NOMAD Laptop

lte rc pla total hours: 88

june 29th:

4 hours

So ive already done a lot of research but today i finally made the repo and im ready to start working on it. Ive done about 20 hours of research so far. Basicly this idea had started a while ago when i learnt about single board computers and was looking to buy better options and cheaper options that just a normal raspberry pi because they are pretty expensive and could be better. This was around the time where i wanted to get more into open source digital fpv and also into 3d printing and well for both of those you need well you dont need but its recomended that you have a sbc like a pi or and for open source fpv the alternative is a bit more expensive. So i read stuff, read articles etc. Saw that most sbc manufactuers also make compute modules or SOMs (System On Module) and also there is a lot more variety that just pis for example the radxa rock lineup based mostly on rockchip chips. After reading about what they could do i was like huh why hasnt anyone made like a laptop with one of these. Turns out 2 weeks later id get recomended bryan's youtube videos and i was like oh wow. Turns out it is fairly doable. He used an arm rk3588 based SOM. Its pretty powerfull for a SOM and its twice as efficient as a pi. Trust me i got an sbc (rock 5c) based on it and wow its crazy how much less electricity it consumes like it straight up stays way cooler. Anyways i decided on using a lattepanda mu. Why? because as far as i know no one else has made an x86 laptop of this type and the lattepanda mu is still fairly efficient and should give usable battery life on a laptop. Also because it should be slightly faster i beleive because although the n100 and the rk3588 have similar benchmark scores, the rk3588 has 8 cores and the n100 only has 4. each core of the n100 is close to 2 times faster which means that for single threaded tasks like loading things ie loading a website the n100 should be faster. And this laptop will be made for light work like coding and browsing so thats perfect.

Goals:

youtube battery life: 6 hours thickness: under 1.6cm modularity: m.2 slot, mxm (mobile) gpu, compatible with n305 lattepanda mu, extra pcie, repairability / workability: ideally make so you dont have to dissasemble the entire laptop to get to 1 part ex batteries peripherals: internal room for pcie fpga and an sdr and other stuff like that ports: usb A, usb C, hdmi, microsd

Also i spent the last 30 minutes working on this concept art lol I like when people make concept art to visualize somthing so thought id do it too

july 30th:

4 hours

I did some more research im looking at what batteries im going to use. At the start i wanted to use 18650 cells but they are 1.8cm thick which is far too much. i want the laptop to be under 1.6cm thick total so instead i think im going to go for some 7mm lithium ion pouch cells probably 3 of them in series or maybe 4. After that i also dissasembled an old 2013 thinkpad to see if it has an embedded display port display and i think it does so I think its a pretty good option to use as a cheap display for the laptop. It could be better but im not very picky when it comes to screens as all my displays are average or under average. Other than my phone. something like this

july 7th:

10h back from camping I started placing stuff in the schematic but im still figuring out what to connect ot waht pins because beleive it or not there are 260 pins in the conector. I'm prob not even going to use half of those but i still need to understand what they do. Lattepanda actually has decent documentation about this and they have a discord that ive joined and also i read this blog? where this guy was making his ownn carrier board and i followed what he did for finding your trace settings and setting up the drc so you have the right impedance but unfortunuately the guy never finished it so it didnt go far. Theres also this guy on youtube who made a carrier but he didnt go super in depth. Ive been going through the example carrier board that is provided in their github but im trying to make sense out of it because they use different pages and ive never used different pages also idk why they have these blocks rn so yeah i still need to do some mroe googling and chatgpting.

I also exported the example board to jlcpcb with the settings i plan on using for the final board and this is the price but keep in mind this doesnt include pcba which ill be using. Its about what i estimated earlier so with pcba it should be around 90usd ideally less. S

july 8th:

7 hours

Today i finished the schematic part of the main pcb. I say main pcb because there will be a second pcb that controls rtl-sdr usb connection sends custom touchpad hid info to the lattepanda mu and keyboard keystrokes and handles the batteries. Disclaimer i did base myself off of the lite carrier board example. The mainboard has 4usb connections 2 usb3 A ports 1 usbc 2.0 port and one internal (broken out to pins) usb 2.0 The purpose of this is to let me make a second pcb (because i want to keep this mainboard pcb under 10cmx10cm to keep the cost down as much as i can) that aside from doing all of what i said earilier will also act as a usb hub to add 2-3 more usb ports 2 of them staying inside the laptop one going to the rtl-sdr inside the laptop and the other one going to human interface devices i.e. the touchpad and keyboard The reason i have to use a usb hub is because the n100 intel usb controller only allows for 4 root usb connections. This pcb will use an rp2040 in hid mode.

Have not been placed yet but they are now inside the pcb editor ready to be placed.

usb schematic sheet

july 9th:

7 hours:

Today I basically placed most components where they should go and whats left is mostly just waiting to be placed on the board but already togheter i.e. boost converters are togheter but not on the board. I also had to edit the schematic a little bit because i realisd some things were missing or had to bbe modified like the power button, it needs to be a connector and not a button. Also i messed up and put everything on the wrong side and had to restart though lucky i was only like 2 hours in at the time so it wasnt that bad. basically the problem was that i forgot this was going to go upside down and i want the usb ports to be on the right of the laptop and the lattepanda mu and its ddr slot need to face the display so i can connect the embedded display port fpc cable dirrectly to the mu. So yeah i basically did it flipped the first time. The reason it has to go upside down is that while the chassis will most likely be 3d printed as of right now. I say most likely because I am considering making it entirely out of aluminum maybe press formed. But say it does end up being 3d printed, the bottom will have a solidaluminum sheet as flush and seamless as possible with the rest of the chassis. This will act as a huge heatsink and heatspreader which if anodized should be able to dissipate around 70 watts and if not it should still be able to dissipate around 40 watts which will definitely be enough. You might be concerned though that the heat will stay in one spot and so am I so between the aluminum and the mu i will also add an intermidiary smaller sheet of copper just to get the heat more spread and less concentrated. Aditionally the 70w and 40 w figures i gave earlier are assuming a 30x30cm piece of aluminum under 60c, if the temp goes over 60c then itl disspate even more heat but i dont want it getting too hot because a laptop who can give you burns is a bad laptop. And you might be wondering what the difference between anodized aluminum and normal aluminum is when it comes to getting rid of heat? As it turns out anodised black aluminum radiates heat in infrared emmisions over 20 times more than normal aluminum. 0.0x vs 0.8x-9x emmisivity. So its basically like one of those radiators they use in space. Pretty cool math.

