Do what Phil said! Watch the recommended video from Rick Hartley! Then watch it again. It's worth your time, and it's mind-blowing. Thanks for the recommendation, Phil!
Can I just say that I very much appreciate you making this kind of content? I am still honing my electronics-skills and trying to build the skills for dealing with SMT-components and custom PCBs and, well, you're practically the only TH-camr who does more complex stuff than "custom PCB blinking an LED, using only through-hole components as well." I don't have the possibility of getting formal education on this, so I am left to rely on learning from whatever free sources I can -- my goal is to eventually design my own, reliable Linux-capable SBCs.
Thank you very much, Nita. I'm glad to hear that the videos have been helping. And yes, I also think it's important to go beyond the whole 'blinking an LED' thing - so hopefully more 'advanced' videos to come :)
You sir are an internet gem. As an undergraduate deep diving into my own personal projects and projects in industry, I cannot express in words how valuable your videos are - they get me hitting the ground running faster and really allow me to start building intuition with way less stress and time consumption. I appreciate you!
Rick's video in differential pairs is very interesting. There he recommends treating them just like regular traces 3H spacing between them, same 50 ohm impedance (or whatever is applicableto the system), and he says he's never needed to make them equal lengths (within reason of course).
Rick Hartley makes some of the best and most informative presentations out there for PCB design. So to everyone watching this video, find a presentation by Rick Hartley and watch it ASAP! You won't be disappointed! Also, check out the presentations by Eric Bogatin, totally worth a watch! I have noted that the layout engineers at work will always add via fencing around the board with quite generous spacing (20h-rule). Also, they often route high-speed signals on an inner layer when they have a 10-40 layer board, keeping the top and bottom layer for components only.
Great job. At last someone talks about electronics in greater detail. I can’t begin to say how frustrating it was to ask questions about these sort of subjects to the teachers and either get an odd look or a change of subject. You are helping me to understand things that I had lost hope in getting a grasp on. Hat’s off to you, sir.
@@PhilsLab more than helpful, they are allowing me to up my game. I’ve got fundamentally nothing against academia, but for my personal purposes I have no use for a diploma. All I’m interested in is the knowledge and to be able to problem solve. Sometimes you go a bit fast for me but I can always go back in the videos and catch up. I see you explaining every detail without going off in useless tangents, which helps a lot to someone who is learning a new subject. It’s mentally draining when a teacher does that. So overall, two thumbs up. The majority of teachers could learn from you, it would help their students to make better use of the time they spend at school. I haven’t gone through the entire collection of your videos but just in case you haven’t approached that, if I may suggest, an explanation about lab methodologies to help your viewers get better organised, what software tools you use, templates for projects, how to determine what is the purpose of the exercise they are conducting and stick to it ( ie you put a device in a jig and run a test. Sometimes people are testing the jig instead of the device, or vice versa). That sort of content can be incredibly useful in helping people to avoid pitfalls which waste time, resources and wear out one’s patience, therefore not being productive at all. Again, thank you for all your contributions, your channel is definitely a keeper.
Thanks again, Alexandre. I agree - my engineering degree taught me next to nothing about practical engineering, let alone PCB design. Therefore my hope was to alleviate that gap a bit with this channel - so I'm very glad to hear you say that it's been useful. Great ideas for future videos as well - I'll make sure to make some content on those topics as well, thank you!
I've spent the past month watching as much SI videos as possible. Though this video doesn't go into detail in literal everything (like Rick's video does), this is an excellent summary containing the most crucial tips for significantly upgrading your PCB design skill.
you mentioned that there are more information in books..what books did you read to be able to do such awesome and professional designs? can you tell me the titles
Hey Phil! Firstly I must say that the content you create is precious and makes a difference for many people, especially for electronics amateurs. Can you explain why we at all need two adjacent ground planes? Would it not be the same if we had 3-layer board with a single ground plane in the middle?
this is a bad habit to use single gnd for 2 signal layers because you enforce mixing return paths of hispeed containing traces, so don't do this. 4 layer is cheap and reliable solution .
Would be cool to see a follow up with a comparison of signal integrity where you include/don't include your techniques! Ok it is a lot of work but would be very interesting. Thanks for the informative video.
