Class D Audio Amplifier Hardware Design - Phil's Lab
Вставка
- Опубліковано 7 чер 2024
- Class D audio amplifier basics, part selection, schematic and PCB design, and test. PCBs by PCBWay www.pcbway.com
[SUPPORT]
Hardware design courses: phils-lab-shop.fedevel.education
Course content: www.phils-lab.net/courses
Fedevel courses: www.fedevel.com
Free trial of Altium Designer: www.altium.com/yt/philslab
Patreon: / phils94
[GIT]
github.com/pms67
[SOCIAL]
Instagram: / philslabyt
[LINKS]
Afrotechmods Class D video: • Class D Amplifier Tuto...
Power amplifier classes: en.wikipedia.org/wiki/Power_a...
Inductor selection guide: www.ti.com/lit/an/sloa242a/sl...
TPA3116D2 IC: www.ti.com/product/TPA3116D2
[TIMESTAMPS]
00:00 Intro
00:27 PCBWay and Git Repo
01:01 Class D Amplifier Basics
05:25 IC Choice
07:36 Schematic
16:18 PCB
22:51 'No Load' Test
25:18 Guitar Demo
26:57 Outro - Наука та технологія
Thanks for watching! Next video well be a full KiCad 7 + STM32WB Bluetooth design!
Any project ideas (similar size to this Class D board) that you would like to see?
why not Altium Phil?
What about a smart guitar amp or something along those lines using a nrf microcontroller? Could be interesting
I've been meaning to update the long-form KiCad tutorials for a while (last one was two years ago), as they're very outdated in my eyes.
I'll also be making a longform Altium one start of next year!
Sounds cool! This class D amp + the previous guitar pedals, I'm actually planing on combining into a custom modelling amp.
Hi..Phil...Really nice design on Class D audio amplifier...Suggestion for next video - Quectel M66 based GPS tracker. For a long time your RF and GPS design videos has not been updated...is eagerly waiting for it...
You really made a half hour video just so you can show off your guitar skilles at the end xD I really liked the design, learnt a lot!
Darn you caught me haha!
Not only you can design the board, but also you can play guitar nicely. Gifted.
Audio projects are always great -thx
Thank you, Harald!
Ive been following your channel for a few years now(as an EE ive picked up quite a few tips) and as an music/audio lover im always stoked to see these videos. Many thanks Phil, cheers!
Thank you - very glad to hear that! Definitely more music + EE content to come :)
Im already in admiration of you EE/PCB design knowledge but hearing you shred at the end! Wow, you have my dream skill set 😂
Haha thank you very much!
Kinda late to the party but just wanted to say: Always a delight to see a notification from your channel Phil. Great as always. More audio projects please! haha can't get enough. Cheers!!
Thank you very much - loads more audio stuff to come!
Super interesting video. Impressed with the results and the demo at the end. Thanks!
Man, the timing of this video is incredible. I just thought about giving pcb design another shot and you happen to have a new video on exactly what i wanted to do.
Awesome, great to hear that!
I really enjoy watching your hardware design videos! Especially the easy to understand schematics, layouts and explanation are fantastic. Your videos have helped me boost my knowledge about electronics alot and got me some internships where i was able to useethat knowledge and show off my altium skills. Thank you so much!
Thank you! I'm very glad to hear that - congrats on getting those internships! :)
Big fan of the UA-cam channel and keep going with the experiments. We could easily replace guitar amplifiers with class D technology as the closed loop bandwidths can go up to 120Khz or more. Then its just a matter of maximizing the performance of the design along with the X-Y transformer/amplifier hysteresis. Then you go further on to decide how many switching levels you want, what switching speed your targeting for, and filter options such as post filter feedback..etc..
That is how I went down that rabbit hole :D
Have recently designed a DC switch power supply useing your vids as baseline advise and it turnd out pretty ok for a first time using kicad for PCB
Lovs your vids man
keep it up
Thank you very much! Great to hear that your switcher design turned out good on the first go!
Thanks for the tips and the great solo!
Thanks for watching!
Hell yeah, SHRED!
As usual great, informative video!
