[Ghoostknight]

Hello

i try to build something similar to the Ifi Idefender

it should have
- 3 usb connectors, one input, one output, one external power
- disconnect V from input and just use external power
- if possible disconnect GND too from input and use GND from external power

now to where it gets problematic... how can we leave ground connected but disconnect it as soon as the usb device is sucessfully connected? (since if gnd is completely disconnected most devices wont connect sucessfully) So in the end just D+ and D- should be connected from input to output, V and GND will come from external power input

- maybe current sensing on the output and disconnect ground after 1 sec?
- Can i use a isolation transformer and just connect ground to the output (and leave V unconnected)? or will this not work? (if just connecting GND wont work, how about a large resistor between secondary V and GND and use the GND this way?)
- Is this even a good idea? even if we get GND disconnected too, will noise be transfered over the data lines? or do most receiver chips reject noise on the data lines?

i know normally we should use usb isolators, but im trying to find a cheap solution/workaround (specially in the case of high speed usb), for example to disconnect my attached SSD to the raspberry pi beside the data lines, since i think connecting V and GND also introduces noise back to the system

 

[dunring]

There were plans for making one on the site for the Khadas Tone Board DAC. It's USB powered, but it was what you're describing so it could be powered by an external source, but also carry the usual USB data to it. They put schematics up and everything so you could construct it yourself. If you search Google for "khadas tone board external power supply" it might save you a lot of time.

 

[Ghoostknight]

There were plans for making one on the site for the Khadas Tone Board DAC. It's USB powered, but it was what you're describing so it could be powered by an external source, but also carry the usual USB data to it. They put schematics up and everything so you could construct it yourself. If you search Google for "khadas tone board external power supply" it might save you a lot of time.

the only thing i found was this https://www.khadas.com/post/tone-1-separating-data-power-connections which is refering to the "normal" split cable, but i try to isolate the GND from the pc going to the DAC (those splitter cable just work if both GND`s are connected, from the external power supply and the PC)

i try to figure out if something like this picture is possible

 

[Thorsten Loesch]

Hi,

> i try to build something similar to the
> Ifi Idefender

I am not sure that is valid DIY projects.

Because of the ways something needs to be designed as commercial product, and specific design objectives for the "idefender" (like an instant and transparent "failover" to bus power if external power drops out) there are both design compromises and a lot of extra circuitry.

Further, the PCB design of the USB traces is nontrivial.

> if possible disconnect GND too from
> input and use GND from external power

Technically speaking this is not EXACTLY how the idefender works.

> now to where it gets problematic...
> how can we leave ground connected
> but disconnect it as soon as the usb
> device is sucessfully connected?
> (since if gnd is completely disconnected
> most devices wont connect sucessfully)

You use the kind of circuit that sits behind this switch:

There is a reason for it being a manual switch on the the original iUSB (Power).

Essentially being this switch is a earth/groundloop breaker, specifically a 100 Ohm resistor. This sits between input and output ground, the switch bridges it out.

If there is no DC power current flowing in the USB lines, the 100 ohm resistor is normally ok to let standard USB devices complete a successful handshake. So this ONLY works if supplying new power to the downstream USB device.

Noise problems with USB are almost 100% down to earth/groundloops and this breaks the part of the loop through the USB Cable is reduced by around 80dB.

The switch shorts out the 100 Ohm in case it is not needed or causes problems.

The "idefender" in effect automates this switch and selects external power if available.

> Is this even a good idea?

If you have problems caused by ground/earth loops involving the USB connection that cannot be solved in other ways (e.g. breaking the loops through mains cable's and interconnects), it is.

It is a sufficiently frequent problem in my experience, that I designed solution into several ifi USB products and it eventually got elevated as standalone product.

> i know normally we should use usb
> isolators

While they obviously work I feel that most are overkill and solving the underlying earth/groundloop's is free and works just as well.

If you consider to DIY something like the idefender, I think you'd have the necessary knowledge to troubleshoot the earth/groundloop's and solve the issue groundup.

All the ifi "fixer" products are really designed to be "plug'n'play" for people with very limited technical knowledge, so it has to be something that comes in a box and can be plugged in.

Anyone with the relevant technical knowledge can probably get equal or better results resolving the underlying issues or making something more optimised, that simply cannot be let loose with good conscience on unsuspecting and technically challenged customers.

