I love it when people go so far into the weeds to solve a problem on their own and then document the process for everyone.
I learned a lot from this writeup, and really enjoyed the read even though I know I would have made it about 10% of the way myself and chickened out and just bought a different product in hopes it would solve the problem.
The CS5213 appears pinout-compatible with the MX929x, interesting. The AG620x is a chip I've had the misfortune of encountering; it was common in Amazon DACs, dropped signal output every 15 minutes or so unless you disconnected the EDID pins from the (monitor?), and interprets input signal in a cursed fashion where 16 or below produces black but only 255 produces full white. If your computer outputs limited range HDMI the analog signal can't reach white, but if your computer outputs full range HDMI the shadows are clipped. This is one of the chips that gave HDMI DACs a reputation for black crush and caused CRT communities to recommend DP.
Interestingly the MiSTer game system community actually patched their FPGAs to output digital signals from 16-255, and have some other way of avoiding the signal dropouts (never talk to EDID?), so this chipset is almost seeked out over incompatible models.
That a person who seemingly is a EE will go on a deep dive like this just to figure out how to do it themselves (rather than just buy something that will work) just warms my heart about the human spirit. It makes me smile and realize that I do this myself, just in the domain of computer based devices and programming. More power to you sir, and I hope you have fun in your quest.
> Annoyingly the input pins had a constant DC offset of over 2 volts, which went away when shorted to ground but reappeared moments after I removed the short.
Mic inputs on audio codecs always have DC bias, because they are DC biasing the microphone inputs so that signals can pass. By default, you would expect ~2-2.7V of DC bias on each microphone channel. This is typically connected through a high impedance (~1kOhm) source impedance. The actual impedance of the input should be extremely high resistance, though, typically in excess of 50kOhm.
The proper way to connect this would have been to a line-in jack on your motherboard. There should be no DC bias on those jacks. Or use an audio input that is 'retaskable' to select the line-in function instead of the mic-in function.
> This large of capacitance is not easily available as C0G. The usual X7R MLCC capacitors are piezoelectric and will change voltage when bumped, which can add interference to audio. Electrolytic caps are an option, but will not fit well on pads meant for 0402 ceramics.
C0Gs are not needed for passing something like this. X5R/X7Rs are commonly what is used for DC blocking capacitors in ADC/DAC solutions and you can still keep extremely good audio performance even with the theoretical microphonic problems for Type 2 dielectric capacitors.
> If we wanted to improve audio filtering, we could design a replacement PCB (four-layer for impedance matching and signal integrity?) with footprints and traces to install a second-order Sallen-Key filter. This would be about the same difficulty as the HDMI2SCART, but likely reusing the existing case and screws unlike the HDMI2SCART's 3D-printed case.
Why not just add your filtering circuit externally via the 3.5mm jack? No limits to what you could do. Cascade as much as you want at that point. Might be a bit ugly, but it wouldn't require re-designing the board.
I didn't realize the filter leaked this much AC until I had been using the "working" board for weeks, without audible distortion (but it's still theoretically there). And I'd have to run another power cable to the filter board, since 3.5mm doesn't supply power. It's possible the cable would result in distortion/radiation in the MHz range if the filter isn't located near the chip, but I don't know if this is a problem in practice. It might be something to work on in the future?
Both of your points were the very first things I thought reading this.
A Sallen-Key filter is trivially easy to design if you work within certain constraints - a 2-pole Butterworth filter has the feedback capacitor exactly twice the value of the "second" capacitor to ground, if both resistors are the same value. If you pick 10kΩ for both resistors, 1nF for the feedback cap, and 470pF for the cap to ground, then you'll get pretty damn near a Butterworth response (Q of 0.707, maximally flat in the passband and then as fast a transition as possible to the stopband) at around 23kHz.
This is perfect.
And guess what? If you want to scale the cutoff, just scale the component values! If you use 15kΩ resistors you get 15kHz, if you use 22kΩ you get 10kHz, and so on. The minor error in Q will not be audible.
The weird edge cases that live around connectors never cease to amaze. I recall one, I think it was also HDMI (maybe displayport or VGA), that showed there's a very low resolution digital graphic or text signal and some connectors even have a screen embedded in them.
