PLL in a DAC explained

Paul explains the function of a PLL and speculates on why it might matter in a DAC.

[@tacofortgens3471]

Thoughts and Prayers for our brothers and sisters in Israel 🙏

[@danobrien3189]

And also for the 2 million equally innocent cousins in Gaza. .. Who are being attacked r a State Army

[@bikdav]

I only heard of PLL from the late 70s FM tuner days. I had no idea what it was supposed to do.

[@danobrien3189]

They only came into prominence then... The point was that digital clocks could count with unbelievable precision ... But they couldn't create a high purity sine wave..which was needed for aLL manner of purposes in RF. . The PLL allowed a low precision but High purity sine wave to be locked to the exact frequency dictated by a digital counter

[@andrefortune8278]

Great to see Paul great setup!

[@Roosville1]

A digital source (EG, a CD player) uses a internal clock to bang out the serial data bits, and as such the transmitted data contains frequency’s that strongly correlate to the source clock. So strong that in the receiving end we can see (recover) the source clock from just the data. Extracting raw clock from data is messy, and you get a lot of pattern related jitter and such. We can re-time the data but we need the receiver clock to do this, and as such the receiving clock needs to _frequency match_ the incoming source otherwise it all drifts out of sync, and we lose data. The PLL sorts this. The PLL compares clock frequency recovered from the data, to the frequency of our internal clock. If there is a difference in frequency (this is the phase part of the PLL) it produces an error voltage proportional to the error amount. This voltage is used to tune the internal clock until it matches the source clock frequency and the clocks become “locked”. The trick is we low-pass filter (slow down) the error signal, so it doesn’t track transient frequency changes from the noisy source, but instead produces an AVERAGE clock frequency. So the incoming clock can move around but the PLL will run an average on this and come up with a steady clock. This clock is used to shift in the data, hence re-timed at a steady rate. We will still get some wander but it is effective at removing a lot of jitter. PLL design is a difficult subject.

[@Ghost-Matrix]

I had a problem with my TV dropping out audio when connected to my dac. It was so annoying cause it was constant. However, I discovered when playing around with the PLL on the dac it cured the drop outs.

[@philiptong4978]

is it media source sensitive? e.g. over the air RF/ internet streaming live broadcast vs pre-recorded local media playback (optical disk/flash drive/hdd/lan)

[@quangmi3402]

Can you tell me How did you fix this problem?

[@karledwards2319]

Say the DAC needs an oscillator that is 16x faster than the data coming in, the PLL counts the DAC clocks for every data clock. Because of jitter on the data clock sometimes it counts 15, sometimes 16 or 17. If over a period of time, like 1 second, say the count averages a bit low, the PLL slowly speeds the DAC clock up, if high it slows it down until the average is 16. This way the DAC clock matches the average data clock but is stable and doesn't jump around with the data jitter.

[@chuckmaddison2924]

I guess says a lot about the lot in UK.

[@xcvbxcvb2179]

Content starts at 1:15

[@audiorick841]

If you have a source with a great clock accuracy and you use Spdif coax or AES (not USB) you don’t need the DAC to work very hard to do its job reclocking. I believe that’s the working principle of the Grimm MU1 and it sounds awesome.

[@analogueman5364]

I'm interested/curious - why not USB?

[@audiorick841]

@@analogueman5364 my understanding is that if you use USB it’s the DAC clock that takes control defeating the aforementioned principle and spending a lot in the clock of the streamer if you go USB. Doesn’t mean usb is bad but may not be worth investing in a crazy expensive streamer clock. At that point if you go USB shift the investment to the DAC clock but arguably the DAC end up “working harder” again. I don’t think there is an absolute truth in either way but understanding how it works can certainly help in making better decision on maximizing where you choose to drop your hard earned money 😊

[@carminedesanto6746]

GM ☕️ I’m going to the upcoming Toronto Audiofest …hopefully you’ve got a bigger room this year😉

[@robaroy2269]

I didn't know he was there. I also didn't know how much they charge for entry to the show. I would like to meet other people who are audiophiles

[@carminedesanto6746]

@@robaroy2269He wasn’t..but they had a great set up ..in too small a room off the main corridor….great but meh all at once .

[@abghg66777]

Regarding Dacs ; PS audio Dacs are no 1 in business . 2nd place are Chord Dacs .

[@ThinkingBetter]

This topic arise when you run a digital audio architecture where you have two clock masters of the audio sample clocking. Much better is when your DAC doesn’t need to bother with other clocks than its own and the audio stream is pulled asynchronously on demand from the source. In fact, modern streaming is working in this way and with a USB connection between player and DAC you can avoid this issue. A single clock master in the DAC allows for the audio data to be buffered and pulled from source when running low without the mess of PLL being needed. In the old days with CD players outputting TOSLINK or SPDIF you ended up with the need for PLL and in some cases it caused audible jitter issues. But let’s be in current time where we can run digital audio without this concern and super precise clock.

[@terriludolf6101]

does the directstream also reclock i²s ?