Thor

 

[Ghoostknight]

Thanks for the insight!

Yes completely mirroring what the Ifi idefender does is probably(pretty much surely) out of my (knowledge)-range but i try to search a solution that is "kinda" the same functionality but much cheaper (maybe i even do my first pcb design, im not sure yet),
i was thinking maybe the switching could be done with a microcontroller but im A. not sure this is gonna work good and B. it will include relais and other stuff and i normally like minimalistic approaches, specially in audio, so i think if i can stay as close to the very simple drawing i made the better

Mainly i wanna do this since A. the raspberry pi has not alot of current to offer on usb`s and B. keep the influence of attached devices to the raspberry pi`s GND and Voltage rail as minimal as possible while still being cheap, a usb 3.0 isolator is far from cheap and not talking about multiples, and maybe even others can benefit from it if i end up selling pcb`s with this very simple but effective design (specially for bus powered dacs too), since i also heared good differences with a diy splitter cable i made a few years ago (with disconnected gnd and bus power) but i didnt know about the potential difference risk back then (it worked out good tho)

There is a reason for it being a manual switch on the the original iUSB (Power).

Essentially being this switch is a earth/groundloop breaker, specifically a 100 Ohm resistor. This sits between input and output ground, the switch bridges it out.

If there is no DC power current flowing in the USB lines, the 100 ohm resistor is normally ok to let standard USB devices complete a successful handshake. So this ONLY works if supplying new power to the downstream USB device.

Noise problems with USB are almost 100% down to earth/groundloops and this breaks the part of the loop through the USB Cable is reduced by around 80dB.

The switch shorts out the 100 Ohm in case it is not needed or causes problems.

i found this discussion https://forum.allaboutcircuits.com/threads/isolating-usb-power-to-prevent-ground-loops.135912/ which suggests that only isolating power and gnd is actually just as dangerous as disconnectig the ground completly, there is risk of huge potential differences between the two grounds then (i guess since data lines are not isolated then) and also this thread suggest coupling the two grounds with a resistor (tho in the thread 10k ohm is suggested) to keep the two grounds the same potential while minimizing ground loop current

i kinda like the idea of a simple resistor it would make things way easier if we dont need automatic switching like the idefender

tho im not sure which value resistor i should go for, was there a reason for ifi to just go with 100ohm? (maybe since there is risk of buspowering still being used while the switch is engaged?) preferable i would want to go with the largest resistor possible while still A. allowing handshaking sucessfully without pushing a button eachtime or other manual solutions and B. still keeping the potential difference between grounds as low as possible to not risk damaging any device

is there a risk that potential differences between grounds with a high value resistor could still "build up" (since the grounds are "kinda" disconnected with a value resistor like 100k) ? if not i would probably just test which high value resistor still allows handshaking

 

[Thorsten Loesch]

Hi,

> was there a reason for ifi to just go
> with 100ohm?

The reasons I selected 100 Ohm for use with all similar circuitry, be it build in or in case of the idefender stand alone (and yes, I designed these) was based on analysis of the USB 2.0 Circuit, then select a safe value that should be 99.9% unproblematic and a standard value.

You could experiment with different values. Again, a commercial product needs to be extra safe.

Ultimately, if the resistor is too the voltage dropped becomes large enough to prevent handshake or cause dropouts.

A better choice would be to deal with the underlying problems.

A simple DIY version needs no PCB, just modify a USB Cable. Get a nice usb3 cable (or USB2 if USB3 is not usable) and DC socket "tail".

Near the device plug (usually type B) open the sleeving, cleanly separate the shield, ground (black) and power (red) Wire.

Wire the DC socket tail to power and ground on the device side.

Use a pair of 100 Ohm 1206 SMD resistors to like both sides of the shield and both sides of the ground wire.

Isolate everything with heatshrink sleeving and tape, apply overall heatshrink sleeving, done. Plug in any 5V supply, done.

Note this simple setup has no over voltage or reverse polarity protection on the DC etc. so make sure polarity and voltage of supply plugged in is correct.

If you really make a PCB, you should add all protection etc. of this but you probably end up making something that costs more than an ID from the shop or better 2nd hand and does less.

