I definitely think a huge step back needs to be taken when it comes to manufacturability. Just because the aim is a high-end router, it shouldn’t have a high-end manufacturing cost (if that makes sense). Most people won’t be screaming “oh my god it’s not milled from a single block of aluminium, I’m outraged!!”. Multi piece design is great if it brings the cost down enough to make it palatable for people to buy :/
I totally agree. I am searching for affordable high-end router but I do not care if it is from plastic with off-the-shelf cooler inside (with maybe small lowspeed fan) or from huge block of aluminium. Tomaž should consider making also cheaper version with just plastic case as I personally will not buy it if 1/3th of the router cost is nice aluminium case.
I guess what I’m trying to get across is it’s not that I’m worried I can’t afford it and it’s too expensive, that’s not the problem I can just choose not to buy it. It’s that he and his partners will be putting in huge sums of money into development and manufacturing and if no one is willing to palate the cost one they’re in production, it’s going to backfire hugely. It would be wise to play it safer initially and then build up to a proper high end router, rather than plunge directly into custom tooling and really expensive manoeuvres needed to mill what is (at the end of the day) a PCB inside a box which will collect dust on a side shelf while managing your network. The functionality and features should be high end, not the box its sitting in.
I don't think I can afford this router, but your top-notch presentation style alone would have convinced me to follow the video series, even if I didn't also enjoy the technical side of it (which I do, I'm an engineer!).
@@muazapuhan I think selling just the naked board is very good idea. Mikrotik does that with huge success. People 3d-print whatever they fancy, or cut enclosures from cardboard.
@@NGC1433 Agree a bare bone version would also be great. It would not have any extra cost maybe a different bare bone package for delivery. It would be like the BanaPI R4. IMHO it should have 3 version: - Bare bone Version (maybe with accessories) - Standard switch/router body for mid range - Premium CNC version I know it would be more work but that way it would have a broader audience. And it would also be much easy to recommend it to different people.
@@muazapuhan we do plan several variants. Premium is the first one just to show we can do this on the,... "top" level, for a lack of a better word. Like all companies that manufacture stuff tend to showcase their most premium models first (think nVidia and their GPUs if you're in the PC space). The second variant will be sheet metal (and because of that, cheaper). Not sure yet about bare PCBs because honestly, sheet metal is quite cheap and doesn't add much cost.
I enjoy seeing the lengths you have gone to in order to achieve a beautiful design. On the other hand, by the time you got to custom tooling for a single usb-c port … ON A ROUTER … I was feeling like the design should change for manufacturing reasons. It sure looks nice tho
Tomaž, I think that this project is unfortunately DOA due to too many complications to get close to the design, e.g. Apples, which unnecessarily raises the price. CNC machining hours are very expensive and it is better to increase the printed circuit board in order to simplify the processing of the case, to have less CNC machining hours, special tools, etc... many people have network equipment in a rack cabinet so that it is also necessary to include possible rack mounting. It might be worth considering if there will be a large series, to make the case with aluminum die-casting... Instead of thermal pad, consider thermal putty. Think about this points it would be a shame if the project fails becaus of pricing...
Agree. I wrote to a similar comment the same: A bare bone version would also be great. It would not have any extra cost maybe a different bare bone package for delivery. It would be like the BanaPI R4. IMHO it should have 3 version: - Bare bone Version (maybe with accessories) - Standard switch/router body for mid range - Premium CNC version I know it would be more work but that way it would have a broader audience. And it would also be much easy to recommend it to different people.
@@muazapuhana plastic case version was promised/planned for sometime later. I don't agree that a bare PCB version would be a good thing. It is not a hardware tinkering platform like Raspberry Pi, only has some M.2 connector, and ports. There would be no hardware ecosystem around it - at best, only some 3rd party cases...
@@cristinelcostachescu9585 I musst have missed the plastic version. I understood from the previous videos that the hardware and software will be open source. I can well imagine that the community will do a lot with it, so I still see that the bare bones will be very popular.
Yeah there is good reason why plastics and bended sheets of metal are being used for basically all networking equipment instead of CNC milled blocks of metal. Faster to manufacture and cheaper.
Blue is connecting/booting, green good, red error(needs attention), orange lost Internet/WAN but LAN is receiving data. Yellow, Violet, use as see fit(you're kindof limited by number of colors as they blend together or get confusing to end users), and then flashing for other information of the current status(such as continuous blue for booting, blue flashing for connecting) or (red flashing pulses = error number, 1 - 8). Just a suggestion as most companies use this same style and I understand it out of the box, but that's just my opinion obviously. Good Luck, I am interested.
@Fortless as long as the manual states the meaning of each color, and there are no ambiguous colors (both yellow and orange, or pink and violet), I don't see this as a problem.
@@nathantron I wish the debug leds would all be visible and configurable. If you don't like the x-mas light tree effect, switch them off in the settings. Multiple leds could indicate more information clearly than a single led. A user configurable small status display would be even better.
And in the end, for Apple all aluminum cases not only made the device more prone to corrosion, but the whole mechanical resilience got affected. So many fractured aluminum standoffs for hinges documented and seen out there. It was only a fancy aluminum case which added nothing to the device. Just like the "magnesium cage" of the last IBM Thinkpads. But at least thinkpad screen hinges had lesser failure rate.
Wouldn't a more cost effective solution for the case to be a 2 part design (top+side)? You could get all the standoff/cooler block features you need on the "top plate" where your cooling mass is relevant, and the complicated undercut features would be achievable with less complicated tooling (and possibly even with a bending "horseshoe" design a substantial reduction in material. Less tooling, less material, faster CNC times in exchange for a seem which could even become part of the visual design if you wanted. I guess it would result in 2 parts in the BOM vs 1 (which isn't nothing) and an extra assembly step (though probably a faster/less fiddly one). I wonder if there are other tradeoffs I'm not considering?
That was my thought exactly. Milling from a single block of aluminium is extremely expensive in terms of machining time and wasted material. Only companies like Apple can afford to make devices from an aluminium billet.
@@faustovieira Well, it depends, if he's making a high-end product for rich people that can afford it. Companies can always afford anything as the cost is passed to the customers. Its the customers that you're targeting that must afford it. Not gonna lie, having a beautifully sexy machined aluminum block is enticing for me, specially compared to crappy "plastic routers", its like a hi-fi audio equipment, but for your internet. There's a public that would buy such things.
@@faustovieira on the other hand I imagine how hard would be to make the thing out of foldable aluminum sheet, like the cybertruck is made (people don't realize that the car isn't the product, the product is the assembly method, if all cars were folded, they would be cheaper to make).
