In a DAC system, the controller and DAC chip are working at 5V or 3.3V. This can be easily powered by USB. But some DAC need a operational amplifier buffer or driving output stage, it need 12V-18V.
A stand alone power supply may be the best solution, but it really make users uncomfortable. A 5V to 12V DC-DC converter may be a solution, but some said the DC-DC will make the sound quality worse. I understand the total power of this solution is limited. For a headphone amplifier it may cause problem, but digital source doesn't need much power.
High frequency noise is another problem. But this can be avoid by using high freq DCDC chip and good shielding.
So, what other side effects of DCDC will meet in a DAC system? How to deal with it?
A stand alone power supply may be the best solution, but it really make users uncomfortable. A 5V to 12V DC-DC converter may be a solution, but some said the DC-DC will make the sound quality worse. I understand the total power of this solution is limited. For a headphone amplifier it may cause problem, but digital source doesn't need much power.
High frequency noise is another problem. But this can be avoid by using high freq DCDC chip and good shielding.
So, what other side effects of DCDC will meet in a DAC system? How to deal with it?
I plan to use a DC/DC (boost) converter chip to power my latest DAC design. I was looking on Taobao yesterday and noticed some fairly cheap PCBs using 'Simple Switcher' parts from TI. High frequency noise is the major issue - eliminating the conducted interference is the first priority, before concerns about shielding. Simple IC regulators have close to zero rejection of HF so passive filters (LC, ferrite beads) are probably the way to go. I'm unsure whether higher or lower freq DC/DC chip is easier to filter the noise from - I'll probably start out with a lower freq (40-50kHz) part because these are more mature designs and cheaper.
I'm afraid 40-50k is a bit low for this application. MC34063 can be used for this frequency. I had design many DCDC modules at 1MHz+, very high effeciency and low profile. It's very common for DCDC now days. you can find many chips work at this frequency. I like to use Richtek's product. For higher performance I recommended Linear's product. Layout is very important for reducing EMI. I think 4 layers PCB should be better if cost is not a problem. Actually I'm going to design with 4 layers PCB for my USB power project.
What's your reasoning behind 40-50kHz being too low a switching frequency? Because its close to the audio range and therefore harder to filter out? I'll give it a try anyway and see how I get on. Cost isn't a problem but I'll always seek out the lowest cost solution that gets the job done because that'll ensure the biggest demand. High efficiency is laudable, but if the DAC takes 1.5W (roughly what mine requires) and the USB port can supply 2.5W then 90% efficiency is overkill.
Yes 40-50K is too closer to audio freq. I had tried 34063 at 65KHz, there was noticeable noise. Nothing can be heard with Richtek chip at 1.2MHz. High freq means smaller inductor, higher efficiency. There's no harm for high efficiency because it means higher power budget that may be useful for high dynamic audio. Not sure, waiting for the testing result.
I've been using fairly cheap SMPSUs to power my chipamps (Mean Well and no-name bricks) - they're going to be running sub-80kHz because they're built down to a price. No noise is noticeable when well filtered but I must admit, the filtering called for is fairly comprehensive.
Higher frequencies are going to mean more radiated emissions, not just conducted because the harmonics are much higher, probably into VHF territory. Yes to smaller inductors but how small is small enough at 2W power levels? Sure higher efficiency isn't harmful but like on my comment in the previous paragraph, have you seen anyone doing offline switchers at 1MHz+ frequencies? Sure they'd be more efficient, but at what cost?
Higher frequencies are going to mean more radiated emissions, not just conducted because the harmonics are much higher, probably into VHF territory. Yes to smaller inductors but how small is small enough at 2W power levels? Sure higher efficiency isn't harmful but like on my comment in the previous paragraph, have you seen anyone doing offline switchers at 1MHz+ frequencies? Sure they'd be more efficient, but at what cost?
I'm convinced that high frequencies only mean higher emissions because they didn't choke the switching elements down as hard, due to high snubber power loss.
otherwise a 500Khz inverter would only have 10Db more noise than a 50Khz element.. but that doesn't seem to be the case.
As a general rule, higher frequency does not mean more efficient, it just means the transformer is smaller. nothing. else. changes. resonant topologies might be more efficient, because the smaller magnetics mean more money can be dumped into them, but that's it. money saved on ceramic, plastic, and low esr capacitors can also be dumped into faster gate drive, more expensive lower inductance circuit boards, and higher quality inductors.. but most of the time it just ends up in the CEO's pocket.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.