Good to see you QA403 arrived. Mine should be here this month :)
Thanks, Michal! Hope you get yours soon!
@@PhilsLab All this Class ABCEDFGHIJKLMNOPQRSTUWXYZ class amplifiers, are these just for audio or they all apply to any sort of signal?
The classification applies to pretty much any type of power amplifier (audio, RF, ..).
@@PhilsLab
Besides making sound louder, I can't think of any other reason why I would ever need to use amplifiers.
Thanks a lot for your videos. They are so informative
Thanks for watching!
Sounds great and you play guitar really well!
Thanks a lot!
Excellent coverage. I followed along while looking at datasheet. Your layout is much tighter than TIs, good job here. You've opted for a different thermal relief strategy based on experience 👍. The sheet says in 7.3.7, "In a single-ended input application, the unused input should be ac grounded at the audio source instead of at the device input for best noise performance." Not really possible with a guitar/pedal combo with various grounds floating around. Thanks for the video, I'll start using pours more often than wide traces.
Thank you very much, Craig!
Yeah, exactly, a bit difficult to implement that grounding strategy if the source isn't entirely controllable.
@@PhilsLab plus, its a guitar amp, so it really doesn't matter much.
Great video thanks, also the usage of the QA403 - mine still on backorder, hope getting soon.
Thank you, Rolf! Hope you get your QA403 soon!
Nice video, well done thanks for sharing it :)
I guess the most precious about audio design is the technique of laying out the pcb tracks in order to reduce unwanted interference to the maximum
I use a TPA3116 as a stereo power amp, which follows an ME-80 signal processor, so far it worked nicely. The only downside is with 50W of power, it's simple not lound enough. Looking at some discrete design now.
I never thought much about it but I guess an 89 dB s/n ratio for a power amp is quite normal irrespective if it's class AB or D or whatever? Certainly it makes you think a bit about fretting about material being captured at 16 bits as opposed to 24 bits..
Nice guitar work in the end :) i used to play with Yngwie in the 80s (played keyboards on the first five or so rising force albums)
Very cool! Thanks for breaking down the operating model for the class-d up front, it really helped with understanding the overall design.
How feasible would it be to combine USB-PD and a DAC + this kind of amp to create computer speakers? Guessing most motherboards don’t supply enough power?
Thanks, Matthew! Glad to hear that.
Some modern desktop PCs I believe have USBCPD-capable ports and most laptops these days as well, so that could definitely be a cool project. If only someone could do the mechanics haha..
Nice shredding!
Thank you!
Great video (and great taste in guitar gear) as always. Do you think you can expand this power amp to a full guitar amp, jfet preamp with cascading gain stages and a real tonestack from a vintage amp? As far as I know, the tonal characteristic of an amplifier (other than cab and speaker) comes from the tonestack. So in theory, you can create an amp that sounds like some historic amp only by implementing its tonestack.
Thank you, Cem! This is definitely gonna be part of a bigger, guitar amp project. Basically combining this with the preamp DSP frontend from previous videos, to make a portable guitar amp. As you say, the tone can then be changed - in my case in software - quite considerably by varying the preamp.
The cross talk and emission of these inductors will be a blast 🤐.
Well, I enjoyed that!
Thank you very much!
very nice video abou the class-d board. One thing I noticed are the thermals on the barrel Jack power supply, they seem too tiny. Is that enough to link the main rail track to the barrel jack? your lovely guitar sounds the cool new entry into the video!!!
Thanks! Given that a single 0.25mm trace can easily carry ~1A with a 10deg temperature rise, and the distance is very short for the spokes, the thermal reliefs for the jack are more than sufficient - also given 4 of them per pad. The total current draw is only around 1A at full power.
Thanks for sharing!
Maybe not so critical for a guitar amp, but interesting nonetheless: TI Paper "Selecting Capacitors to Minimize Distortion in Audio Applications" (Zak Kaye), describing how a cheap X5R input cap on these class D amp chips can create 30+ dB more THD+N instead of a through-hole film cap. Depends on frequency and other parameters of course, but highlights details that distinguish an audiophile amp from a standard amp.
