Critiquing My Own PCB Designs - Phil's Lab #43

My first PCB design was in 2019 - let's look back and see how bad these were and how I've improved over the last few years! Featuring KiCad 6.0 and Altium Designer. (Apologies for the less-than-optimal audio!)
[SUPPORT]
Free trial of Altium Designer: www.altium.com/yt/philslab
PCBA from $0 (Free Setup, Free Stencil): jlcpcb.com/RHS
Patreon: / phils94
[LINKS]
GitHub: github.com/pms67
PCB Design Course Survey: docs.google.com/forms/d/e/1FA...
[TIMESTAMPS]
00:00 Introduction
00:24 Altium Designer Free Trial
00:47 PCB Design Course Survey
01:13 First PCB (Arduino Due Shield)
07:01 Second PCB (Update on Arduino Due Shield)
10:05 Third PCB (Teensy Shield)
11:15 First Four-Layer PCB (First STM32 Board)
15:53 Hades Flight Control System (Second STM32 Board)
21:43 My Current PCB Designs (2021)
ID: QIBvbJtYjWuHiTG0uCoK

[@JLCPCB]

We love your designs! Also your videos are very informative and helpful to many young engineers and students! 🥰👍

[@albertoprogida8977]

Can’t even put into words how much your work has helped me learn! I’m studying mechanical engineering but work on lots of avionics stuff and your videos have been an incredible resource! Excited for all the new tips coming next year. Cheers!

[@ZainAhmed]

@@alejandroperez5368 it's always good to have a wide array of skills. I know test engineers that learned to design PCBs just to make their wiring harnesses and sensor setups cleaner. Same way that learning python is helpful to people in many engineering sub fields.

[@PhilsLab]

Thank you very much, Alberto! Many more electronics/PCB design videos to come this year.

[@MrAnanas997]

See you at Euroc!

[@artrock8175]

Videos like these where you reflect on your own designs of the past, are extremely valuable for the "young players falling for the never ending traps" Dave Jones is always talking about. Thank you for taking the time to put this together and sharing it with the world! 🙏🙏

[@josephsanchez08]

This is where I’ve been lucky to have good mentors to help me review my work and learn from their mistakes. I appreciate that you review your work because it shows that you care and aren’t too prideful to learn.

[@InterFelix]

I would love to see a video of you going over do's and dont's when starting out with PCB design, essentially things to look out for, good practices and bad practices etc. Could also be something like "things I wish I knew when starting out with PCB design". This would be tremendously helpful to anyone starting out in the field and I haven't yet found any good video on this topic.

[@gammaleader96]

It is always a good idea to take a moment and look at what you did previously. It always helps me to see where I came from, what I did in the past and also to understand what effects some changed had. After all I think this is a great way of maximizing the learning effect, thanks for the video.

[@rahuls7039]

Again, such an in-depth video. Please make these types of videos every interval of a few months or so. Thank you for doing this, Phil. I am learning a lot, slowly and steadily.

[@tamasv1290]

Thank you for this video, quite inspiring to see how you started back then. I am also happy to hear you are working on the video course, I thought you abandoned that idea!

[@yaghiyahbrenner8902]

Thanks for this very good to look into the past and talk about improvements, honestly for a first time design your work is really good probably uphill from there.

[@MrZomhad]

Thanks a lot for making these videos, they help a ton! Would love to see a bit more on the ESC, it seems quite neat! Happy new year🙌

[@deangreenhough1178]

Great to see the improvement. You've become one of my favorite channels with Altium and the STM Cube IDE. Happy NY to you

[@PhilsLab]

Thank you very much, Dean!

[@adaminsanoff]

You have made incredible progress in such a short time. Bravo!

[@PhilsLab]

Thank you very much, Adam!

[@Cptnbond]

Thanks, Phil, an excellent idea to review your progress. I'm kind of on the same journey, challenging myself with more advanced designs. Cheers.

[@PhilsLab]

Thank you very much! Yeah, it’s always interesting to look back and see how one’s skills have changed (and hopefully improved).

[@piotrlenarczyk5803]

Thank you for video. Detailed revisiting of own inventions disadvantages prove the highest skill of professional design. Even the best design possess some drawbacks, and making them publish is quite courageous:).