If it has balanced xlr inputs a DI will completely eliminate the earth hum. That the signal is dropped to mic level doesn’t matter much as the impedance is correct, the connection balanced and the mixer preamp designed for the purpose. // I am somewhat professional sound tech :)
3.5mm, I now know why conventional consumer audio devices don't use mixers, they're a ground loop party
It's also that I don't know how to build a differential input not limited to power supply range, I do think they are sold and I have a mixer at home but it's somewhat bulky for my computer desks.
Back in the good old days of analog audio we made differential audio inputs with
Iron and copper. :-)
Transformers are amazing at reducing common mode noise like induced hum, but rather expensive compared to the price of these boards.
Fun fact: The entire landline telephone system used to rely on balanced lines with transformers for in/out. The hum nulling is so good that trunk lines had no shielding inside buildings.
The best way to resolve ground loops is to... resolve the ground loops[1].
But to test, for you? You can try a regular/non-special, cheap ground loop isolator. The car audio section on Amazon is full of them. You ideally want one that just has transformers inside of it; these will be wound with a 1:1 ratio. They don't cost much; it's generally cheaper to buy them pre-assembled in a box with connectors than to pick up the parts individually and DIY it.
What you've read about transformers is correct: Transformers are imperfect. All of them are imperfect. They're measurably flawed. But if it works and you don't hear a problem with it, then: There is no problem.
Transformers are also the simplest way to add balanced audio inputs or outputs for...well, anything. :) Using different turns ratios (as DI boxes do) instead of 1:1 also allows conversion of impedance and voltage, which is useful.
(To cover another thing to mentioned in your writeup: HDMI audio is part of the video bitstream. It's not separate at all. So, yeah: The audio extractors that exist definitely operate on all of the bits. They don't get a choice in this matter. They also, ideally, intercept the EDID line so they can request only the formats that they can decode: This can be a fun way to encounter what I'd like to call "unexpected learning opportunities".)
---
[1]: One of the things I used to do was help install residential AV systems for a small local shop. We'd sometimes have customers with surround systems that worked great for years, until they had us back to add a powered subwoofer. And then it was 60-cycle hum central.
In those cases, what I invariably found was this: The new subwoofer was the first component in the system that had a 3-prong plug, and this introduced a new connection to the home's electrical ground where there had been none before for that AV system.
Which should not ever be a problem at all, except: The cable TV or satellite line was grounded elsewhere (sometimes to a water pipe or whatever, sometimes a separate ground rod), and not bonded to the normal ground at all except through the AV system.
That's a violation according to the NEC, but it was often this way anyhow.
The correct fix was to eliminate the problem, and there's plenty of wrong and grey ways to do that... but the Right Way to eliminate it is to call the TV service provider and throw a fit until they finally send someone out to make their ground connection something other than illegal. :)
After that, the hum is gone (and the AV system is also better-protected against lightning damage, but that's a different topic).
Extron made great audio extractors and they can be found for cheap. Also nice is it provides balanced stereo output. I have a bunch lf them for getting audio to my mixer from whatever with hdmi.
Definitely something to look into, but I checked on eBay and they seem to have captive screw connectors I'd have to assemble a cable from. Not impossible but another thing for me to look for.
It seems like this person is incredibly intelligent and resourceful. Why not just design a new PCB for this instead of modding one? Surely the component cost of 2 connectors and a little PCB could still be in the < $20 range and actually output good quality analog signals? $20 is the same as $3 if you’re also buying CRTs just to play Switch games. This is such a hobby level activity I’m sure people would love a kit + a guidebook, could even provide a 3d printed case.
I think you misunderstand the obsessive, blissful, motivations of a nerd. :)
I was like this when I was younger, where a goal like this would entirely consume me, regardless of practicality, and more of a "see if I can". Some of the best days of my life, where I was most alive, were those times of complete, detrimental, obsession. As I grew older, and patterned by employment, it slowly faded, and sometimes I miss it so much.
(funny enough, I was once nerds sniped about a grids of resistors, to understand PCB trace resistance better, for high currents, because an EE said my intuition was wrong. I made a paint program where you could paint/paste an image, with red being + voltage sources, blue being negative, and greyscale being resistance, and it would show you the current flow through whatever you drew. I was able to prove that my intuition was actually correct!).
> I think you misunderstand the obsessive, blissful, motivations of a nerd. :)
Kind of the point of a board called Hacker News. Doing it yourself while someone else holds your beer is on a totally different level compared to someone buying something off the shelf at a fraction of the cost.