Thanks for watching! Yes, which is why I'm using tantalum caps here on the input (although not optimal, but significantly less expensive than film). As you said, given that this is a guitar power amp, this selection isn't too critical in any case. In case you're interested, I have a video going in more detail on the up- and down-sides to class 2 MLCCs here, which describes where some of these distortions come from: ua-cam.com/video/kUruN6WBgSw/v-deo.html
As always very nice video.. I've done recently a TPA3255 based class-d amplifier with a PFFB and the sound was AMAZING! But I've a question please.. would it be better to put the electrolytic capacitors on the top of the board and place the IC on the bottom side so your power loop inductance will be minimized even further? Because I've done this trick on my board.. thank you!
Thank you! Not sure how much you actually gain from doing that - also, then you're paying for double-sided assembly, making testing/debugging more difficult, etc.. Maybe this could make sense for far higher power - but (at a guess) same layer will probably be sufficient in most cases.
Can you use kicad for pcb design?
Phil out of curiosity, do you manually place the crosses over the DNF components in the normal schematic view, or is there a way to show it without entering the "variant editor"?
For very small projects, such as this one, yeah I place them manually and set parts to No BoM. For larger projects, I use variants or dedicated DNP library components.
Great project! As always, appreciate the breakdown. Are you into symphonic metal? :D
Thank you! I used to play quite a bit of metal, these days more 80s hard rock, some Malmsteen, some Dream Theater (I try at least 😅)..
Thanks for another excellent video.
Question for you regarding power supply. What is your take on using one of these usb-c chargers - that start to accumulate unused at home - together with a PD chip on the board for an amplifier like this? The common ones can supply 60W and up to 20V that seems about right... Would such power supply be too noise? Would the filtering caps that you used in your design be enough to get decent results?
Thanks, Marco! For a 'proper' USB-C PD-powered amplifier design, I would make provisions for line filtering before feeding to the pre-/power-amp. This exactly is actually part of a larger project (digital guitar amp) that I'm working on, which will also feature on the channel. I'll be able to go into more detail in that video!
damn , that play was epic .
Thank you very much!
Looking forward to the discrete Class D amplifier design video because I mainly work with high-power audio amplifier designs and Class D amp chips for high power do not deliver what they promise.
Thank you - should definitely make for an interesting project!
I have a question that popped into my mind watching the start of the video explaining the basics.
I get the Low Pass filter, but in order to fully utilize the swing of the speaker coil you'd want also a high pass filter (let's assume -6dB@50Hz) to get rid of the DC component.
If you i.e. run it form 0 to 15V you'd lose the -15V range but if the resting potential was -15V any silence would burn up your speakers. I know that i also could be very wrong, bc. i'm stoopid :-D
For clarification, i do electronics professionally, but i never had any in depth contact with audio stuff.
@PhilsLab Did I miss the input impedance being set to 1M+ for the guitar input, or are you loading the pickups a bit?
Guitar is plugged into my audio interface (high Z instrument in), as shown in the video, through Guitar Rig, out again via the audio interface to the Class D amp board.
@@PhilsLab ah right, fair enough. If you were going to directly input the guitar would you have to preamp the power amp ic?, or can it take the 700mv or whatever the swing is on your guitar output, and amplify it without causing distortion, I have zero experience with class D. Thanks for responding!
Fantastic
Thank you very much!
Once I played around with simple class D on FPGA. It is really fascinating that any signal can be generated by single digital pin.
However main frequency should be far more than actual signal frequency, which is limiting factor.
Definitely a cool project to try out on an FPGA!
@@PhilsLab how?
I'd really like to see a power delta-sigma modulator!
@@VEC7ORltthat’s basically how D-class amplifiers work, especially the better ones are mostly just sigma delta modulators
Hi, does anyone know how to display not fitted components in editor's tab as in Phil's video at 9.47? He also doesn't have any variants for this project, I am a bit puzzled how he was able to do that. Anyways awesome video
If you want to use it as a guitar amp you should have shown on the scope how it distorts when overloaded with a pure sine wave. Another good test is how it amplifies a square wave.
On sine it just chops off the tops, and square wave is kinda meaningless.
