposedly) would read their instrument's user manual? I guess not.

Wrong:
View attachment 371754

Here's Audio Precision's own documentation (https://www.ap.com/blog/signal-gene...udio-analyzers-part-2-regulation-measurements) regarding the test.

View attachment 371755

And the APX-500 user's manual showing the panel:

View attachment 371761
View attachment 371760

reading further into this, I’m assuming this is done to replicate music more and the power needed for short bursts where more power is drawn for lower frequencies?

and supercapacitors can be used to provide that additional power quick enough when needed?



thanks

 

[solderdude]

There is nothing 'quick' about a low frequency burst (except the start and stop).

Look at a music signal up close of a bass or other low frequency signal. It does not start nor stop at a dime. It swings in, has a very, very short sustain (at max power) and then decays again.

It is an interesting test for speaker amps with non regulated 'linear' power supplies.

Headphone amps use regulated power supplies. Switchers are regulated (well... most of them are)

this is what he replied to me




External Power supply - high current, high peak current. Supercaps and linear from AliExpress at a grand total well below 100 Bux delivered is what I recommend.

Linear power supplies with dual chamber EI core transformers have the lowest coupling of noise from mains.

Linear power supplies with off the shelf toroidal or R-Core transformers without special construction are much worse. They are often worse than the iFi' s switching supplies for which I developed a special transformer design that allows to pass EMC testing without needing the usual "Y" capacitor.

Common "cooking" switching power supplies are worst of all with a 2200pF or greater capacitor coupling mains noise into the audio ground.

Yep totally correct.

Whether or not leakage or mains injected common mode noise becomes an audible problem depends on how the amp is wired internally and the grounding path as well as the audio path before the amp.
In most cases this is good enough. In some cases you can hear hum, noises, ticks, weird sounds or even radio stations. Only then you should look from where this comes.
The power supply of the headphone amp usually isn't the problem.

No.

It has an on board switching power supply operating at 1.2MHz (typical plugtop units run at 0.067MHz) that turns 5V into +/-12V and has it'd own regular build in.

The noise on the +/-12V depends only on this system, external noise on the 5V input is not really relevant.

I bolded the relevant part. The noise on the 5V is not relevant simply because the internal voltages are regenerated (by a switcher which also produces noise).
The 'no' from him is in regards of common mode noise. Differential noise (between + and - of the 5V) is 'smoothed' by an input cap and removed (and replaced by) noise from the internal switcher which feeds 2x 2000uF which can supply 'power' to peaks in music.
About the max capacitance the internal switcher can handle without additional inductors or other measures.

Momentary power needed for music is drawn from these caps. The 5A limit of the internal switcher determines how fast those caps are replenished.
As it is a MHz switcher this can be quite fast (so low ripple).

EMI is a "grab bag", there are different aspects.

Any power supply connected to mains will allow a small amount of mains noise (not just 50/60Hz but quite wide band) to "leak" into the circuit ground.

This noise (current) wants to flow to earth, somehow. In the process it can create as much as 1/2 mains voltage as noise Voltage between different pieces of equipment, depending on design, construction and system setup.

So it's system dependent.

All else being equal, less noise leakage from the mains is better. Note that this is not the same as "power supply noise" which is between +/- but it is in effect a separate noise between signal ground and earth.

Yep correct.

We "need" nothing. But if improved sound quality is desired and we want to also "hedge" against potential sound quality problems, inexpensive chinese linear power supplies offer currently the best choice at a price that compares favourably with most "audiophile" options.

Yep, that's why I recommended some.

Adding Supercapacitors can help in other areas.

Thor

Yep.

It could even help in this case... not for improved sound quality but for increased power output (with music, not with test signals) so you can keep using a 2.4A power supply but get 5A current peaks when needed.

