I enjoyed the video! The whiteboard is helpful for clearly stating the logical process as well as the math. The discussion of how the heatsink's thermal resistance varies with airflow was useful too. As noted, that can get messed up with a poor enclosure design.
Yes, this can help - but to get the full power out of these - you would need a very large heatsink;) The problem is - the heat is generated in such a small spot - that it it hard to spread it enough to get a heatsink by itself to do the trick. Conduction and relying on the chimney affect is probably not going to do it.
Thank you! They use RMS to relate it to DC, so for power RMS current and RMS voltage is used to find Power - but that is not RMS power it is just power at that point;)
Ti recommends to ground conductive heatsinks, so thankfully anodized parts are non conducitve. Usage of thermal pads is acceptable (0.5 mm thickness) or simply non conductive paste. It would be interesting, if an antenna effect of coupling heatsinks could be measured and if there's audible / electrical effect. I prefer CPU heatsinks with many fine vertical fins, those chunky heatsinks with small surface area and horizontal orientiation will be sufficient too, but not ideal. Just allow good passiv airflow in the case. Take a look at the poor thermal design of many tin desktop amps with upside down mounted chip screwed to the case somehow. Example: The A07 max is a warm one, it works, but it's not a nice long term ambient for electrolytics in this box. For harder driven PBTL @ 2-3 Ohm very slow running silent fans were added to the case. In stereo BTL config you will seldom use more than 5 W average power output in real life, cause this is the region when dynamic peaks quickly begin to sound harsh / distorted and become audible in the upper frequencies (5W +10 dB headroom for dynamics = more than 100 W impulse needed!) . So all that numbers in the specs are meaningless for Hifi applications. Anyway, most times you merely need more than 2 W RMS for private home usage and then there's plenty of headroom.
I'm really interested in peltier devices for cooling! I have couple devices that I plan to do a video about them coming up in the next couple of weeks.
More heat sink is always more betterer. I've never heard of an overclocker wishing they'd opted for a smaller heat sink. 10% distortion isn't the sound we want?! That's the point where an electric guitar amp just starts to get "lively," and there are actually quite a few genres of music that depend on much more distortion for their characteristic timbre. Try to imagine Modern Country music without a crunchy rhythm and chimey lead guitar, or the Star Spangled Banner without a Fuzz Face, or Big Band music without muted brass, or Death Metal without that vocal _growl_ that comes primarily from vestibular vocal folds going smacka-smacka-smacka. Of course I know you're coming at this from the perspective of sound reproduction and loudening; just pointing out that the music production side of things loves distortion. Also, some people seem to appreciate about 10% harmonic distortion that emphasizes even order harmonics and perceive it as "warmth."
@@petersage5157 the wrong kind of harmonic distortion at 10%. It’s not all that nice warm tube, honey sound of second order distortion . It’s across the broadband can be described more like the white noise you would hear on an old TV when you would turn on the TV and there is no station broadcasting just black-and-white dots on the screen was a lot of scratchy static noise. It would be nice though if the more you turned a solid state piece of equipment up the more distortion, the more sounded like a tube amplifier . With second order distortion.
@@coldfinger459sub0 That's why we often use snubber capacitors to simulate Miller capacitance. A lot of high gain choob amps also have snubbers to tame the higher frequency distortion.
We all know fans fail. And usually to make them so they are not loud but yet give you a lot of airflow.. you usually have to pick a bigger fan than you then turn down the voltage to turn down to RPMs. To get the airflow that you need at a reduced fan speed to keep the sound level down. ( db.) So let’s just cut to the chase use some thermal probes attached to the hottest spot on the base of the chip to the heat sink. And let’s just figure out what size passive radiator heat sink. Do we need to operate this thing at max with no fan? Then add a fan turned on low . And when it fails, you never have to worry.. Sounds like a plan 🤔 Basically just treat this thing as if it was a Class A amplifier . And put the biggest heat sink it would need 100% duty cycle in the worst case scenario . Never have to worry have plenty of headroom.
Thanks for this feedback. It is true that the electromechanical parts will fail first. Basically anything that has moving parts is less reliable than semiconductor parts;) But to be fair - there are fans in everything - and I don't think I have every had to replace one;) The Fans I choose for these projects are pretty quite, and especially if there is any music playing - you just can't hear them. If you buy a powerful fan - than you can run it at a lower voltage to run at lower speeds where the noise is lower. But I choose fans that run quieter and are less powerful;)
Some of the pc fans have an external temperature probe, so that can speed i proportional to heat generated, i have used theese in audio amps with great Silent result
Eddie, outstanding video, thank you. The math was on point.
Thanks so much for your support!!
Hi Eddie, I really appreciate your knowledge and enthusiasm. Thank you for this informative video.😀
Thanks for the kind words!
Nice one, very informative as usual.
Thanks so much!
I enjoyed the video! The whiteboard is helpful for clearly stating the logical process as well as the math. The discussion of how the heatsink's thermal resistance varies with airflow was useful too. As noted, that can get messed up with a poor enclosure design.
Thanks for the feedback! That's the goal - clear and easy to follow.
Good info!
I appreciate you watching!