@@monad_tcp as @originaljws mentioned, I would split the design into top and sides, with milled top, keeping the same design, and folded aluminium for the sides with the seam at the back. The bottom could be stamped out of an aluminum sheet.
please add support for HSGMII for the sfp cages so that we can use sfp ont with it. and would be cool if it could be rack mounted as well. that would be really nice
You guys going the extra length to make sure the device is repairable is such a chad move, Tomaž. That step alone takes extra effort. And the fact that it's both functional and beautiful is a cherry on top! Massive multi billion/trillion dollar companies who intentionally put repair blockers into their designs should be ashamed.
Thank you for the kind words. Not only do we want to make sure it's repairable, my ultimate vision is that the whole OS is 100% open source as well. We need to pass on the knowledge to future generations!
DO NOT SANDBLAST THE INSIDES!! Sandblasting will significantly reduce thermal transfer with the thermal foam, so it’s actually making your product worse, not better!!
@ OpenWRT based but it’s basically running Alta labs system. In the forum you can see changelogs with updates from just 4 days ago and I'm pretty sure they'd need to publish source code too but didn’t see anything there. Don’t own it but from just the hardware it looks very compelling even ignoring the price
@@meco I agree what hardware looks great. But there is number of devices which use closed source fork of OpenWRT and at some point abandoned. Or stay closed for custom pkgs. I started with OpenMoko back then and for now completely lost interest in such devices...
@@meco It is very fast and stable, but I worry about the same as others that eventually they will get bored of maintaining it, and at that point it is a brick since they have locked it down.
@@justinrutledge1221 Qualcomms SDK for the hardware seems to be available (mirrors on GitHub and at Linaro). I guess it only takes one motivated dev that reverse engineers the hardware specific files to have at least that type of openwrt working
Don’t worry too much about the internal surface, it might even be a bragging feature when you see the machining - as long as thermal interface and assembly is unharmed. (But of course optimize the non-frivolous case versions more for numbers :)
You could add some design elements to the sides of the case that hide some cuts allowing you to get a tool in sideways where you currently relying on ones from above. This might allow you to easily get square cuts under the corner overhangs instead of round ones. You might just end up with a distinctive design that you like better than the monotony of pure aluminum block. Or if you feel your customers want to pay for the pure block look IDK.
Great project can wait for buying one, btw: One simple idea that can be made with software or just adjusted resistor value, just make option somewhere to make led barrely light, because sometimes I have my router in same room where I sleeping, so it one small feature I missing on ready to buy solution.
Has to be the most beautiful Router ever made, amazing stuff all round i love seeing your updates, the amount of work you have put into this project is impressive and i will have to get one when they are done and available. Keep up the great work.
You could use the RGB LED both for the boot and the normal usage: - Boot: Red - startup ; orange - BIOS ; yellow - OS ; green - System config ; blue - Starting of Services ; indigo - User Auth ; violet - placeholder or - in use: (red - power error; orange - stand-by/loop of death; yellow - network problem; green - nominal ; blue - nominal 10G ; indigo - overheat ; violet - overpower) Or you could leave it in "open source" state so that people can tinker with it later (as long as they don't reset, in which case, the "factory settings" apply).
If you are going to make this case this nice and perfect, make it bigger and the board and use a mounting plate that goes on the board as well as a changeable backplane for heatsink so that when you come out with faster boards or different IO in the future the customer can just buy a new board and they can use the case again.
Thank you for the update. It's really fascinating to follow your journey. On the "even if it's possible" comment - your enclosure look very similar to the previous mac mini enclosure, which has lots of weird things in it. Mind you, they didn't make it with all 4 sides in aluminium, probably because it way way simpler to manufacture :) Looking forward to the next steps and seeing the final product.
It does look pretty. But it looks way too expensive for a router. Guesstimating but the Aluminum block is at least €40 plus machining of this is certainly over €30 and these would be volume prices. Why not just laser cut sheet metal that's than bent up to form the bottom case. And than a single pice on top.... Could probably get the case cost down to €10-15 You'll be creating an "Apple" router without the ability to promise support/warranty into the future. If you'd hit a price close to Cisco or even Ubiquity, speaking for myself, I'd just choose the enterprise option that has a history of user support/ability to take a warranty claim within days not months I still hope for the best for the project. But personally seeing the manufacturing of the case being over engendered to a point of unnecessary... I don't know if I'd get one
The case is used as a heat sink for passively cooling the CPU. I doubt sheet metal has enough thermal mass for this purpose. Tomaz doesn't want a fan in there, and personally I agree. Though, I am pretty sure there are better ways to machine that case to bring the cost down while keeping efficiency...
Depends on the real world throughput of this device. Mikrotik can handle up to 300 MBps with less than 1000,--. Cisco 1000 up to 200 MBps for 1500,--, 4xxx up to 400 MBps for 4500,-- I don't know about Ubiquiti, i only use theiur WiFi solutions. That is why i use general purpose PC hardware for my firewalls. They are much cheaper. If this thing can handle more than 500 MBps fully encrypted for less than 1.500,-- i do not care about 10,-- more or less. Or any repairability.
@@cristinelcostachescu9585 thermal mass isn't as relevant as thermal conductivity and emissivity; if the heat load is constant, thermal mass won't matter because the enclosure will "heat soak" and reach equilibrium. What's more important is being able to radiate/conduct the heat out of the enclosure. Sheet metal is fine if the thermal conductivity is still high enough that the energy can still be dissipated without too much temperature rise.
That is the reason pcengines APUs put the CPU on the bottom which allowed them to use sheet metal cases. Maybe you can simplify the case design with some inspiration from aluminium Raspi Cases which solved many of the problems
I am so hyped about this, even so much more because it's a local product. Sadly, I am a complete noob in terms of networking (but I am software engineer), but I can't wait to get engaged with the project. If nothing else, as a user!
5mm thermal pad is crazy! Was it not feasible to shrink that gap down to 2-3mm? Was this done to make the assembly process easier? your thermal performance will suffer substantially.
Of course! I just meant to state that the gap actually already is 3mm and not 5, which I thought you were assuming, but smaller would always be better when only regarding thermals. I myself am sceptical of the thermal conductivity of any foam, since pockets of any kind should be rather unideal for that, but since apparently something like thermal foam exists, it seems at least feasible. Surface contact is a very important factor as well after all, and maybe, taking tolerances into consideration, a compression of 2mm is necessary to account for that - but I'm just guessing in that matter, I have more experience with thermal foam in particular
Yep, 3mm gap is a bit exagerrated. Since planarity tolerances at this design are quite low, the gap can be 0.5 - 1mm. Lower than this would require spring tensioners. But at the end it all depends on how much heat you need to transfer. Also SFP cages need proper cooling for 10GBASE-T modules (4W each).
