Hi In a recent video which I'll link in where I explained all of the PCB manufacturing options that a typical manufacturer would give you when you go and check out and get your board manufactured and I asked in that would people like to see one on flex or PCBs and other people said yes. So a Flex a full on Flex PCB video will have to follow this because it'll be quite in-depth But as it turns out, just yesterday, we had a requirement come up for a Flex a PCB adapter that we required for the new micro supply project. So I thought we'd just take a look at rather than just me do it and just send it away. I'll press record and we'll talk about the various options here so this won't be an in-depth flex.

PCB Tutorial This just happens to be a real-world thing that we need for the micro supply. If you want to see the micro supply today, here it is. Doesn't it look sexy? Here we go. Oh look at that micro supply.

Oh Beautiful thing of beauty. Enjoy! Forever doesn't have the LCD in there, but anyway, um yes, this actually really does exist and we have it working and where As it turns out, we just got the USB power delivery circuit of it working the other day and herein lies the problem. And if this video, we'll start out with a rant about St As in St. Semiconductor because well we've had an issue with this so let's take a look at it before we get to the Flex piece of bball bloody out him design our first time I'm using version 19 actually and I get our object reference not set to instance of an object.

Whatever sender do I have to send and close Can I just close and I want to send and close? Here we go anyway. this is the art schematic for the USB portion of the the isolator USB side of the micro supply. and yes, micro supply videos will come in due course. And what we've got over here is we've got an Stm32f4 no F6 p6t our micro controller because all those letters on the end matter that's rant.

number one. The Stm32 Fo 7o micro is very different depending on the letters you put on the end of it and it's really freakin' annoying. Anyway, So in the hardware prototypes that we've built, we chose this relatively low-cost Stm32 Micro because we're also using Stm32 Micro as the main micro on the isolated side as the main control element as well and it's a little 20 pin T sub a package and it's got 32 K of flash and we thought how 32k of flashes? plenty because we won't. You know we don't need to do much.

All we need to do is some USB a power delivery configuration type stuff and of course USB our comms like serial type comms, HID interface stuff and you know things like that and it seemed at the at the time it seemed like a good choice. 32k was plenty, but as we've found out through great a personal anguish mostly on David solid personal anguish of the development of this thing. USB PD Libraries that come from the manufacturer are actually huge. They're enormous and we actually chose a rich tech part down here.

It's the RT 17 16 and that's the USB PD controller. So you can see these lines here CC 1 and CC 2 their control lines which go over to the USBC connector here over here CC 1, CC 2 and they actually configure the USB power delivery specification. They negotiate how much power you can you know the host can deliver and how much the load needs and all that sort of stuff in one wanted to do it properly. By the way when we chose this micro as we'll look into another St micro that had USB power delivery built-in was not available that well the cheaper version of it wasn't available as we'll see at the time of making that decision.

but and the one that did I think was was grossly more expensive so it was a cheaper solution to go for this to chip solution here for. So the Stm32 of course has USB built in here. it is here. You be there, it is.

There's the twisted-pair up there and so it's got the USB controller. but the USB power delivery is handled via the I Squared C bus down here in this little tiny 8 pin RT 17 Rich Tach RT 17 16. As it turns out, which we didn't know at the time, the Rich Tech Library for USB power delivery is enormous. So as the St Micro one the St micro one is I think it's like 2 megabits or something enormous like that and the Rich Tech one was huge as well and David's actually and it wouldn't fit.

Neither of them would fit into the 32k of flash on our Stm32 micro, let alone all the other stuff that we wanted to put in. the serial comes, the Hid and Us regular USB stuff and you know, all sorts of stuff. Just the Power Delivery library was absolutely enormous. So yeah, David's had to work with the design engineers at Rich Tech to get the the library down and we've finally got it pruned down where we can actually fit the Rich Tech USB PD library into the 32k of flash on this ST micro that we're using.

Unfortunately, we can't fit anything else and we can't fit all our other stuff and maybe if we keep working with them for longer, we might be able to get it down further where we can just squeeze everything this 32k. But anyway, we're decided bugger that we're just going to put in a larger micro and of course Murphy's Law says that. The package we chose this 20 pin t SOT package the largest part is a 32k part. so to get say 128 K part we have to move to an entirely different package and at this stage we've got like half a dozen prototype boards build up.

Now you know they cost a fair bit to get these manufactured and they're all working so we don't want to just scrap those and have to rebuild boards from scratch on a type timeline at the moment. And like we just want a little flex adapter board that converts 80 SOT package like this into a larger footprint for the larger part right? So here's a photo of the bored here and this is the offending chip right here that we have to convert. So what we need is a Flex PCB Because you can't really do this with a rigid, this is where a Flex PCB comes in real handy. What we need is some sort of flex PCB which goes like this up here like this and goes down there like that.

