Important! My own designs here were not EMC tested. I will explore that topic in a future video ;-) There we will see whether my boards follow guidelines and how we can possibly change the design to make that happen. Also we will test the Ali Board as well concerning EMC. That will be fun. I hope you are looking forward to that ;-) Also if you are interested in Würth Elektronik products/services then you can contact them directly if you are a company or business: www.we-online.com/en/products/components/ordering-form But if you are a without a company then you can contact their distributors about their products: www.we-online.com/en/support/network/distribution
Please make a video on rda5991ic it is a multipurpose ic with features like it integrates 802.11b/g/n MAC, PHY, 2.4Ghz radio, PowerAmplifier and antenna switch into one chip and is optimized for mobile applications. WLAN, Bluetooth and FM can work simultaneously and independently, with low power consumption levels target to battery powered devices. I have one lying around but couln't find its pinout or schematic. Please help.
If there's time, it would be fun to bring in a Pololu converter board as well - they're a lot more expensive than Ali Express fare but are nominally backed by a name brand and are the smallest off the shelf converter boards I'm aware of (albeit still a bit bigger than your design)
Yeah, whenever I've experimented with these kinds of regulators, one of my first tests is doing a conducted emissions test. And doing that test at multiple current draws. Without an input filter (which none of the reference schematics ever have), I wouldn't be the least bit surprised if all the designs fail.
Another idea to get it even smaller: This is a non-isolating converter, so you don't need a GND pin at output. That enables you to just put all the contacts on one edge. GND in the center, and capacitors on the sides in between. I bet you can make it smaller if you place more components on the backside too, but for manufacturing all components on the same side are prefered.
Great idea! Making it pin-compatible with a 7805 for example. Also, if you need to solder-blob the voltage, why not just use a different resistor (and loose one). Perhaps you can even get everything on a single side (no hand-soldering required)
Experienced engineer here... Looking at your PCB design i noticed that you make so called via-stitching equally spaced, which is a bad habit. Especially in fast switching circuits. That's because those places often become nodes of the high frequency standing waves, which in result amplifies the noise, energy loses, and destroys the electromagnetic compatibility of your circuit. This is quite unintuitive, but randomly placing vias (which for an untrained eye may seem unprofessional and ugly) in fact makes your circuit better!
While browsing TI's product page for the TPS63020 I came across the TPS631000, which TI itself describes as 'Upgraded product with 8-uA IQ and smaller SOT package.'. It would be really cool to see you make a board design using this chip, especially since i couldn't find any boards using this IC online.
TPS639011 is a good and affordable option especially with the new technology… however, if you want better transient response with less passive components, tps63807 is a recommendation. - a TI employee
Nice little chip. I have considered using it for backup power for some home-grown smart home devices but ultimately didn't go through. I missed one thing from the video though, that can make or break a power supply and that is capacitor derating. Class II ceramic capacitors are pretty nasty and change their capacitance if they are DC biased and me not taking that into account actually cost me weeks of troubleshooting because of unstable power caused by the lack of capacitance at the input. TI has a pretty nice online design tool that helps you find derated values and most capacitor manufacturers have a public database. P.S. Thanks for reminding me about this IC, I have a project where it should serve me perfectly (Edit: corrected grammar)
@@Jefferson-ly5qe I heard it and while it is a good idea to oversize a bit, every capacitor has its own unique voltage-capacitance curve. I have recently looked at two 10uF 50V caps, one was 1206, the other 1210. The latter one had half of its rated capacitance at 25V, the other was already under 1uF. And that's not even accounted for changes because of temperature. And it made a huge difference
This was very timely for me. I would love to see the lipo charging ic built into a tiny board like this - that’s something I need for lots of small projects
Wow! That’s amazing! Also, those inductors are incredible for their size! I’ve been working on a PCB which needs to put out 5V with as much current as possible, as small as possible, which runs on a single 18650. I ended up landing on a TI chip as well that’s just amazingly small for the power it can handle. It’s a fun challenge!
Pcb designing for switching converter is actually pretty intense and interesting The way you place your current loops become very important I recently did first attempt at pcb designing and have it fabricated ( Not good but usable) Useing kicad I highly recommend Phil's lab vids if anyone wants to know more about it
Multiple different size (and or type) capacitors are better at filtering noise than just one large one. Inductance might also play a role in choosing multiple capacitors.
Trying to find a good combo of types & sizes for filtering noise is quite difficult - as it is very dependent on the frequency response characteristics of the types of capacitor you use. Most manufacturers won’t even give you accurate ESR/ESL values to model with so you have to measure them yourself. A single large capacitor (especially an MLCC) probably will also have detrimental impact on the frequency response of the controller. I wonder what the transient response is like for the boards.
Cute, BUT! Check out the TI's "power modules", which have the inductor integrated. They are spectacular! Tiny, very efficient and absolutely great EMC-wise, as the noisiest switched node doesn't even leave the module. For your application, TPSM83100 looks like a great solution and could easily get you to 1/2 size of your final board.
