I don't know about a legend... in many areas I'm just slightly more experienced than a beginner. And I was a complete beginner in this video. But I appreciate the comment! Thanks for watching.

Yeah... I think as I mentioned in another comment, a few folks on Discord that were helping me out mentioned that it was a pretty ambitious first design. And it was a LOT of soldering between the LEDs and the resistors for each LED. The worst part? I'm just finishing up soldering all the components on the 'upgraded' version with higher resistors. So I'm doing it all over again! But I'm glad that my video might have inspired you to give it a shot. I have a few other designs (quite a bit simpler) in store for myself as well. Thanks for watching and taking a few moments to post a comment. I do appreciate it.

Congrats! It's kind of exciting receiving that first box, isn't it? I've learned a lot since my first design as well, and could have done it a lot better and more efficient (and likely with only two layers instead of four) if I had known then what I've learned since. But thanks for watching and taking time to comment. And congratulations again on your first PCB. If you are like me, it won't be your last (I've already got two more designed... I just haven't ordered them yet).

@@ResinChemTech you are so correct. I have already designed 2 more as well. This is sooooo addictive.

Yeah... I've done a lot of LED installs using WS2812b LEDs, and the controller is much simpler with the single data line and addressable LEDs. But the original purpose of this project was to experiment and show the use of the MC23017 port expanders. A clock with 72 outputs and 3 inputs from just 2 pins on the ESP was just a way to demonstrate that. The PCB became "phase 2" of that project as I kind of wanted to keep the clock that I built and the breadboard wasn't feasible. And as I mentioned, this isn't the best way to build a clock. In fact, I have 4 other clocks that are done using WS2812b LEDs. I guess I got lucky in the fact that my first stab at the PCB actually worked (no one was more surprised than I was). Thanks for watching and taking time to share your experience.

Great! A few folks that helped my along the way said I was being pretty ambitious for my first board. And I have learned a lot of things I'd do differently/better if I were to create this same board again today (like using vias!). But hey... it worked on the first shot! I really enjoyed the design process, so just be forewarned that if you are like me, your first will likely not be your last! Good luck... and let me know how it turns out.

You are welcome. Hopefully my goal came across that creating a PCB isn't really as difficult as I originally believed it to be. Thanks for watching... and thanks for the comment.

Yeah... as I've mentioned, I have numerous other clocks and they use WS2812b LEDs. This one primarily started as a way to demonstrate the use of the MCP23017 port expanders and I just used a clock as an example because it allowed me to control 72 different outputs and accept 3 different inputs all from just two GPIO pins (I2C) in the ESP. And like I said in the video, this is not the best or most efficient way to build a clock. But I did sort of like it and wanted to keep it, so I also used the original breadboard design for my first PCB project. But in hindsight... or if I were to start over again.... PWM would be great to be able to control the brightness. Instead, I've taken a second one of my 5 PCBs and remade the clock using higher resistors for the minute LEDs. It is much better now... but it is still a fixed brightness. Given the fact I've now built around a half dozen clocks with everything from WS2812b, the TM1638, ILI9341 and now individual LEDs... I'm probably done with clocks for the foreseeable future. There's pretty much no place left in the house where you can't see a clock! But thanks for the tip and feedback.

Thanks. I considered trying my hand at SMD, but was persuaded to stick with through hole for this first board. Especially since I had ample space on the board since it had to be a fixed size to match the clock face. There's still lots of blank space on the board as it is. I'm a very long way from being able to produce any sort of Fusion 360 video. I didn't dwell on it in the video, but I really struggled since this was my first time using Fusion 360. It was definitely more of a struggle than the PCB design.. at least for me. I had to watch a lot of videos on the different functions (how to create a circle, how to create a hole, etc.) to create that clock face. I eventually got it, but I'm sure not far enough along yet to show someone else how to do it! But I appreciate the recommendation... maybe some day when I get more experience with it. Thanks for watching... and for the comment!