[@ThinkingBetter]

Brilliant question. I very much prefer my DAC to be the precise clock master of the timing of which the samples are translated to analog rather than any external source where the risk of jitter can arise. Surely adding a PLL to mitigate such issue is inferior to running the audio data buffered in the DAC and pulled on demand e.g via a modern USB connection where such asynchronous audio transfer is trivial. There really should be only one clock master in an optimally working digital audio playback system and it better be the DAC. If PS audio clocks the I2S bus from the DAC side it’s still ok as the DAC won’t need PLL in such case.

[@philiptong4978]

another solution is to use a dedicated master clock to sync up all independent digital devices in the chain

[@terriludolf6101]

i think if the dac reclocks everything thats pointless@@philiptong4978

[@ThinkingBetter]

@@philiptong4978 It’s a bit hilarious how the audiophile business is making money out of ridiculous claims. Yes, you can use an external clock source but the entire notion of this topic is ridiculous considering nowadays your DAC can run a very precise clock costing very little pulling the samples on-demand via USB completely making this topic non-existing. I stream my music from Amazon Music HD in up to 192kHz 24 bits lossless better than CD quality and my DAC clock is the clock master with no such issue. This problem is only a topic when using some source that is also trying to be clock master such as an old CD player/transport or media streamer using old standard one-way output like SPDIF.

[@donpayne1040]

Errr

[@chungang7037]

relax bro, you were first!

[@cengeb]

Where where you when Philips invented the CD, and all things laser digital optical....now you use it, cus you don't have to pay royalties to the now off patent stuff....then they invented SACD DSD, where where you 20 years ago?

[@artyfhartie2269]

Hifi manufacturers have hit the mother load when digital audio came on the scene. There is always a newer and better DAC and streamer on the market like every month.They are just scams IMO. But a fool and his money are easily parted.

[@davidfromamerica1871]

They are not fools. They are hobbyists. 😀😀😀😀😀

[@Pete.across.the.street]

What's the scam?

[@chungang7037]

@@Pete.across.the.streetthat dacs matter

[@artyfhartie2269]

@@Pete.across.the.street Would anyone really notice the difference between new and old models? I bet no one can tell the difference by just using a CD player. The manufacturers are just playing with the obsession of people who are sound junkies. Nothing seems to satisfy them

[@Pete.across.the.street]

@@artyfhartie2269 that's not really a scam, the manufacturer is actually making improvements or diffrances in the models. They are following the smart phone playbook

[@sidesup8286]

Most things which sound technically apt to leave other things in the dust, like an old dinosaur, turn out to be only verbally impressive or impressive on paper. One example is SACD, with it's million times sampling rate, near 25 times higher than regular cds. "Wow oh wow" people were thinking, when SACD was announced. Yet when they came out and we actually got to hear them, there was only a slight to moderate improvement usually. Some redbook cds that were mastered well, actually sound better than their SACD counterpart. The format failed and now it's mostly relegated to audiophile labels who still come out with SACD discs. Or maybe in Japan or somewhere they still do a bit. So I don't get too excited over so called technical advances or specs. If you got some Conrad Johnson equipment from the 1990s, a good turntable and speakers from back then, you're still sitting pretty I'm sure, (some say even prettier), than most of the new expensively priced stuff that they manufacture today. With all this talk of countless technical advances, you'd think the equipment of today would sound far better than equip. of yesterday. In my experience; it doesn't. The biggest difference between current equip. and equip. of yesteryear seems to be the orbital pricing that has taken place. I remember back around 1981, my dealer tried to get me to buy a $150 cable. I thought that sounded nuts. To my ears, cables have made the biggest sonic gains throughout the years. I remember back to a time when Petersen Litz cables were the only truly high end cables available for those seeking a lift in sound quality. Of course most audiophiles don't know what they are doing. They had their tone controls taken away from them decades ago, which made getting just the right tonality a real crap shoot. If ypu get just the right balance without oh so many expensive tries; you're lucky. People who have warm sounding systems; some of them know that their warmth is masking detail higher up. But they realize that tonality is more important than a thin bright sound. Seek out a sound that is warm, big and highly open and tell me if you still care about micro details or hearing cymbals a little more distinctly.

[@analogueman5364]

Not being critical of your comment what's it got to do with PLL?

[@sidesup8286]

To Analogue Man. My comments don't have to have anything to do with Paul's subject. My comment had more to do with Arty, Chang's and another guy's comment on here. Many commenters always diverge from Paul's subject on a routine basis. When I'm buying carrots in a store; I'll almost never be talking carrots. Talk progresses to other things. Some of the other commenters diverged into a much more interesting topic. The topic of "Is there that much difference between dacs sonically", and if not, then PLL doesn't matter much. I hope I answered your question.

[@analogueman5364]

I see, thanks. I had to go back and re-read the comments from those you mention and it made more sense 👍

[@chungang7037]

Wow, I clearly need a new DAC 5X the price in order to make sure this does not affect me.