You can see what's inside the ID here:

Thor

 

[Ghoostknight]

Thanks! Yes the idefender looks way more spohisticated but its nice to know that i can do essentially all the idefender does with a little resistor even if it will not be as "universal" as the idefender, so i can save a few bucks, i hope you dont get in trouble with ifi posting this here tho! but i always was curious what the ifi idefender really does, the marketing of ifi to "break" groundloops is kinda missleading a bit if its still connected over a resistor but maybe they just dont wanted to go into details

after some thinking you are probably right, i should just modify a cable, is there a reason you recommend SMD resistors in this case? (since i have generic throughhole ones laying around and they are a bit easier to work with)

And do you have a estimate resistor value i should -not- go beyond? like 1k or 10k ohm?

i hope/guess that protection is built in into most if not all devices i plug those cables in

since you designed the idefender, did you also design the ifi isilencer? im kinda curious if it also filters the datalines, i just bought some china usb filter which just applys some capacitors on the power lines, im asking since i have a ifi isilencer (and heared good improvements) but im curious if it filters the datalines (or does anything at all to the datalines) too

 

[Thorsten Loesch]

Hi,

> i hope you dont get in trouble with ifi posting this here tho!

I have nothing to do with iFi for several years now.

> but i always was curious what the ifi idefender really does, the marketing of ifi to "break" groundloops is kinda missleading
> a bit if its still connected over a resistor but maybe they just dont wanted to go into details
Actually, I wrote that "break the ground loop" piece because it is what it does. Well, if we want to be super technical it "ONLY" reduces the ground/earth loop noises 1,000 times in a common setup, though it may be more (depends also how big the problem is).

> is there a reason you recommend SMD resistors in this case? (since i have generic throughhole ones laying around and they are a bit easier to work with)

Parasitics, SMD resistor work better at higher frequencies.

> And do you have a estimate resistor value i should -not- go beyond? like 1k or 10k ohm?

I recommend 100 Ohm. I would not go above 100 Ohm, though I found 330 Ohm to work in most cases.

> i hope/guess that protection is built in into most if not all devices i plug those cables in

Yes, it is.

> since you designed the idefender, did you also design the ifi isilencer?

Yes. I designed pretty much every iFi product until 2019 and some that came out afterwards.

> im kinda curious if it also filters the datalines,

It does, using a special common mode filter choke. The other key feature is an active noice cancellation circuit on the +5V line.

When all of this was being developed I actually pushed for a single product combining the two and selling for 29.99 US Street price (which would be fair pricing based on development effort, BOM Cost and sensible profit margins).

But it ended up as two pieces selling together for three times as much. And aving to stack them is a major pain in the proverbial.

I thought it would have sold way more as the combined product for 30 Bux than now and thus overall generated ore income and profit. But apparently I'm an idiot who no clue about business.

Something similar happened with the mains Products, respectively the Earthloop breaker and DC Blocker. All in one for 50 or maybe 80 USD would have been fair. two pieces that inconvenient to stack and want tripple the money, many people who actually need the products will look at the cost and be yeah, right....

> i just bought some china usb filter which just applys some capacitors on the power lines, im asking since i have a ifi isilencer
> (and heared good improvements) but im curious if it filters the datalines (or does anything at all to the datalines) too

Yes, iS filters data (passive) and power (active & passive).

The USB iPurifier includes the same passive USB Signal filtering and USB signal regeneration (active) plus Power Filtering (active & passive).

BTW, IF I was doing a lot of DIY on USB, I would probably buy a sub 50 USD Orico USB 3.0 Hub and modify it to accept a separate power supply for downstream devices, include a switchable USB ground lifter etc. Such a unit will make sure that you get the most out of USB Audio device. It would be very close to this:

https://www.tnt-audio.com/sorgenti/ifi_iusb3_e.html

And it would look better and be more permanent.

FWIW, my own system has a Zen Signature 6XX stack (with EQ in the Amp modified to suit my modified Fostex T50RP MK2 headphones) which has a similar power system as in the iDefender build into the DAC, a iPurifier USB in the back of this, an iFi Mercury USB Cable and an Orico USB3 Hub. Works very well and removing each USB element has a small but audible effect.

Thor

 

[Ghoostknight]

> is there a reason you recommend SMD resistors in this case? (since i have generic throughhole ones laying around and they are a bit easier to work with)

Parasitics, SMD resistor work better at higher frequencies.