[@PhilsLab]

Thank you for your comment, Piotr!

[@brus54per]

Very helpful to see and hear your comments and how your skills have improved. The "don't:s" are very helpful - as well as the "do:s". Thank you and my best wishes for a bit more normal 2022!

[@PhilsLab]

Thank you!

[@bartek153]

Vey honest and interesting Phil. I can see already where I can improve too! :D

[@vilius20001]

Excellent videos. Thanks for all of your hard work lad.

[@PhilsLab]

Thank you very much!

[@andershestad284]

Great insightful video! A cool video idea for the future could be to review viewer submitted PCB designs :)

[@ashnad1499]

Bro we love your design it is helping me a lot And thanx for all the free resources

[@RaymondBarakat]

It took me a second to realize how you set up the enable pins on the LTC3370 for your most recent Hades Z edition board. Really simple and elegant way to do power sequencing, just daisy chaining each voltage rail as they come up. Gonna have to remember that trick. Great work! Lots to learn from your channel!

[@PhilsLab]

Thank you very much, Raymond!

[@mouradmkhakh8903]

Good luck, i appreciate your work. Thanks for sharing.

[@PhilsLab]

Thank you, Mourad!

[@michajoko7534]

I would be very interested to see something about ESCs or about the design of H-Bridges and the thermal aspects.👍

[@mallikarjunse]

Looking forward to ESC! 😄

[@johnferrenby7486]

These videos are gem!!! Thanks a lot for work.

[@PhilsLab]

Thank you for watching, John!

[@coderebel3568]

I've learned a lot from you Phil, and still learning. Thanks a lot! One request is if you could dive a bit into the setup and debugging of audio codecs, like PLL and sampling frequency, and differential outputs and such. I've played around a bit with an AD1937 audio codec, which I chose because of it's many input and output channels. I also provided it with a 12,288MHz crystal I believe to get a precise sampling frequency. But I can't seem to get it to work. I discovered the I2S lines were ringing on the first design, so I looked at the schematic for the AD1937 discovery board and provided the same termination on the new version, but it's still not working. I've made the output and input stages modular, so I can provide differential inputs and single ended inputs as well as differential buffered outputs and a headphones output using a headphones amplifier IC (TPA6111A2). All it does is "cracle" a lot and giving some weird looping noise. I think one reason could be my separated ground planes for the analog and digital sections, but I did provide a jumper and ferrite bead to connect them together. It's a bit better if the ferrite bead is shorted so the grounds are connected together, but I'm hearing (or seeing on my scope) none of the sine wave I'm having my STM32F446 (on a Nucleo board) send to the DAC. One thing I did learn from your videos though (which I wasn't sure about though I must have tried it) was that I need to use signed integer values. But perhaps I should have tried with 16 bit on a 16 bit frame first.. Not sure if I've done that. I suppose you must have gone through a similar learning curve. I couldn't solve this puzzle with the AD1937 datasheet anyway.. So would you please consider making a video about some debugging and pitfalls when working with audio codecs?

[@mcorrive12]

These videos are amazing! Thank you for sharing!

[@PhilsLab]

Thank you very much, Matt!

[@ianthehunter3532]

What were the approximate dimensions of your first board? Also at 14:17 you said you didn't like the clearance between plane and traces, so how much clearance do you recommend for that?

[@maxim25o2]

I try my first designs, because I noticed that connecting all arduino modules apart, and dealing with cables, can be messy, I start trying design one board ready to use. Your videos are so acknowledging, and rather to make mistakes, like You did at beginning, because of Your videos, You helping me to make better design and learn good design practice.

[@Program2live]

Really helpful video, great progress! In the future or on the course can you go into CE, UKCA, RoHS, in particular self-certification (tools to use, dos and don't etc.), rather than sending it off to get it tested?

[@rick_er2481]

Yes do this please, would be awesome!

[@PhilsLab]

Thank you very much! I'd definitely like to go over something like that in future videos, especially w.r.t. CE marking.