Hi Scott! I know I’m a bit late to the video but if you do end up seeing this and you would like to try using smaller components like 0402 and don’t want to pay for a metal stencil from JLC I would recommend getting some thin Mylar that is about double the thickness of a zip lock and a bit stiffer and then using a laser cutter to make a stencil by importing the easy eda stencil files into its software. I use this all the time for rapid prototyping and it works great even for panels
I didn’t get a good look at your layout, but adding some additional ground layers in the middle of your board can help with reducing output ripple. It is also helpful to “isolate” the feedback node by ensuring the feedback trace is thin and doesn’t go anywhere near the switching node. Same thing with the ground reference for the feedback divider, have it be a trace that goes straight to the return of the chip and ensure the power paths are “separate” if that makes sense. Great video! Thanks for sharing!
Yep ... and placing bulk cap on the bottom layer of a thick 1.6mm two layer pcb will render them almost useless for high f noise surpression 😢. For driving Leds it will work, but ur emc engineer will jump out of the next window 😂. Pro Tip READ THE FRICKING DATASHEET ...
The vias and the ground plane also serve as thermal management like a heatsink. I have been afraid to make some boards smaller that it may raise the operating temperature of the components and shorten their life.
I think when we see overspecced components on aliexpress stuff like this it's mainly cost based. Not that cost can't be lower with the right components, but a lot of the time I think what happens is they say "oh we have a bunch of on hand, lets use that" or "Our other factory makes this part so we can get a better rate on it".
For low power applications I would actually recommend LM3671. It is small (SOT-23), it only needs 3 external components, it is 16uA quiescent current. Pairs really well with a 0806 form factor power inductor from Murata.
Tiny SMD boards are great. You’re minimizing trace distance, they’re very easy to stencil, and you can just drop it on a heated board to set the solder. It’s a win-win-win!
Great video overall and a good option for many applications. Would love to have seen you go down in resistor size to 0402 since the ones you used are pretty bulky. Also, would recommend using flux when hand soldering as it would result in much better solder joints.
Maybe some shielding to lower the noise? Even just a secondary ground plane over the top may help. You could probably test it with foil until you find a solution before making something rugged, but it's a thought. Also, using more of the back of the board, especially to try to isolate the inductors.
Shielding is very unlikely the issue, we’re not talking about coupled noise here (it’s power source, it’s a very low impedance node). The noise is probably resonance based, the caps and inductor having a particular noisy resonance at that frequency.
It can surely be made even smaller - the in/out pads are huge in the final design and they take space on both sides. Change them to the single-sided smaller soldering spots. Same about the voltage selection jumper: you can reduce number its pins from 3 to 2 by making it short a resistor in series (or not short for 5V). Or you can eliminate it completely by just soldering different resistors depending on desired board output voltage. And you can put more capacitors on the back side to bring the noise back to acceptable level.
In my opinion the Vias job here is to act as a heatsink and keep the IC as the ambient temperature because the FR4 material is not good when it comes to heat transferring. Overall good video as always, keep up the good work
At around minute 5:28 where you connect pin 13 to GND and pin 12 to VCC. The spec sheet at around minute 2:38 all 3 pins are connected to GND via C3 100nf. My question: Is there an issue if you do not use a capacitor?
Speaking of really small electronics, I've been looking for an excuse to use the Renasas DA14531. The SOC somehow includes a buck/boost and ble radio in a 1.7x2.05mm package. So if you make miniaturizing into a series it might be one to check out.
you could omit that route obstruct on the sides of L1 pads and go into them from the direction of IC. That would make it even smaller and you could move capacitors closer.
Great video 👍 A few suggestions that could improve performance, and keed all components on one side. Move all important text to bottom side. Get rid of designators and component outlines. Then you will have more space for decoupling capacitors. Not all capacitors are created equal, especially not MLCC class 2. So you may want to look into the delta C as function of delta V. This depends on package size, dielectric and vendor. If only area is important, the use one 1206 instead of two 0603, as you can find taller 1206, which is more volume for capacitance. Perhaps a 1206 polymer tantalum. Move input and outputs to the eges of the design, perhaps even castellated holes. Share input and output ground. Resistors could be 0402.
Very neat. Receiving new boards is so exciting. I use tiny 7805 pin compatible switching regulator boards in my designs at 1A. They're still smaller than yours, but the 2A is impressive.
I love videos that show design and iteration! Thanks. I would be good to know the design lifecycle from starting design to the EMC tested final product. An evaluation on component cost is probably a factor that influenced the size.
Nice project, but in the end one should approach it as an engineer would and list all of the requirements of the circuit and see if it is feasible at all and then trade (space, components, count...) to achieve the goal. Strictly going to "small" (as shown) compromised noise and current. Then of course need to consider the operating temperature range and likely other issues (supply side noise, vibration, etc). This is a great channel!
Most awesome...and the fact it can put out fairly stablish 5vdc at 2A (might need a wee bit more output filtration)...it could be used in a DIY UPS (with a charging IC) to build a pretty decent battery backup/UPS for a SBC... Very cool on the "miniaturization" of the board! Great job! Keep em coming!!!!
@@greatscottlab Well, i will almost certainly look forward onto the next video topic too, keep the good work man, i wish i had started an analog electronics channel too
You can still make it smaller by scaling the components down. Use 0402 resistors (you don't need hand solderability when you use a reflow plate anyway, and 0402 is still not that hard to solder), and smaller in / out pads, these are huge! Also, the tracks / polygons going in and out of the coil don't need to be this wide at the top, this only increases the boards performance... as an antenna.