Please excuse the crude D of the model, didn't have time to build it to scale out of pain it. There's our new chip which is a quad flat-pack and then we will have pads on here like this with little via holes in them so that we can actually solder down to this board here to the existing path. So we remove this chip we get our Flex PCB which is very thin Polly put the kettle on material polyamide material that as it'll be like a regular double-sided board, but it'll just be flexible so that we possibly fold it up or something like that. Although you can see that this connectors in the way here so we could go so we could actually and I do it something like this.

Perhaps that might work and but then you've got. it's quite difficult cuz these are very thin. you got to get down and solder it onto the pads down here. We might have to physically remove this capacitor here, but that's okay.

That little tiny BGA there that's that rich tech package is a little pain in the ass. Anyway, it's cheap and it does USB micro USB power delivery negotiation stuff in it. So yeah, that's you know. So we need some sort of flex board.

So at this stage I'm thinking something like that. But the issue here is okay. this is a real fine pitch I think it's I have to double check I think it's point five millimeter pin pitch. so it's a real pain in the ass.

So you know, really, if there's a couple of ways we can do it either to have little vias on the board on the Flex like that, then you just put your iron on top of that and put your solder on and then it just flows down through the vias and attaches to the pads underneath. and you could actually stagger your veers like this. Perhaps you know that's a common technique used on, you know LCD ribbons and all sorts of you know, commercial products like that. that just gives you a bit more clearance there to do those pads so we could do that or what we could actually do is have our flex going like this through the pads like this and then have little castellated pads on here.

right? little Halfmoon castellated pads and I've showing those on I'll do it exaggerated here. Okay, so imagine that's one little pad and then you slice through them. just let you would on a regular Fr4 fiberglass. PCB you can do that on flex as well and then you could like sort of like just solder in like that and then the board could actually be like this if we didn't have room up the top here like cuz you know we've got this in the ways as large package so that the Flex would have to flip up like that and we could.

You know it gets a bit messy so we could have a flex shape like that for example. and then all of your traces just run around here like this and then you know now chip can just be fluffed around in the breeze over here. no problems whatsoever. and then we don't need to fold the Flex but there's lots of things.

This is what I'll go into in if I do a specific flex PCB but I'll just touch on something like this like you wouldn't do right angles like this on a flex. you you round everything on flexes, you round everything, including you wouldn't do your traditional right angle. trace it like 45-degree traces like that on flexes. You want to round your traces like that.

so when it flexes, there's less. You know, mechanical stress on sharp junctions and stuff like that. So you want to. Just you know the radii.

Everything like that just because it's nice for flick because they flex and you don't want sharp corners when things flex. All the mechanical engineers can tell you all about that. So even though this board that we're doing for this particular application won't actually flex, so to speak, it'll just sort of. You know, sit relatively flat like this on top of you know the existing chips and things like that.

Just be careful if you've got exposed pads on the bottom. I'd make sure they don't short out towards the tops of any other components than things like that. so you'd have to put some sort of insulated layer on the back so you can actually have another layer of Polly put the kettle on material stuck on the bottom of your board so that none of your Vias are exposed. And of course, you could do tented fears and you know it's stuff like that.

But just be careful that you don't nothing shorts out because if you had a Via, that just happens to be here. and it was an exposed one and then that's sitting flush on this on the top of this capacitor here, which is going to be larger than their physical height of the resistors there up so you can come a gutter and it's short and it might be intermittent and that have really ruined your day. Murphy Didn't sure that'll be intermittent at the worst possible inconvenient time. Yes, I'm not actually sure how I'm going to run this one whether or not I'll shape it I think I'll put the check because there is.

Well, there's not room for it here. but here, there's room for a before it gets to this planar transformer up up the top here. So there's certainly room for the chip because as we'll see in a minute, it's not a huge amount bigger than this one here. and even though it's a quad flat-pack but you know, as I said, it's got a light go on the top of this and it's got a sort of Bend upwards and oh, I don't know.

six of one, half a dozen, the other like III haven't chosen yet. I'm about to lay this out. they haven't chosen which way to do it. Like whether or not I go for like little holes here like this.

Whether or not I just go for little holes like this. Little vias like that, staggered vias? that'll probably and just just have the board shaped like that I think I think that'll probably do the business I Yeah I've just run with that because there we go we can remove that capacitor and we can either put it on top of the flake so we can do some unseen and physically take out the capacitor because it's not. You know it's not a big deal, it's just a bye Pass Cap'n Well, you know you can put it on Flex if you want. I could put a pad down for it, but yeah, it'll work.

So anyway, that's the story behind what we're doing today. We got a man. You flex your flex and manufacture man your flat. Sure, that's a new word.

I'm gonna run with that. we're gonna menu Fletcher A new flex piece PCB Just a modern PCB Literally has just the - well, one chip on it plus a pad to solder down onto the top of this chip here and you would design this like you would any other piece of me. There's nothing special about Flex PCB design here unless it does actually flex. And then as I said, you want to use curved traces.