Sometimes it's fun to mod something that other people can also easily buy and follow your mod recipe. It's more of a network effect than publishing Gerbers and a BOM.
For example, there are (or maybe were) lots of cheap LED controllers that use proprietary phone apps, but if you crack them open, you find that they're ESP-8266-based, and if you can find the right locations on the PCB, you can reflash them with WLED.
> I'm not sure where to source straddle-mount VGA connectors. Alibaba has a product listing that describes the same part found in my DAC, but has no pictures of the actual product.
> I also found straddle-mount VGA to break off the board easily
> All of this is possible... but it's a lot of work for an incremental upgrade that doesn't resolve the jitter and audio dropout problems. Compare this to the HDMI2SCART which required more manufacturing effort, but fulfilled a previously unmet use case; even then the author stopped selling them himself because he couldn't find a better source for chips.
There are photos at https://www.aliexpress.us/item/3256806185045998.html (I didn't insert one into the article), and a cross-section near the end. I could take a photo of the dongle with heatsink, but not unmodified dongle.
I love it when people go so far into the weeds to solve a problem on their own and then document the process for everyone.
I learned a lot from this writeup, and really enjoyed the read even though I know I would have made it about 10% of the way myself and chickened out and just bought a different product in hopes it would solve the problem.
>> A photo revealed the chip was a NX3303X; I could not find any datasheets
Not much information out there but these are similar chips, possibly it’s a clone?
Capstone CS5213 datasheet:
https://file.elecfans.com/web2/M00/2B/BD/pYYBAGHVUbyARd_-AAU...
And also the AlgolTek AG6200 / AG6201 Series, datasheet:
https://assets.yoreparo.com/attachments/28SGHqLW2Z2MfSzEBJB4...
The CS5213 appears pinout-compatible with the MX929x, interesting. The AG620x is a chip I've had the misfortune of encountering; it was common in Amazon DACs, dropped signal output every 15 minutes or so unless you disconnected the EDID pins from the (monitor?), and interprets input signal in a cursed fashion where 16 or below produces black but only 255 produces full white. If your computer outputs limited range HDMI the analog signal can't reach white, but if your computer outputs full range HDMI the shadows are clipped. This is one of the chips that gave HDMI DACs a reputation for black crush and caused CRT communities to recommend DP.
Interestingly the MiSTer game system community actually patched their FPGAs to output digital signals from 16-255, and have some other way of avoiding the signal dropouts (never talk to EDID?), so this chipset is almost seeked out over incompatible models.
Gemini suggested to me that a number of these chips are “industry standard drop in replacements” for each other effectively clones.
But finding more information to settle it would take googling in a desktop for a few hours chasing down weird Chinese forum rabbit holes.
That a person who seemingly is a EE will go on a deep dive like this just to figure out how to do it themselves (rather than just buy something that will work) just warms my heart about the human spirit. It makes me smile and realize that I do this myself, just in the domain of computer based devices and programming. More power to you sir, and I hope you have fun in your quest.
> Annoyingly the input pins had a constant DC offset of over 2 volts, which went away when shorted to ground but reappeared moments after I removed the short.
Mic inputs on audio codecs always have DC bias, because they are DC biasing the microphone inputs so that signals can pass. By default, you would expect ~2-2.7V of DC bias on each microphone channel. This is typically connected through a high impedance (~1kOhm) source impedance. The actual impedance of the input should be extremely high resistance, though, typically in excess of 50kOhm.
The proper way to connect this would have been to a line-in jack on your motherboard. There should be no DC bias on those jacks. Or use an audio input that is 'retaskable' to select the line-in function instead of the mic-in function.
> This large of capacitance is not easily available as C0G. The usual X7R MLCC capacitors are piezoelectric and will change voltage when bumped, which can add interference to audio. Electrolytic caps are an option, but will not fit well on pads meant for 0402 ceramics.
C0Gs are not needed for passing something like this. X5R/X7Rs are commonly what is used for DC blocking capacitors in ADC/DAC solutions and you can still keep extremely good audio performance even with the theoretical microphonic problems for Type 2 dielectric capacitors.