> And do you have a estimate resistor value i should -not- go beyond? like 1k or 10k ohm?

I recommend 100 Ohm. I would not go above 100 Ohm, though I found 330 Ohm to work in most cases.

alright, i will go with your recommandation and 100 to 1000 times less already sounds plenty, thanks

> im kinda curious if it also filters the datalines,

It does, using a special common mode filter choke. The other key feature is an active noice cancellation circuit on the +5V line.

ah, good to know, thank you, specially with usb filter its kind of mysterious if you dont know the details to compare different usb filters on an objective technical standpoint, big thanks for helping me out

When all of this was being developed I actually pushed for a single product combining the two and selling for 29.99 US Street price (which would be fair pricing based on development effort, BOM Cost and sensible profit margins).

But it ended up as two pieces selling together for three times as much. And aving to stack them is a major pain in the proverbial.

I thought it would have sold way more as the combined product for 30 Bux than now and thus overall generated ore income and profit. But apparently I'm an idiot who no clue about business.

Something similar happened with the mains Products, respectively the Earthloop breaker and DC Blocker. All in one for 50 or maybe 80 USD would have been fair. two pieces that inconvenient to stack and want tripple the money, many people who actually need the products will look at the cost and be yeah, right....

yes, i also dont like having multiple devices, one device which combines all would have been a bless, and there is also the downside that contact resistances and data lines length just add up, best case there is no difference but since usb cable length can make also a audible difference im not sure about that

profit markups in hifi audio are generally kinda high so i wouldnt blame on Ifi really and compared to other manufacturs ifi`s pricing seem "somewhat" reasonable (tho for consumers cheaper is always better and for 30€ for a combined device it would have probably become THE goto device compared to others, now it just seems to be a reasonable priced alternative), i guess you also have to count for lower sell volumes since many dont believe in this stuff making a difference and ASR/objective measurements doesnt really help here either

on the other hand ifi offers so many "gadgets" that it seems to be built around the idea of let you buy more and more
in case of the DC Blocker, you have to buy one for each device where you could have a 16A one in front of your power strip with the same effect as far as i know

BTW, IF I was doing a lot of DIY on USB, I would probably buy a sub 50 USD Orico USB 3.0 Hub and modify it to accept a separate power supply for downstream devices, include a switchable USB ground lifter etc. Such a unit will make sure that you get the most out of USB Audio device. It would be very close to this:

yes a neat idea, i actually wanna implement it in a somewhat similar way, putting my raspberry pi with the hdmi/usb c side on the backpanel of a case and usb/ethernet inside the case and connecting the usb ports with a short cable to the backpanel and on those cables i will do the modifcation with a dc barrel connector on the backpanel to feed the usb cables with external power

FWIW, my own system has a Zen Signature 6XX stack (with EQ in the Amp modified to suit my modified Fostex T50RP MK2 headphones) which has a similar power system as in the iDefender build into the DAC, a iPurifier USB in the back of this, an iFi Mercury USB Cable and an Orico USB3 Hub. Works very well and removing each USB element has a small but audible effect.

yes sounds like a good setup, im still thinking about getting the ifi ipurifier usb OR the holo titanis (which has reviews for being one of the best reclockers, tho its not many reviews), i would probably pair it with my topping HS01 ( PC -> usb cable -> topping HS01 -> usb cable -> reclocker -> (as short as possible usb cable or directly connected) -> DAC

i actually want the modified cables to just decouple usb devices like wlan stick and ssd from the raspberry pi to some degree but i will probably try my dac on them too

 

[Thorsten Loesch]

Hi,

> yes, i also dont like having multiple devices, one device which combines all would have been a bless,
> and there is also the downside that contact resistances and data lines length just add up, best case
>there is no difference but since usb cable length can make also a audible difference im not sure about that

Actually, in the "best case" you can take 5pcs of 5m USB 2.0 Cables that do not permit reliable USB 2.0 High speed connections (audible dropouts, high error rate in the debug window of the USB driver).

And you can get an "active" 25m of USB cable that runs DSD512 and/or 768kHz/32Bit with no errors. That is using 4pcs of the "Type A-A Repeater" USB iPurifier and one "Type B" USB Purifier. I demoed this at several HiFi Shows. With the original 99 USD USB iPurifier this could even be a quite cost effective solution for a long USB connection from PC to DAC at "audiophile" grade.