By the way, Bend Radius of Flex PCBs You As a general rule of thumb, you want to keep the bend radius to ten times minimum ten times the thickness of the material. So if the material is, you know half a millimeter, you're the bend radius you want at least five millimeters. That's just a rule of thumb. So I think with the pin pitch, we're probably going to have trouble with the cast elations on here.

I Don't like our chances of getting cast elations with 0.5 millimeter pin pitch on a flex. I've never actually tried it at that pitch before, so yeah, I don't know. I think I think the safer option is just to go for staggered the arrangements like that and we'll just let the solder flow down through the vias like that. I think we'll give that a go, right? So let's go over to the adapter schematic here and what we've got is we've just got the two chips.

The new one is an STM if oh, seven. Oh again, but it's the C BT six as opposed to the F6 P6t chip. God OTR But you don't need the TI t I was just taping real. That's just the package that it comes in after the whole video on that have an eye I'm sure I have hmm anyway, but this new chip has more pins which we don't need but in and comes in a quad flat-pack hence why we need this little flex adapter board.

but it's a functionally identical chip except it's got 128 K of flesh. So let's actually just for a rent. let's go look at the Esti website. So and David has started.

He used to like Esti, it was a bit of an ST fanboy and then he started to greatly dislike them because they would announce all these chips without having like a real proper support for them and stuff like that. And this Stm32 G0 series which is a new one. As I said before, this wasn't out, this wasn't available when we originally our chose the chip for the micro supply. The developments have been quite some time so it is a new one.

Bloody Modern website slew this. USB PD USB Power Delivery the STM it's the 32 G this is their new value series line has USB PD in so we can completely get rid of well a couple of Bonham Bomb items here. not only the rich tech tip but also these resistors here for the pull ups. A couple of caps here we can consolidate our bomb and it's cheaper.

but AHA it ain't that easy because if you're if you were just reading that, the top level here you might think are fantastic. USB PD But let's go have a look at the datasheet shall we? Well, you're searching the datasheet for power delivery. You search for USB and what there's no mention of USB Wah-wah-wah-wah Not only if you actually go into the correct data sheet, not only does it say for the USB power delivery just says our data coming soon or whatever. Um, there is no USB controller inside this thing so if comer guts are right there.

but if you are looking at sort of like the top level things like this, just trying to pick a micro app like first Shot, you might say oh it's got USB PD but it doesn't have an actual USB transceiver in it so it's absolutely useless so we can't use that value line series. what we have to do you go to if you go I'll save you the time going through all the parametric searches and stuff like that. Well, you have to actually go to the G-force series down here before you can find one that has USB interface with power delivery including the physical layer like that and that one is about three times the cost of the dual chip solution with the rich power delivery controller and the F-series micro that we're going to choose. So yeah, it's no way.

Well we've come a guts are there any way they've got a way key video over here I won't bother Planer, but it's how to create a USB power delivery sync in less than 10 minutes and it goes through and you got all there. It doesn't explain well, it doesn't explain no music and it just you know it's it's using their of their cube software or whatever and it just it's not explaining anything, it's just so you do all this. do all their store this and you can implement a USB power delivery in ten minutes but it it ain't that easy but you have to actually I think you have to contact them to get this library. You might even have to pay for it or something like that.

I don't know I don't wanna know the details but yeah it just isn't this easy they they don't explain anything so it's anyway. that's the that's their new Cube software. Apparently you know they make it out to be. You know you can just snap your fingers and you've got USB Power Delivery.

Know and talk to David about this if you want it is a USB Power Delivery. Implementing it correctly and thoroughly is a pretty horrific experience. Okay, so let's actually go down here: USB PD Dead Battery Support The content of this section will be provided later. Thanks a lot st like yeah, you buy into these chips, they advertise them, You know you might design it into your design and then you go.

If you didn't release, you may go well. how do I implement my USB PD yeah? Do I follow some little YouTube tutorial video with all wanky music on top and that's it. Like it's just no, no, no no no no. So yeah, st pain in the ass.

but handling a lot of manufacturers rich tech hasn't been ideal. For example, the Rich Tech one is unfortunately not open source. Come on, Rich Tech Open Source! USB PD Anyway, we're trying to work with them so that all of our firmware and implementations will be open source. but I think we're going to have to rely upon some rich tech by in libraries unfortunately.

but they're trying to strip out or whether we're working with them to strip out all the crap so that we can get a minimal implementation of us BPD that we need. So yeah, otherwise it's just it's too big and they've got undocumented registers and we've had to get special spreadsheets from them which document these undocumented registers. And well, to document these undocumented registers. there's a phrase anyway, which detail these registers that we didn't know anything about because we found it in the software.

they're accessing these registers that wasn't in the data sheets and that notes and things like that and we had to ask them what what's going on and they do have the details. but yeah, they don't make them relatively public. so like a lot of manufacturers are like this, they just man the devil's always in the detail. Anyway, what are we 30 minutes into the video already? Sorry.