@3:36 be careful to consider the voltage rating of those caps. Capacitors have a capacitance de-rating proportional to the DC bias. Large caps have higher voltage rating and hence are less susceptible to said derating.
I think that you should try adding back some of the smaller output capacitors. That should limit the output noise (voltage swing) at higher frequencies. Also I have a question on your measurements - have you enabled a 20MHz BW limit in the scope? This is the "industry standard" when measuring power supply noise.
Really nice! Awesome circuit for the really low power stuff, where the DC/DC-converters own current consumption becomes a significant part of total power draw.
This video reminded me that in 2014 I bought an Accurite Temperature/Humidity monitor. I stuck a brand new Energizer battery in it. This device is still being used 10 years later with the original battery in it. The device is always on and always displays the temperature and humidity and cannot be turned off unless you remove the battery.
Thanks i love electronics since i was a kid i dissasebly my toys and radios and stuff now is my hobby im a welder iron worker but i love electronics .i used that board to mod my gameboy advance with a lipo battery
As others have said, watch out for capacitor de-rating. Caps with larger packages generally have better de-rating characteristics. Also consider that more caps in parallel will decrease ESR, which can greatly help reduce noise on a switching regulator. Lastly, keeping components on the same side of the PCB can often minimize the size of current loops, thus minimizing unwanted inductance and noise. I'm really curious to see the comparison of EMI/EMC on these boards. I've designed a few boards with switching regulators, but have only recently started paying more attention to EMI/EMC.
Yes, EMC is important but for me the noise is important too. For a regular microcontroller application, noise may be less of an issue but as soon as you start building analog (audio or HF) circuits, the noise on the power line is for a large part responsible for the bad quality of the output signal. Different capacitors have different specs (a.o. ESR), resulting in different behavior for different frequencies, that is the reason why you see different capacitors (value and type) in schematics. It's not just about the 100 nF + 22 uF but also about the types of capacitors being used. Same goes for inductors. Different core materials have a different behavior on frequencies so a coil suitable for a converter using 200 kHz may not be suitable for one working at 2 MHz.
There is an even smaller TPS the 63900. I have used it in Esp32 projects. The only limit is that only gives 1A. But is tinier and has 50nA quiscent current
I find your video very informative as I'm currently designing a boost converter based on the Texas Instruments LM5158. On a two layer board 0.8mm might offer better emc with the same cost as the 1.6mm. looking forward to your next video on this.
Been working on a 3.3v reg for a work project to power a small SoC, which has been an interesting dive into TIs catalog. If you ever want an interesting design project, something similar is extremely interesting.
Would be nice to mount inductor on opposite side, but that would make soldering more difficult. Maybe might be possible with two separate thin boards and then soldering them together? But size is not everything :)
Talking about wurth, you should try two components glue, it drying time is about max 10min and it holds like metal. Yess it’s expensive but drying time less than 24h like you said in one of your videos and strength is incredible, worth the money, at least here in Netherlands 👍
The right size of this kind of circuit depends on switching frequency! As higher is the frequency smaller will be the components and more eficient must be the output filter!
How did the smaller boards handle the heat generated by the components? One of the reasons why you use large ground planes on a pcb is to use the copper layer as a heatsink
You can still utilise 0402 resistor or capacitor package or 0201 if you're master in soldering (Or do machine soldering). 47uF 6.3V SMD capacitor is available in 0805 package. You can replace multiple 22uF capacitors with it or use Tantalum capacitors because they have highest capacitance density. along with 1 big 47uF 0805 capacitor, one 100nF capacitor is recommended to reduce ESR and improve EMC. If you're doing double sided assembly, 4 layer board is recommended for max. performance and better component placement with routing flexibility. With this recommendations you can build a PCB just a little bigger than the IC itself. Or Contact me, I'll design it for you 😊🇮🇳
Yeah pretty awesome stuff :-) I bet we could go even smaller though with some converters that switch with GaN FETs. But that is a topic for another video ;-)
Fantastic, this could be the kernel for a while family of useful parts added on because it's so small. Usb-c recharging, very portrait and very square form factors for pen/cell or low profile spaces, single sided and double sided for a flat surface or smaller size, a prototype board with lots of breakouts and led indicators... 😁
Hi you have to check PFM based DC converters. TI and Toradex have pretty cool ones. They can operate as low as 1 uA output power. When I have developed an ultra low power circuit / controller ( sleep current 3 uA / run current 200 uA ) this was the solution. Quiesen current where around 500 nA of the DC/DC
looks like they had used tantalum yellow capacitors, the speed is required for MCU like esp8266 s to work stably or have a lot of capacitance like 1F on output.
Great job! I've used boards that were _kind_ of similar (boost converters), except that they also had a USB-C port and a charging circuit. Connect the battery on one side, the MCU on another, and optionally plug in USB power to charge the battery. Disconnecting from USB which the battery still attached does not cause the MCU to lose power. Have you used those? I'd be interested to see what you think of them, I'm mainly curious to hear if there are issues with it that I don't know about. The main documented downside I know of is their claim that the circuit shuts off if the power draw is under 50mA, which is fine for my use case but doesn't really work for MCUs that will work on a cycle of long low-power sleeps + periodic wake-ups, a common approach.