A large capacitor bank could do something similar but most SMPS don't like to see a high capacitance and may not even start (goes in protection mode).
This is where the super cap helps. It'll also help with 5V devices that draw too much power from a USB bus.
What these cap banks do is charge the capacitors slowly (with a charging current lower than the USB output) and when momentary a current is demanded above the USB bus rating it draws it from the cap bank.
See them as a small battery on a slow charger.

I mean he is an engineer with a degree. He isn’t some fake.

so am I.

The difference is Thor still believes in 'magic' in audio (my perception of Thor, a guy I like b.t.w.) where as I see audio as just electronics which acts measurable and to laws of physics.
The 'magic' in audio happens in the brain. That is the unpredictable one along with acoustics and transducers.
The latter 2 of course also answer to laws of physics but also depend on local factors which are never the same in test conditions of a different copy or the DUT.

Whenever you get around to using the headphone amp in ways they were not designed to be used (any headphone other than HD650/HD6XX) which are lower impedance than 50ohm and are so insensitive/inefficient that they require a few Watt peak levels then you could look for a (not audio expensive) 5V/4A (to 10A max) SMPS to feed it.

Don't expect a 'better sound quality' though.

 

[Nkam]

There is nothing 'quick' about a low frequency burst (except the start and stop).

Look at a music signal up close of a bass or other low frequency signal. It does not start nor stop at a dime. It swings in, has a very, very short sustain (at max power) and then decays again.

It is an interesting test for speaker amps with non regulated 'linear' power supplies.

Headphone amps use regulated power supplies. Switchers are regulated (well... most of them are)

Yep totally correct.

Whether or not leakage or mains injected common mode noise becomes an audible problem depends on how the amp is wired internally and the grounding path as well as the audio path before the amp.
In most cases this is good enough. In some cases you can hear hum, noises, ticks, weird sounds or even radio stations. Only then you should look from where this comes.
The power supply of the headphone amp usually isn't the problem.



I bolded the relevant part. The noise on the 5V is not relevant simply because the internal voltages are regenerated (by a switcher which also produces noise)



Yep correct.


Yep, that's why I recommended some.


Yep.

It could even help in this case... not for improved sound quality but for increased power output (with music, not with test signals) so you can keep using a 2.4A power supply but get 5A current peaks when needed.

A large capacitor bank could do something similar but most SMPS don't like to see a high capacitance and may not even start (goes in protection mode).
This is where the super cap helps. It'll also help with 5V devices that draw too much power from a USB bus.
What these cap banks do is charge the capacitors slowly (with a charging current lower than the USB output) and when momentary a current is demanded above the USB bus rating it draws it from the cap bank.
See them as a small battery on a slow charger.

thanks so much man. You don’t know how much I appreciate the knowledge and help. Thank you!

I think the best ‘ upgrade’ ( god I hate that word cause it usually implies a snake oil salesman, lol) would be the supercap which I listed here. its well made and I can box it if I want to look half decent. It seems like it would solve any issue with current peaks when needed.
it looks like it’s pretty low ESR as well. Can you confirm Please?

from the measurements Amir provided I don’t see the use for even lower noise since the amp measures extremely well in that regard.

cheers.

below is the supercapacitor I’ll most likely buy for the zen can.
https://github.com/iancanada/Docume.../UcConditioner/UcConditioner5VUsersManual.pdf

 

[solderdude]

thanks so much man. You don’t know how much I appreciate the knowledge and help. Thank you!

I think the best ‘ upgrade’ ( god I hate that word cause it usually implies a snake oil salesman, lol) would be the supercap which I listed here. its well made and I can box it if I want to look half decent. It seems like it would solve any issue with current peaks when needed.
it looks like it’s pretty low ESR as well. Can you confirm Please?

from the measurements Amir provided I don’t see the use for even lower noise since the amp measures extremely well in that regard.

cheers.

below is the supercapacitor I’ll most likely buy for the zen can.
https://github.com/iancanada/Docume.../UcConditioner/UcConditioner5VUsersManual.pdf

The supercap will do nothing for this amp (also what Thor stated)
Those devices are handy for USB fed devices that momentarily draw more current than the USB port can deliver.
Say... a 'high power' DAC/amp that draws power from a single USB 2.0 port or phone/tablet/laptop.
It will do nothing for the 'sound quality' and won't filter common mode noise either.