On my zyxel device, both the RJ45 and the SFP+ ports are fixed in place by the front of the enclosure - and that's not even so nice aluminum. I wonder if by replacing the USB ports for something that can get through the enclosure you might not be able to save significant amount of money on both milling and assembly - removing all the screws. When the enclosure holds and fixes the ports in vertical and horizontal direction, you need just the screws around CPU to fix the lateral movement and at the same time provide the assurance of sufficient pressure for thermal contact. The only thing I learned form my short engagement with physical device manufacturing is - use as little features as possible, everything cost a fortune if you multiply it by 1.000 devices made. Each saved screw, internal thread, etc, makes a lot of difference, not only during manufacturing, but also during assembly, QA, ...
Good progress. I'm looking forward to be able to order and support the innovation. I personally don't care about the looks of the device, and even its size is not my main worry: I won't be carrying in my pocket. This device will be sitting (possibly 2 of them) in a basement or network cabinet. Will the sfp+ cages be sufficiently cooled by being in contact with the case ?
I think faster milling would be in order, needs a thorough sandblasting on the outside to feel nice, the inside should be done quickly but don’t forget it. The bigger focus would be the pcb quality and software, because most people won’t switch the software it comes with from you
Not in the market for a router, but will by this one anyway!!! Awesome performance and it looks good! Will happily throw out my generic plastic wrapped router!
Looks pretty. What if the bottom screws were the feet? Then you could unscrew the thing with hands. And those who dont like it can use normal screws? Greetings from Germany
Perhaps if you wanted to reduce the price (maybe a ‘lite’ version in the future?) look at using additive manufacturing? An sls/ mjf process would allow for a nylon 6/12 body with little to no modification to the current design.
I don't see any attempt to make the router rack cage mountable. That is, no provision for additional front panel brackets to screwed to side of router.
Do we know what the max IDS/IPS throughput will be for this router? I am looking for something that can get 10Gbps IPS, mainly just for the hell of it. Unifi's Enterprise gateway accomplishes this at 12.5Gbps IPS but their firewall implementation is just now adding in firewall aliases, or "Zones" as Unifi calls them and I feel like I will be missing other features like that, if I switch from OPNsense to Unifi. So this rack mounted would be perfect if it has enough processing power to handle 10Gbps IPS.
Having used other passively cooled devices and knowing how hot these chips can get, especially after age. You're going to want to consider cooling fins, similar to passively cooled mini pc cases, not sure how much they actually help w/ cooling but at least the fins would provide a standoff that would be some degree cooling to touch. That or some kind of rubber casing or handle so you don't burn your hands.
Dude! Can you consider anodizing if possible without sandblasting and leaving the raw cnc cut look in it? Personally I think cnc milling cuts look awesome! Maybe do some logo engraving after the fact?
I think sandblasting is a required step for anodizing in order to prepare the surface (mainly for cleaning off oxides). Though, I am not educated on the topic, I might be wrong.
If the enclosure will have a lid, could you not avoid having to machine what you call the "overhangs" with expensive opertations / custom tooling by changing the lid geometry?
Just curious, did you consider chaining the design of the bottom of the case to not requires the overhang milling in the corners? I would think eliminating that overhangs, you would probably save a good amount of money on the manufacturing.
idk, seems a bit expensive. nice migration to bluesky and good idea with the LED RGB, I need to keep mi switch, rpi and ap in a drawer so i don't get blinded with the lights at night time.
Im concerned on electric shorts happening internally since aluminum hopefully the anodization is also internally leaving only parts that need to be conductive exposed. Also a full range thermal test is probably wanted to see what the operating temperatures are like can it perform well at -18C or 0F to 100F?
Electric shorts are not an issue. There's plenty of space between components and the enclosure. And thermal tests are next on our to-do once the bringup is done.
Molds are stupid-expensive, and considering this is a sort of prototyping stage, this makes (money-wise) most sense. After everything's done, slicing the case into multiple parts for either quicker CNC or other methods, like half plastic half aluminum (like Apple's Mac Mini).
I love the enthusiasm and the desire for high quality output. OTOH, if the plan is to make a business out of this and sell to ISPs, then I think this is going in the opposite direction. ISPs (e.g.: Comcast/KPN/Reliant, etc...) buy devices from OEMs and if they’re going to buy from small startups, it’ll be for as little as possible; I know this bc we sold to both OEMs and ISPs. ISPs aren’t selling high end "hey look how expensive my router looks" status symbols. Besides, when it comes to failure modes, it's a router and it will likely sit somewhere no one will ever look at. So the failure rate due to bending/warping is part of the failure pareto & cost model. I love your excitement & the desire to perfect it but I could never invest in such a startup…
Reducing finishing passes on the inside makes sense to optimize cost. I would not be able to wait over the weekend to get the boards. As an American (West Coast), a 2 hour drive for anything is routine. I have done that for a casual dinner with a friend.
The original estimate was around 600$ for the machined aluminium version. Check out some of the first videos on the series for more details on cost estimation.
1) Why not direct case-to-IHS contact, with thermal paste in-between? That thick thermal pad will affect performance significantly. 2) Also, wouldn't it make more sense to machine a flat bed and then use standoffs like motherboards in pc cases? I am pretty sure that this reduces the machining time and the number of different bits to use. 3) The machining through the USB-C cutout is ingenious, not gonna lie, but I am pretty sure it is the wrong outcome. I am not sure how much trouble it is, but you could swap the reset button with the outer USB-C port, thus, eliminating the need for an additional step in machining. 4) I would also ditch the double bottom, completely unnecessary ( you could, again, use standoffs for some mechanical support for M.2 cards) I don't want to be the Grinch of this project, but some parts of it don't seem to be the best choice even before the optimization step. But maybe the design choices need more explanation, so maybe, a deep insight video is necessary...
I am not sure if there is a need to CNC the top of the case 7:38 If you CNC the top you have to sandblast the top and if you don't CNC the top of the Case than you can save the cost for Sandblasting the top. The edges most likely still need CNC to be a bit smother. The inside does not need sanding in my opinion. For the foot with magnets it should be grabble and not a hassle to pick it. Maybe grabbing grooves or holding from the foots it will have any.
Any raw block isn't straight so it's a must to make it flat. But instead of a normal end mill I would use a fly cutter (of something large with max 2 passes) to make it flat.
The first operation is not just for aesthetics. The raw stock is not perfectly square, an initial facing pass is done to make sure the stock sits perfectly square in the machine vice after it’s flipped (to mill out the internal features). Still I think milling the enclosure out of a solid block of aluminium is a mistake, a sheet metal box and a normal (off the shelf) extruded heatsink ought to be fine for a networking appliance. Producing this part on that DMG Mori machine is not going to be cheap.
@@just.oblivious @Damicske I see top needs to be flat so the internier can be CNCed precisely. That makes sense. I thought that the block on right side here 0:00 is already a flat surface. Any way it looks pretty and pretty heavy. I am curious if we will also see a version with a more affordable case like standard switch case body.
The thermal stuff we're using from Wurth doesn't even have a commercial name yet, it's that new. I'm planning a dedicated video on it (and show/run tests). From our initial tests that we did on mockup coolers it performed well enough, but we didn't want to bother with precise numbers and methods yet. We will in a month or two.