Now that you got a taste of making your own boards, you are probably going to make more of them to cut down on wiring when connecting different modules. You could make an even smaller Power board, which you can just plug into these, with just 3 connections and an angled pin header for a vertical connection to your custom PCB. And try using flux. It is what seperates soldering beginners from pros.
One project idea that I would like to see explored. Is the use of Milwaukee branded tool batteries for a powerwall system? Where they discharge and charge while in the same slot.
Given the price and lack of a proper speck of the aliexpress board, i'm 100% sure they were not tested for EMC compliance at all. EMC labs are expensive and you'd have to sell way to many of those cheap boards in order to break even. Its funny how in the "hobby DC-DC convertors" they tend to use ancient ICs and designs (mostly because of costs) and when a board is made with a newer part from 2010 its a giant leap forward, but its still a part from 2010. TI have better DC-DC convertors from the last couple of years, but they tend to cost a bit more.
I wonder whether you could solder a cap on top of another cap. It would be twice as tall, but take up the same board space. Achieve tiny boards without sacrificing input filtering!
When I searched for good dc-dc converter some time ago if found one module, which I thought is perfect one. But later during testing I found that it has strange "low voltage mode". Like if your battery is already discharged, its voltage can drop significantly during dc-dc start which pushes IC to open mosfet more and more until it completely shortened. Unfortunately such behavior maked it totally unusable with simple batteries. So, I'm wondering if this TI IC has the same issue?
WOW 😲 Nice work! What i observed though and will be great to know, was the mini ocb solder reflow unit. That will be ideal for my current design. Could you please let me know a link? Very much appreciated. Peter
I know that Wurth sponsored the video, so you couldn't avoid using their inductors, but may I suggest you to take a look at the TPSM82903? It's a low-Iq TI buck converter with integrated inductor that requires only a couple of external components and I think it might fit on a 10x10 mm2 board.
Important! My own designs here were not EMC tested. I will explore that topic in a future video ;-) There we will see whether my boards follow guidelines and how we can possibly change the design to make that happen. Also we will test the Ali Board as well concerning EMC. That will be fun. I hope you are looking forward to that ;-)
Also if you are interested in Würth Elektronik products/services then you can contact them directly if you are a company or business: www.we-online.com/en/products/components/ordering-form
But if you are a without a company then you can contact their distributors about their products: www.we-online.com/en/support/network/distribution
Thanks for useful board!
Please make a video on rda5991ic it is a multipurpose ic with features like it integrates 802.11b/g/n MAC, PHY, 2.4Ghz radio, PowerAmplifier and
antenna switch into one chip and is optimized for mobile applications. WLAN, Bluetooth and FM can work simultaneously and independently, with low power consumption levels target to battery powered devices. I have one lying around but couln't find its pinout or schematic. Please help.
Looking forward to it. You're such an inspiration.
If there's time, it would be fun to bring in a Pololu converter board as well - they're a lot more expensive than Ali Express fare but are nominally backed by a name brand and are the smallest off the shelf converter boards I'm aware of (albeit still a bit bigger than your design)
Yeah, whenever I've experimented with these kinds of regulators, one of my first tests is doing a conducted emissions test. And doing that test at multiple current draws. Without an input filter (which none of the reference schematics ever have), I wouldn't be the least bit surprised if all the designs fail.
Another idea to get it even smaller: This is a non-isolating converter, so you don't need a GND pin at output. That enables you to just put all the contacts on one edge. GND in the center, and capacitors on the sides in between. I bet you can make it smaller if you place more components on the backside too, but for manufacturing all components on the same side are prefered.
Thanks for the feedback. I guess there are many way to go even smaller. But still, I was pretty happy with the outcome.
pololu does it this way, with 3 pins, like it
Why bother with PCB? 3D assembly, drag a copper thin foil for heat dissipation and blob everything in resin...
Here's trick to minimize components: but a high tech factory that will allow you to make transistors the size of few atoms
Great idea! Making it pin-compatible with a 7805 for example.
Also, if you need to solder-blob the voltage, why not just use a different resistor (and loose one).
Perhaps you can even get everything on a single side (no hand-soldering required)
Experienced engineer here... Looking at your PCB design i noticed that you make so called via-stitching equally spaced, which is a bad habit. Especially in fast switching circuits. That's because those places often become nodes of the high frequency standing waves, which in result amplifies the noise, energy loses, and destroys the electromagnetic compatibility of your circuit.
This is quite unintuitive, but randomly placing vias (which for an untrained eye may seem unprofessional and ugly) in fact makes your circuit better!
Exactly the same for mechanical noise - for example when making strengthening ribs
@@martinfisker7438 Is that because of resonance?
Thanks for your comment! I always want to make it look neat. Can you name more pitfalls? ❤
top tip, thanks for sharing!
That is the cutest reflow plate I've ever seen!