Thanks! If you look closely near the end of the video where I quickly show a couple of additional boards I've designed, you'll notice the one in the top left corner is exactly that.... a PCB for the WLED controller that I normally build, with some additional breakout pins for things like a button, microphone, IR receiver, etc. But if there is enough interest and I decide to do a live stream of building a PCB from beginning to end, it probably will be my WLED controller that I will create in that video. Thanks for watching and taking time to comment!

I'd like a video on the led controller PCB. Not so much a live stream but a structured one or even a couple. I think something smaller like the led controller would be a good project to cover the basics and be a good reference.

Yes That sounds a good idea. 😃

Thanks for the feedback. I'm still on the fence about a live stream vs. a 'normal' video (which probably would have to be two or three parts due to the length). Of course a live stream can always be watched later.... and for me, I can get feedback and questions as I'm going along if I'm not describing something adequately or understandably. But I completely agree about a smaller controller. This clock PCB was pretty ambitious for my first attempt (and I had a few warn me that it was a pretty complex design for my first PCB). And as I clearly show in the video, I was actually shocked (but excited) that it actually worked after all the work I put into it. But yes, if I do a video (live stream or not), it will be something significantly less complex than the clock PCB. Thanks again for the feedback. It really will help me make a decision as to whether to do a video... livestream or not.

This project may seem complex in the number of connections, it's pretty basic when you look at it in blocks. Most of the connections are localized to the individual expanders, which you can logically think of as individual boards. What's cool about the overall design which you already figured out is most errors in hardware design, such as wiring being out of order can be fixed in firmware

Yeah... that would be pretty easy to do. For me, however, I'm much more likely to use the Home Assistant dashboard to change the settings. Partially because the clock isn't sitting "at my fingertips" to be able to reach the touch screen. I actually only included the ILI9341 because I had an extra one left over from another project and thought it would be nice to have the date displayed. I added the weather forecast really as an example to show that it was touch-enabled... I'll likely never use it to display the weather (again, since that is available on my HA dashboard). I was hoping to just provide some examples of how others might use it. Since it really wasn't part of the clock PCB, I almost didn't include it in the video at all... but since my version had it, I knew I'd get asked about it if I didn't at least mention it in the video. But you are correct... and that would probably be a better use of the touch screen. It would also allow me to eliminate the buttons on the back that, as I stated, are problematic for hanging the clock on the wall. Thanks again for a great idea!

Thanks. I appreciate it. I hope that I showed that designing a PCB isn't as difficult as many (myself included) believed it to be.

@@ResinChemTech exactly, I'm gonna try myself too.

Thanks for the comment. And that's a good note on the buttons. I think that @danman32 actually had a good suggestion to use the ILI9341 touch for controlling the features of the clock and then the buttons could be eliminated altogether... also solving my wall mount problem. 99% of the time I'll use Home Assistant for controlling the clock. But like all my projects, I like to have more than one method of controlling a device, including a local option that doesn't rely on any external systems. That was the original intent of the buttons here... but they could likely be eliminated via the use of the touch panel. And yeah... the silkscreen labels (and adding new ones) was something I learned how to use better after this initial design. But that is a good idea of having them in both locations. Thanks.

Thanks... As I mentioned a couple of times in the video, this was my first time and since then, I've learned quite a bit more since, including the use of vias (which were not used at all in this initial design). I agree that knowing what I know now, I could easily have gotten away with only two layers. There are many other things I'd change for this design with the experience gained by creating other designs. I was just really happy that it worked at all! Thanks again for the feedback.

Yeah... something I noticed when building the second version. I've already learned a bunch since the original board. I'm just glad the first one worked as expected! Thanks for the tip.

Ha! Thanks. I think I had held off on creating my first PCB because it seemed so complicated and overwhelming (and I expected to be a lot more expensive). What I hoped to show in this video was that even with zero experience, I was able to create a working PCB. Although I've already learned a lot since this first board and there are a number of things I'd do differently/better if I were to design this board again. But it's pretty much that way with most things... you continue to learn and get better. And I still haven't decided about a live stream. I am by no means an expert at the design yet, so it would definitely be from the perspective of 'beginner to beginner'. And yes, if a live stream, I think an edited/cut down version of that live stream could also be created. As always, thank you for taking time to leave what is always an amusing comment that makes me smile!