The kind of 5m USB Cable audiophiles would use are likely more than 100 USD each anyway. I don't really know any 25m "Audio Grade" USB Cable. The optical cables that have several issues are usually 500 USD as well.

> profit markups in hifi audio are generally kinda high

That was already designed into my original design target price.

> im still thinking about getting the ifi ipurifier usb OR the holo titanis (which has reviews for being one of the best reclockers, tho its not many reviews)

All "USB reclockers" are strictly speaking USB Hubs. I actually prefer the term "repeater".

While there are differences between individual Chipsets and implementations, these are not very large.

Getting a commercial USB hub and modifying that, IF you are able to moderately simple DIY, to add a ground loop breaker and a feed in for clean external power 2X (one to operate the Hub and one to operate the attached device) would pretty much give you something that is a complete single device with around 50 Bux for a nice hub and a rainy afternoon worth of work.

> topping HS01

If you have this, simply put it in front of the DAC, with the shortest possible tail. It should be functionally identical (no idea about insides) to the ifi iGalvanic.

I would not add another repeater downstream, unless the HS01 use an isolator with poor output drive. If you do need one, use a Hub and modify to feed in clean USB power.

> i actually want the modified cables to just decouple usb devices like wlan stick and ssd from the raspberry pi to some degree

This will probably not have the result you expect. As neither of these have their own PSU there is no ground loop. The best solution for this would be an externally powered USB 3 Hub. This removes the PSU load from the R-Pi and will have some separation if you use the USB3 ports on the R-Pi for "system" devices and leave the USB2.0 Ports for Audio.

Then use the HS01 Isolator between DAC and R-Pi and use the "clean power" feed-in to feed in a clean power.

I would recommend a fairly inexpensive commercial linear power supply with an added Super Capcitor Bank. Like this:

Crucial here is the 2-chamber transformer that maximises the separation between mains and output.

The iFi iPower has been designed to have very low coupling between mains and output. A common switching power supplies has 2,200pF, the iFi iPower is 100pF, while this simple cheap and simple linear supply has 16pF:

I also modified the stock supply as follows:

Replace rectifier Diodes with Schottky types.

Fit 10uF/50V/X7R Capacitors (1210 SMD format) across rectifier output, filter capacitor and the input and output of the regulator IC.

These capacitors are effective to over 1MHz and have a lathe capacitance, it help's lower noise.

The regulator IC is a generic 7805 equivalent. It could be replaced with something lower noise, however for most DAC's it will be quiet enough especially once the next step is added.

Left "naked" the noise is in the region of 10uV RMS. That is about 10 times (or 20dB) higher than iFi's iPower SMPS but as much as 1,000 time (60dB) less noisy than common "cooking" SMPS.

Cut the power cable and insert a Supercapacitor bank near the plug.

This forms a RC filter with the DC Cable (say 0.3 Ohm) and the Super Capacitor Bank.

I used 0.5F/5.5V X 20 for the DAC, so we get 10F with 15mOhm ESR for the capacitors I use. With these values the Super Capacitor Bank filters noise from the 7805 regulator by around 26dB (twenty times) above 0.05Hz, for around 0.5uV total noise for the complete linear supply with Super Capacitor Bank.

Total cost is well below 100 USD for 5V/1A.

The AMP incidentally uses iFi iPower and 7.5F/5.5V X 20 so there it's 150F/5.5V with 5mOhm ESR for the Amp.

Thor

 

[Ghoostknight]

> i actually want the modified cables to just decouple usb devices like wlan stick and ssd from the raspberry pi to some degree

This will probably not have the result you expect. As neither of these have their own PSU there is no ground loop. The best solution for this would be an externally powered USB 3 Hub. This removes the PSU load from the R-Pi and will have some separation if you use the USB3 ports on the R-Pi for "system" devices and leave the USB2.0 Ports for Audio.

well i kinda just wanna make sure since im adding a own power supply for them also through the mod cable (i wanna add a power supply mainly because the pi just can deliver around 1A and its too little for SSD and other stuff, but i thought it may be beneficial to decouple them a little, i dont expect audible differences here, i just thought it would be good practice so to say)

