[Thorsten Loesch]

Hi,

> You are missing that the LM317 in the device I referenced allows the voltage output to be adjustable.

No, I'm not. But you don't noisy LED's on the ADJ pin to adjust the voltage.

> We've found that voltage variances in the many PC/laptop/USB-cables can be significant when attempting
> to regulate prior to a USB-powered DAC.

Yes, cables can be an issue, especially where there is higher current draw.

The correct route of action, if we consider a need to compensate should be peruse the USB.org specifications for USB cables and to then create a circuit with negative impedance that compensates cable resistance. All modern battery charger IC's include such a function.

> A 7805 regulator can't create voltage greater than what it is supplied.

Neither can a LM317.

I did not suggest to use the 7805 in my post, where I discussed modifying a commercial OTS ultra cheap (13 USD) linear power supply.

I suggested that given the final intended setup with a super capcitor bank after the 7805, there was little point to make an effort to replace it with something else.

Finally, if I want 5.25V from a 7805, I put a 1A low voltage schottky diode into the ground pin connection. Or I could use a 3.3 Ohm resistor in the ground line and a 62...68 Ohm resistor between output and ground pin.

Not that I would do that in a ground up design, be it LM317 or LM7805 etc.

> The USB standards themselves allows for voltages as low as 4.4VDC, even from "quality" hubs. The problem is that most quality DACs have onboard
> regulators for 4.75V analog and 3.3V digital, among other things. They're all designed and produced in theory for 5VDC supplies in mind, but potentially
> not very tolerant to lesser voltages for best noise-reduction performance.

That's up to the DAC designer.

I tend to use a > 1MHz switcher (with lower audio band noise than a LM317 and the ability to work down to 3V input) to produce 5.7V for internal use. It's actually part of a single chip power SOC marketed for digital cameras that includes a number of liner and switching regulators on board.

The 5.7V is just enough shy of the 5.8V rating of a specific IC I frequently use directly from this rail to guarantee worst case tolerances do not exceed this rating. That IC is a charge Pump that produces a negative line and also operates around 1MHz, so I can run the headphone amplifier on +/-5.7V and have enough overhaed to regulate the rail down to 5V for a 5V DAC using a low noise circuit.

Of course, such an approach is often beyond what can be done DIY.

> Anyway, a case can be made that a slightly higher USB voltage from the power supply will result in less overall system noise. The LM317 gives you that flexibility.

The LM317 gives that flexibility with less noise if the LED's are replaced by a resistor and capacitor as suggested in the datasheet. Moreover, it is not specifically a function of the LM317 that cannot be achieved with any other circuit.

So it is not enough of a reason to use the LM317 and then specifically in a way that adds noise over the datasheet application.

Thor

 

[Thorsten Loesch]

Hi,

> i probably go with trough hole for easier soldering and buy like a 50 pack to get a discount
for power systems multi-layer boards with ground and power planes are preferred as solution, specifically because this dramatically improves the high frequency behaviour.

> i saw that 16V 470uF is probably the way to go, i also saw that the SEPG line is a little better than the SEPC line, tho the difference is quite small
> and will probably add 5-10 to each voltage rail

I use the SMD SEP... line myself.

> i saw that the ian canada filter board (shieldpi pro) uses also little smd capacitors, probably something like 0,1uF ?
> is there a benefit to adding them too? (can you recommend one here too?)

Yes, there is a POTENTIAL benefit, however one needs to be careful in the design, not just parts choice but also mechanical arrangement.

The principle is called "bypassing" and it can be beneficial, but needs care. Just randomly placing different capacitors in parallel can do more harm than good:

Bypass capacitor resonances

So in your case it will probably best to avoid adding then and rely on the designer of the CM4 board to have done a good job and have them on board.

> i think in the example of the 4 voltage rails of the CM4 itself i can make a board which sits directly above it with wire lengths of approxiametly 4-5cm,
> a other approach would be to solder the through hole legs directly, but this way i can only add one capacitor, maybe two

The longer the wire, the less the benefit at higher frequencies.