Why you choose 1u tantalum caps at the audio inputs? What about the polarity?
I try to avoid larger class 2 MLCCs in series on the audio path, as I've experienced problems with microphonics in the past. For guitar, tantalums have been a good/cheap compromise, so I don't have to go with film caps for example.
Polarity is that at DC the 'anode' is attached to the higher voltage (in this case the DC biased input pins).
An Aluminium substrate PCB might be the way to couple the heat into a heatsink.
Cool ❤ both electronics and guitar. You should try some guitar staff on UA-cam too. 😂
Thank you haha yeah maybe have to pivot the channel to guitar :D
What I missed or do not understand is; you are connecting a balanced signal to a mono jack, right? This goes to you speaker cabinet. I thought these are speaker connections where mono unbalanced. At least if it was balanced wouldn't you use a balanced signal with ground over a TRS?
What about some DSP implementation in FPGA video?
Definitely on my list of videos to make!
I've been meaning for a while to design a class D amp with I2S digital input and feed it directly from FPGA. I wonder what kind of latency this system will have.
Sounds like a cool project! Let me know how it works out, when you get round to it!
Very nice playing and pcb design! IMO, damping factor of class-d chip amp is way too high for guitar to sound good, way less “punch” than a push pull pentode tube amp. Do you have any idea to lower the damping factor of 3116? Cheers!
Thank you very much! To be honest, I love valve amps but I quite like the 'simplicity'/ease-of-use of solid-states. Regarding damping factor, I believe it's rather difficult to measure output impedance and thus damping factor, as these are non-linear, switching devices. You'd have to linearise about some operating point, calculate output impedance, and then compare that to other ICs (or what other factors influence output impedance).
Amplifier damping factor (DF)is defined as “the ratio of the load impedance (loudspeaker plus wire resistance) to the amplifier internal output impedance.” This basically indicates the amplifier's ability to control overshoot of the loudspeaker, i.e., to stop the cone from moving.
I don't see why you'd want to lower it, but the TPA3116 has an output Rdson of 0.12Ω meaning about 0.25Ω total, which gives the damping factor of 16 with 4Ω speakers and 32 with 8Ω speakers.
If you'd want to lower the damping factor, you can always put a resistor in series wit the output.
If you put a 3.75Ω in series, you'll get a damping factor of 1 with 4Ω speakers.
@@dedamarsovac "the amplifier's ability to control overshoot", right, you are very correct. That's what guitarist call the "punch", it's basically speaker cone overshoot. We want to lower the DF to simulate the tube amp overshoot. Most guitar amplifier has no negative feed back thus the overshoot is wild, for example, Fender 6V6 amplifiers and ALL EL84 push pull amplifiers.
can make esp32 a1s with tpa3116 pcb?😁😁
in case of it will drive 50W on 8 ohms speaker - output AC amplitude should be about 28 V.. how will it bring 15V to 28V? is there integrated boost coverter in IC? if yes - where is the bulky inductor coil?
This is the 15W version of the IC. The higher power versions (e.g. 50W, different part number) use higher supply rails and in any case, will also use bootstrapping. The datasheet has more details.
Phil, when do you sleep? Great content as always...
Thanks, Isaac - I'm afraid not much these days haha..
Thanks, always get some loose ends
Thanks for watching!
I notice you use a lot of thermal reliefs. One that stuck out to me was a thermal relief on a powerpad ... I dont know that i've ever seen that before. I suppose you are trying to keep the board hand solderable, right? Great video, but I found that intriguing, so thought i'd ask, as knowing you, it is not without reason.
Yeah, if you're just gonna reflow, then the thermal relief on powerpad shouldn't be there. However, given my limited soldering equipment (iron, hot air gun - but no reflow oven or hot plate), I wanted to keep my options open if I do wanna go the hot air gun route. Given the amplifier's high efficiency and comparatively low power output, in this case, having the thermal relief wouldn't be particularly problematic in either case.
If you prefer EL34 to KTxx you might like MA12070 more than TDA...
Thanks for the tip, I'll check those out!
Steve Vai's hands there!