RGB front led is nice for the basic debug indicator (and complex ones too). Optimize the inside as much as possible, because it can be a bit rought with larger bits when it'll get somewhat sandblasted/anodized at some level anyways. I mean we ahve the PCB on the inside anyways and it has more prickly edges due to components & rough surfaces anyways, why bother with inside when at most it just needs to not remove skin or cause irritation after handling it.
I really like this series, and the idea, but switch from Unifi to something else would be a big ask, will there be continued software updates? Features? And why is this better than say Unifi?
Regarding debugging: I really like old 7 segment displays instead of LEDs. They are easier to read and debug but maybe pricier? Is a UART Terminal planned?
I really hope to see this router pop in the homelabs of tech youtubers once in a usable state :) Do you plan on sending few for review? I feel like Level1Tech, NetworkChuck, TechnoTim and Christian Lempra would live to completely terrorize you by testing your product to the limit in front of millions of people hahahaha
Ok, that seems like a not so good decision. This adds so much cost for so little benefit, there must be a better way. And I say that as an enclosure maker! Also, that chamfer bit for the USB port - nothing custom needed, front/back chamfer bits are quite common and you can even buy them on Aliexpress.
Why don’t you make the curve start later so you don’t have to machine the USB-C port that way? So instead of a „normal“ curve, you could use a compound curve and shorten it that way.
Because the curve is machined with an inside cut, and they are already using the smallest bit possible for that. There is no more machining space for a smaller inside curve... This being said, I seen this as one of the reasons (among many others) why the case design is flawed and I suspect the price is going to be way higher than expected...
That's a really good question. One that I don't have an answer to, really because clock precision was never on the very top of my list. Actually, scratch that, I've never even thought about it from this angle. But it's definitely worth testing out!
11:30 - Oh my god... You have no idea how many people I had to explain this to... Yes, magnets CAN destroy electronics, but order for it to happen they must be moving relative to the electronic parts to induce currents that flip a bit or fry a delicate circuit, just like in an electric motor/dynamo... It doesn't produce electricity when it's not moving? Come on...
There has been a previous survey / call-to-enthusiasts, for a "beta" or testing version. Check back some videos in the series, I don't remember the details. However, the offering was not a final product at a lower price (early adopter), so don't expect the best performance or fully functional product. Though, it has been promised that in the future there will be a cheaper version with a plastic case instead of aluminium. You might want to look forward to that.
I am searching for affordable high-end router for ages now, but I personally do not care if it is from plastic with off-the-shelf cooler inside (with maybe small lowspeed fan) or from huge block of aluminium. Tomaž, you should consider making also cheaper version with just plastic case as I personally will not buy it if 1/3th of the router cost is nice aluminium case. Selling just the board itself would be also way I think.
5) I also want to ask about Wi-Fi. The router comes without Wi-Fi, but M.2 cards could be installed - right? Then, does the case provide mounting holes for external Wi-Fi antennas? Can't see this in the video...
Its going to look amazing but to mass produce it might make a dent into your profit. But it could be marketed definitely for a posh high end version for sure. Nice
Nice, but if you add some generic rubber feet to this thing, you won't escape the firing squad. "Prekmurska jelša" crosses my mind. Oak is also "prekmurski" enough. :-)
While I'd personally prefer the anodized look, just out of curiosity, is there anyone that would prefer the raw/industrial look if there was a choice? If there are such people, you could make a small run of limited editions.
Laptops with aluminum housings are usually smooth on the outside, and full of machining marks on the inside. So it makes sense to do it the same way, and make the outside look better.
I definitely think a huge step back needs to be taken when it comes to manufacturability. Just because the aim is a high-end router, it shouldn’t have a high-end manufacturing cost (if that makes sense). Most people won’t be screaming “oh my god it’s not milled from a single block of aluminium, I’m outraged!!”. Multi piece design is great if it brings the cost down enough to make it palatable for people to buy :/
this for practicality
I totally agree. I am searching for affordable high-end router but I do not care if it is from plastic with off-the-shelf cooler inside (with maybe small lowspeed fan) or from huge block of aluminium. Tomaž should consider making also cheaper version with just plastic case as I personally will not buy it if 1/3th of the router cost is nice aluminium case.
I guess what I’m trying to get across is it’s not that I’m worried I can’t afford it and it’s too expensive, that’s not the problem I can just choose not to buy it. It’s that he and his partners will be putting in huge sums of money into development and manufacturing and if no one is willing to palate the cost one they’re in production, it’s going to backfire hugely. It would be wise to play it safer initially and then build up to a proper high end router, rather than plunge directly into custom tooling and really expensive manoeuvres needed to mill what is (at the end of the day) a PCB inside a box which will collect dust on a side shelf while managing your network. The functionality and features should be high end, not the box its sitting in.
@muhammadazlanshah777 Agree. I am kind of sad that they've chosen path of total overengineering...
Heatsink case similar to what PCengines used for the various APU series devices is a great example.
I don't think I can afford this router, but your top-notch presentation style alone would have convinced me to follow the video series, even if I didn't also enjoy the technical side of it (which I do, I'm an engineer!).
I don't think i can afford just the case...
It looks pretty and pretty heavy.
I am curious if we will also see a version with a more affordable case like standard switch case body.
@@muazapuhan I think selling just the naked board is very good idea. Mikrotik does that with huge success. People 3d-print whatever they fancy, or cut enclosures from cardboard.
@@NGC1433 Agree a bare bone version would also be great. It would not have any extra cost maybe a different bare bone package for delivery. It would be like the BanaPI R4.
IMHO it should have 3 version:
- Bare bone Version (maybe with accessories)
- Standard switch/router body for mid range
- Premium CNC version
I know it would be more work but that way it would have a broader audience. And it would also be much easy to recommend it to different people.
@@muazapuhan we do plan several variants. Premium is the first one just to show we can do this on the,... "top" level, for a lack of a better word. Like all companies that manufacture stuff tend to showcase their most premium models first (think nVidia and their GPUs if you're in the PC space). The second variant will be sheet metal (and because of that, cheaper). Not sure yet about bare PCBs because honestly, sheet metal is quite cheap and doesn't add much cost.
I enjoy seeing the lengths you have gone to in order to achieve a beautiful design.
On the other hand, by the time you got to custom tooling for a single usb-c port … ON A ROUTER … I was feeling like the design should change for manufacturing reasons.
It sure looks nice tho
Priorities, dude! You've got to know them, so you you can get them hilariously wrong....
I wonder if it'd be better to diecast the shell.
Tomaž, I think that this project is unfortunately DOA due to too many complications to get close to the design, e.g. Apples, which unnecessarily raises the price. CNC machining hours are very expensive and it is better to increase the printed circuit board in order to simplify the processing of the case, to have less CNC machining hours, special tools, etc... many people have network equipment in a rack cabinet so that it is also necessary to include possible rack mounting. It might be worth considering if there will be a large series, to make the case with aluminum die-casting... Instead of thermal pad, consider thermal putty. Think about this points it would be a shame if the project fails becaus of pricing...