Great video, it will be very useful when I will start routing my ADC board. Any reason why you chose a microstrip design over a CPWG for impedance control?
Great Video! I get wildly different values for trace width and spacing between online impedance calculators. For instance, EEWeb using the settings for the JLC7628 4 layer stackup microstrip gives me Width: 0.41mm and Space: 0.2mm for 90ohm diff. Yet JLC gives me, Width: 0.254mm and Space: 0.19mm. Vastly smaller than EEWeb suggest. Thats a huge discrepancy! . EverythingRF gives closer to JLC, Width: 0.267mm, Space: 0.19mm. I guess I'll ignore EEWeb even though thats been suggested to be the better calculator.
Is it possible to make some video on how to handle PCB design where you have to deal with multiple voltages on one PCB board, i.e. DC/DC converters, BLDC drivers etc. In such applications very often you have to integrate high power and logic parts into one board.
Great video. Im currently making my own USB 3.0 Hub as a hobby project. I saw that you placed test pads directly on the signal path. I know I should avoid stubs any other recommendation for test pads so I can later measure signal integrity? Also at 9:45 you have a Note about time and delay. I which cases does the physical length actually differ to timing, delays?
(not my words, but i thinks spot on for what your asking) Determining length requirements requires an understanding of flight time, electrical length vs. physical length, loading and signal quality. Those elements are vital in determining what the length really needs to be, as well as in determining the allowable trade-offs to meet system timing goals. Search web for eetimes and "meeting-timing-specs-on-boards-with-picoseconds-of-margin" some good info here
Thank you, Jonas! The test pads I placed only on 'low-speed' lines. Timing differences/delays occur when routing on different layers (i.e. inner layer vs outer layer) or when going through vias.
Hi Phil, what about for an LED display panel with 48 LED drivers and 760 LEDs with LEDs on one side and surface mount components on the other? It is difficult to have a clock trace going to each LED driver without switching layers and keeping the clock trace short as the LED drivers need to disperse across the PCB, a power plane would also be useful for LED voltage.
Good luck with your interview! Yes, I'd like to make more RF videos - however, I usually base the videos on small, example projects (and I haven't had many RF ideas so far!).
great video. I tried to use the same calculator but could not find out the Prepreg material type and its Df value on the JLCPCB calculator page. Can you please share you have found those two pieces of information?
Awesome channel dear. Go ahead! I’m very excited to get the training sessions with you about advanced PCB routing skills (on Udemy for example). Greetings from Lebanon
Hi Phil, very good video. Is it possible to design a microcontroller interfacing with RS485 chip, and tell about crosstalk, differential routing, and impedance matching with the layout design
Amazing video! When you talked about 50 Ohms for single ended signals: can this be used as a starting point for all kinds of signals? The reason I'm asking is that I've been trying to find out what I should aim for to get the maximal performance on a relatively long SPI connection for ensure signal integrety.
Thank you, Jacob! Yes, that's actually my starting point for 'all' signals. I'll typically route signal traces as 50 Ohms (even things like I2C, UART, SPI, etc.).
The 50 ohms calculator assumes the skin depth of the wire is approximately on the surface, so I2C is a little slow for 35um 1oz copper. SPI can definitely get fast though. Doesn’t hurt on 100Mbps, but you are only going to lose 1.2dB/meter…
Two questions: The design Phil uses is called a microstrip line. But there are alos designs that uses ground planes on the top, like coplanar waveguides. Is it better to use microstrip lines and use multiple layers for the signal ground? Or would it make sense to use the coplanar waveguide and add GND areas on the surface?
Hello Sir, can you make a video about the Flight Controller? Like a 2 3 hour video going through the schematics and PCB routing. Vielen Dank für die bisherigen Videos und danke im voraus.
Thanks for uploading another great video! Phil do you think you could make a video of how to use CircuitMaker 2.0? It was released just a few weeks ago and appears to be very similar to Altium Designer, just more limited.
Thank you very much! I actually haven't used CircuitMaker yet. I'm afraid for the forseeable I'll just be sticking to Altium and KiCad (also for these videos).
@@PhilsLab I downloaded it and I'm currently testing it. Altium says it uses the same AD engine, but of course they have stripped some of the nice features of AD. But being free, its not bad at all.
impedance for transmission lines assume infinity length. For example impedance between signal and ground might be 50 ohms assuming the transmission line continues forever. You then terminate it with a 50 ohm resistor which looks exactly like an infinite long transmission line.