The audio peaks are drawn from the internal reservoir caps (after the internal switcher). There is nothing that can be done externally to 'improve' that.

You should realize that the internal DC/DC converter is limited to 25W so feeding it with 1 Jigga Watt will not increase the 25W of the internal regulator.
Feeding it with a power supply less than 25W (the 2.4A/5W is just 12W) is the limiter here for continuous power requiring more than that.
With music this could only happen with very low impedance and low efficiency headphones (for which the amp is not designed) playing at loud levels and only with short peaks.

When you plan to misuse the amp (use it where it is not intended for, i.e. very low impedance and insensitive headphones (with internal EQ for HD650/6XX ONLY) then the best upgrade you can do is buy a 5V/6A switcher which will make the amp (with HD650 EQ !) go a tiny bit louder at very high listening levels.

When that is needed (using any other headphone than HD650/HD6XX you should buy another amp (could even be an ifi) that is purpose designed.


That said... (its how the human brain works) when you do buy any 'audio power supply upgrade', regardless of what it is, you may hear and increased 'sound quality' anyway.
Just know the signal coming from the amp is not 'improved'.. what improves is how one perceives the sound after the brain is 'informed' that 'an audible improvement' has been inserted into the audio chain.
This effect is what the whole audio industry (especially the snake-oil part) relies upon. That is the brands that use superlatives and tell (potential) customers they 'know' what is needed for super audio quality and other brands don't.

 

[wwenze]

reading further into this, I’m assuming this is done to replicate music more and the power needed for short bursts where more power is drawn for lower frequencies?

and supercapacitors can be used to provide that additional power quick enough when needed?



thanks

Higher frequencies actually... because nothing has lower frequency than DC. Music is AC, so it can work on "peak" power supplied by capacitors.

The idea is easier to understand if you look at power supply design and loading in the frequency domain. Suppose you have an ideal power supply which provides a low output impedance across all frequencies. To be specific, this power supply has a non-constant output impedance vs output power, because output impedance rises to infinity when current limit is reached. (In reality: The output impedance rises very high as you lose regulation). But let's try to limit output current so we don't hit the point where impedance goes high. Worth noting that a regulated power supply will have a lower output impedance than unregulated, reflected in tighter voltage regulation.

The ideal power supply's performance should be identical across all frequencies. In reality, the output impedance increases at higher frequencies, and also increases at lower frequencies

We need the power supply's output impedance to be low, because impedance results in voltage drop when there is load current.

Your load draws an AC current of a certain frequency range. Your power supply provides low impedance across all frequencies, which is expensive. Now, what if, there is a cheaper way to reduce power supply output impedance for just a certain frequency range, without the need to increase the capacity at DC? From the viewpoint of the load looking into the power supply, you just need to add a low impedance in parallel. And that low impedance is... yup, the power capacitor.

 

[Nkam]

Higher frequencies actually... because nothing has lower frequency than DC. Music is AC, so it can work on "peak" power supplied by capacitors.

The idea is easier to understand if you look at power supply design and loading in the frequency domain. Suppose you have an ideal power supply which provides a low output impedance across all frequencies. To be specific, this power supply has a non-constant output impedance vs output power, because output impedance rises to infinity when current limit is reached. (In reality: The output impedance rises very high as you lose regulation). But let's try to limit output current so we don't hit the point where impedance goes high. Worth noting that a regulated power supply will have a lower output impedance than unregulated, reflected in tighter voltage regulation.

The ideal power supply's performance should be identical across all frequencies. In reality, the output impedance increases at higher frequencies, and also increases at lower frequencies

We need the power supply's output impedance to be low, because impedance results in voltage drop when there is load current.