If I had a test bench like that, I would sleep in that room!
LOL - how do you know that I don't;)
One way people have got around not needing fans for cooling is to mount the PCB's vertically and having ventilation slots below, and above the PCB's.
Yes, this can help - but to get the full power out of these - you would need a very large heatsink;) The problem is - the heat is generated in such a small spot - that it it hard to spread it enough to get a heatsink by itself to do the trick. Conduction and relying on the chimney affect is probably not going to do it.
@@KissAnalog Maybe you can have a go at using a CPU watercooler as a heatsink. Might make a cool video.
You have read my mind;) I actually have one to show how well they work.
Love all your videos but especially musical amplifier related. Awesome explanations, just wondering is that why they use rms?👍
Thank you! They use RMS to relate it to DC, so for power RMS current and RMS voltage is used to find Power - but that is not RMS power it is just power at that point;)
Ti recommends to ground conductive heatsinks, so thankfully anodized parts are non conducitve. Usage of thermal pads is acceptable (0.5 mm thickness) or simply non conductive paste. It would be interesting, if an antenna effect of coupling heatsinks could be measured and if there's audible / electrical effect.
I prefer CPU heatsinks with many fine vertical fins, those chunky heatsinks with small surface area and horizontal orientiation will be sufficient too, but not ideal. Just allow good passiv airflow in the case. Take a look at the poor thermal design of many tin desktop amps with upside down mounted chip screwed to the case somehow. Example: The A07 max is a warm one, it works, but it's not a nice long term ambient for electrolytics in this box.
For harder driven PBTL @ 2-3 Ohm very slow running silent fans were added to the case. In stereo BTL config you will seldom use more than 5 W average power output in real life, cause this is the region when dynamic peaks quickly begin to sound harsh / distorted and become audible in the upper frequencies (5W +10 dB headroom for dynamics = more than 100 W impulse needed!) . So all that numbers in the specs are meaningless for Hifi applications. Anyway, most times you merely need more than 2 W RMS for private home usage and then there's plenty of headroom.
Thanks for the great feedback! I'll be testing a bunch of amps very soon and we will check their thermal designs;)
ever considered peltier devices for cooling.. it would need an internal thermostat naturally to avoid condensation
I'm really interested in peltier devices for cooling! I have couple devices that I plan to do a video about them coming up in the next couple of weeks.
More heat sink is always more betterer. I've never heard of an overclocker wishing they'd opted for a smaller heat sink.
10% distortion isn't the sound we want?! That's the point where an electric guitar amp just starts to get "lively," and there are actually quite a few genres of music that depend on much more distortion for their characteristic timbre. Try to imagine Modern Country music without a crunchy rhythm and chimey lead guitar, or the Star Spangled Banner without a Fuzz Face, or Big Band music without muted brass, or Death Metal without that vocal _growl_ that comes primarily from vestibular vocal folds going smacka-smacka-smacka.
Of course I know you're coming at this from the perspective of sound reproduction and loudening; just pointing out that the music production side of things loves distortion. Also, some people seem to appreciate about 10% harmonic distortion that emphasizes even order harmonics and perceive it as "warmth."
@@petersage5157 the wrong kind of harmonic distortion at 10%.
It’s not all that nice warm tube, honey sound of second order distortion .
It’s across the broadband can be described more like the white noise you would hear on an old TV when you would turn on the TV and there is no station broadcasting just black-and-white dots on the screen was a lot of scratchy static noise.
It would be nice though if the more you turned a solid state piece of equipment up the more distortion, the more sounded like a tube amplifier . With second order distortion.
@@coldfinger459sub0 That's why we often use snubber capacitors to simulate Miller capacitance. A lot of high gain choob amps also have snubbers to tame the higher frequency distortion.
Thanks for pointing that out! I guess I need to at least give a shout out to the musicians who might want to use this amp;)
We all know fans fail. And usually to make them so they are not loud but yet give you a lot of airflow.. you usually have to pick a bigger fan than you then turn down the voltage to turn down to RPMs. To get the airflow that you need at a reduced fan speed to keep the sound level down. ( db.)
So let’s just cut to the chase use some thermal probes attached to the hottest spot on the base of the chip to the heat sink.
And let’s just figure out what size passive radiator heat sink. Do we need to operate this thing at max with no fan?
Then add a fan turned on low . And when it fails, you never have to worry..
Sounds like a plan 🤔
Basically just treat this thing as if it was a Class A amplifier .
And put the biggest heat sink it would need 100% duty cycle in the worst case scenario . Never have to worry have plenty of headroom.
Thanks for this feedback. It is true that the electromechanical parts will fail first. Basically anything that has moving parts is less reliable than semiconductor parts;) But to be fair - there are fans in everything - and I don't think I have every had to replace one;) The Fans I choose for these projects are pretty quite, and especially if there is any music playing - you just can't hear them. If you buy a powerful fan - than you can run it at a lower voltage to run at lower speeds where the noise is lower. But I choose fans that run quieter and are less powerful;)
Some of the pc fans have an external temperature probe, so that can speed i proportional to heat generated, i have used theese in audio amps with great Silent result