Anyway, we've got a PLCC chip and we want to convert that to an STM 32 like package over here. so we've got the adapter. PCB Now here's where you can come a guts up real easy. So this is our chip.

obviously the red. It's on the top layer there as you can see and then we've got a bottom layer down here so we only got a two layer. PCB Like I said, we won't do anything different to a regular. We wouldn't design this any different to a regular.

PCB It's going to be exactly the same as you would layout, except you when you go to the checkout of the manufacturers website, you'll choose Flex PCB instead of this now because this is a hundred percent flex. PCB It is not what's called a rigid flex PCB which I'll show you. this is an example of a rigid flex PCB It's like I can't make these Bend that I'm aware of. Anyway, you can see that these they've got some caps and I presumably they're they're LEDs right? So yeah, those caps and LEDs are on this flex PCB section.

and then you've got two other rigid boards hence the name flex Rigid and that's got a battery on the bottom like that I don't know what this board does. just an example which came without him. When I do my video I might show a more complex example, but you can see how the flex material in this case usually it's sandwiched in the middle. So if you've got a four layer board for example, then your flex material will be the two inner layers and it extends into the PCB so that you can then fold the board's up like this and create all sorts that you can make.

as many boards as you want, but you basically lay these out as one big PCB like you've got multiple PCBs but it's effectively done as one large board including all the separate. You can have as many of these as you want. You can have 10 different rigid PCBs like this, all connected with all these flex functionality. I Don't have an example.

Now here's an example of a traditional PCB that's connected with a flat flex and you can see the ribbon connector cable there and on the other end of the board over here. Okay, and of course you need the connectors and well, that's you know. expensive, extra bomb items, more things that can go wrong or you can integrate I sorry I don't have the best example. We've got a better example next time, but in this one here, you can see that this PCB if I flip it over is actually flat flex on the bottom of the PCB.

It's not in the inner layer as I said, but it's on the bottom. and then they've got this extra bit over here which then goes off to another board. so it's all integrated so you don't need the connectors on there and that's what they're doing here. You don't need any connectors cuz this flex PCB here extends as a layer inside the PCB.

Now as I said, ordinarily, you would wedge it, sandwich it in the middle so to be the inner two layers of the PCB because I do ordinarily I Wouldn't recommend that you put it on the outside like this one because it doesn't flex, there's more stress, and the adhesion can come off and stuff like that, it is more rigid no pun intended I'm here a week the when it's sham sandwiched in the middle but you can put it on the bottom like this and the manufacturers will do whatever you want. You can see that the bottom layers like that are the Flex PCB and that is called a rigid flex because it contains combines rigid Fr4 PCBs with Flex. But in this but tequila case we're not going to do this. We're just going to have flex.

And if you want to see how that works in the layer stack manager here like here it is. Here's the bottom layer here. They've actually specified the bottom layer as Polly put the kettle on Polly amide material so the the dielectric one, dielectric twos. So this is actually a four layer board but it's got so they could have put it in the middle.

but they've actually sandwiched it and they've specified the polyamide on the bottom. Now, this isn't the sort of data that would be automatically given to the PCB manufacturer. this is the sort of data that you might. You could just specify on the garba.

You could just say you know, is it like instructions on the Gerber or instructions in an email or whatever you can just say write. My bottom layer is the polyamide material. so the Flex on the bottom or if you wanted the Flex on the top, you wouldn't actually have to change. So if we go into single layer mode here, you can see that that's the layer on the cut on the top layer, then the middle layer is there.

There's no copper on the Flex at all because there is no flex in the middle layer. It's only when you get to mid layer two or the one from the bottom that this one is flex so you can see like the bottom layers or flex. it's all ground plane by the way. Another recommendation is not to use solid ground planes like this on flex wherever possible.

Usually you would dark crosshatch them so you might put your plane out here like this, but then I'd start my crosshatch from inside here and then all of this would be a crosshatch ground fill. and you might if you tear down products you might see that commonly is that they'll use a crosshatch ground fill. and that's for thermal and flex and stress reasons which all the mechanical engineers can no doubt tell you about. But once again, that's another life rule of thumb.

Unless you absolutely need a solid ground plane, you would do flexes as a crosshatch, but when it goes inside the board like this, you can certainly make it solid because it's it ain't flexing when it's sandwiched inside that fiberglass. And just here's an examples of rigid flex: R PCBs So this one here, you can see that they've got the two copper layers in the middle. Then there's an easier polyamide insulation. With that, he's even there.

There's another adhesive in there. Then you got the Fr for prep egg. So these are your rigid if our falls over here and this is your Flex material sandwiched in the middle. But as we shown on the Altium example, we've got that flex on the bottom two layers.

So this big hopper here and this copper here are done and manufactured as a flex. Then they manufacture the Fr for fiberglass ones separately and then they'll stick those on the top and it's generally nicer if you do it as a big panel like this. For example, that one's only got two separate boards, but you can see how they've designed those as a flex and that one's got a little like a LCD connector or something. you know, some sort of ribbon II thing at the bottom say only got some Mouse bites on there.