Agree. I wrote to a similar comment the same:
A bare bone version would also be great. It would not have any extra cost maybe a different bare bone package for delivery. It would be like the BanaPI R4.
IMHO it should have 3 version:
- Bare bone Version (maybe with accessories)
- Standard switch/router body for mid range
- Premium CNC version
I know it would be more work but that way it would have a broader audience. And it would also be much easy to recommend it to different people.
@@muazapuhana plastic case version was promised/planned for sometime later.
I don't agree that a bare PCB version would be a good thing. It is not a hardware tinkering platform like Raspberry Pi, only has some M.2 connector, and ports. There would be no hardware ecosystem around it - at best, only some 3rd party cases...
@@cristinelcostachescu9585 I musst have missed the plastic version.
I understood from the previous videos that the hardware and software will be open source. I can well imagine that the community will do a lot with it, so I still see that the bare bones will be very popular.
Yeah there is good reason why plastics and bended sheets of metal are being used for basically all networking equipment instead of CNC milled blocks of metal. Faster to manufacture and cheaper.
Blue is connecting/booting, green good, red error(needs attention), orange lost Internet/WAN but LAN is receiving data. Yellow, Violet, use as see fit(you're kindof limited by number of colors as they blend together or get confusing to end users), and then flashing for other information of the current status(such as continuous blue for booting, blue flashing for connecting) or (red flashing pulses = error number, 1 - 8). Just a suggestion as most companies use this same style and I understand it out of the box, but that's just my opinion obviously. Good Luck, I am interested.
Making the light orange will make people think its negotiating as 100baseT, possibly not a good idea?
@Fortless as long as the manual states the meaning of each color, and there are no ambiguous colors (both yellow and orange, or pink and violet), I don't see this as a problem.
@FortlessI also wouldn't think so as it's not the ports/LEDs next to the RJ45 jack. He was referring to a singular front RGB led on the case.
@@nathantron ah, my bad!
@@nathantron I wish the debug leds would all be visible and configurable.
If you don't like the x-mas light tree effect, switch them off in the settings.
Multiple leds could indicate more information clearly than a single led.
A user configurable small status display would be even better.
That looks SO expensive to manufacture. It wasn't too long ago that Apple was the only company designing cases this way.
Nothing is expensive as long as you find someone willing to pay for it. Same as Apple. The catch is finding enough rubes.
It was decades ago since companies designed cases like this. Do not assume that because you are not aware of them.
@@firedeveloper what are you even talking about?
@@firedeveloperRemember the Mac Pro or all the new Mac’s?
And in the end, for Apple all aluminum cases not only made the device more prone to corrosion, but the whole mechanical resilience got affected. So many fractured aluminum standoffs for hinges documented and seen out there.
It was only a fancy aluminum case which added nothing to the device. Just like the "magnesium cage" of the last IBM Thinkpads. But at least thinkpad screen hinges had lesser failure rate.
I agree with the repairability initiative and not creating a Christmas tree appliance :D
Leave the CNC marks!!! That's the best heat sink solution I've seen, very nice. 😎
Wouldn't a more cost effective solution for the case to be a 2 part design (top+side)? You could get all the standoff/cooler block features you need on the "top plate" where your cooling mass is relevant, and the complicated undercut features would be achievable with less complicated tooling (and possibly even with a bending "horseshoe" design a substantial reduction in material. Less tooling, less material, faster CNC times in exchange for a seem which could even become part of the visual design if you wanted.
I guess it would result in 2 parts in the BOM vs 1 (which isn't nothing) and an extra assembly step (though probably a faster/less fiddly one). I wonder if there are other tradeoffs I'm not considering?
That was my thought exactly. Milling from a single block of aluminium is extremely expensive in terms of machining time and wasted material. Only companies like Apple can afford to make devices from an aluminium billet.
@@faustovieira Well, it depends, if he's making a high-end product for rich people that can afford it. Companies can always afford anything as the cost is passed to the customers. Its the customers that you're targeting that must afford it.
Not gonna lie, having a beautifully sexy machined aluminum block is enticing for me, specially compared to crappy "plastic routers", its like a hi-fi audio equipment, but for your internet. There's a public that would buy such things.
@@faustovieira on the other hand I imagine how hard would be to make the thing out of foldable aluminum sheet, like the cybertruck is made (people don't realize that the car isn't the product, the product is the assembly method, if all cars were folded, they would be cheaper to make).
@@monad_tcp as @originaljws mentioned, I would split the design into top and sides, with milled top, keeping the same design, and folded aluminium for the sides with the seam at the back. The bottom could be stamped out of an aluminum sheet.
The unibody case (machined block) decision is because the router is passively cooled, and the case itself acts as a heatsink.
please add support for HSGMII for the sfp cages so that we can use sfp ont with it. and would be cool if it could be rack mounted as well. that would be really nice
You guys going the extra length to make sure the device is repairable is such a chad move, Tomaž. That step alone takes extra effort.
And the fact that it's both functional and beautiful is a cherry on top!
Massive multi billion/trillion dollar companies who intentionally put repair blockers into their designs should be ashamed.
Thank you for the kind words. Not only do we want to make sure it's repairable, my ultimate vision is that the whole OS is 100% open source as well. We need to pass on the knowledge to future generations!
DO NOT SANDBLAST THE INSIDES!!
Sandblasting will significantly reduce thermal transfer with the thermal foam, so it’s actually making your product worse, not better!!
Oh, absolutely, we didn't plan to sandblast any surfaces that'll be used to transfer heat!
Have you seen the Alta Route10 which is a 200$ Router that has 2x SFP+ 4x 2.5GbE (2 with PoE+)? It's based on a Qualcomm IPQ95XX SoC
OpenWRT?
@ OpenWRT based but it’s basically running Alta labs system. In the forum you can see changelogs with updates from just 4 days ago and I'm pretty sure they'd need to publish source code too but didn’t see anything there. Don’t own it but from just the hardware it looks very compelling even ignoring the price
@@meco I agree what hardware looks great.
But there is number of devices which use closed source fork of OpenWRT and at some point abandoned. Or stay closed for custom pkgs.
I started with OpenMoko back then and for now completely lost interest in such devices...
@@meco It is very fast and stable, but I worry about the same as others that eventually they will get bored of maintaining it, and at that point it is a brick since they have locked it down.