That would be ideal, but as I said in the original video on the MCP23017 port expanders, I'm not sure PWM would work with these due to the precise nature of the timing required by PWM. PWM would be great if not using the port expanders, but then you sure couldn't have 72 outputs + 3 inputs from a single ESP board. I didn't test PWM with the port expanders though, so I'm not sure if it would or wouldn't work. Thanks for the recommendation though.

@@ResinChemTech I've done something similar for a large 7seg clock (using led florescent type tubes, THAT big! ;) ) I did a similar way to expand the IO this was quite a few years ago. You could try a quick and dirty method by just sending a all off LED and then update the display alternating a few times a second. Anyway.. I love clock projects 🙃 great to watch your process

Wow... that would be a BIG clock! I too have a number of clocks that I've built. My favorite is probably be my LED matrix clock using WS2812b pixels. The nice thing about it is that it is also a scoreboard, countdown timer and can display text. It also has a secondary controller with WLED and now that WLED supports 2D matrix effects, it looks great when not being used as a clock. I have two of those and a "mini" version in my office that also flashes reminders, like when it's time to pay the bills. Naturally, those are dimmable... and the colors can be changed as well. So, I'm right there with you and love clock projects!

Thanks for the tip. But this project was originally developed to demonstrate the use of the MCP23017 port expander. I just selected a clock as a way to have 72 outputs and 3 inputs from just two pins on the ESP. As I mentioned in the video, it is definitely not the best or most efficient way to build a clock. I do have other clocks that I've built with addressable LEDs... and you are right, the circuits are much simpler! But I actually ended up sort of liking the clock that I built to demonstrate the port expanders, but couldn't keep it on the breadboard version... hence, I decided to turn it into my first PCB. So it kind of grew organically from the other video. But you are certainly right... there are much better ways to build a clock with LEDs.

@@ResinChemTech Don't get me wrong, it's a really cool project and a good reason to design and make a custom PCB for it. I enjoy your videos and blog posts. I've made some of your projects and aspirations to build others.

Thanks... and not a problem. I do appreciate the recommendation. I've just received a number of comments making recommendations about a better way to build a clock... which really wasn't the purpose of the video. And I do like those thru-hole addressable LEDs... I've used individual RGB LEDs, but not the addressable type and they would be ideal for something like this... or maybe a future project. So really do appreciate you mentioning them. Something else to go into my bin of "components I want to play with"!

Hmm... that has to be annoying. I've started adding 'last boot' to the output on many of my projects that I can see in Home Assistant. It may not help you, but at least I can see if something is causing the device to restart. Another thing... and I'm sure you've already checked on this, do you possibly have a retained value in MQTT that would cause the clock to turn back on if it disconnects/reconnects to the broker? I had a light that was "ghosting" like that.. seeming to just turn itself on and it ended up being due to a retained MQTT power settings.

Yes... you are correct. But as I mentioned in this video, the original intent was to show the use of the MCP23017 port expanders. I just happened to select a clock as way to show how you could have 72 individual outputs and 3 inputs using only to GPIO pins on the ESP board. I also mentioned, in both this and the prior video, that this wasn't the best or most efficient way to build a clock ( 01:18 ). I do have at least four other clocks that use addressable LEDs... and yes, the controller was much simpler. But that was not the intent of this video.

@@ResinChemTech it's because you are talking about improvement for the next generation of the clock it why I said that, also it's first video I watch from your channel (KZfaq recommendations) 😊

No problem... sorry if I was a bit snippy in my response. I've have done a lot of projects (and videos) using addressable LEDs, so I am pretty familiar with them. It's just that I often get comments from some people that have only watch a minute or two and then rush off here to the comments to make a remark without understanding the context of the video they are commenting on. That doesn't seem to be the case with your comment... so I apologize for my initial reply. I do appreciate the recommendation and that you took the time to send it.

@@ResinChemTech I only reply if I think it can be useful, I generally comment in the middle or end of the video 😊