So that's you know. That's quite quite a nice design. And there's just two boards with some flexors sandwiched. and then there's ones with multiple that might that might have it look that looks like it's got a like an RF connector on there or something like that.

You know needed more advanced ones like this and you know it can get really cool. But we're just going to manufacture a simple and solder directly onto a Flex PCB So you leave out the word rigid. So we're basically doing like a Flex PCB Like that, we're going to sort out our components directly on to the polyamide material instead of fibreglass. It's the only difference I'm So from a like a design point of view, it it really is.

If you're just doing flex, there's no difference. You just specify flex in your checkout. That's it. You might have to make it.

Have a few design considerations for the Flex mechanical flex part, but apart from that, the design process is exactly the same. So there you go. it's going to fold. This is Altium Design here is just going to like flex around like this.

It folds up and then this are the two balls like that, fold up like that and then the flexi bits. They go around the outside into it like a like a hockey puck puck configuration or something like that. And once again, this isn't really an extreme flex board like there's nothing really inherently like. 10 years ago, getting one of these manufactured very specific, yet to find a manufacturer who was willing to do it.

Now it's practically a shopping cart checkout at some manufacturers. Anyway, basically all you got to do is tell them and you lay a stack up on your Gerber's or wherever you tell them it. Which layers are Polyamide Flex material, Which layers a fiberglass material and the other two just do it. Easy-peasy They'll just choose the material and get to get the job done.

And of course and then though, when they sell medium order together, all of the veers from the flex layers will automatically line up with the and be bonded to the vias on the rigid PCBs. But from a PCB manufacturers point of view, the construction is exactly the same as a multi-layer PCB They don't care that you're using the Flex material instead of a rigid material. There might be a few little process you know things in there, but basically they don't care whether or not they're making an eighth layer rigid board or whether they're making an eighth layer rigid flex combination. It's just that one of the inner layers is made out of polyamide instead of fiberglass.

Anyway, I'm getting way too much into that. So now here's we can come a gutter. As I said, right? You got to be very careful here. Now this is of course all your PCB designs are always done looking through the board.

So this is the top layer. By definition it's the top layer, so it's the red layer here. Okay, so this is pin one over here now. ordinarily this one pin ones in the correct configuration for the chip, but you'll notice it's blue.

It's on the bottom, so this one's actually been flipped twice. Not only is the component been flipped from the top side the red side over to the blue side, but then it's been mirror imaged like that because we have to connect down to another board underneath which has this particular pin configuration. So if you didn't realize that you can completely come a gutter and get your footprint back to front. So anyway, we're just going to have a chippy on there having to find the board outline yet.

But yes, we're gonna have a pair. This won't be a chip, this will just be some Vias. So what we need to do now is go in here. Geez.

Look at that expansion on those pads. That's a bit how you're doing as touching. Look: There's no solder mask live. Oh, if we go into 3d view this this.

There's no. there's no soda mask. no sort of mass between pads there. I Could easily breed your solar out.

So our soda mask expansion is too big. so we have to change that solar mass expansion for Mills So we'll change that. So that's coming from the rules. We could have changed the rules, but you know, like, who cares? doesn't matter, we're not worried about this.

There we go. One fouled mask expansion. that'll do it. So now if we go in here like this bloody era.

Marcus There we go. We got sold. Um, us between our pads. Now it's a Christmas miracle so we can flip that over.

You can see there's a chippy there, but of course there won't be a chippy because that will just have pads anyway. what we want now is we just want to go in and just put some manual vias. So let's do a pad that's aa. You know a point three millimeter hole for example.

That's all you know because you wanted this order to be able to flow through you. you don't want it to be too small that the manufacturers gonna charge you more for example. And then our pad is 0.5 millimeters are here. We go there we go and not pointing go to Metric.

So I do Metric for whole sizes and for dimensions and boards and stuff like that. but for trace and space I still like to do Imperial Sumi Now it's going to give a clearance constraint there because I haven't set up the rules I haven't done like anything. we could actually put two V's here. There's no reason why we couldn't have two like that.

just in case, like one doesn't make doesn't make adequate connection. Here you go. Those holes look enormous, don't they? But their point three millimeters. We can do What? as I said, you could do one that's staggered.

but like I I Think that's but then again, you've got to also get your soldiering iron. You've got to have enough exposed copper on there to use solder in I and on. So then you can heat up the pad and heat up the via that it goes through and then the bottom of the the bottom side of this. if your flippity doo-dah the bottom side of this has to then transfer the heat also through to the pad which he tried to solder to on the bottom of this on the micro supply board.

so you know as a bit of a bit of trickery required there. So I I don't know. Should I Go. People are prescribing.

Use one. Dave used to go for the cast elation ones. Now of course if I wanted to do the castellated holes for example, I would just specify my board outline as going right through the center of those pads and then they would manufacture it. They'd cut it, get the laser out, and look straight through, and then I'd be left up with the left off with the castellated pads.

But anyway, I think two like that will probably do the business. Let's get rid of those error. Marcus Now of course, when we've just got the pads like this, of course the soldering iron has to. The only contact it's got is on the annulus ring around that's called the annulus ring.