Hey Phil. thanks for the content. I'm confused by you not having the third layer as a power plane. Would you route power or the signal layers in the stack you described here?
Thank you, Brian! I'll try to bring more detailed videos in the future. Still currently working on the PCB design course which will include a lot of info!
Hi Phill, Great video as usual! Are you going to make a course which we could buy (for example on udemy or orther sites). I feel like there are a lot of us that would like to learn something from you. Not only course for beginners but maby also something more advanced. You mentioned about it a few months ago so i'm looking for an update :)
Hi Kamil, Thank you! Yes, I'm still working on a course that will cover many aspects of hardware design. It's a lot of work so will probably take me a few more months I'm afraid :)
Hi Phil Love the Video! Would you be able to post board in the video on GitHub? I’ve been meaning to get into high speed design. Also when are the audio board/ fpga videos coming out? No rush just curious.
Hi Pietro, Thank you! I'm afraid it's a company board, so I cannot post it on GitHub. I've recently designed a Spartan 7 board with DDR2 memory that I still need to order/assemble. But once that's done, then I'll make a video about it :)
@@PhilsLab Thank you Phil, I have one more question too. I recently designed a board with a 5V regulator where the 5V gets filtered through an inductor coming from a buck regulator before it is dissipated as 5V. I need to draw about 2.3A but now I am a little worried because one of the traces I used coming off of the buck was 0.5mm. I used a 0.75mm to go into the inductor and a big copper pour and a 5V inner layer but I am just worried that the current coming right out of the inductor, 2.3A max would be a lot for the 0.5mm trace. Do you think that trace is too thin or do you think it’ll be alright considering the other 0.75mm trade and the copper pour at the output of the inductor?
@@pietrogagliano4484 I think you'll be fine. It may run a bit warm as according to a calculator (for example, the KiCad track width one), a 0.5mm wide trace with 2.3A running through it will give a temperature rise of 'only' 30 degrees.
@@PhilsLab Ok thank you! Yeah I have beefed up the traces for a later iteration, it’s a board I’ve been designing for work and these are prototypes so the final iteration will be rather beefy with the power traces to provide a good safety measure I just forgot to look that over as I was designing rather quickly for testing stage
Does the JLCPCB impedance calculator actually use a field solver? Or is it a conventional formula-based calculator with a ripped off image from Polar? ;)
@@PhilsLab No problem, I messaged you on you web site about the headphone amp and haven't heard anything, is the site still active or did I mess up ? !
Do what Phil said! Watch the recommended video from Rick Hartley! Then watch it again. It's worth your time, and it's mind-blowing.
Thanks for the recommendation, Phil!
Can I just say that I very much appreciate you making this kind of content? I am still honing my electronics-skills and trying to build the skills for dealing with SMT-components and custom PCBs and, well, you're practically the only TH-camr who does more complex stuff than "custom PCB blinking an LED, using only through-hole components as well." I don't have the possibility of getting formal education on this, so I am left to rely on learning from whatever free sources I can -- my goal is to eventually design my own, reliable Linux-capable SBCs.
Agree
Thank you very much, Nita. I'm glad to hear that the videos have been helping. And yes, I also think it's important to go beyond the whole 'blinking an LED' thing - so hopefully more 'advanced' videos to come :)
You sir are an internet gem. As an undergraduate deep diving into my own personal projects and projects in industry, I cannot express in words how valuable your videos are - they get me hitting the ground running faster and really allow me to start building intuition with way less stress and time consumption. I appreciate you!
Thank you so much for your kind comment. I'm very glad to hear that the videos have been helpful!
I would recommend Rick’s other video ‘What your differential pairs wish you knew”
That was a revelation
I watched this on your suggestion. Between the grounding video and the differential video I feel like I have had a revelation in high speed circuits!
Am I the only one that wishes that not too many people should watch this video so I can stay ahead of the competition? ;)
Rick's video in differential pairs is very interesting. There he recommends treating them just like regular traces 3H spacing between them, same 50 ohm impedance (or whatever is applicableto the system), and he says he's never needed to make them equal lengths (within reason of course).