--
thanks for your insights

yes im also considering some power supply options ( i actually have a Ifi iPower X 5V already here, its a nice inexpensive improvement over normal SMPS, tho many claim its not that good, i cant agree/disagree here since i dont have other power supplys tried )

but i read alot recently about supercapacitors, i read first about it as i saw the ian canada modules, tho they are quite expensive (still cheap compared to "hifi" power supplys i guess), ESR seems quite crucial as far as reviews go and i like the idea of supercapactiors even being superior to batterys

Do you have a "easy" schematic for implementing super capacitors by myself? (for example, to make a box for the Ifi power supply or any other power supply to add it downstream to the chain), i think the loadbalancing is probably best implemented with resistors since its totally passive then, tho im not sure how to limit current for first charging them since they would pretty much overload any power supply if current isnt limited

im also quite curious how the uptone super capacitor power supplys sound since its alternating between two super capacitors banks, no upstream connected to the currently used bank

i kinda think supercapacitors are really the way to go for hifi if we can trust most reviews, linear power supplys seems to be not that great measure wise (atleast the somewhat cheaper ones) but many report improvements anyway

one option i really consider right now is using one of ian canadas battery modules + supercapactiors, i kinda like the idea of being completely off the noisy power grid

 

[Ghoostknight]

Actually, in the "best case" you can take 5pcs of 5m USB 2.0 Cables that do not permit reliable USB 2.0 High speed connections (audible dropouts, high error rate in the debug window of the USB driver) and get 25m of USB cable that runs DSD512 and/or 768kHz/32Bit with no errors. That is using 4pcs of the "Type A-A Repeater" USB iPurifier and one "Type B" USB Purifier. I demoed this at several HiFi Shows.

ah well, maybe i was wrong then, good to know!

All "USB reclockers" are strictly speaking USB Hubs. I actually prefer the term "repeater".

While there are differences between individual Chipsets and implementations, these are not very large.

Getting a commercial USB hub and modifying that, IF you are able to moderately simple DIY, to add a ground loop breaker and a feed in for clean external power 2X (one to operate the Hub and one to operate the attached device) would pretty much give you something that is a complete single device with around 50 Bux for a nice hub and a rainy afternoon worth of work.

imo each usb hub i tried influenced the sound in some way, tho its hard to say which sound the "best" and the influences are generally speaking very small, i kinda hoped reclockers are "made for" audio purposes with the "right" soundquality in mind

one thing that is also on my list is trying the Rpi4 OTG port which is directly connected to the SOC/Cpu (with the topping hs01 behind it) to kinda elimate any hub in between, but i have to still see how it compares/sound

> topping HS01

If you have this, simply put it in front of the DAC, with the shortest possible tail. It should be functionally identical (no idea about insides) to the ifi iGalvanic.

I would not add another repeater downstream, unless the HS01 use an isolator with poor output drive. If you do need one, use a Hub and modify to feed in clean USB power.

hmm what made me curious about reclockers is the fact that the ifi isilencer behind the HS01 still makes a little difference, tho i should try placing it closer to the dac since i now have it close to the PC, thanks

 

[Thorsten Loesch]

Hi,

> i actually have a Ifi iPower X 5V already here, its a nice inexpensive improvement over normal SMPS, tho many claim its not that good,

Well, it is MUCH better than a generic SMPS and better than most generic linear power supplies. It is relatively inexpensive to product (a lot of the internals match decent grade "commodity" SMPS). At a somewhat more reasonable price point it would be a very good choice.

> i read alot recently about supercapacitors

They are not "magic". They do something. And if implemented correctly they do it well. That "thing" is to provide very high capacity at reasonable ESR and thus allow that attached device to draw hgh peak currents, plus they reduce noise with even just wire resistance as "ballast".

> ESR seems quite crucial as far as reviews go and i like the idea of supercapactiors even being superior to batterys

Correct. That is why I use 20pcs of fairly cheap "CDA" Brand Super Capacitors in parallel. It lowers ESR a lot.

If we compare energy storage, battery, supercap, electrolytic cap, film/ceramic cap, then battery stores the highest energy per volume, but releases it slowest and is most current limited. Supercapacitors are much "faster" and allow greater peak current in the same volume than batteries, but they cannot store as much energy. As move to conventional electrochemical capacitors the capacitor get's even "faster" and handles more peak-current relative to capacitance. This continues moving to solid dielectric capacitors and ceramic/film capacitors. Energy Storage or capacitance per volume declines, relative peak current handling and "speed" of energy release increases.