While browsing TI's product page for the TPS63020 I came across the TPS631000, which TI itself describes as 'Upgraded product with 8-uA IQ and smaller SOT package.'. It would be really cool to see you make a board design using this chip, especially since i couldn't find any boards using this IC online.
I can put it on my to do list :-)
@@greatscottlab Awesome! I look forward to it, thanks for the great vids
If you care about Iq you can go even below that. TPS63900 has only 75 nA, even though the current limit is a bit lower.
TPS63900 is brilliant, I used it in my low power system, it's incredibly versatile.
TPS639011 is a good and affordable option especially with the new technology… however, if you want better transient response with less passive components, tps63807 is a recommendation. - a TI employee
Nice little chip. I have considered using it for backup power for some home-grown smart home devices but ultimately didn't go through. I missed one thing from the video though, that can make or break a power supply and that is capacitor derating. Class II ceramic capacitors are pretty nasty and change their capacitance if they are DC biased and me not taking that into account actually cost me weeks of troubleshooting because of unstable power caused by the lack of capacitance at the input. TI has a pretty nice online design tool that helps you find derated values and most capacitor manufacturers have a public database.
P.S. Thanks for reminding me about this IC, I have a project where it should serve me perfectly
(Edit: corrected grammar)
Thanks for the feedback. I will look into that.
I believe the usual rule of thumb is to order MLCC caps with at least twice the voltage rating you need.
@@Jefferson-ly5qe I heard it and while it is a good idea to oversize a bit, every capacitor has its own unique voltage-capacitance curve. I have recently looked at two 10uF 50V caps, one was 1206, the other 1210. The latter one had half of its rated capacitance at 25V, the other was already under 1uF. And that's not even accounted for changes because of temperature. And it made a huge difference
I am designing something that does use this IC, I have never designed anything of the kind before and this video was indirectly a great help
ty man
You are welcome :-)
This was very timely for me. I would love to see the lipo charging ic built into a tiny board like this - that’s something I need for lots of small projects
Wow! That’s amazing! Also, those inductors are incredible for their size!
I’ve been working on a PCB which needs to put out 5V with as much current as possible, as small as possible, which runs on a single 18650. I ended up landing on a TI chip as well that’s just amazingly small for the power it can handle.
It’s a fun challenge!
Thanks for the feedback. Sounds interesting on your side as well.
that sounds like a good thing for raspberry pis!
Pcb designing for switching converter is actually pretty intense and interesting
The way you place your current loops become very important
I recently did first attempt at pcb designing and have it fabricated
( Not good but usable)
Useing kicad
I highly recommend Phil's lab vids if anyone wants to know more about it
Thanks for the feedback. And yes, Phils videos are definitely more in detail and super useful.
Multiple different size (and or type) capacitors are better at filtering noise than just one large one. Inductance might also play a role in choosing multiple capacitors.
In what application you would want to do that? And when do you know when you need multiple small ones? And how to determine their values?
@@cryptophoenix3031 I was reacting to Scott saying he could use one instead of three capacitors.
Trying to find a good combo of types & sizes for filtering noise is quite difficult - as it is very dependent on the frequency response characteristics of the types of capacitor you use. Most manufacturers won’t even give you accurate ESR/ESL values to model with so you have to measure them yourself.
A single large capacitor (especially an MLCC) probably will also have detrimental impact on the frequency response of the controller. I wonder what the transient response is like for the boards.
I would love to see more miniaturization videos! This was great.
The only pcb I've seen that makes that hotplate look big!
I am amazed anyone can solder that tiny, i for sure cannot. Good video as always.
0805 is definitely possible. 0603 also. Tinier than that and it becomes a guess work. Maybe I should do a challenge video about this ;-)
@@greatscottlab I find 0402 to not be that difficult. 0201 takes some patience though. Haven't tried 01005 yet.
Not difficult when using good solder paste (Sn/Pb) and hot air. Very difficult if one insists on lead free solder and an iron.
@@greatscottlab YES! I would love to see that. I can't even do 0805, let alone smaller th an 0603
Cute, BUT!
Check out the TI's "power modules", which have the inductor integrated. They are spectacular! Tiny, very efficient and absolutely great EMC-wise, as the noisiest switched node doesn't even leave the module.
For your application, TPSM83100 looks like a great solution and could easily get you to 1/2 size of your final board.
Also, 8uA Iq for the mentioned chip ;)
Wow, those are really good to know about. Thanks for posting about it.
Hi Scott! I know I’m a bit late to the video but if you do end up seeing this and you would like to try using smaller components like 0402 and don’t want to pay for a metal stencil from JLC I would recommend getting some thin Mylar that is about double the thickness of a zip lock and a bit stiffer and then using a laser cutter to make a stencil by importing the easy eda stencil files into its software. I use this all the time for rapid prototyping and it works great even for panels
I didn’t get a good look at your layout, but adding some additional ground layers in the middle of your board can help with reducing output ripple. It is also helpful to “isolate” the feedback node by ensuring the feedback trace is thin and doesn’t go anywhere near the switching node. Same thing with the ground reference for the feedback divider, have it be a trace that goes straight to the return of the chip and ensure the power paths are “separate” if that makes sense. Great video! Thanks for sharing!
