How Do DACs Work? - The Learning Circuit

[@MacXpert74]

7:00 What you show and say here about 'small steps' in the DAC output is actually incorrect. Although the digital information is 'quantized' the output of the DAC will still look smooth. The reason for this is the use of low-pass filtering in the output. The 'small steps' would represent frequencies beyond the maximum recordable frequency possible with the used sample rate. These frequencies are filtered out with the low-pass filter. So no, the 'audiophiles' are NOT right when they say the waveform isn't smooth or 'real analog'.

[@NeonGreenT]

'steps' don't exist anyway, so wiggly-modulation upon the signal would pretty much always exist independent of how much a low pass damps them out, you just have to zoom further and further in. If they are actually audible though is a different story.

[@akh345]

At 3:15 and onward, the description is incorrect. The Least Significant Bit (LSB) should be next to 8k resistor and the MSB should be next to 1k resistor. Not vice versa.

[@RicPerrott]

Even a vinyl LP is a compromise from an original analog signal due to the need for the RIAA curve. LPs have a limited tolerance for low frequency signals because the stylus can jump out of the groove if it is too wide. So to compensate, the RIAA curve was created. This means that the analog signal coming off of an LP still needs to go through processing before it is reproduced. Even if this processing is purely analog, it will still not be a perfect representation of the original signal. Great video!

[@techobsessed1]

This is wrong. Audio DACs (and ADCs) are bandwidth limited to an appropriate range for the sample rate. The reconstruction filters low-pass the output of the voltage stage, eliminating high-frequency content (the stairsteps). The output signal accurate represents the input signal with the addition of some amount of quantization noise. Nevertheless the signal to noise ratio ( and dynamic range) of digital audio exceeds that of vinyl.

[@hraharahra]

yes, this is common misconception, but in a tutorial video, it's a sin. Hope more people see your comment

[@pracheerdeka6737]

Sir do dac resample the sound 🔊 through transistor switch

[@bananaear23]

What video do you suggest I watch to understand dacs

[@jeffjuhre1494]

Also Nyquist's theorem states that you can recreate an analog signal accurately by using a sampling rate of twice the frequency you need to capture. Human hearing only goes up to about 20 Khz (in young ears), so sampling at 44.1 KHz is good enough. The part of the signal lost is above the range of human hearing, so for audio puposes it doesn't matter.

[@jamiepearson532]

Thank you.

[@BenjaminNelsonX]

Element14 videos with KAREN are the best!

[@BenjaminNelsonX]

Mentions of vinyl records and PWM motor controls were both things going through my head right before they were then mentioned in the video!

[@TboneLoyal]

Sure wish you did more of these. The way you explain them is very easy to understand. Miss the show also. But thankful for these learning circuits. Thank you. Ive learned alot in the last year watching all these and all the old one. Always had repaired when I was a kid And now full time working as its the only thing i can do now. Glad to be learning again. Love electronics. Keep up the great work.

[@9augustine]

Awesome delivery, easy to understand, learned a lot. Make more. You’re awesome!

[@DarthTyroler]

There is a slight mistake on the weight of bits at 2:45, the first bit has a weight of 1 not zero, if a bit had the weight of zero would make it useless. Otherwise good video, keep it up.

[@st200ol]

Yes, there is a missing "2 to the power" in that diagram so it should read 2^0=1. :-)

[@Rob_III]

@@st200ol No there's no missing "2 to the power", then the values would be 0, 1, 2, 3 (e.g. 2^0, 2^1, 2^2, 2^3). But when the values as they are used in the video (2, 4, 8, 16... ) are used then the LSB should be 1 (2^0 = 1).

[@akh345]

Yes and there is a bigger mistake at 3:15, the least significant bit is the one next to 8k resistor, not next to 1k resistor.

[@OMNI_INFINITY]

I really don't think the output wave is stair stepped on a proper audio DAC. Had a look at the output on a scope?