As you can see most things are aglomerated or on the board already. I'm going to route the high speed traces before i place mostof the other components too since thats the most important thing to get right.

july 10th:

3 hours went fishing today so didnt have nearly as much time but i finished placing everything where it should be and now im finishing the setting up of the board settings because in kicad, when you are making impedance matched traces and differential pairs and stuff, you set up the board to havve 4 layers etc and then you set up the nets and tell them the right sizes so you get everything right and the impedance is right. Also ive started making sure to commit my kicad files everytime i update because im scared ill mess something up later and wont be able to come back. Also cool thing i learnt is that kicad seems to also have a version control system system but im still just commiting with vs code because why not.

july 10th:

2 hours:

july 17th:

6 hours: I did more research and i think ive finally settled on the display im going to use and how im going to use it, at first I was planing on using a 40 pin edp display that fit the mu but they i very expensive and the one specifically for the mu is even more expensive. The worst part is that other than the one for the mu there is no guarantee they are compatible. Hence why I've decided to go with a cheap laptop display with 30 pin edp and a hdmi to that specific model controllerr since this way i can be 90% percent sure it will work. I also restarted the pcb routing yet again. I've decided to go with a much larger board so that the mu is not half hanging off even if it might be like 10 $ more expensive also this way i have more space and can be more sure that the components will work as planned. The old routing was very bad now its looking a bit better

image is mostly for myself

routing i think is looking cleaner

july 18th

6 hours: Im writing this from my phone because i fell asleep so i wont have images till i get back home.

yesterday i almost got all the high speed routing. ompletely done and i think im gonna stick with it this time instead of restarting for a 4th time. It took a while because i have to get the length matching very precise in the same differential pair and somewhat precise between different pairs like rx tx and the 3hdmi lanes plus their clock. also i have to make sure i have no or very small stubs because or else yoy get reflexions. then theres also the issue of having a lot of traces that i need to cross.

new routing

july 20th

6 hours: today i continued routing the high speed traces, turns out i forgot some. I also had forgotten to add ground stitching vias near high speed signal vias so i did that, had to move some stuff around to make space for new stuff. then i also placed more components on the boardd like the power supplies (5v, 12, 3) and power leds, bios battery. i2c and uart ports and debug among others. Earlier i talked about the high speed traces. It turns out some of them were too close to each other which would have caused cross talk so i also had to change that. Im not sure how sensitive usb and pcie and hdmi really are but I'm gonna play it safe so im trying to make stuff the best i can even if i means stuff takes up more space on the board and looks less good, or a lot of gpio ports are under the som. Though i dont think thats a big inconvenience. Also the thing i made the most sure to get right so for this is made sure to only route on the front layer was hdmi because without the other io i have alternatives but with the display not really unless i spend more which i cant really do xd. thats why youll see its one of the only ports that doesnt have l3 routing. There are 2 other ones ig but they are the usb 2.0 ports, one is 4 2.54mm pitch pins and the other is the usb c port you'll see in the image. they only have 1 differential pair so they are really easy to route. I dont think ive mentioned this but, for the stackup i'm going with l1 high speed signals priority then low speed and power, l2 gnd plane, l3 high speed signals some low speed, l4 gnd plane. this is because each high speed differential pair needs to have a gnd plane to keep impedance constant. at the start i thought itd do l4 as a power plane and have l3 use l2 for its gnd plane but it turns out its not such a good idea because the thickness of the material between l2 and l3 is pretty big so its not very ideal. so that means ill just have to find a way to route power around that limitation even though it woudl have been great to have an entire level for power.

july 21st

X HOURS

my ground plane settings for me later

july 22nd

4 hours Fixed the drc errors there were like clearances and a lot of vias and a lot of false alarm. final pcb

jlcpcb settings:

july 23rd

4 hours I finished the BOM on jlcpcb exept for one usb that i cant find so im now looking for one on digikey Finally found one on digikey downloaded the .step file of both my pcb and the usb connector to check if the usb receptacle fits into the holes and it does!

https://www.digikey.ca/en/products/detail/molex/0483930003/4701504 this one

july 26th

6 hours

I must admit i skipped a bit thats because it was mostly tweaking stuff i already wrote about and because i only had time to work on it in litte bits because i was busy.

Anyways today im focusing on the cad for the chassis, at the moment im working on modelling the parts like the pcb and the keyboard and screen then ill be able to make a case around those

note to self: total laptop width is 32cm atm

well its now officially the next day cuz its 1 am but im continuing this log,

as you can see, this is an assembly with the main laptop parts here, once i get the parts im going to have take some measurements that i couldnt find online because im not buying well known things then ill have to tweak the cad which shouldnt take long and then i should have a laptop! I'll add mods to it and upgrade somethings in the future and i will probably update this journal too as this is a really cool project that i really like.