@@justinrutledge1221 Qualcomms SDK for the hardware seems to be available (mirrors on GitHub and at Linaro). I guess it only takes one motivated dev that reverse engineers the hardware specific files to have at least that type of openwrt working
Don’t worry too much about the internal surface, it might even be a bragging feature when you see the machining - as long as thermal interface and assembly is unharmed. (But of course optimize the non-frivolous case versions more for numbers :)
You could add some design elements to the sides of the case that hide some cuts allowing you to get a tool in sideways where you currently relying on ones from above. This might allow you to easily get square cuts under the corner overhangs instead of round ones. You might just end up with a distinctive design that you like better than the monotony of pure aluminum block. Or if you feel your customers want to pay for the pure block look IDK.
Great project can wait for buying one, btw:
One simple idea that can be made with software or just adjusted resistor value, just make option somewhere to make led barrely light, because sometimes I have my router in same room where I sleeping, so it one small feature I missing on ready to buy solution.
Has to be the most beautiful Router ever made, amazing stuff all round i love seeing your updates, the amount of work you have put into this project is impressive and i will have to get one when they are done and available.
Keep up the great work.
You could use the RGB LED both for the boot and the normal usage:
- Boot: Red - startup ; orange - BIOS ; yellow - OS ; green - System config ; blue - Starting of Services ; indigo - User Auth ; violet - placeholder
or
- in use: (red - power error; orange - stand-by/loop of death; yellow - network problem; green - nominal ; blue - nominal 10G ; indigo - overheat ; violet - overpower)
Or you could leave it in "open source" state so that people can tinker with it later (as long as they don't reset, in which case, the "factory settings" apply).
If you are going to make this case this nice and perfect, make it bigger and the board and use a mounting plate that goes on the board as well as a changeable backplane for heatsink so that when you come out with faster boards or different IO in the future the customer can just buy a new board and they can use the case again.
Thank you for the update. It's really fascinating to follow your journey.
On the "even if it's possible" comment - your enclosure look very similar to the previous mac mini enclosure, which has lots of weird things in it. Mind you, they didn't make it with all 4 sides in aluminium, probably because it way way simpler to manufacture :) Looking forward to the next steps and seeing the final product.
It does look pretty. But it looks way too expensive for a router. Guesstimating but the Aluminum block is at least €40 plus machining of this is certainly over €30 and these would be volume prices.
Why not just laser cut sheet metal that's than bent up to form the bottom case. And than a single pice on top....
Could probably get the case cost down to €10-15
You'll be creating an "Apple" router without the ability to promise support/warranty into the future. If you'd hit a price close to Cisco or even Ubiquity, speaking for myself, I'd just choose the enterprise option that has a history of user support/ability to take a warranty claim within days not months
I still hope for the best for the project. But personally seeing the manufacturing of the case being over engendered to a point of unnecessary... I don't know if I'd get one
The case is used as a heat sink for passively cooling the CPU. I doubt sheet metal has enough thermal mass for this purpose. Tomaz doesn't want a fan in there, and personally I agree.
Though, I am pretty sure there are better ways to machine that case to bring the cost down while keeping efficiency...
Depends on the real world throughput of this device.
Mikrotik can handle up to 300 MBps with less than 1000,--.
Cisco 1000 up to 200 MBps for 1500,--,
4xxx up to 400 MBps for 4500,--
I don't know about Ubiquiti, i only use theiur WiFi solutions.
That is why i use general purpose PC hardware for my firewalls. They are much cheaper.
If this thing can handle more than 500 MBps fully encrypted for less than 1.500,--
i do not care about 10,-- more or less. Or any repairability.
@@cristinelcostachescu9585 thermal mass isn't as relevant as thermal conductivity and emissivity; if the heat load is constant, thermal mass won't matter because the enclosure will "heat soak" and reach equilibrium. What's more important is being able to radiate/conduct the heat out of the enclosure. Sheet metal is fine if the thermal conductivity is still high enough that the energy can still be dissipated without too much temperature rise.
That is the reason pcengines APUs put the CPU on the bottom which allowed them to use sheet metal cases. Maybe you can simplify the case design with some inspiration from aluminium Raspi Cases which solved many of the problems
I am so hyped about this, even so much more because it's a local product. Sadly, I am a complete noob in terms of networking (but I am software engineer), but I can't wait to get engaged with the project. If nothing else, as a user!
5mm thermal pad is crazy! Was it not feasible to shrink that gap down to 2-3mm? Was this done to make the assembly process easier? your thermal performance will suffer substantially.
The gap is around 3mm, but they're using 5mm thermal foam to assure sufficient contact pressure
@@magnuslubkowitz Me speaking with my physicist phd in mind: the smaller the gap the better! Even 3 mm seems a lot, unless you fill it with diamonds
Of course! I just meant to state that the gap actually already is 3mm and not 5, which I thought you were assuming, but smaller would always be better when only regarding thermals.
I myself am sceptical of the thermal conductivity of any foam, since pockets of any kind should be rather unideal for that, but since apparently something like thermal foam exists, it seems at least feasible. Surface contact is a very important factor as well after all, and maybe, taking tolerances into consideration, a compression of 2mm is necessary to account for that - but I'm just guessing in that matter, I have more experience with thermal foam in particular
Yep, 3mm gap is a bit exagerrated. Since planarity tolerances at this design are quite low, the gap can be 0.5 - 1mm. Lower than this would require spring tensioners. But at the end it all depends on how much heat you need to transfer. Also SFP cages need proper cooling for 10GBASE-T modules (4W each).
Yeah, anyone who ever spread some Arctic Silver as thin as humanly possible is screaming their heads off right about now.
On my zyxel device, both the RJ45 and the SFP+ ports are fixed in place by the front of the enclosure - and that's not even so nice aluminum. I wonder if by replacing the USB ports for something that can get through the enclosure you might not be able to save significant amount of money on both milling and assembly - removing all the screws. When the enclosure holds and fixes the ports in vertical and horizontal direction, you need just the screws around CPU to fix the lateral movement and at the same time provide the assurance of sufficient pressure for thermal contact.
The only thing I learned form my short engagement with physical device manufacturing is - use as little features as possible, everything cost a fortune if you multiply it by 1.000 devices made. Each saved screw, internal thread, etc, makes a lot of difference, not only during manufacturing, but also during assembly, QA, ...
You are a hardcore nerd about every little thing regarding this router. Well done bro being passionate! Love the energy
Good progress. I'm looking forward to be able to order and support the innovation. I personally don't care about the looks of the device, and even its size is not my main worry: I won't be carrying in my pocket. This device will be sitting (possibly 2 of them) in a basement or network cabinet.
Will the sfp+ cages be sufficiently cooled by being in contact with the case ?
Was it possible to swap the order of SPF ports and reset button so you could have a space for mounting hole in the corner?
I think faster milling would be in order, needs a thorough sandblasting on the outside to feel nice, the inside should be done quickly but don’t forget it. The bigger focus would be the pcb quality and software, because most people won’t switch the software it comes with from you
Not in the market for a router, but will by this one anyway!!! Awesome performance and it looks good! Will happily throw out my generic plastic wrapped router!