As such Super Capacitors are great to "buffer" either batteries or a moderate sized DC power supply for music. The reason is that music is not a continuous sinewave, but dynamic. Generally an audio amplifier playing music needs 16 times the peak power of the average power consumed. A battery may limit available peak current, as may a modest size of line power supply.

To take an example. The power supply inthe Zen Can can actually use around 4.5A @ 5V input (22.5W).

The Class A output stages and other losses account for around 2A @ 5V (10W).

It ships with a 5V/3A PSU, leaving only 5W between 2 Channels for power output. Being Class A / ClassA/B efficiency is max 70% theoretical. So 1.75W absolute maximum per channel, realistically less and the PSU will start limit current for any more.

In some ways the choice to ship an under-powered PSU was deliberate to limit output power and leave room for upgrades or for a higher range product. The other reason was that 5V/5A power supplies of decent quality are large, relatively costly and rare.

So the Zen Can could actually drive more power into low impedance Headphones (e.g. DCA Ether or indeed my own T50RP) than the power supply allows. Adding Supercapacitors to the mix gives a lot of extra peak power using the same (low noise, but limited power) powersupply. One could build a very large external power supply with at least 5V/5A, but that is a lot more work and cost.

Originally I had a few interesting extra products planned for the Zen range, including among others a supercapacitor PSU "Enhancer", but these came to nothing.

> Do you have a "easy" schematic for implementing super capacitors by myself?

A super capacitor is just a capacitor with a very large capacitance relative to size. So you apply it like any other capacitor. Observe voltage ratings. Or kind of treat as a battery that is constantly charged by the existing power supply and delivers extra current to the load when the power supply runs out of steam.

> for example, to make a box for the Ifi power supply or any other power supply to add it downstream to the chain

Just take a large enough number of low ESR Supercapacitors that wired in parallel will give you desired capacitance (say 10-50F per Ampere). If you use higher voltages than 5V you need to use series/parallel connection to get higher voltage ratings, for USB the 5.5V rated units off the shelf are fine.
Then wire with +of the whole bank to + of the PSU Wire and - to minus. There is no need for anything else.

> im also quite curious how the uptone super capacitor power supplys sound since its alternating between two super capacitors banks,
> no upstream connected to the currently used bank

It will sound the same, EXCEPT if the power supply that powers this power supply has significant noise leakage from the mains (almost all SMPS have). If this noise causes problems the switching between dual banks approach will disconnect this noise.

That is why in my case I started with supplies that have low noise leakage to start with.

> i kinda think supercapacitors are really the way to go for hifi if we can trust most reviews, linear power supplys
> seems to be not that great measure wise (atleast the somewhat cheaper ones) but many report improvements anyway

For items with external power supplies (especially smaller ones) adding a block of super capacitors can be a very efficient and easy DIY way to lower noise and add capacitance. There is no reason why a linear or switching power supply cannot be engineered to have similar low noise and high peak current capability as what I use.

But I don't want to design as a one-off.

> imo each usb hub i tried influenced the sound in some way, tho its hard to say which sound the "best" and the influences are generally speaking very small,
> i kinda hoped reclockers are "made for" audio purposes with the "right" soundquality in mind

Much of the "sound" relates to power supplies. In the end the clock noise is also mostly power supply noise.

PCB layout and parts quality are a second factor. But here a commercial hub of good quality may be comparable.

> one thing that is also on my list is trying the Rpi4 OTG port which is directly connected to the SOC/Cpu (with the topping hs01 behind it)
> to kinda elimate any hub in between, but i have to still see how it compares/sound

Do try this.

> hmm what made me curious about reclockers is the fact that the ifi isilencer behind the HS01 still makes a little difference,

It may be that the cable after the HS01 causes degradation, maybe something else.

Thor

 

[tomb]

The DIY DoodleBug has been around for a while:
DoodleBug USB Isolator and Power Supply

 

[Thorsten Loesch]

The DIY DoodleBug has been around for a while:
DoodleBug USB Isolator and Power Supply

This is limited to USB Audio Class 1, meaning it will not work at all with any modern USB Audio Class 2 DAC that does not include a UAC1 fallback option. Pretty noisy regulator for the power as well.

Other than that, sure. Pretty good.

In my system it would not work at all though.

Thor