Rick Hartley makes some of the best and most informative presentations out there for PCB design. So to everyone watching this video, find a presentation by Rick Hartley and watch it ASAP! You won't be disappointed!
Also, check out the presentations by Eric Bogatin, totally worth a watch!
I have noted that the layout engineers at work will always add via fencing around the board with quite generous spacing (20h-rule). Also, they often route high-speed signals on an inner layer when they have a 10-40 layer board, keeping the top and bottom layer for components only.
Great job. At last someone talks about electronics in greater detail. I can’t begin to say how frustrating it was to ask questions about these sort of subjects to the teachers and either get an odd look or a change of subject.
You are helping me to understand things that I had lost hope in getting a grasp on. Hat’s off to you, sir.
Thank you very much, Alexandre - glad to hear that these videos are helpful!
@@PhilsLab more than helpful, they are allowing me to up my game. I’ve got fundamentally nothing against academia, but for my personal purposes I have no use for a diploma. All I’m interested in is the knowledge and to be able to problem solve. Sometimes you go a bit fast for me but I can always go back in the videos and catch up.
I see you explaining every detail without going off in useless tangents, which helps a lot to someone who is learning a new subject. It’s mentally draining when a teacher does that. So overall, two thumbs up. The majority of teachers could learn from you, it would help their students to make better use of the time they spend at school. I haven’t gone through the entire collection of your videos but just in case you haven’t approached that, if I may suggest, an explanation about lab methodologies to help your viewers get better organised, what software tools you use, templates for projects, how to determine what is the purpose of the exercise they are conducting and stick to it ( ie you put a device in a jig and run a test. Sometimes people are testing the jig instead of the device, or vice versa).
That sort of content can be incredibly useful in helping people to avoid pitfalls which waste time, resources and wear out one’s patience, therefore not being productive at all.
Again, thank you for all your contributions, your channel is definitely a keeper.
Thanks again, Alexandre. I agree - my engineering degree taught me next to nothing about practical engineering, let alone PCB design. Therefore my hope was to alleviate that gap a bit with this channel - so I'm very glad to hear you say that it's been useful.
Great ideas for future videos as well - I'll make sure to make some content on those topics as well, thank you!
Hi Phil, love your videos. Waiting for your CAN Bus Firmware Design, as I'm having some issues regarding CAN Bus.
So happy to see someone else using Altium. This makes it easier for me to follow what you are doing.
Many more Altium videos to come very soon!
I've spent the past month watching as much SI videos as possible. Though this video doesn't go into detail in literal everything (like Rick's video does), this is an excellent summary containing the most crucial tips for significantly upgrading your PCB design skill.
Thank you very much, Jako!
New video...I dance round the room :)
Haha, thanks Mark! :)
Altium has a nice automated stitching mechanism too , when you want to stitch a plane or a copper pour.
These videos are awesome. Such a great learning resource!
Thank you very much, Guillaume!
Thanks Phil Phil's Lab always gives the best ttip
Still hopeful for the course you gauged interest on earlier this year. Great education as usual.
you mentioned that there are more information in books..what books did you read to be able to do such awesome and professional designs? can you tell me the titles
Your videos are just amazing. Please keep up with your great work! I'm looking forward to your content and PCB design course
Hey Phil! Firstly I must say that the content you create is precious and makes a difference for many people, especially for electronics amateurs.
Can you explain why we at all need two adjacent ground planes? Would it not be the same if we had 3-layer board with a single ground plane in the middle?
this is a bad habit to use single gnd for 2 signal layers because you enforce mixing return paths of hispeed containing traces, so don't do this.
4 layer is cheap and reliable solution .
where do you even get 3 layer PCBs? a 3 layer pcb is always just going to be a 4 layer board with an entire layer etched away.
Thank you. Very useful. Love those understandable tips that come with proper reasoning.
I sense Rick Hartley may have inspired some of this content ;)
Would be cool to see a follow up with a comparison of signal integrity where you include/don't include your techniques! Ok it is a lot of work but would be very interesting. Thanks for the informative video.
Great video, it will be very useful when I will start routing my ADC board.
Any reason why you chose a microstrip design over a CPWG for impedance control?