The exposed copper around the video like that, so probably what you want to do. Once again, you could do this as a slotted hole for example you can like, but no. Anyway, so what I'm just going to do is just put a fill. So what we can do is we can actually put a fill on the top like this and of course this won't.

We've got the fill there. There's a copper fill. Okay, but the problem is we've got we haven't removed the solder mask expansion. So we need to go to our solder mask, our top solder mask layer and put in another fill like that.

and Bingo! We now have this nice big pad that we can get out a nice chisel tip iron onto and then the solder will flow down through these vias here. And of course we've got the matching pad on the bottom like that. So I I think that's a good solution and we just place that there. You'll notice how it takes up all the nicknames.

Alright, so this is looking pretty nifty. I Haven't died to find the board outline yet, but there you go. We've got our nice pads like that and on the bottom side there, we've got our matching pads down there. We're going to solder mask between pads so it should be all hunky-dory Yeah, I'm sure the anal-retentive out there Dave pull back, pull back just a tad on those on the annulus ring there.

Alright, so I will just go veer now about Noir. Point Four five. Well I'd do it. Oh yeah, good enough restrain.

In fact, it's bang on. Is it winner winner chicken dinner? So there you go. Um, that is how above decide to implement. As I said, there's many ways to skin this cat.

You could do castellated holes, you could do staggered vias and all sorts of stuff you could do as a slide. I put my thinking cap on I can probably come up with half a different way dozen ways to do this and now all we got to do is route that and we'll just define our board outline as I said I think we'll do it like that I think cuz otherwise you've got these resistors in the way. If you try and bend it up this side up here, you've got the resistors in the way and you've got these, you know? I can't take those out. So really? yeah.

I think I'll just have it coming out like this because there's a bit more gap between that rich tech tip and those pads than there is between these resistors and these pads down here. So and then as I said, if you put the board and you put the flex coming up here like this, then it's gonna flip up on top of that. That's quite a large thick package there, so it's really gonna flip the chip up like this and I don't really want that. So I think I'll have it coming out the side like that.

Oh, hang on, have I got pin one around the right way. No pin ones over here I was gonna come a gutter. Let's just double check on our micro supply board here. Yep, pin number one is there so I need to orient it I Of course I I had it oriented that way because I don't really correct way on the camera anyway I think you know what I mean if you've been playing along at home then you know I was gonna come a guts are there so you need to double check stuff like that.

So pin one. so I want the Flex to be coming out the bottom like this? so because I don't think there's enough on the top? Yeah I could no cousin got the crystal over here? Cool. could have it there. Oh I don't think there's the physical size of the chippy I can just measure it out line a bit.

You got the switch as well, but we can put it off to one side. That's no problem. We're talking about 8.3 millimeters. nine point nine millimeters.

So there you go. We don't have the physical distance you know I always measure, don't assume. So even if we buttered the pins right up against here, we do not have the room to fit in the chip in there and likewise we don't have. You can just tell you can just physically we're not gonna have here.

Okay so let's allow for some flexi flexi because once you've got the chip on there the chips gonna be like relatively like the chip doesn't really Bend So this flex underneath this chip isn't going to bend. so we want to add you know some Bend radius so we can can just come up over that resistor and then and then we can like put and some insulating stuff on the bottom and then just stick it down with some silastic or some double-sided tape or something like that. like just leave it flatten around in the breeze. Doesn't matter, it's just a prototype.

This is not high speed stuff at that. We do actually have a differential pair on here and it's It's designated by these little differential pair markers because it is USB. So technically we should do out these as it. You don't have to have controlled impedance, but if you would, this was high speed stuff.

Yes, you would look at controlled impedance, differential pair length matching, or alternate trace length match in these and and stuff like that. but this like we're not transferring huge amounts of data. It's basically Rs-232 type rates here, so we don't have to worry about just it's still round them as a pair just because you can. and you should.

But you don't have to worry about the nitty-gritty of that. Yeah, so what we'll do is we'll just define an outline here. we could. So we just put our little radii in there like that.

That's we can just make this bigger to get all our traces out. of course. and we've won our radii in there because that's the part that might flex. L is going to a little bit.

So yeah, you just want to take the edge off that these corners you don't have to worry about them, you can just keep them sharp. doesn't really matter. of course that's that's showing up as like thick I could go in there and physically set the three-dimensional thickness of that to the polyamide material to make it look a bit more realistic. but who cares, that will there be a little board, but instead of being made out of fiberglass, it's gonna be made out of Polly Put the kettle On! So I don't want this to be a routing tutorial, but here's where you move your chip around, use your rats nests and also rotate your component like this to see you can see.

Oh look, these are all swapping over like this. This is probably not going to be pretty so you can see that there's all these ones down here which have to get all the way over to here Oh Dull I haven't switched I'm still pad one over here God I'm dumb I'm gonna have to do let's take all of this and flippity doo-dah that around. look like in see those always rats Nessa Crichton those Nets are crossing over so you know that's going to be ugly. These ones down here.

you have to connect up with these ones over here. Those go over the couple over the couple over here. go over to there things like that. So that's actually just place a trace here.