Haha thanks, I wish!
Ever done any tube audio stuff?
That's actually next on my list - have purchased some low voltage valves and will be making a guitar preamp!
@@PhilsLab Awesome! What tubes are you using? I used some 12AU7s on my preamp stage. I built my tube amp with point to point wiring ( sorry no PCB :( ). Very satisfying project and made me think a little differently on electronics than I usually do, especially when it came to what I consider "aesthetic".
Electronics is right, but the real question is: wich string gauge and string action u use
Nice video as always!
Haha thank you! Lowest action possible, and typically 9-42 gauge strings.
@@PhilsLab I go for 9-46. Hopefully the amplifier work for that gauge 😂
Play at 0.75% to understand
👍🙏❤
woah, wen collab w/ Max Ostro? :p
Hadn't actually come across him before - will need to check his playing out!
You should've placed the guitar part in the beginning 😅
Haha yeah, I should someone else to do the editing.. my choices are often questionable :D
Deerfield I was wondering if you could do me a favor and maybe help me with an issue I'm thinking about upscaling E-Waste repurposing and recovery as opposed to just burning it up and recovering the metal because I live in a country where the E-Waste is thrown away are fully functional MacBook pros and New Generation in micro laptops and desktops and I think it's absolutely ridiculous to burn them for their metal and it would be much better to repurpose what can be repurposed and then you know dispose of the rest but most of the stuff can still run manex super fast and is absolutely good stuff for countries where the classrooms need computers just to be competitive in the world. So I was thinking of an SSD design where you recover all the storage from old thumb drives and old sticks and your repackage them into strips by type and buy a generation and then have an SMA automatic calculator place them on a PCB that has a small fpga and USB connector as his required by the EU and use a double-sided bored and have the fpga to a diagnostic of all the chips that have been placed and don't map any of the faulty or potentially failing sectors. In addition run kind of a raid array on what is there just so that they're reliable and it's all accomplished on board on the fpga. If you can order me the removal and replacement as well as the sorting and packaging it could really save and help and reducing what gets burned and what can be reused. The FBA GA would run the diagnostics and form a raid array of what's there to provide redundancy and resiliency considering their second hand. Also to buy a any cheap but reliable copies from any Source already enrolls to produce reliable maybe not the fastest or the best on the market but good solid working drives made from E-Waste and where I live they throw away the best stuff in the world I mean I recovered out of one bin that my neighbor who has a company lets me fish in and I recovered two working Apple MacBook pros and all kinds of stuff from usable and security resilient switches and Poe switchesand I even recovered a security device that I think will accept pfSense as a firmware upgrade. So just for me putting this stuff to use instead of having to end up in a waste disposable or put in a kiln or ship to a third world country where it talks about the people and the environment just to recover the small amounts of metal that remains in them is still going to happen but the last thing we send in the morning material we can recover and place back into use for education in schools and you know the power supply from The Towers can easily be turned into very safe and very accurate and precise desktop power supplies with a simple component from of course China but it could easily be redesigned here but they just haven't you know the connector for the board right there you snap it in the even include included fuses and then I'm going to take off the switch they have on there and put a proper switch and put a voltage regulator
Cool, but connectors on 3 sides - come on, also whats the point of thermal relief under the ic - during reflow its all the same, but now we are not using the top layer for cooling, my choice would be moving vias just beyond the pad, not using relief - no wicking problems and probably better cooling.
15v power is weird as well - 24v is better and more common and you can still use anything 15-24v.
Connectors on 3 sides, although not optimal, was just for aesthetics (or maybe I've designed too many guitar pedals...) - in this case for a test PCB, it doesn't particularly matter.
Agreed on thermal relief, I forgot to remove the rule on top at the EP. Regarding vias in the EP, I've had no problems with this method and wicking (0.25mm-0.3mm drill) - done many, many boards and ICs that way.
15V vs 24V - sure, 24V is more common but then again this is a test board for a work-in-progress USB-C PD based guitar amp running specifically off 15V.
@@PhilsLab USB PD would be very interesting to see.
#DemoTooShort:(
class-D amps sound like crap, especially for guitar