> and one question just to make sure, if there is a LC filter right behind the regulator which consist of a Inductor+Capacitor
> is it alright to solder the capacitor bank directly to the inductor output pin, or will this mess with the LC filter properties
> (since i solder the wire to the capacitor bank inbetween the capacitor and inductor)?

If it really is a FILTER inductor, yes. If it is instead the inductor and output capacitor of a buck (step down) DC-DC converter, better not. In this case adding another inductor in series is probably a good idea, even if it means cutting traces. Care must be taken to not disturb feedback loops and compensation circuits.

> is there a LT3045 board you can recommend?

No, I do not tend to use expensive "regulator du jour" chips. They are often single vendor which can cause supply problems and expensive.

> my plan was to go with boards from ldovr.com (https://www.ldovr.com/product-p/lt3045-1a0g.htm)
> but there are also a few different boards on aliexpress like this one: https://www.aliexpress.com/item/1005003787293400.html

Both boards look to have very limited thermal capacity. not sure if this is sufficient for your application.

> Can i hit you up with a little tip for the help? (like 25€) through paypal? just PM me your paypal (or bank) address if you want
> since you save me quite a few bucks through DIY anyway!

Thank you very much.

If it does not mortally offend your political sensibilities, could you donate to a charity that helps children in Ukraine, e.g. savethechildren.org or if you find this politically disagreeable perhaps medicines sans frontiers (mfs.org)? I am sure there are people who need this little tip more than I do.

I would even prefer if instead you could consider giving some of your time to a local charity that works with children in need.

Thor

 

[Ghoostknight]

I use the SMD SEP... line myself.

Ah good, i just bought the SEPG 16V 470uF ones

Tho the price differences are kinda ridicoulus, specially in the german available shops over 1 Euro for the most part... the cheapest was 70 cents, i finally found HQonline.com after some digging (well it was also one of the few that had the SEPG line in stock) where the capacitors cost about 35 cents, imo there are huge price differences but im glad i stumbled over HQonline.com in the end

i also bought some 1206 smd resistors (probably way overkill with 0,1% 25ppm) for the usb cables

Yes, there is a POTENTIAL benefit, however one needs to be careful in the design, not just parts choice but also mechanical arrangement.

The principle is called "bypassing" and it can be beneficial, but needs care. Just randomly placing different capacitors in parallel can do more harm than good:

Bypass capacitor resonances

So in your case it will probably best to avoid adding then and rely on the designer of the CM4 board to have done a good job and have them on board.

Ah understood, yea probably best to rely on the 0,1uf smd capacitors that are already on the board then

I also stumbled across this board https://www.dfrobot.com/product-2484.html , im not sure what the different capacitors do on this board but would you say only using the panasonic os-cons are superior to such design? since i guess i could also use this board for my different voltage rails with a little modification (and not use it as HAT, since its just designed for 5V, but i guess there isnt much wrong with using it for other voltages too)

The longer the wire, the less the benefit at higher frequencies.

hmm ok, i will keep them as short as possible, maybe i can squish it to 3cm

I guess then its also beneficial to design the PCB with the OS-Cons in such a way that i dont have just one line of caps after eachother but two/three lines (phyisical in parrallel) where i actually connect the wire "in the middle" of the lines since the furthest away cap will be closer then

No, I do not tend to use expensive "regulator du jour" chips. They are often single vendor which can cause supply problems and expensive.

hmm true, specially if you can design "worse" regulators to the same performance, tho it doesnt matter in my case (since i only need a few) and its probably the easiest way to get good performance for me

Both boards look to have very limited thermal capacity. not sure if this is sufficient for your application.

hmm i think the LDOVR board has a heatsink included, atleast there is a picture that shows connectors and heatsink attached
probably best to go with the LDOVR board since the chinese board will cost about the same if i go with two (to get 1A capability)

i will buy/attach a heatsink myself if it isnt included

Would it be beneficial to ground the heatsink of the LDO and CM4 ?

Thank you very much.