Just just willy-nilly right. This is a five foul trace. You know you don't want to go much smaller, but are you are going to be able to get that through that pad? No, you're not. See, that's the problem, right? We're we're constrained.

so really, we've probably got no option I Really don't have an option really unless you want to go to really fine tolerances. Why? Because she's gonna pay more I'll just extend this all the way up to the top traces all up there because there's nothing on the top side like I Don't care if that's all flap and Bend and up or whatever done doesn't really matter I Want to be able to get my damn traces out of there like that? Alright, I Do believe that is our finished design I'm certainly not not gonna write on to my mum about that one like it's just it's just slapped together I found five our traces point three millimeter holes. the the V is uneven tinted for example, so you might want to tent over those. but but as you can see, that is the basic.

oh I could pull back the width of that and then we'll just solder our chip onto the directly onto the Flex because you have solder and directly under copper and temperatures are a different you know thing to Fr4. Not as tolerant but no problems whatsoever. I should I don't off? I'll put the bypass cap on there or not. but anyway, um, that should do the business all right.

So that's our finish board and we just generate our Gerber's exactly the same way as we would for a regular board. I've done our videos on that. It's just assumed that we generated our Gerber's no problems whatsoever. Then you go over to your PCB manufacturer.

Then you choose FPC rigid in whatever shopping cart options they're got, you don't want to Richard flex PCB Of course, because we're not manufacturing rigid flex, we're just manufacturing a flexible PCB that tells them you want polyamide material, probably put the kettle on instead of fiberglass. Fr4 PCB We've got a two layer board. We just want single pieces. We don't want a panel.

We don't care. You can get them to penalize it. they might penalize it for you, but no. I'm just happy with the little single pieces.

so it's absolutely tiny. It's 22 by 12 millimeters and like we don't even need silkscreen thing. Let's just say we want ten of these little puppies. FPC thickness.

We can choose our thickness like we want. like point one. It looks like point One is their standard, so that's fine. You don't want to deviate from that because you made that, may increase your lead time and stuff like that because you may not be able to share a panel with everyone else.

Yes, they'll still do these flex things as a pen or a rigid flex would be very custom specific, so you'd almost certainly be buying your own panel there. And their minimum tracks space? actually 0.06 millimeters. That's like under three fell. That's pretty good.

So we've got our 5 5, so no problems whatsoever. Minimum hole size our 0.35 Oh, we've come a gun sir. I should have checked this before I laid out my board dope from this manufacturer. Anyway, it needs to be 0.35 millimeters.

So yeah, they're just going to come back and reject that. So I can change into 0.35 how much we're looking at now? Five to six days Quantity: Ten hundred and eleven bucks. It's not particularly cheap, is it What? If we want five, it's not going to matter. Five or ten because they're so tiny.

They've got to manufacture this and are they doing it just for us like somebody else may not like? We may not be sharing a panel of flex material. for example, with them. You test. No, Oh, it's a bit cheaper There you go.

Without the E test, it's cheaper. We don't need the electrical test. I I Wouldn't bother. Seriously.

I wouldn't bother with the electrical test for this. I Just just take my chances. So we get our silkscreen for thrift for free so you know, might as well put a picture of platypus on there or something. Man, now there is some concern about this.

Is that. okay, Our resistors. If we go look at our board over here, then we've got our resistors. These resistors are fairly close.

Okay, so we're going to have a bend in the Flex going down like this. so you've got to flip it around. So we've got a bend like that and then it's got to bend up over this little. This little BGA down here is not.

You know, it's not that high. It may not even be as far as the resistors I don't know, it's a tiny little thing. but anyway, a Bend line across there right just after that chip. So just whole thing will try and flex up like that for example.

and then this chip overlays that. So you could argue that I probably should have put the chip down below the bender line of the top of that and then you've got it Bend in here. but I'm not currently concerned about that. so you've got two intersecting Bend lines like that.

So maybe with hindsight, I could have put I should have put the chip down the pins down around like no, no higher than the other chip because then the whole thing could flex along that axes like that if you can see my cursor if you know what I mean if you know what I'm talking about. So yeah, I I think it did fit I've had things got to be a problem once again. for production. You'd fuss over this sort of stuff.

Alright, just to avoid that jewel bending radius like intersecting Bend radius at this point here I might go with what I originally proposed might don't I'll turn it one where I just route all the traces up around here. like half on the top, half on the bottom. They should be able to get around there. that's only four point, three millimeters.

Look, yeah, yeah, don't realize how small this is and measure that from there. Over to this component over here, that's six. L that's almost seven millimeters. So you know the flex only comes to here.

so we've got. You know, we can make it much, much longer. But as I said cuz it's gonna Bend it's gonna Bend along that pretty sharp because those resistors are, you know, close to that chip. You could argue that you know there was heaps of room here.