Yep ... and placing bulk cap on the bottom layer of a thick 1.6mm two layer pcb will render them almost useless for high f noise surpression 😢. For driving Leds it will work, but ur emc engineer will jump out of the next window 😂. Pro Tip READ THE FRICKING DATASHEET ...
The vias and the ground plane also serve as thermal management like a heatsink. I have been afraid to make some boards smaller that it may raise the operating temperature of the components and shorten their life.
Thanks GREAT Scott. I like this video, just like all the others you have made. You and Captain Dissolution are 2 of my favourite creators
Excellent work! I try to use thinner PCBs for smaller projects. 1.6mm is fine for large boards, but small PCBs look "nicer" when they are thinner 🙂
I really like these old school circuit design videos!
I think when we see overspecced components on aliexpress stuff like this it's mainly cost based. Not that cost can't be lower with the right components, but a lot of the time I think what happens is they say "oh we have a bunch of on hand, lets use that" or "Our other factory makes this part so we can get a better rate on it".
For low power applications I would actually recommend LM3671. It is small (SOT-23), it only needs 3 external components, it is 16uA quiescent current. Pairs really well with a 0806 form factor power inductor from Murata.
Tiny SMD boards are great. You’re minimizing trace distance, they’re very easy to stencil, and you can just drop it on a heated board to set the solder. It’s a win-win-win!
Great video overall and a good option for many applications. Would love to have seen you go down in resistor size to 0402 since the ones you used are pretty bulky. Also, would recommend using flux when hand soldering as it would result in much better solder joints.
0402 resistors are so fun to use... soooo tiny! but certainly helps shrink stuff down!
your a legend, iv been watching your videos from when i was 5
Oh boy. Very early. Crazy to believe that. Awesome for you to stick around and learn more :-)
Yo so how old r u right now?
@@Abodino im 15
@@Lyrical-lounge-music Wow
I think I started following him when I was 12.
Learned so much from him.
This channel is making me want to get back into the electronics hobby.
I watch all your videos and learn something new everytime thanks 🙏.. love from india
Maybe some shielding to lower the noise? Even just a secondary ground plane over the top may help. You could probably test it with foil until you find a solution before making something rugged, but it's a thought.
Also, using more of the back of the board, especially to try to isolate the inductors.
Shielding is very unlikely the issue, we’re not talking about coupled noise here (it’s power source, it’s a very low impedance node). The noise is probably resonance based, the caps and inductor having a particular noisy resonance at that frequency.
It can surely be made even smaller - the in/out pads are huge in the final design and they take space on both sides. Change them to the single-sided smaller soldering spots. Same about the voltage selection jumper: you can reduce number its pins from 3 to 2 by making it short a resistor in series (or not short for 5V). Or you can eliminate it completely by just soldering different resistors depending on desired board output voltage. And you can put more capacitors on the back side to bring the noise back to acceptable level.
In my opinion the Vias job here is to act as a heatsink and keep the IC as the ambient temperature because the FR4 material is not good when it comes to heat transferring.
Overall good video as always, keep up the good work
At around minute 5:28 where you connect pin 13 to GND and pin 12 to VCC. The spec sheet at around minute 2:38 all 3 pins are connected to GND via C3 100nf. My question: Is there an issue if you do not use a capacitor?
Speaking of really small electronics, I've been looking for an excuse to use the Renasas DA14531. The SOC somehow includes a buck/boost and ble radio in a 1.7x2.05mm package. So if you make miniaturizing into a series it might be one to check out.
ST had or has some SOCs with built in switch drive signals too. Last one I saw even had two.
Always a pleasure to see your projects.
I'm excited for that EMC testing video. Great video, Scott!
you could omit that route obstruct on the sides of L1 pads and go into them from the direction of IC. That would make it even smaller and you could move capacitors closer.
Your "solderable by hand" threshold clearly is something else :o
Great video 👍
A few suggestions that could improve performance, and keed all components on one side.
Move all important text to bottom side.
Get rid of designators and component outlines.
Then you will have more space for decoupling capacitors.
Not all capacitors are created equal, especially not MLCC class 2. So you may want to look into the delta C as function of delta V. This depends on package size, dielectric and vendor.
If only area is important, the use one 1206 instead of two 0603, as you can find taller 1206, which is more volume for capacitance.
Perhaps a 1206 polymer tantalum.
Move input and outputs to the eges of the design, perhaps even castellated holes.
Share input and output ground.
Resistors could be 0402.
Very neat. Receiving new boards is so exciting. I use tiny 7805 pin compatible switching regulator boards in my designs at 1A. They're still smaller than yours, but the 2A is impressive.
Truly an inspiration as an aspiring engineer. Excited for the next video!
I love videos that show design and iteration! Thanks. I would be good to know the design lifecycle from starting design to the EMC tested final product. An evaluation on component cost is probably a factor that influenced the size.
That giant wurth elektronik case of parts looks amazing.
I really enjoyed this video. Showing the progression as you removed components, especially capacitors. Very informative.
Glad you enjoyed it
Nice project, but in the end one should approach it as an engineer would and list all of the requirements of the circuit and see if it is feasible at all and then trade (space, components, count...) to achieve the goal. Strictly going to "small" (as shown) compromised noise and current. Then of course need to consider the operating temperature range and likely other issues (supply side noise, vibration, etc). This is a great channel!