Looks pretty.
What if the bottom screws were the feet? Then you could unscrew the thing with hands. And those who dont like it can use normal screws?
Greetings from Germany
Wow..... amazing progress milestone to reach! 🥳 Bravo! 🎉🎊
Perhaps if you wanted to reduce the price (maybe a ‘lite’ version in the future?) look at using additive manufacturing? An sls/ mjf process would allow for a nylon 6/12 body with little to no modification to the current design.
Would you put the power button on the bottom?
To finish the Apple look? 😂
Routers don't need power buttons.
Exactly. No power button planned.
I can redesign this case it will look exactly the same but drastically cheaper by breaking up the design.
This is what I do
Getting the PCB model shared I guess would allow others to show their ideas for an enclosure or how to simplify the enclosure
Are the final PCB dimensions/measurements available somewhere?
In a previous video in the series I believe there are some details about dimensions, but I can't remember much. Maybe have a look at those videos
HAPPY NEW YEAR from over the northbound mountains (;
respect and appreciation for all your hard work! looking forward for this little beast
I don't see any attempt to make the router rack cage mountable. That is, no provision for additional front panel brackets to screwed to side of router.
Do we know what the max IDS/IPS throughput will be for this router? I am looking for something that can get 10Gbps IPS, mainly just for the hell of it. Unifi's Enterprise gateway accomplishes this at 12.5Gbps IPS but their firewall implementation is just now adding in firewall aliases, or "Zones" as Unifi calls them and I feel like I will be missing other features like that, if I switch from OPNsense to Unifi. So this rack mounted would be perfect if it has enough processing power to handle 10Gbps IPS.
Looking great already, What about ears for rack mounting?
Having used other passively cooled devices and knowing how hot these chips can get, especially after age. You're going to want to consider cooling fins, similar to passively cooled mini pc cases, not sure how much they actually help w/ cooling but at least the fins would provide a standoff that would be some degree cooling to touch. That or some kind of rubber casing or handle so you don't burn your hands.
Dude! Can you consider anodizing if possible without sandblasting and leaving the raw cnc cut look in it? Personally I think cnc milling cuts look awesome! Maybe do some logo engraving after the fact?
I think sandblasting is a required step for anodizing in order to prepare the surface (mainly for cleaning off oxides). Though, I am not educated on the topic, I might be wrong.
Sandblasting isn't required, but in my opinion makes the surface nicer-looking. But yes, it's a subjective thing.
If the enclosure will have a lid, could you not avoid having to machine what you call the "overhangs" with expensive opertations / custom tooling by changing the lid geometry?
I'm so glad the LED's won't be lighting up the room. I would make the drive also LOL Plus you get to try it out. Yes I'd drive also.
Oh yes! I hate blinking LEDs
I trust it'll have rack mount ears and fit in 1U?
Not initially, but they are being considered/planned.
Just curious, did you consider chaining the design of the bottom of the case to not requires the overhang milling in the corners? I would think eliminating that overhangs, you would probably save a good amount of money on the manufacturing.
idk, seems a bit expensive. nice migration to bluesky and good idea with the LED RGB, I need to keep mi switch, rpi and ap in a drawer so i don't get blinded with the lights at night time.
6) Would you consider adding holes for screws for wall-mounting it?
(please also see my other comments)
I love it and would buy it. CNC machined enclosure is so dope.
this case is far to expensive to be any usefull. will it be available wihtout this case?
He told in the past that will be plastic version too
Im concerned on electric shorts happening internally since aluminum hopefully the anodization is also internally leaving only parts that need to be conductive exposed. Also a full range thermal test is probably wanted to see what the operating temperatures are like can it perform well at -18C or 0F to 100F?
Electric shorts are not an issue. There's plenty of space between components and the enclosure. And thermal tests are next on our to-do once the bringup is done.
@@tomazzaman Thanks!😁
Why not using castaluminum / molding often known as die-Casting and then milling ?
It’s less material and cheaper.
Molds are stupid-expensive, and considering this is a sort of prototyping stage, this makes (money-wise) most sense. After everything's done, slicing the case into multiple parts for either quicker CNC or other methods, like half plastic half aluminum (like Apple's Mac Mini).
die-casting + milling costs roughly the same when you factor in mold manufacturing and maintenance.
I love the enthusiasm and the desire for high quality output. OTOH, if the plan is to make a business out of this and sell to ISPs, then I think this is going in the opposite direction.
ISPs (e.g.: Comcast/KPN/Reliant, etc...) buy devices from OEMs and if they’re going to buy from small startups, it’ll be for as little as possible; I know this bc we sold to both OEMs and ISPs.
ISPs aren’t selling high end "hey look how expensive my router looks" status symbols.
Besides, when it comes to failure modes, it's a router and it will likely sit somewhere no one will ever look at. So the failure rate due to bending/warping is part of the failure pareto & cost model.
I love your excitement & the desire to perfect it but I could never invest in such a startup…
Fair enough, thanks for the feedback!
Are you not aiming them at ISP's? Surely this pushes this even further out of the range that ISP's will consider.
Reducing finishing passes on the inside makes sense to optimize cost. I would not be able to wait over the weekend to get the boards. As an American (West Coast), a 2 hour drive for anything is routine. I have done that for a casual dinner with a friend.
What's the commercial sell price point you are aiming for?
The original estimate was around 600$ for the machined aluminium version. Check out some of the first videos on the series for more details on cost estimation.
The "if you have to ask, you can't afford it" tier.
@@AttilaAsztalos there was promised/planned a cheaper version with plastic case. Maybe you're more interested into that :)
Will the final product keep the very esthetical finish on the top surface ? or will it be polished to look like a boring mirror ?
Sandblasted and anodized. It will look like the blue case in this video, just smaller.
1) Why not direct case-to-IHS contact, with thermal paste in-between? That thick thermal pad will affect performance significantly.
2) Also, wouldn't it make more sense to machine a flat bed and then use standoffs like motherboards in pc cases? I am pretty sure that this reduces the machining time and the number of different bits to use.
3) The machining through the USB-C cutout is ingenious, not gonna lie, but I am pretty sure it is the wrong outcome. I am not sure how much trouble it is, but you could swap the reset button with the outer USB-C port, thus, eliminating the need for an additional step in machining.
4) I would also ditch the double bottom, completely unnecessary ( you could, again, use standoffs for some mechanical support for M.2 cards)
I don't want to be the Grinch of this project, but some parts of it don't seem to be the best choice even before the optimization step. But maybe the design choices need more explanation, so maybe, a deep insight video is necessary...
this is amazing :D i really want one when its ready
I am not sure if there is a need to CNC the top of the case 7:38
If you CNC the top you have to sandblast the top and if you don't CNC the top of the Case than you can save the cost for Sandblasting the top.
The edges most likely still need CNC to be a bit smother.
The inside does not need sanding in my opinion.