Great Video! I get wildly different values for trace width and spacing between online impedance calculators. For instance, EEWeb using the settings for the JLC7628 4 layer stackup microstrip gives me Width: 0.41mm and Space: 0.2mm for 90ohm diff. Yet JLC gives me, Width: 0.254mm and Space: 0.19mm. Vastly smaller than EEWeb suggest. Thats a huge discrepancy! . EverythingRF gives closer to JLC, Width: 0.267mm, Space: 0.19mm. I guess I'll ignore EEWeb even though thats been suggested to be the better calculator.
Thanks for another valuable video! I can't wait for full PCB design course, you promised.
Thank you, Marek! The PCB design course is still in the works :)
Thanks a lot Phil for videos like these! Much appreciated 👍
Thank you very much for watching!
great stuff, thanks for the video
Thank you 2 years later
Is it possible to make some video on how to handle PCB design where you have to deal with multiple voltages on one PCB board, i.e. DC/DC converters, BLDC drivers etc. In such applications very often you have to integrate high power and logic parts into one board.
Great video. Im currently making my own USB 3.0 Hub as a hobby project. I saw that you placed test pads directly on the signal path. I know I should avoid stubs any other recommendation for test pads so I can later measure signal integrity?
Also at 9:45 you have a Note about time and delay. I which cases does the physical length actually differ to timing, delays?
(not my words, but i thinks spot on for what your asking) Determining length requirements requires an understanding of flight time, electrical length vs. physical length, loading and signal quality. Those elements are vital in determining what the length really needs to be, as well as in determining the allowable trade-offs to meet system timing goals.
Search web for eetimes and "meeting-timing-specs-on-boards-with-picoseconds-of-margin" some good info here
Thank you, Jonas! The test pads I placed only on 'low-speed' lines.
Timing differences/delays occur when routing on different layers (i.e. inner layer vs outer layer) or when going through vias.
This is a great video; thank you Phil!
Thank you very much!
thank you very insightful 👏
Hi Phil, what about for an LED display panel with 48 LED drivers and 760 LEDs with LEDs on one side and surface mount components on the other? It is difficult to have a clock trace going to each LED driver without switching layers and keeping the clock trace short as the LED drivers need to disperse across the PCB, a power plane would also be useful for LED voltage.
Thanks Phil , I Learn a lot from you . I 'm going to have an interview with Aerospace and defense company. Can you make more video on RF ? please :)
Good luck with your interview!
Yes, I'd like to make more RF videos - however, I usually base the videos on small, example projects (and I haven't had many RF ideas so far!).
great video. I tried to use the same calculator but could not find out the Prepreg material type and its Df value on the JLCPCB calculator page. Can you please share you have found those two pieces of information?
Awesome channel dear. Go ahead!
I’m very excited to get the training sessions with you about advanced PCB routing skills (on Udemy for example). Greetings from Lebanon
Thank you very much, Omar!
Hi Phil, very good video. Is it possible to design a microcontroller interfacing with RS485 chip, and tell about crosstalk, differential routing, and impedance matching with the layout design
Nice video, thanks :)
Thanks a lot Phil totally usefull tips!
Thank you, Manolis!
A fantastic video. Thanks very much!!
Thanks, Scott!
Amazing video!
When you talked about 50 Ohms for single ended signals: can this be used as a starting point for all kinds of signals?
The reason I'm asking is that I've been trying to find out what I should aim for to get the maximal performance on a relatively long SPI connection for ensure signal integrety.
Thank you, Jacob! Yes, that's actually my starting point for 'all' signals. I'll typically route signal traces as 50 Ohms (even things like I2C, UART, SPI, etc.).
@@PhilsLab Thanks!
The 50 ohms calculator assumes the skin depth of the wire is approximately on the surface, so I2C is a little slow for 35um 1oz copper. SPI can definitely get fast though. Doesn’t hurt on 100Mbps, but you are only going to lose 1.2dB/meter…
Hi. Is this video a summary of Rick Hartley's 2-hour video that you attached in the description? Or are you talking about different topics? Thank you
another great video as usual
really nice and very helpful video.
Thank you, Paul!
Two questions: The design Phil uses is called a microstrip line. But there are alos designs that uses ground planes on the top, like coplanar waveguides.