I like this style of video, keep it up!
Most awesome...and the fact it can put out fairly stablish 5vdc at 2A (might need a wee bit more output filtration)...it could be used in a DIY UPS (with a charging IC) to build a pretty decent battery backup/UPS for a SBC...
Very cool on the "miniaturization" of the board! Great job!
Keep em coming!!!!
Great job, i am very courius about their EMC tests, hope it comes out soon!
I do not want to ruin your day; but it might take a little while.....
@@greatscottlab Well, i will almost certainly look forward onto the next video topic too, keep the good work man, i wish i had started an analog electronics channel too
You can still make it smaller by scaling the components down. Use 0402 resistors (you don't need hand solderability when you use a reflow plate anyway, and 0402 is still not that hard to solder), and smaller in / out pads, these are huge! Also, the tracks / polygons going in and out of the coil don't need to be this wide at the top, this only increases the boards performance... as an antenna.
Man I really started missing this type of videos of yours, thanks for returning to theme 💘
No problem. My main topics do vary from time to time. Always electronics though ;-)
@@greatscottlab glad to know 😂
Truly it's crazy tiny 👍🏻
Thanks :-) I bet you could squeeze it all a bit tighter together. But all in all very happy with the result.
@3:36 be careful to consider the voltage rating of those caps. Capacitors have a capacitance de-rating proportional to the DC bias. Large caps have higher voltage rating and hence are less susceptible to said derating.
I love your videos. There explained so well and I always learn, some videos I bookmark for future reference in case I need them. Thank you!
Hey, you should use a bit more flux when soldering by hand, to have the solder balls a bit rounder without the spike sticking out.
I think that you should try adding back some of the smaller output capacitors. That should limit the output noise (voltage swing) at higher frequencies. Also I have a question on your measurements - have you enabled a 20MHz BW limit in the scope? This is the "industry standard" when measuring power supply noise.
Oh this is definitely just getting started. Can't wait to see the followup.
Very neat, looking forward to the follow-up video!
Thanks! 👍
Really nice!
Awesome circuit for the really low power stuff, where the DC/DC-converters own current consumption becomes a significant part of total power draw.
Every day you upload a video is a good day. Thanks for your work! Keep it up!
Thanks, will do!
Very timely, I am searching for compact supplies like this. Looking forward to the EMC chapter!
This is such a great video, because I am currently in need of a tiny PCB with six different power outputs
This video reminded me that in 2014 I bought an Accurite Temperature/Humidity monitor. I stuck a brand new Energizer battery in it. This device is still being used 10 years later with the original battery in it. The device is always on and always displays the temperature and humidity and cannot be turned off unless you remove the battery.
Well done. Nice work. I hope that you sell a million of them!
Really nice work and i would like to see more like it also great utilization of the sponsor
Thanks i love electronics since i was a kid i dissasebly my toys and radios and stuff now is my hobby im a welder iron worker but i love electronics .i used that board to mod my gameboy advance with a lipo battery
As others have said, watch out for capacitor de-rating. Caps with larger packages generally have better de-rating characteristics.
Also consider that more caps in parallel will decrease ESR, which can greatly help reduce noise on a switching regulator.
Lastly, keeping components on the same side of the PCB can often minimize the size of current loops, thus minimizing unwanted inductance and noise.
I'm really curious to see the comparison of EMI/EMC on these boards. I've designed a few boards with switching regulators, but have only recently started paying more attention to EMI/EMC.
Yes, EMC is important but for me the noise is important too. For a regular microcontroller application, noise may be less of an issue but as soon as you start building analog (audio or HF) circuits, the noise on the power line is for a large part responsible for the bad quality of the output signal.
Different capacitors have different specs (a.o. ESR), resulting in different behavior for different frequencies, that is the reason why you see different capacitors (value and type) in schematics. It's not just about the 100 nF + 22 uF but also about the types of capacitors being used.
Same goes for inductors. Different core materials have a different behavior on frequencies so a coil suitable for a converter using 200 kHz may not be suitable for one working at 2 MHz.
There is an even smaller TPS the 63900. I have used it in Esp32 projects. The only limit is that only gives 1A. But is tinier and has 50nA quiscent current
I find your video very informative as I'm currently designing a boost converter based on the Texas Instruments LM5158. On a two layer board 0.8mm might offer better emc with the same cost as the 1.6mm. looking forward to your next video on this.
Brilliant, dude! Fantastic work!!! 😃
Looking forward to the EMC tests!
Stay safe there with your family! 🖖😊
Been working on a 3.3v reg for a work project to power a small SoC, which has been an interesting dive into TIs catalog. If you ever want an interesting design project, something similar is extremely interesting.
Would be nice to mount inductor on opposite side, but that would make soldering more difficult. Maybe might be possible with two separate thin boards and then soldering them together? But size is not everything :)
Talking about wurth, you should try two components glue, it drying time is about max 10min and it holds like metal. Yess it’s expensive but drying time less than 24h like you said in one of your videos and strength is incredible, worth the money, at least here in Netherlands 👍
Nice video. I know you said you wanted to use 0805 components, but since you were reflowing, 0402 resistors would save you some space
The right size of this kind of circuit depends on switching frequency! As higher is the frequency smaller will be the components and more eficient must be the output filter!