If it does not mortally offend your political sensibilities, could you donate to a charity that helps children in Ukraine, e.g. savethechildren.org or if you find this politically disagreeable perhaps medicines sans frontiers (mfs.org)? I am sure there are people who need this little tip more than I do.

Sure will do no problem with this

 

[Thorsten Loesch]

Tho the price differences are kinda ridicoulus, specially in the german available shops over 1 Euro for the most part... the cheapest was 70 cents, i finally found HQonline.com after some digging (well it was also one of the few that had the SEPG line in stock) where the capacitors cost about 35 cents, imo there are huge price differences but im glad i stumbled over HQonline.com in the end

Are you sure they are genuine?

I would buy from Mouser or a similar shop.

Ah understood, yea probably best to rely on the 0,1uf smd capacitors that are already on the board then

As said, if done RIGHT, it can improve things, if wrong, it will cause more problems. So not a case of just randomly sticking a few pcs in.

I also stumbled across this board https://www.dfrobot.com/product-2484.html , im not sure what the different capacitors do on this board but would you say only using the panasonic os-cons are superior to such design?

Yes. You can fit a lot more Os-Con's onto the same PCB. The cap's on there are a typical "audiophile" foolery of "mixing flavours" quite random.

The silver/blue Cap's are cheap chinese Os-Con imitations, the yellow bricks are tantalum which have less capacitance and higher ESR than the others, so they are pretty much not present.

The Brown bricks are ceramics. Together with the Os-Con's they have very low ESR and will thus cause resonances.

Better to use only one type, unless the interactions between multiple types are well analysed and controlled.

since i guess i could also use this board for my different voltage rails with a little modification (and not use it as HAT, since its just designed for 5V, but i guess there isnt much wrong with using it for other voltages too)

Certainly, this kind of style.

I guess then its also beneficial to design the PCB with the OS-Cons in such a way that i dont have just one line of caps after eachother but two/three lines (phyisical in parrallel) where i actually connect the wire "in the middle" of the lines since the furthest away cap will be closer then

Best is to use "power" and "ground" planes. Basically solid copper pours.

Would it be beneficial to ground the heatsink of the LDO and CM4 ?

IN PRINCIPLE - yes.

In practice it may be necessary to check if there is any electrical connection (intended or random) to anywhere else to avoid causing problems.

Thor

 

[Ghoostknight]

Are you sure they are genuine?

I would buy from Mouser or a similar shop.

atleast i think so, i hope they are, nextpcb/hqonline is quite a big company with somewhat of a good reputation

The cap's on there are a typical "audiophile" foolery of "mixing flavours" quite random.

Better to use only one type, unless the interactions between multiple types are well analysed and controlled.

ah, i kinda thought so, since it doesnt look particular "designed" beside some banks of capacitors, thanks
i will just stay with the oscons then

Best is to use "power" and "ground" planes. Basically solid copper pours.

my plan was to use a normal 2,54 grid pcb, solder the capacitors on the right positions to the pcb, dont cut wires and add something like a 1,5mm² solid copper wire where i solder the wires to, since i also thought its not a good idea to make traces with solder for example

optionally i could probably design a simple pcb for it (this is probably what you mean with planes)
or make just a pack of capacitors like you did with the supercapacitors

IN PRINCIPLE - yes.

In practice it may be necessary to check if there is any electrical connection (intended or random) to anywhere else to avoid causing problems.

Thor

ok thanks, i will check but i think it should be no problem, atleast with the CM4 heatsink, since the copper around the screwholes is removed

is there a risk that the heatsink acts as sort of a antenna? (maybe it makes the ground "more dirty"?) i think i read something along those lines a while ago

 

[novcze]

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

Hi Thor,

just registered to say thanks, I had problems with a ground loop involving USB DAC/headphone amp (K5 Pro), PC, PS5 and monitor.
Your advice helped a lot and I can now use USB connection to DAC instead of optical that I was using to break the loop.