> If we wanted to improve audio filtering, we could design a replacement PCB (four-layer for impedance matching and signal integrity?) with footprints and traces to install a second-order Sallen-Key filter. This would be about the same difficulty as the HDMI2SCART, but likely reusing the existing case and screws unlike the HDMI2SCART's 3D-printed case.
Why not just add your filtering circuit externally via the 3.5mm jack? No limits to what you could do. Cascade as much as you want at that point. Might be a bit ugly, but it wouldn't require re-designing the board.
I didn't realize the filter leaked this much AC until I had been using the "working" board for weeks, without audible distortion (but it's still theoretically there). And I'd have to run another power cable to the filter board, since 3.5mm doesn't supply power. It's possible the cable would result in distortion/radiation in the MHz range if the filter isn't located near the chip, but I don't know if this is a problem in practice. It might be something to work on in the future?
Both of your points were the very first things I thought reading this.
A Sallen-Key filter is trivially easy to design if you work within certain constraints - a 2-pole Butterworth filter has the feedback capacitor exactly twice the value of the "second" capacitor to ground, if both resistors are the same value. If you pick 10kΩ for both resistors, 1nF for the feedback cap, and 470pF for the cap to ground, then you'll get pretty damn near a Butterworth response (Q of 0.707, maximally flat in the passband and then as fast a transition as possible to the stopband) at around 23kHz.
This is perfect.
And guess what? If you want to scale the cutoff, just scale the component values! If you use 15kΩ resistors you get 15kHz, if you use 22kΩ you get 10kHz, and so on. The minor error in Q will not be audible.
The weird edge cases that live around connectors never cease to amaze. I recall one, I think it was also HDMI (maybe displayport or VGA), that showed there's a very low resolution digital graphic or text signal and some connectors even have a screen embedded in them.
IIRC HDMI/VGA is I2C and can drive an I2C display, but it's normally only used for EDID display identification and DDC brightness control.
These chips appear to be 8051s with multi hundred MHz DACs.
Be fun to program them but I can’t find any programmable devboards with the Capstone CS5213.
Theres been some work on the firmware for the Realtek RTD2660H LCD driver that's also 8051 based.
I believe pine/the community worked on that. They may have some how tos / pointers.
What kind of mixer do you use?
If it has balanced xlr inputs a DI will completely eliminate the earth hum. That the signal is dropped to mic level doesn’t matter much as the impedance is correct, the connection balanced and the mixer preamp designed for the purpose. // I am somewhat professional sound tech :)
3.5mm, I now know why conventional consumer audio devices don't use mixers, they're a ground loop party
It's also that I don't know how to build a differential input not limited to power supply range, I do think they are sold and I have a mixer at home but it's somewhat bulky for my computer desks.
Got it. In my studio I have mostly unbalanced cables but the gear is all good quality so I never had any issues.
I think there are ICs for that specific purpose? If I understood your second paragraph correctly.
https://nl.farnell.com/en-NL/pulse/pls00547/ground-loop-isol...
Also this should work.
Back in the good old days of analog audio we made differential audio inputs with Iron and copper. :-)
Transformers are amazing at reducing common mode noise like induced hum, but rather expensive compared to the price of these boards.
Fun fact: The entire landline telephone system used to rely on balanced lines with transformers for in/out. The hum nulling is so good that trunk lines had no shielding inside buildings.
there are transformers and transformers.
A lundahl or murata, the kind used in Studer, Neve & Yamaha mixer preamps can cost a whole lot.
There are pretty decent cheap ones from china these days too.
And don't forget AMI Tabfunkenwerk from USA.
The best way to resolve ground loops is to... resolve the ground loops[1].
But to test, for you? You can try a regular/non-special, cheap ground loop isolator. The car audio section on Amazon is full of them. You ideally want one that just has transformers inside of it; these will be wound with a 1:1 ratio. They don't cost much; it's generally cheaper to buy them pre-assembled in a box with connectors than to pick up the parts individually and DIY it.
What you've read about transformers is correct: Transformers are imperfect. All of them are imperfect. They're measurably flawed. But if it works and you don't hear a problem with it, then: There is no problem.
Transformers are also the simplest way to add balanced audio inputs or outputs for...well, anything. :) Using different turns ratios (as DI boxes do) instead of 1:1 also allows conversion of impedance and voltage, which is useful.