[@youtubeviewer7077]

You mentioned you don't do thermal reliefs anymore. I understand they're to prevent toombstoning, but I've only ever hand-assembled PCBs (manually or with solder paste/heat gun) and haven't had those issues. I assumed it must be needed for professional assembly, but you get PCBs assembled on the regular, so I guess that's not the case? I plan to order some assembled PCBs soon and wondering if I can change that Kicad default to not have thermal reliefs.

[@snivesz32]

Looking at a data sheet I see the common mode choke specified close to the can transceiver followed by the terminator, does that make a difference?

[@Retinatronics]

Hi Phil, In your video around 14:23 you said thermal reliefs doesn't do anymore for smd components only for through hole. Can you elaborate on that point please?

[@JeanLucCoulon]

Thanks for the video. Twice, you said "the GPS didn't work". Can you tell why ?

[@warlock5784]

very useful drive shaft

[@frankbose544]

we all start somewhere i actually pulled up some of my fist designs and boy oh boy they were bad im looking forward to getting your udemy pcb design course tho so i can improve even further

[@abdhullahjr3822]

very nice stand

[@jenshaldrup4952]

With a 3.3V supply, a 100 ohm resistor is reasonable for a through-hole LED. They need more current than surface mount LEDs to look bright, especially if you want it to be clearly visible outside under bright sunlight

[@diskdrive123]

Floating decoupling caps in schematic is fine as long as it’s well defined which IC it’s for. Unless you make very spacious symbols.

[@dhruslab9563]

When are you releasing the EKF part ?

[@dhruslab9563]

Hello sir Your work is just amazing. I have a question You have connected NVIC IC to FLIGHT CONTROL IC through different communication protocol, I2C, Uart and few GPIOs also , could you please mention what data are you sending / recieving through I2C and what data are you sending/ recieving through UART and why you have connected few GPIOs in between NVIC to fc .waiting for your answer! Regards Dhruti

[@Jonas_Meyer]

14:20 Don't you get tombstone without thermal releases on SMD components?

[@GilsuShin]

Nice review!

[@PhilsLab]

Thanks!

[@robertmacharia3141]

Something i think i am a culprit of...the clearance between the traces. What criteria do you use to calculate it?? Or recommend a video

[@Piccodon]

You learn more when things go wrong.

[@robertpieper5142]

At 14:22 why don't you use thermals for SMD pads and just through hole? I would think the cap would tombstone if you had a thick non thermally-relieved plane tied to one end and a small trace on the other.

[@couplingconstant]

Thank you very much, Phil for all the videos you make. I somehow thought that you were doing PCBs for 10+ years, so I'm very surprised to hear that you started about 3 years ago. But I must say your first PCB looks pretty advanced to start with. :) I wanted to ask you about the JST connectors that you use. Could you please give the exact part numbers for the 2-4 pins ones that you use? They don't seem to have locks. Are they safe enough to use when there is a lot of vibration? Do you have any recommendations for designing PCBs that will be used in high vibration applications, other than the type of connectors to use? Do you solder things differently for example? Thank you.

[@PhilsLab]

Thank you very much for your comment! On the 'Hades' flight controller board, I'm using Molex PicoBlade connectors. JST connectors seem to be a bit harder to get - especially these days. For those I usually go with JST GH. Anything locking will usually do the trick, even though the PicoBlade only have a 'friction lock'. Haven't had any problems so far, we've even been using them in high-vibration environments for quite some time.

[@couplingconstant]

@@PhilsLab Thank you, that's very helpful.

[@jagadishk4513]

Few days ago i was frustrated with Kicad not being able to draw tracks with arc, was planning on using nightly build but what a Christmas present it is to get official stable version 6 with lots of features and one of them being able to draw tracks in an arc shape.

[@martinmckee5333]

@@alejandroperez5368 In what way? I have heard from many people who do RF that curved tracks are helpful when working at extremely high frequencies.

[@xue-yitian2596]

Thanks for your viedo. It's helpful to me.

[@PhilsLab]

Thanks, glad to hear that.

[@DiegoColl44]

Thanks for sharing..!!

[@rajdip2836]

i love your projects ❤️ can you make a tutorial on high amp bldc motor driver using stm32 with usb programable option. Thanks in advance.