How did the smaller boards handle the heat generated by the components? One of the reasons why you use large ground planes on a pcb is to use the copper layer as a heatsink
You can still utilise 0402 resistor or capacitor package or 0201 if you're master in soldering (Or do machine soldering). 47uF 6.3V SMD capacitor is available in 0805 package. You can replace multiple 22uF capacitors with it or use Tantalum capacitors because they have highest capacitance density. along with 1 big 47uF 0805 capacitor, one 100nF capacitor is recommended to reduce ESR and improve EMC. If you're doing double sided assembly, 4 layer board is recommended for max. performance and better component placement with routing flexibility. With this recommendations you can build a PCB just a little bigger than the IC itself. Or Contact me, I'll design it for you 😊🇮🇳
Danke Scott für ein weiteres super Video. Beste Grüße aus Hannover (:
Danke fürs zuschauen :-)
Impressive how such a small board can deliver this much current. Technology is improving more and more :-)
Yeah pretty awesome stuff :-) I bet we could go even smaller though with some converters that switch with GaN FETs. But that is a topic for another video ;-)
Fantastic, this could be the kernel for a while family of useful parts added on because it's so small. Usb-c recharging, very portrait and very square form factors for pen/cell or low profile spaces, single sided and double sided for a flat surface or smaller size, a prototype board with lots of breakouts and led indicators... 😁
Hot take: 0603s are easier to solder than 0805s once you get the hang of it. You just have to take advantage of surface tension.
Hi you have to check PFM based DC converters. TI and Toradex have pretty cool ones. They can operate as low as 1 uA output power.
When I have developed an ultra low power circuit / controller ( sleep current 3 uA / run current 200 uA ) this was the solution.
Quiesen current where around 500 nA of the DC/DC
Thanks a lot. I’ll design another pcb like yours but includes tp4056 with bms components to have em all in a single board!
Thank you so much for another great video. It was really helpful
Awesome video idea! Very interesting! Keep them coming😊
looks like they had used tantalum yellow capacitors, the speed is required for MCU like esp8266 s to work stably or have a lot of capacitance like 1F on output.
This is brilliant! Could you share details of that small hot plate pls? It would be so useful to me right now. Thanks in advance.
What is that cool probe used at 7:41? Looks super useful.
ECM tests would be nice. From time to time I visit EMC lab, it looks crazy :D good job as always!
Great job! I've used boards that were _kind_ of similar (boost converters), except that they also had a USB-C port and a charging circuit. Connect the battery on one side, the MCU on another, and optionally plug in USB power to charge the battery. Disconnecting from USB which the battery still attached does not cause the MCU to lose power. Have you used those? I'd be interested to see what you think of them, I'm mainly curious to hear if there are issues with it that I don't know about. The main documented downside I know of is their claim that the circuit shuts off if the power draw is under 50mA, which is fine for my use case but doesn't really work for MCUs that will work on a cycle of long low-power sleeps + periodic wake-ups, a common approach.
Excited for the emc video!
Now that you got a taste of making your own boards, you are probably going to make more of them to cut down on wiring when connecting different modules.
You could make an even smaller Power board, which you can just plug into these, with just 3 connections and an angled pin header for a vertical connection to your custom PCB.
And try using flux. It is what seperates soldering beginners from pros.
One project idea that I would like to see explored. Is the use of Milwaukee branded tool batteries for a powerwall system? Where they discharge and charge while in the same slot.
Brilliant and so, so useful. Thank you again.
Given the price and lack of a proper speck of the aliexpress board, i'm 100% sure they were not tested for EMC compliance at all. EMC labs are expensive and you'd have to sell way to many of those cheap boards in order to break even.
Its funny how in the "hobby DC-DC convertors" they tend to use ancient ICs and designs (mostly because of costs) and when a board is made with a newer part from 2010 its a giant leap forward, but its still a part from 2010. TI have better DC-DC convertors from the last couple of years, but they tend to cost a bit more.
I wonder whether you could solder a cap on top of another cap. It would be twice as tall, but take up the same board space. Achieve tiny boards without sacrificing input filtering!
This in fact makes such board not manufactureable.
01:28 most important part of the video: THE SMILE
When I searched for good dc-dc converter some time ago if found one module, which I thought is perfect one.
But later during testing I found that it has strange "low voltage mode". Like if your battery is already discharged, its voltage can drop significantly during dc-dc start which pushes IC to open mosfet more and more until it completely shortened.
Unfortunately such behavior maked it totally unusable with simple batteries.
So, I'm wondering if this TI IC has the same issue?
At last something exceptional.
WOW 😲 Nice work! What i observed though and will be great to know, was the mini ocb solder reflow unit. That will be ideal for my current design. Could you please let me know a link? Very much appreciated.
Peter
I know that Wurth sponsored the video, so you couldn't avoid using their inductors, but may I suggest you to take a look at the TPSM82903? It's a low-Iq TI buck converter with integrated inductor that requires only a couple of external components and I think it might fit on a 10x10 mm2 board.