DAC itself is powered externally so it doesn't draw any power over USB so I made this:

 

[Thorsten Loesch]

my plan was to use a normal 2,54 grid pcb, solder the capacitors on the right positions to the pcb,
dont cut wires and add something like a 1,5mm² solid copper wire where i solder the wires to

This can work, better several thinner wires in parallel than one fat wire. Or use self adheasive copper foil to make wide traces.

optionally i could probably design a simple pcb for it (this is probably what you mean with planes)
or make just a pack of capacitors like you did with the supercapacitors

I suggest a PCB. The super capacitor pack is acceptable because it is meant for very low frequencies, essentially the "envelope" of the music waveform.

You want something that is effective at frequencies from switching power supplies. Hence Os-Con's and not something else.

ok thanks, i will check but i think it should be no problem, atleast with the CM4 heatsink, since the copper around the screwholes is removed

Maybe removed because the metal slug on the CPU is connected to something else?

is there a risk that the heatsink acts as sort of a antenna? (maybe it makes the ground "more dirty"?) i think i read something along those lines a while ago

Yes, heatsinks that are "floating" can act as antennae and cause noise issues.

As a general rule to combat EMI "no floating metal". But rules have exceptions.

Thor

 

[Thorsten Loesch]

Hi Thor,

just registered to say thanks, I had problems with a ground loop involving USB DAC/headphone amp (K5 Pro), PC, PS5 and monitor.
Your advice helped a lot and I can now use USB connection to DAC instead of optical that I was using to break the loop.

DAC itself is powered externally so it doesn't draw any power over USB so I made this:

Excellent job.

Next version keep the twist on the green/white wires and solder them directly to the connector, as intended.

Thor

 

[novcze]

Excellent job.

Next version keep the twist on the green/white wires and solder them directly to the connector, as intended.

Thor

I will remake this one, this version was made to see if it works for me and is made to be easily modified.

I'll twist green and white together, USB socket is actually surface mount type (plug is modified cable mount type) and I'll secure the wires with electronic grade Dowsil.
Would you bother with extending the shield from the plug side over the wires? Thanks.

 

[Thorsten Loesch]

I'll twist green and white together,

No, retain the original twist, it is part of the impedance of the cable. That is why originally I suggested to modify a cable, as this avoids all the breaks and impedance discontinuities.

Would you bother with extending the shield from the plug side over the wires? Thanks.

I'd probably wrap self adhesive copper foil around the assembly obviously isolated on one side.

Thor

 

[novcze]

FYI after more testing I went with real galvanic USB isolator (aliexpress.com/item/1005004899278027.html) ...not DIY but will work 100% and is cheaper than parts I could buy on farnell or elsewhere, bandwidth is only 12 Mbps, but it's for PS5 so it will do 2ch. 16bit/48kHz just fine.

edit: fixed link to isolator

 

[Ghoostknight]

FYI after more testing I went with real galvanic USB isolator ...not DIY but will work 100% and is cheaper than parts I could buy on farnell or elsewhere, bandwidth is only 12 Mbps, but it's for PS5 so it will do 2ch. 16bit/48kHz just fine.

good to know, yes i expected this (i also use a Topping HS01 on my dac) but the resistor tweak is very cheap compared to usb 3.0 isolators which begin at >300€ and probably atleast helps a bit compared to nothing, noise can still get through datalines without galvanic isolation (and from testing even with isolation but it helps further)

if you are able to use those full speed usb isolators DO IT they are very cheap but some dacs wont connect with it, even if you set 16bit/48khz, it seems like even the handshake (or communication overall) is high speed usb with most newer dacs and in these cases the resistor tweak is worth a try if you dont want spending much money
well, the topping HS01 high speed 2.0 isolator is also "just" 70€ and should work with all dacs
but in my case i wanna use the tweak to atleast improve things by using it on "real" 3.0 devices like SSD`s connected to my raspberry pi, im curious myself if this does make a difference but lowering noise "anywhere" seems to help (atleast a bit)

i will report back once i tried this, but it will probably take 1-3 months more, since im building a new RPI4/CM4 Streamer with some tweaks