Your load draws an AC current of a certain frequency range. Your power supply provides low impedance across all frequencies, which is expensive. Now, what if, there is a cheaper way to reduce power supply output impedance for just a certain frequency range, without the need to increase the capacity at DC? From the viewpoint of the load looking into the power supply, you just need to add a low impedance in parallel. And that low impedance is... yup, the power capacitor.

im a bit confused.
so that supercapacitor will work in providing the extra power peaks if needed?

this is what Thoren said as well as he uses a supercapacitor in the same manner. But he built it himself.

This is what he said about it below ( in italics) (Im including @solderdude as well)


As is I used the iFi iPower 5V I had and added a truly massive Supercapacitor bank, 7.5F/5.5V/0.1Ohm X 20 adding 25 USD to the cost of the base power supply. This gives a 150F/0.005Ohm capacitor and prodigious peak current capability. It takes even longer to charge this bank from the 3A iFi iPower.

But once the super capacitor bank is full and plugged in, the difference is way bigger with my Fostex than on the DAC. I am not sure if the "supercapacitor charged" Zen or the Pro iCAN in solid state mode would win, on sonic memory it would be at least a fight.

Bottom line, Signature ZEN DAC and ZEN CAN with supercharged power supplies go a long way to bridge the gap in sound quality to the Pro Stack. Of course, the Pro Stack has many features and options the Supercharged ZEN Signature stack misses. They are nice to have but not essential for me.

 

[solderdude]

This is marketing talk for ifi customers and audiophools.
It is NOT relevant for the HD6xx amp also not when used with 250 ohm headphones.
Again, Thor is talking about Zen Can (the regular one, not the HD6xx version) and the pro-ican (which is a totally different design)

What a super cap will do is give a 2.4A power supply the chance to feed devices that require more than 2.4A without sagging the 5V.

The limit the amp can draw is 5A anyway so no need for 60A momentarily as the internal DCDC converter will limit the current to 5A anyway... in this particular amp.
So any power supply that can deliver 5A or more continuous is more than enough and do the same for a fraction of the cost.

In your case the amp will never draw more than 2A anyway, not even when at max power using the headphones you connect to it.
There is NO improved sound quality to be had.

Of course you should buy it as you believe it IS an upgrade and you will HEAR improved sound quality as a result.
You would be buying it for the wrong reason though. It is wasted money in YOUR case.

 

[wwenze]

it looks like it’s pretty low ESR as well.

Not special, not terrible, ESR is largely dependent on enclosure construction anyway. CHQ-2R7407R-TW has ESR of 1.2mohms at 1kHz. About there for the package size. The module uses 2 in series so we have to double the impedance, but we're still at the same order of magnitude.

Which brings us to talking about impedance. So basically if the impedance of the capacitor + ESR is too high, then it won't be doing the job of filtering well. The impedance of 2 farad @ 20Hz is 4mohm... A value so low that ESR is going to start bottlenecking. Heck, board trace will bottleneck. The capacitance value provided by that module in post #1 is 2 x 400F in series = 200F. It has enough capacitance to keep impedance low for minutes.

 

[Nkam]

This is marketing talk for ifi customers and audiophools.
It is NOT relevant for the HD6xx amp also not when used with 250 ohm headphones.
Again, Thor is talking about Zen Can (the regular one, not the HD6xx version) and the pro-ican (which is a totally different design)

What a super cap will do is give a 2.4A power supply the chance to feed devices that require more than 2.4A without sagging the 5V.

The limit the amp can draw is 5A anyway so no need for 60A momentarily as the internal DCDC converter will limit the current to 5A anyway... in this particular amp.
So any power supply that can deliver 5A or more continuous is more than enough and do the same for a fraction of the cost.

In your case the amp will never draw more than 2A anyway, not even when at max power using the headphones you connect to it.
There is NO improved sound quality to be had.