[@ArtemShoobovych]

Out of curiosity: why some of the traces on your latest board (as shown at 24:02) are wavy? Is this for noise filtering purposes?

[@Foxhood]

I can explain. These are known as "Meanders" and they serve to delay the signal. See. Signals are not instant, they travel through the copper to their destination which introduces inconceivably tiny delays (it travels at 2/3rd the speed of light). When dealing with components communicating at extremely high speeds via parallel traces these delays can be big enough that you get errors as one signal arrives a fraction of a nanosecond earlier then the other. So normal practice when dealing with stuff that might be sensitive to such delays is to snake parallel signal traces till they are identical in length. aka meandering. nearly all PCB CAD software got built-in meander tools to keep traces identical down to a micrometer. Best example of where Meanders are used a lot is your PC's own motherboard. You can spot meanders by the hundreds in-between the CPU, Memory and PCI-E slots.

[@anotherwingover]

Is there a pcb design book you would recommend? What learning ressources did help to improve your skills?

[@thedrunknmunky6571]

I’d like to know about a good book too. Hopefully there’s someone who can recommend.

[@frankbose544]

in my experience there isn't a lot of books on pcb design its one of those things that is hard to teach other than books on emi and emc

[@PhilsLab]

My main recommendation is just to design as much as you can and have other people review those designs. Try to always go one step ahead - even if you're never gonna work with FPGAs, try laying out and routing such a board. Then going back to 'only MCUs' will be a doddle. W.r.t. book recommendations and other resources, I can say anything by Rick Hartley, and 'EMC for Printed Circuit Boards' by Keith Armstrong.

[@RectSouth]

Thanks for sharing！ Altium and Kicad, which one is better for use?

[@PhilsLab]

Thanks for watching! I prefer Altium (despite being sponsored) - but then again KiCad is great and most importantly, free.

[@user-tt1yl3ws4j]

You should keep fottprints in project directory so that you don't lose them

[@waynekeenansvideos]

Cool, thank you

[@PhilsLab]

Thanks for watching, Wayne :)

[@SpaceWolf011]

Found this really interesting, especially when comparing with my own first PCB design, which, thinking back about it was a completely insane endeavour. It was a flight control board (autopilot) with two STM32H7 controllers, IMU, GPS, a pretty powerful datalink module (1W RF output), input power regulation, SD card, a ton of peripherals (PWM input, output, RS232, RS422, RS485, CAN, Ethernet), ADC channels, barometric and differential pressure sensors, extra memory... Basically, it was an attempt to cram 70% of the electronics of an existing small UAV onto a single PCB for size/weight reduction and reliability improvement (less connections etc). The attempt included porting a previous 20000 line bunch of flight software (written for 4 MCUs) to the new 4-layer board (with 2 MCUs) and adding a bunch of improvements. Attempting this as a first PCB would be a bit overkill, but for me as a mechanical engineer (aeronautics) with some tinkering experience with microcontrollers, but none with designing electronics it was hugely optimistic, unrealistically so. And it worked! There were a few mistakes in the first version (footprint errors and one missed connection which was fixed with a wire later), the second version was OK. The software took the longest, with board bring-up being perform by me also for the first time. It was a very stressful experience, with every peripheral bugging me for a short (or longer) time, at the end finding out that the problem was of the sort:"It works better when you turn it on first". Took a while, learned a ton, got stressed even more, but it runs fine. P.S. The inner layers were done properly. The GPS active antenna in-line resistor too. But I still place holes in silkscreen occasionally.

[@JA-ur8ob]

Sounds like you have done exactly what I intend to... Great work man. Glad to hear it worked. -fellow mech engineer

[@alfaalfa99]

nice

[@DharmaSindhu]

SUPERRRRRRRRRRRRRRRRRRRRRRRR

[@hakanersoy9210]

Love from turkey:)

[@PhilsLab]

Thanks Hakan, greetings from Germany :)

[@sourabhkumawat4863]

that's me looking at my facebook posts from year 2012 and thinking why tf i posted it?😝😂

[@pugglez4798]

I am a mechanical engineer student just starting to work on my first PCB. and your videos are so helpful! I would be interested in seeing more of the firmware behind your projects. Keep it up! 👍