 

[novcze]

good to know, yes i expected this (i also use a Topping HS01 on my dac) but the resistor tweak is very cheap compared to usb 3.0 isolators which begin at >300€ and probably atleast helps a bit compared to nothing, noise can still get through datalines without galvanic isolation (and from testing even with isolation but it helps further)

if you are able to use those full speed usb isolators DO IT they are very cheap but some dacs wont connect with it, even if you set 16bit/48khz, it seems like even the handshake (or communication overall) is high speed usb with most newer dacs and in these cases the resistor tweak is worth a try if you dont want spending much money
well, the topping HS01 high speed 2.0 isolator is also "just" 70€ and should work with all dacs
but in my case i wanna use the tweak to atleast improve things by using it on "real" 3.0 devices like SSD`s connected to my raspberry pi, im curious myself if this does make a difference but lowering noise "anywhere" seems to help (atleast a bit)

i will report back once i tried this, but it will probably take 1-3 months more, since im building a new RPI4/CM4 Streamer with some tweaks

Dunno if you are familiar with the tragic state of the PlayStation 5 audio, but it doesn't support USB audio class 2, so newer DACs like my FiiO K5 Pro do not work and I need USB for PC connection anyway.
So I need to use this USB audio class 1 DAC it uses C-Media HS-100B chip and datasheet states it's Full Speed compliant, so it shouldn't be a problem. FiiO K5 Pro then does amplification and it's needed because for some unknown reason USB audio on PS5 is like 10 dB lower than on HDMI, but this solution will allow use of HRTF 3D audio for headphones which is not possible over HDMI.

 

[Ghoostknight]

Dunno if you are familiar with the tragic state of the PlayStation 5 audio, but it doesn't support USB audio class 2, so newer DACs like my FiiO K5 Pro do not work and I need USB for PC connection anyway.
So I need to use this USB audio class 1 DAC it uses C-Media HS-100B chip and datasheet states it's Full Speed compliant, so it shouldn't be a problem. FiiO K5 Pro then does amplification and it's needed because for some unknown reason USB audio on PS5 is like 10 dB lower than on HDMI, but this solution will allow use of HRTF 3D audio for headphones which is not possible over HDMI.

this sounds indeed tragic and i wouldnt have expected this by a PS5 (this would be a main argument for me to go with xbox if its different)
does it even not have toslink?

maybe this here is something you can take a look at: https://www.amazon.com/FiiO-Headphone-Amplifier-Resolution-Computer/dp/B09B35TCC7/ref=sr_1_1?crid=2GPMT99AV4OV8&keywords=fiio+k3&qid=1673375823&sprefix=fiio+k3,aps,391&sr=8-1
its the only dac i saw that has a switch to switch between usb class 1 and usb class 2

a other solution would be to use some of these hdmi audio extractor boards that output spdif

 

[novcze]

this sounds indeed tragic and i wouldnt have expected this by a PS5 (this would be a main argument for me to go with xbox if its different)
does it even not have toslink?

maybe this here is something you can take a look at: https://www.amazon.com/FiiO-Headphone-Amplifier-Resolution-Computer/dp/B09B35TCC7/ref=sr_1_1?crid=2GPMT99AV4OV8&keywords=fiio+k3&qid=1673375823&sprefix=fiio+k3,aps,391&sr=8-1
its the only dac i saw that has a switch to switch between usb class 1 and usb class 2

a other solution would be to use some of these hdmi audio extractor boards that output spdif

HDMI audio extractor is not good, because PS5 can't output HRTF 3D audio for headphones over HDMI and because HRTF actually works very well, I want to keep using it.
No toslink on PS5 or XBX either.
That small external USB soundcard I'm using is sufficient for gaming, quality is like any other DAC in TVs or monitors, I only need that signal ground loop gone, hopefully aforementioned USB isolator (aliexpress.com/item/1005004899278027.html) will not introduce more problems.

edit:
galvanic isolator from aliexpress arrived, everything works and ground loop is gone...