Of course you should buy it as you believe it IS an upgrade and you will HEAR improved sound quality as a result.
You would be buying it for the wrong reason though. It is wasted money in YOUR case.

he actually has the Sig 6xx not the regular one.

i have absolutely zero issues with telling anyone if I buy the supercapacitor and am wrong and it does nothing. I don’t have those hang ups. I’ve been known to be wrong before. We all are sometimes. You learn that hopefully at 55. It’s no shame.
I am learning a lot from you and others here, that’s for sure and I am grateful.

and this is what he says about the power draw.


Due to the use of a switched Mode stepup, the Zen Can draws a lot of current under "no signal" conditions.

As a result the Zen Can (my own one anyway) draws ~2A (10W) quiescent.

Use a 3A PSU and you have not a lot of current left to drive transducers. Add balanced mode which doubles current (and voltage) and which people insist on using despite the fact it sounds worse and things become ok for high impedance can's only.

The SMPS inside can handle minimum 4A and typically 5A from 5V.

The typical iPower is a 12W design, so it is a bit challenegd to power this.

 

[Nkam]

Not special, not terrible, ESR is largely dependent on enclosure construction anyway. CHQ-2R7407R-TW has ESR of 1.2mohms at 1kHz. About there for the package size. The module uses 2 in series so we have to double the impedance, but we're still at the same order of magnitude.

Which brings us to talking about impedance. So basically if the impedance of the capacitor + ESR is too high, then it won't be doing the job of filtering well. The impedance of 2 farad @ 20Hz is 4mohm... A value so low that ESR is going to start bottlenecking. Heck, board trace will bottleneck. The capacitance value provided by that module in post #1 is 2 x 400F in series = 200F. It has enough capacitance to keep impedance low for minutes.

sounds good man thank you.

@solderdude , dude, it’s only audio and it’s only $100. I get it, it won’t be an addition where I’ll say “ omg the soundstage opened and it sounds so much more relaxed with chocolaty mids !” Lol.
Come on we can have fun as well And learn at the same time.

I would get the supercapacitor for future proofing the amp so I can use lower impedance planars or such.

we aren’t curing cancer here. We are just talking about a measly headphone amp. It’s all good.

 

[wwenze]

Our point is that if you buy a proper 5A PSU, linear-regulated or even perhaps a cheaper well-built SMPS, you get to output 5A continuous

versus

If you buy that supercapacitor but stick to the stock 2.5A PSU, all you get is the same 2.5A continuous, with the option of getting 5A output for 1 minute or 1 hour (and thereafter the equipment distorts disastrously because there is no power control), with no audible difference at below 2.5A either.

 

[Nkam]

Our point is that if you buy a proper 5A PSU, linear-regulated or even perhaps a cheaper well-built SMPS, you get to output 5A continuous

versus

If you buy that supercapacitor but stick to the stock 2.5A PSU, all you get is the same 2.5A continuous, with the option of getting 5A output for 1 minute or 1 hour (and thereafter the equipment distorts disastrously because there is no power control), with no audible difference at below 2.5A either.

I actually tried two ipowerX in parallel and it didn’t sound good for some reason. Muddy bass.
but that might just be with it in parallel.

I get the point man. No worries. And thanks for educating me. i appreciate it very much

 

[solderdude]

he actually has the Sig 6xx not the regular one.

i have absolutely zero issues with telling anyone if I buy the supercapacitor and am wrong and it does nothing. I don’t have those hang ups. I’ve been known to be wrong before. We all are sometimes. You learn that hopefully at 55. It’s no shame.
I am learning a lot from you and others here, that’s for sure and I am grateful.

and this is what he says about the power draw.


Due to the use of a switched Mode stepup, the Zen Can draws a lot of current under "no signal" conditions.

As a result the Zen Can (my own one anyway) draws ~2A (10W) quiescent.