They shouldn't have been that close to begin with, but you know, and we've already got those boards, so there's no point worrying about that now. and and so then we don't have them. Then we don't have to remove the bypass cap there, for example. It's no very effective bypass cap anymore with the traces, doesn't It doesn't matter.

She'll be right, no worries. And yeah, I think that will that will do it so that will remove like this crinkle, right? You know when you got two intersecting radio I Like that because this has to bend. at that point this has to bend here. so this will all be crinkly cut.

Smith's chips crinkly cut in here. So and to avoid our Smiths crisp quickly cut, we will add a bender radius in here because this is where it's all doing. Although I might even move this down a little bit down here something like that I'll just reroute that and it'd take me ten minutes. and because we can't get across there any more you might see I might flip it.

Actually, you could argue that whoa no yeah, most of them are coming from here. I'm just going to leave it like that I think maybe move the chip over here cuz look there's only four pads there that occur now I know those six over here but a couple of those go over there. so I'm like I might leave it say there ish that's working out nicely. already know that I like that I didn't even like trial that it was just like yeah just got feel yeah that Bobby Dazzler and I'm just going to be bang on on the width of that to just like guesstimating that a winner.

ah why can't they go to there Now You could argue that I can maybe increase the the width of this for example because I had that available and then route some extra traces around here like this from here. but really, when you look at this I'm looking at this guy I don't know these ones have to go over to here anyway. Then they're going to have that same routings squeeze constraint that I had through there before in there's three of them and I'd look, couldn't be bothered I'll just route. they're all these out on the bottom layer.

By the way, sometimes when you're routing traces like this often, you will offset them you wouldn't have. You may not have one trace above the other because there can be slight mechanical things there when you're really getting too tight Bendis and things like that. So a technique which is sometimes uses to offset your traces like that. but for something like this doesn't matter.

a rat's actually. I've got my guts are here a little bit cuz like I can't obviously route them out in this direction. So I've rated these ones lowest point down here cuz I can't I might be able to sneak one up there perhaps but my sneaking one on the side look at that. and but so these ones will come around here like this and what do you do? You got to go all the way around here, all the way with LBJ all the way right around here.

So you've got mm-hmm so it's Murphy you'll get every time the life of a PCB designer. No, that's actually the power ground. So there you go. Once again, like you could leave it to later.

and then you know, try and flood fill your way out of your routing problem. But like let me just get the other traces around and we'll see what's what. This is where I could actually go to the top layer. You know, power and ground here.

You could argue that. Okay, you could maybe here out the top layer out like this and around. So and push all those traces up. Mm-hmm You could argue that I could even move it.

One more like that, extend my board out. This is what you can do when your constraints aren't really. you know major constraints. Yeah, we got a nice fat looking power trace coming around there now.

Wow I Already run that trace over there. Ah, there are from there to there. That's all. Okay, yeah, no workers there.

we go. that's better. So we actually got a we got a nasty, nasty cross in there. So yeah, we ran our 3.3 volt USB supply around there like that and we can probably do a similar thing to our ground.

and of course I could you know route multiple ones at the same time? Of course I could select them and then we'll bring them up all over and all that sort of jazz. But yeah, and here's our differential pair. And as it turns out, we don't actually have to connect this on top because this would already be connected on our board. So there's absolutely no reason to connect those two pads there.

But just in case, this one doesn't make connection. and this one does when you're soldering. but just for completeness and for design rule checking of course. you would then get this across here like this: No workers there, we go.

Almost done, just got to join up some grounds. There we go, we are. We are done. We are fully routed.

I didn't define my I stretched my board outline just a tad there. So I'll redo that, but we are done that. there's our flex once again. I'll just tent those fears.

but yeah, um, that is that. looks neat. That looks like a winner winner chicken dinner so we avoid that flex point in there. So yep, I like that beauty so there's not really much left to do there.

I Forgot to change this whole size as they will to 0.35 millimeters and and basically all we need to do is add our shopping cart, upload the Gerber's and Bob's your uncle. We should just get that flex board because they won't They shouldn't have any problems with that as long as you meet all the clearance guidelines, the whole guidelines, and stuff like that, and it's obvious in your gobo what your board outline is on one of your mechanical layers, it is. There's nothing fancy, we're not doing a fancy rigid flex and they need to know the layers stack up and all sorts of jazz like that, they'll just manufacture it as a two layer. PCB Pretty simple these days.

When I was a boy, that. was rocket science anyway. um, that's it I Hope you enjoyed that. Don't know how long this videos been.

It's been quite lengthy, but if you enjoyed that, please give it a big thumbs up. And as always you can discuss it down below and we'll have to do a follow-up video. When we actually get this and assemble it on the board and and see if it works, it should. Fingers crossed I don't know, it's a rush job.

um, but it should. Can't see why not? I don't think we've coming outs or anywhere, but you never know, never know your luck in the big city. so I'm anyway I hope you liked it. Catch you next time.