(To cover another thing to mentioned in your writeup: HDMI audio is part of the video bitstream. It's not separate at all. So, yeah: The audio extractors that exist definitely operate on all of the bits. They don't get a choice in this matter. They also, ideally, intercept the EDID line so they can request only the formats that they can decode: This can be a fun way to encounter what I'd like to call "unexpected learning opportunities".)
---
[1]: One of the things I used to do was help install residential AV systems for a small local shop. We'd sometimes have customers with surround systems that worked great for years, until they had us back to add a powered subwoofer. And then it was 60-cycle hum central.
In those cases, what I invariably found was this: The new subwoofer was the first component in the system that had a 3-prong plug, and this introduced a new connection to the home's electrical ground where there had been none before for that AV system.
Which should not ever be a problem at all, except: The cable TV or satellite line was grounded elsewhere (sometimes to a water pipe or whatever, sometimes a separate ground rod), and not bonded to the normal ground at all except through the AV system.
That's a violation according to the NEC, but it was often this way anyhow.
The correct fix was to eliminate the problem, and there's plenty of wrong and grey ways to do that... but the Right Way to eliminate it is to call the TV service provider and throw a fit until they finally send someone out to make their ground connection something other than illegal. :)
After that, the hum is gone (and the AV system is also better-protected against lightning damage, but that's a different topic).
Extron made great audio extractors and they can be found for cheap. Also nice is it provides balanced stereo output. I have a bunch lf them for getting audio to my mixer from whatever with hdmi.
Definitely something to look into, but I checked on eBay and they seem to have captive screw connectors I'd have to assemble a cable from. Not impossible but another thing for me to look for.
It seems like this person is incredibly intelligent and resourceful. Why not just design a new PCB for this instead of modding one? Surely the component cost of 2 connectors and a little PCB could still be in the < $20 range and actually output good quality analog signals? $20 is the same as $3 if you’re also buying CRTs just to play Switch games. This is such a hobby level activity I’m sure people would love a kit + a guidebook, could even provide a 3d printed case.
I think you misunderstand the obsessive, blissful, motivations of a nerd. :)
I was like this when I was younger, where a goal like this would entirely consume me, regardless of practicality, and more of a "see if I can". Some of the best days of my life, where I was most alive, were those times of complete, detrimental, obsession. As I grew older, and patterned by employment, it slowly faded, and sometimes I miss it so much.
https://xkcd.com/356/
(funny enough, I was once nerds sniped about a grids of resistors, to understand PCB trace resistance better, for high currents, because an EE said my intuition was wrong. I made a paint program where you could paint/paste an image, with red being + voltage sources, blue being negative, and greyscale being resistance, and it would show you the current flow through whatever you drew. I was able to prove that my intuition was actually correct!).
> I think you misunderstand the obsessive, blissful, motivations of a nerd. :)
Kind of the point of a board called Hacker News. Doing it yourself while someone else holds your beer is on a totally different level compared to someone buying something off the shelf at a fraction of the cost.
Sometimes it's fun to mod something that other people can also easily buy and follow your mod recipe. It's more of a network effect than publishing Gerbers and a BOM.
For example, there are (or maybe were) lots of cheap LED controllers that use proprietary phone apps, but if you crack them open, you find that they're ESP-8266-based, and if you can find the right locations on the PCB, you can reflash them with WLED.
> I'm not sure where to source straddle-mount VGA connectors. Alibaba has a product listing that describes the same part found in my DAC, but has no pictures of the actual product.
> I also found straddle-mount VGA to break off the board easily
> All of this is possible... but it's a lot of work for an incremental upgrade that doesn't resolve the jitter and audio dropout problems. Compare this to the HDMI2SCART which required more manufacturing effort, but fulfilled a previously unmet use case; even then the author stopped selling them himself because he couldn't find a better source for chips.
To reduce e-waste? Not everything has to be optimised on time or money spend.
Maybe the chips aren't available, or maybe they just wanted to tinker.
As long as it was done in a weekend. Otherwise, they'd suck as a hacker as that seems to be the qualifying time allotted to be cool
Linux Torvalds?
This post just makes me realize that one day I need to buy myself an oscilloscope.
A picture of the adapter would have been nice.
There are photos at https://www.aliexpress.us/item/3256806185045998.html (I didn't insert one into the article), and a cross-section near the end. I could take a photo of the dongle with heatsink, but not unmodified dongle.