Use a 3A PSU and you have not a lot of current left to drive transducers. Add balanced mode which doubles current (and voltage) and which people insist on using despite the fact it sounds worse and things become ok for high impedance can's only.

The SMPS inside can handle minimum 4A and typically 5A from 5V.

The typical iPower is a 12W design, so it is a bit challenegd to power this.

The Zen Can is not the same as the Zen Can signature (which has a certain EQ the Zen Can hasn't).
Different use case.
Of course, the amp inside and power supply is the same.

Now... if you had the Zen Can and would be intending to use that amp to drive low impedance and low efficiency headphones to loud levels I would say just use a 6A power supply.
That would technically be preferable over the 2.4A power supply + expensive caps and much cheaper.

ONLY the ESR of the internal capacitors matter. NOT the ones of the device feeding the internal DC/DC converter which passes through wire, connectors and PCB traces and basically nullifies the low ESR. This is exactly why decoupling capacitors are used as close as possible to the power devices and is of no use in circuits preceeding the actual parts that deliver the power (output stage).

You must realize there is a complete SMPS between the external power supply and the actual audio circuit.

The fact that you hear differences is not because of technical reasons but the change of mindset.

When the idea that a super cap (because of the word super and current draw capabilities that can NEVER occur except when plugging in the power supply to the amp, including a harmful spark in the connector and possible damage to the input cap of the amp) will improve things for merely $ 100.- then it will. But not for technical reasons.

You should ask yourself the question why one would want to apply the EQ for HD650/HD6XX to all headphones anyway.
It is ONLY valid for those headphones.

 

[Nkam]

I did connect two ipowerX together in parallel using a Y cord as Thoren told me to. But it made the sound mushy for some reason.
so I sent the one back.
or the two together increased capacitance and why it sounded noticeably mushy?

so I probably don’t need extra power

 

[solderdude]

so I probably don’t need extra power

Indeed and when you are going to use it what it is not intended for you also would not need it and when you are lacking power in inefficient very low impedance headphones just buy yourself a 5V/6A SMPS.

 

[Nkam]

Indeed and when you are going to use it what it is not intended for you also would not need it and when you are lacking power in inefficient very low impedance headphones just buy yourself a 5V/6A SMPS.

wonder why it sounded muddy with two. ‍

 

[wwenze]

wonder why it sounded muddy with two. ‍

Because it's all in your mind.

But don't take our word for it. Ask the seller why it sounded muddy with two, and ask him for replacement since it isn't working. And watch his reply.

 

[MC_RME]

And his name is Thorsten.

 

[solderdude]

wonder why it sounded muddy with two. ‍

My suspicion is what @wwenze said.
The DT1990 and HD650 do not draw any substantial currents so the incoming DC to the internal DC/DC converter was the same so the output voltage will also be the same.
Still you can hear a difference, it just is not of technical origin.
I am sure you knew what was connected.

And yes, Thorsten (Thor for short)

 

[solderdude]

Thor's solution as well as Atomic Bob's recommendations will set you back a lot more than $ 100.-
And ... it would still have HD650 EQ only.

Marv is right:

You can't predict anything. Heck, people here don't even agree on the various LPS, noise-nuke, for the Pi2AES. For instance, I happen to prefer the Pi2AES / Mercury with the cheapo stock switcher. Many others prefer fancier LPS power supplies. Some like the noise nuke with the switcher.

He does not give you the reason why some prefer/like this and other people that.
The reason is simple.... because it is all in your/their minds.

So I will say it again.... When you believe Thoren and suspect that the 'supercap' will help with better sound it certainly WILL. After all there is the word 'super' in it and the promise of fast and high current delivery.
The audible improvement are just not for the reasons you think (better quality 5V, lower ESR, higher peak currents, less ripple).
In the end it does not matter what the reason is (placebo or real)... it will work for you and can almost guarantee it. Simply because you believe it will.

When you ask 10 different 'experts' you are gonna get 10 different opinions.