As a dac designer? Kind of has some of that look to it. You did ask about what would sound different if using linear power supplies. We're like, you don't already know that???
Last edited:
Well, as I said.. I'm a mechanical engineer. This is my second ever PCB design. So ye I'm very much a beginner. I just try to get as far as I can in the first attempt. Thats also why I said its going to eveolve over the next years, as I try different things and learn from my mistakes 😉
I have build a qouple of speakers by now, but all have been driven by Hypex modules so I only know that sound as a reference. Thats SMPS and AKM. If I can ever get to that level I will be waaay more than happy 😀 But i'm aware I wont get there in my first try. probably also not the 10th 😉
Okay, understood. But we would hope you would like the design in progress to sound reasonably close to what the chosen simple dac chip is capable of. To me if you continue with the SMPS ideas, you are just repeating the same errors many of us already made. Maybe everyone's biggest mistake is to believe that datasheets exist to tell you everything you need to know to get great results out of a product. Not quite that simple. We all studied the datasheets too, and then some of us started to realize the main purpose of datasheets is to help sell ICs. That's what application engineers who write app notes are being paid for as well. For the most part, nobody in any business feels obligated to tell you why you shouldn't buy their product. That would be insane from a business perspective, and often quick grounds for getting fired. So, like with anything else its, buyer beware.
Short story: SMPS in dacs tends to be asking for trouble, even though the datasheets show you how to do it. IMHO, time to learn about that is probably now, not later. If you want, why not design the dacs so you can do an A/B test? Put an extra connector and some jumpers so you can cut off incoming SMPS power and bring in clean +- power? Then you will have done a hopefully useful lab class at this stage of your learning. And part of the lesson is to learn how to hear the difference. Do you have any idea what veiled sound is? Would you recognize it if you heard it?
If you find it to be a small difference, then I would suggest to listen to clean power only for a few months, then switch back to SMPS. Has your brain processing learned to recognize it yet? Because if you want to get good at audio design, it can help a lot develop more skill at listening.
Short story: SMPS in dacs tends to be asking for trouble, even though the datasheets show you how to do it. IMHO, time to learn about that is probably now, not later. If you want, why not design the dacs so you can do an A/B test? Put an extra connector and some jumpers so you can cut off incoming SMPS power and bring in clean +- power? Then you will have done a hopefully useful lab class at this stage of your learning. And part of the lesson is to learn how to hear the difference. Do you have any idea what veiled sound is? Would you recognize it if you heard it?
If you find it to be a small difference, then I would suggest to listen to clean power only for a few months, then switch back to SMPS. Has your brain processing learned to recognize it yet? Because if you want to get good at audio design, it can help a lot develop more skill at listening.
Last edited:
Your overall design is excellent and clean. You already placed a back-emf diode across the relay coil (D1) — good catch! Keep relay control lines separate from analog lines as much as possible to avoid noise coupling. Use good-quality caps for the RC — preferably film capacitors if possible for best audio clarity (electrolytics are fine if size is a problem). Since you are working with PCM5102A, here is another device based on PCM5102A, that may interest you. It's a Raspberry Pi HAT. You can connect your Raspberry Pi to your HiFi system using this HAT.
https://www.pcbway.com/project/shareproject/BassWasp_HAT.html
RPi is a notorious source of conducted and radiated EMI/RFI. It can easily couple into hat dacs. Moreover, RPi GPIO bus I2S clocking is known to be quite jittery.
Absolutely!
Good point on the datasheets.. Never thought of that perspective :O
The external SMPS can of course easily be replaced with whatever. The input power connector is a 6 pin connector. The idea is to be able to supply 5Vdigital, 6Vanalog reference, +14VAnalog, -14Vanalog for op amp power, a gnd pin and a mute control pin. How these voltages are supplied is something I will experiment with in the future.
The on-board SMPS dual rail thingy can be de-soldered and then the plus minus voltages can be supplied from the 6 pin connector directly.
So hopefully I can have one DAC board with the on-board supply and one with external supply. This should allow some sort of A/B test. Where one DAC is powered without any SMPS at some point.
Good point. And I certainly want to do more listening and experiments. I just started the journey 🙂
That being said, I'm not sure my goal is the ultimate HiFi performance. The most important thing for me is that I like the sound. Those things can probably be the same but my goal is that I will be happy for the sound not that it's the most HiFi possible.
The current China prototype sounds okay but it is plauqed by noise. I think most of it is power supply related as the noise is there even when the dsp is turned off (only amp powered but wired to the dsp via RCA). When I power it off I currently only cut the positive wire. So the gnd wire from the not-so-great 12v wall wart currently used is still connected to everything through a filter though.
If I unplug the 12v wall wart completely but leave the amp connected then the speakers go dead silent.
Once this noise problem is solved then I can better compare the sound of my dsp vs. My Hypex modules 🙂
Thanks!
Rodger!
I just changed the design to allow both NOG and caps with 5mm pin spacing. Such as wima caps.
Thanks for the inspiration link 🙂
Thanks for your help! However, as said earlier in the thread, I had to put the project on hold. This gave me plenty of time to think and learn, and my future project will be more ambitious and based on another chip. To be continued 🙂