Is it better to use microstrip lines and use multiple layers for the signal ground? Or would it make sense to use the coplanar waveguide and add GND areas on the surface?
Hello Sir, can you make a video about the Flight Controller? Like a 2 3 hour video going through the schematics and PCB routing. Vielen Dank für die bisherigen Videos und danke im voraus.
Thanks for uploading another great video! Phil do you think you could make a video of how to use CircuitMaker 2.0? It was released just a few weeks ago and appears to be very similar to Altium Designer, just more limited.
Thank you very much! I actually haven't used CircuitMaker yet. I'm afraid for the forseeable I'll just be sticking to Altium and KiCad (also for these videos).
@@PhilsLab I downloaded it and I'm currently testing it. Altium says it uses the same AD engine, but of course they have stripped some of the nice features of AD. But being free, its not bad at all.
great video! 1 question: when you calculated the tracewidth for given impedance, why doesn't the tracelength matter?
I would also like to know this!
impedance for transmission lines assume infinity length. For example impedance between signal and ground might be 50 ohms assuming the transmission line continues forever. You then terminate it with a 50 ohm resistor which looks exactly like an infinite long transmission line.
Hey Phil. thanks for the content. I'm confused by you not having the third layer as a power plane. Would you route power or the signal layers in the stack you described here?
Thank you, Brian! I'll try to bring more detailed videos in the future. Still currently working on the PCB design course which will include a lot of info!
Useful tips, cheers!
Thank you!
What voltages did you have on this board and I'm assuming the power plane sizing wasn't as critical to your design here?
Hi Phill, Great video as usual!
Are you going to make a course which we could buy (for example on udemy or orther sites). I feel like there are a lot of us that would like to learn something from you. Not only course for beginners but maby also something more advanced. You mentioned about it a few months ago so i'm looking for an update :)
Hi Kamil, Thank you!
Yes, I'm still working on a course that will cover many aspects of hardware design. It's a lot of work so will probably take me a few more months I'm afraid :)
Hi Phil Love the Video! Would you be able to post board in the video on GitHub? I’ve been meaning to get into high speed design. Also when are the audio board/ fpga videos coming out? No rush just curious.
Hi Pietro, Thank you! I'm afraid it's a company board, so I cannot post it on GitHub.
I've recently designed a Spartan 7 board with DDR2 memory that I still need to order/assemble. But once that's done, then I'll make a video about it :)
@@PhilsLab Thank you Phil, I have one more question too. I recently designed a board with a 5V regulator where the 5V gets filtered through an inductor coming from a buck regulator before it is dissipated as 5V. I need to draw about 2.3A but now I am a little worried because one of the traces I used coming off of the buck was 0.5mm. I used a 0.75mm to go into the inductor and a big copper pour and a 5V inner layer but I am just worried that the current coming right out of the inductor, 2.3A max would be a lot for the 0.5mm trace. Do you think that trace is too thin or do you think it’ll be alright considering the other 0.75mm trade and the copper pour at the output of the inductor?
@@pietrogagliano4484 I think you'll be fine. It may run a bit warm as according to a calculator (for example, the KiCad track width one), a 0.5mm wide trace with 2.3A running through it will give a temperature rise of 'only' 30 degrees.
@@PhilsLab Ok thank you! Yeah I have beefed up the traces for a later iteration, it’s a board I’ve been designing for work and these are prototypes so the final iteration will be rather beefy with the power traces to provide a good safety measure I just forgot to look that over as I was designing rather quickly for testing stage
Great video!
thanks for everything❤💕❤
Thanks for watching!
Does the JLCPCB impedance calculator actually use a field solver? Or is it a conventional formula-based calculator with a ripped off image from Polar? ;)
New video, instant like!
Thank you very much!
Great tutorial !...cheers.
Thank you, Andy!
@@PhilsLab No problem, I messaged you on you web site about the headphone amp and haven't heard anything, is the site still active or did I mess up ? !
I'm thinking of using a csfBGA-121 device.... but have never used a BGA... so if you ever feel the need to make a BGA layout video... ;o)
watch this after watching all signal integrity videos from Robert ferenec
Hoy much does cost allium, how to finance it, I lived videos on kicad.
Интересное видео.
Спасибо!
Why are there 2 dislikes?????