[@flurng]

THANK YOU SO MUCH for pointing out the fact that dac's can never produce a "true" analog output! I try to explain this to some of my audiophile friends & they either don't believe me or just don't understand - now I can just show them this video to convince them that I 'm NOT crazy! (Or, at least I'm not wrong.....) In fact, I've argued with my fellow electo-nerds that a dac is not really a digital-to-analog converter, but a binary-to-decimal converter, since it will only output integer values, rather than floating point, so there's always some discrete "step" between values! Anyway, thanks again for this, and all your other AMAZING videos! You should know that just a few seconds in, I immediately SLAMMED the subscribe button & I can't wait to see MORE of your incredible content! Cheers!

[@roberttiroyy916]

Thank you! I like your series to explain electronics in a simple and informative way. Looking forward to see more.

[@phpn99]

This is idiotic : DACs actually smooth out the reconstructed signal with various anti-aliasing techniques. Vinyl is not a perfect analogue of the recorded signal ; vinyl records are cut on a lathe and they have severe frequency an amplitude-dependent electro-mechanical limitations that means they have a high level of distortion and a limited dynamic range.

[@robinpage2730]

I love how the system got converting digital values to a continuous output value signal is itself an ANALOG of the digital values being converted

[@Sebastian-fc1rb]

am I wrong or the MSB - LSB at the 3:24 are in inverted order - 1kOhm should be MSB in my opinion

[@ONRIPRESENCE]

Another application of DACs are quantum computers, since the fundamental logic devices (physical qubits) are excited by analog pulses, based on the resonant frequency of the device :)

[@andywest5773]

After watching this, I now realize that the DAC I created for a new project is actually a binary-weighted DAC and not a ladder DAC like I thought it was. Good thing to know!

[@prakashs538]

Explained well useful for me to know more about dac conversion

[@theragingdolphinsmaniac4696]

Except with the case of audio, you aren't hearing the stepped waveform because filters are used to smooth the signal back into its "original" form. The better the filter circuits the closer the converted signal will be compared to the analog input signal.

[@utkuserin9750]

is using High resolution dac only way to creating smooth and real analog output ? Is there any other ways to get real analog signal from a device?

[@techobsessed1]

Video is wrong. Before output from DAC signal is low-pass filtered at nyquist frequency (1/2 sample rate), this produces a smooth and accurate waveform.

[@samithetechsavvy6485]

thank you very much , this so instructive

[@wangleimail]

There is an obvious error from 3:12 to 4:15. You should flip the MLS and LSB. In the other word, 1K resistor should be MSB, and the highest resistor should be LSB.

[@saviranjankumar4865]

Hi mam ❤️ I am from India 🇮🇳🙏 very helpful your video

[@OMNI_INFINITY]

Actually supposedly dacs have a lerp style wave smoothing (maybe using capacitors). The output is supposedly smooth.

[@davidemmanuel8600]

Excellent n Well explained. Thank You!

[@prcr]

To the point and very informative. Just what I was looking for!

[@pracheerdeka6737]

What is the use of chip and crystal Osilator on dac if only op amp and resistors are used. No one tells how software works with the frequency with the dac. How switching is done.

[@309electronics5]

Maybe there are some more (digital) functions in the chip or it needs perfect timing

[@casabombn]

Great video, thanks.

[@mesfinhaftu6206]

The MSB is that have the smaller Resistance I think it is flipped here

[@georgeakejelu8049]

Please any one can help with VT25-373-99X9 data sheet

[@jonathan_herr]

0' s place?

[@flywittzbeats4008]

Awesome tutorial ❤ in my opinion, modern audio dsp has gotten the sample rate algorithms and interpolation algorithms almost perfect. But the amplitude and saturation produced by variance algorithms, is just not there to my ears. Which is an awful thing to say since dsp engineers have done monumentally amazing work haha

[@freedomisfood6966]

Thanks❤❤❤❤❤❤

[@supremebohnenstange4102]

That this video is wrong on the audio part is truly sad

[@truesonic669]

I have many expensive dacs thx onyx creative witch is the best answer that question for me

[@hendriagustian7082]

Try using a multiplexer (analog)

[@sudheerkumar5966]

Very good

[@parttimedude5440]

Any audiophiles here ???

[@synapticmemoryseepage4447]

Most excellent!

[@great.Indian.culture]

Thank you mam.