For the foot with magnets it should be grabble and not a hassle to pick it. Maybe grabbing grooves or holding from the foots it will have any.
Any raw block isn't straight so it's a must to make it flat. But instead of a normal end mill I would use a fly cutter (of something large with max 2 passes) to make it flat.
The first operation is not just for aesthetics. The raw stock is not perfectly square, an initial facing pass is done to make sure the stock sits perfectly square in the machine vice after it’s flipped (to mill out the internal features).
Still I think milling the enclosure out of a solid block of aluminium is a mistake, a sheet metal box and a normal (off the shelf) extruded heatsink ought to be fine for a networking appliance. Producing this part on that DMG Mori machine is not going to be cheap.
@@just.oblivious @Damicske I see top needs to be flat so the internier can be CNCed precisely. That makes sense.
I thought that the block on right side here 0:00 is already a flat surface.
Any way it looks pretty and pretty heavy.
I am curious if we will also see a version with a more affordable case like standard switch case body.
Don't forget to have provisions for rack mounting :)
Wouldnt PTM7950 would be better or not needed ?
The thermal stuff we're using from Wurth doesn't even have a commercial name yet, it's that new. I'm planning a dedicated video on it (and show/run tests). From our initial tests that we did on mockup coolers it performed well enough, but we didn't want to bother with precise numbers and methods yet. We will in a month or two.
RGB front led is nice for the basic debug indicator (and complex ones too). Optimize the inside as much as possible, because it can be a bit rought with larger bits when it'll get somewhat sandblasted/anodized at some level anyways. I mean we ahve the PCB on the inside anyways and it has more prickly edges due to components & rough surfaces anyways, why bother with inside when at most it just needs to not remove skin or cause irritation after handling it.
Is there a future Rack mount / more poe out version?
Wouldn't it have just been easier to make the PCB a little wider so that there would be room for a stand-off on the corner the cages are at?
Why not high quality plastics with a small & quiet fan?
I really like this series, and the idea, but switch from Unifi to something else would be a big ask, will there be continued software updates? Features? And why is this better than say Unifi?
Regarding debugging: I really like old 7 segment displays instead of LEDs. They are easier to read and debug but maybe pricier? Is a UART Terminal planned?
Yes, one of the USB-C ports is a dedicated UART/console port.
I really hope to see this router pop in the homelabs of tech youtubers once in a usable state :) Do you plan on sending few for review? I feel like Level1Tech, NetworkChuck, TechnoTim and Christian Lempra would live to completely terrorize you by testing your product to the limit in front of millions of people hahahaha
You forgot a few: Jeff Geerling, David Bombal and Justin Garrison.
Maybe throw in LTT for the budget
Ok, that seems like a not so good decision. This adds so much cost for so little benefit, there must be a better way. And I say that as an enclosure maker!
Also, that chamfer bit for the USB port - nothing custom needed, front/back chamfer bits are quite common and you can even buy them on Aliexpress.
They plan to have a cheap bulk enclosure as well
Why don’t you make the curve start later so you don’t have to machine the USB-C port that way? So instead of a „normal“ curve, you could use a compound curve and shorten it that way.
Because the curve is machined with an inside cut, and they are already using the smallest bit possible for that. There is no more machining space for a smaller inside curve...
This being said, I seen this as one of the reasons (among many others) why the case design is flawed and I suspect the price is going to be way higher than expected...
So eager to being able to get one of these into my homelab!
My homelab data centre rack in my study has the most blinken lights! Need RGB and status indicators!!!!!!
I'm not sure if it was answered somewhere else, but will the router have precise enough clock to be considered Stratum 2 Time Server?
That's a really good question. One that I don't have an answer to, really because clock precision was never on the very top of my list. Actually, scratch that, I've never even thought about it from this angle. But it's definitely worth testing out!
Looks expensive and awesome. but please consider making it rack mountable.
No fins?
11:30 - Oh my god... You have no idea how many people I had to explain this to... Yes, magnets CAN destroy electronics, but order for it to happen they must be moving relative to the electronic parts to induce currents that flip a bit or fry a delicate circuit, just like in an electric motor/dynamo... It doesn't produce electricity when it's not moving? Come on...
Will there be a early adopter option for a non anodized version?? Please please
There has been a previous survey / call-to-enthusiasts, for a "beta" or testing version. Check back some videos in the series, I don't remember the details.
However, the offering was not a final product at a lower price (early adopter), so don't expect the best performance or fully functional product.
Though, it has been promised that in the future there will be a cheaper version with a plastic case instead of aluminium. You might want to look forward to that.
Yep! We plan to release evaluation kits that'll come in a semi-transparent plastic. It'll be a limited edition, though.
wow, all this was beautiful.
I am searching for affordable high-end router for ages now, but I personally do not care if it is from plastic with off-the-shelf cooler inside (with maybe small lowspeed fan) or from huge block of aluminium. Tomaž, you should consider making also cheaper version with just plastic case as I personally will not buy it if 1/3th of the router cost is nice aluminium case. Selling just the board itself would be also way I think.
5) I also want to ask about Wi-Fi. The router comes without Wi-Fi, but M.2 cards could be installed - right?
Then, does the case provide mounting holes for external Wi-Fi antennas? Can't see this in the video...
With every monthly update, I want it more. This is so on my 2025 wishlist! The wife (who is an IT gear head too) wants it too 😀
100% The to shoud hae smooth rounded over fins to dispate the heat ! And anodized different colors !! MMMM. Sezy !
I don't need a router like this but I really want it
The build up through the videos is getting climactic. I hope all turns out as you hoped for on Friday.
umm, some holes on side would be nice for Rack mount
It has been so amazing watching you design this router, explain it along the way and also be so excited about it. Thank you!
Where I can get that mainboard with sfp+
Hello Tomaž, what about rack ear mounts ?
Coming!
Its going to look amazing but to mass produce it might make a dent into your profit. But it could be marketed definitely for a posh high end version for sure. Nice
awesome progress ! keep rockin
Nice, but if you add some generic rubber feet to this thing, you won't escape the firing squad. "Prekmurska jelša" crosses my mind. Oak is also "prekmurski" enough. :-)
While I'd personally prefer the anodized look, just out of curiosity, is there anyone that would prefer the raw/industrial look if there was a choice? If there are such people, you could make a small run of limited editions.
For the algorithm! Wish you good luck for your endevour!
Thank you!
For me you don't even have to sandblast the outside.
I never look at my router, I just need it to work.
Laptops with aluminum housings are usually smooth on the outside, and full of machining marks on the inside. So it makes sense to do it the same way, and make the outside look better.
i want to have one of these, but i wont even be able to afford the case xD
I can't wait to buy the first router from you :)
Great Video. the routers pcb reminds me of the Banana pi BPI-R4
Make a special edition that is all blinged out and charge an additional fee for it (maybe include some bonus items to make it worthwhile)