WRTnode Opened for $25 | Hacker News Reader

WRTnode Opened for $25

wrtnode.com

180 pointsNoel_lo93 comments

Great effort, lovely little device. Building your own hardware is not easy to do cost effectively!

But a few niggles:

- Quoting RAM and flash in Mbits? Whilst that may be the norm for buying components, I'd argue most people familiar with OpenWrt and the Raspberry-Pi generation are going to miss this detail and think "Oh wow, 512MB RAM and 128MB flash!", when in fact it's 1/8th of that.

- There's some mention of a standard shield, but no information on what it includes or when it will be available.

- Like others have said, that scrolling on the website is extremely annoying.

Regardless, I will definitely be buying one anyway.

I once saw an SD card sold as "8gb" and thought it terribly expensive for that much capacity.

At least it wasn't "mb". I am not even sure how that would work.

The Mbit thing is quite common in the embedded world. When I saw this my first thought was "oh wow, twice the specs of my current favorite OpenWRT hack platform (the TL-703N) at about the same price and with gpio exposed. Awesome.)"

Hmmm I'd agree it's common when you're talking about components with hardware developers, but when you're pitching it to software developers/users then RAM/flash is almost uniformly referenced in megabytes. Take for example:

- OpenWrt's table of hardware page: http://wiki.openwrt.org/toh/start

- Raspberry Pi specs: http://www.raspberrypi.org/product/model-b/

- Minnowboard Max specs: http://www.minnowboard.org/meet-minnowboard-max/

- Beaglebone Black specs: http://beagleboard.org/black

Oh and I love the reference to TL-WR703N, that little thing was pretty special! I was always disappointed by the MR3020, which is basically identical but a lot larger for no obvious reason.

I never see Mbit in the embedded world when talking about RAM or Flash. I always see KB or MB, never Mbit.

Agreed, bits should only be used for bandwidth.

It feels like theres like a Lagrangian point out there in device space that these devices are converging on. It doesn't quite exist yet, but we are getting tantalizingly close. It networks like a Cisco, runs linux like a server, has GPIO like an Arduino, and costs as much as a Latte.

Once it exists, its almost like computorium, becoming the center of nearly every maker project and then staying there because its so cheap you don't have to choose which projects get your precious dev boards.

This is a fantastic step towards that goal.

and it's sold in packs of 6 or 12 at your local discount store or walmart, right next to the pens.

Serious makers want this price point with an atheros chipset. In lieu of that, the open-mesh.com routers are a good choice if you are trying to do low level wireless innovation.

I hope this trend of really annoying scroll-skipping goes away like flashing "under construction" .gifs and framed pages.

Absolutely. It's all very well and clever until you can't read all the text on the final page because it flows off the bottom.

Website is totally obnoxious on a mobile device - no pinch zoom, no scrolling, must press little dots to "scroll".

It's just as obnoxious to try scroll through on desktop.

Any electronics-related web page that uses "shield" other than as meaning "RF emission/reception blocking element" is automatically obnoxious, mobile or not.

it is a very common and well known term in the Arduino world, and the physical appearance of the device with all of the header pins would immediately suggest shield applications to anyone who has spent time with Arduino

Why is 'shield' more appealing than 'daughterboard' anyway? Gender neutrality?

Linguistic economy.

"Card" is two characters less than "shield", and has a long history in computing denoting a circuit board that is plugged into another circuit board. (Whereas "daughter" is usually not redundant in "daughterboard", it can almost always be dropped from "daughtercard".)

Examples: network card; disk controller card; video card; SD card.

"Shield" is such an entrenched term in electronics, from long before computing, that it's completely stupid to try give it another meaning. It's also not from some specialized sub-domain whose rare terminology can be reused; shielding is employed in all kinds of devices.

Maybe not if you take into account all the people who will need an explanation of why this item is called a 'shield'.

Opened the web site, tried to scroll on my touchscreen, couldn't and immediately closed it. Then I read these comments and went back.

A lot of really bad ideas are being used on web sites now to make them look sophisticated. Ironic considering all of the simple to use tools developers now have to measure conversion rates and user engagement.

"Designers" trying to show off and that dont understand anything about basic usability.

To the "designers" :

Next time do a simple web page without stupid scrolling hacks,for god sake.I'm sick of it ,really. Dont mess up with basic usability just because it looks good on your Macintosh.I guess you didnt even bother testing on mobile.

Sorry for that rant,but we are in 2014,and I cant take more of that stuff.

Nope, can confirm it sucks on mac desktops too. At least with an Apple magic mouse.

I bet it looks brilliant in Lynx

A couple of years ago, your comment (as well as mine) would have been downvoted. Now - only my comment will be downvoted. What's HN going to be in another couple of years?

On desktop, if I scroll, I get jumped to a random part of the screen (Chrome), so it's not just you. I found that if I barely scroll (just a little tap in a direction on the mouse), it might go the same direction I'm scrolling. Otherwise, it goes to a random part of the page.

Website is also obnoxious on a desktop device.

Even worse on a desktop. No scroll bars, no scrolling at all. That looks like it is because the idiot designer put "overflow: hidden" everywhere.

Wow, thanks, I didn't even realize there was more than one page because there was no scroll bar.

Sorry for the bad webpage design, we will change it but may take a while, actually there's no web front-end person in our small team, all the site was done by our-self while coding Linux or pcb lay-out.

Could I have this please with 802.11ac WiFi, 5GHz, preferrably supporting a 4x4 MIMO setup. I would love to have an open access point for my network. I like to have more than 300Mbps and to avoid the 2.4GHz network though even more. I understand this owuld be more expensive, but would be willing to pay a higher price.

I don't think networking is the market of this thing, despite the name. In fact, it could probably use a different name. It's a mips24k; I doubt it can do software networking at the 300mb speed the wifi runs at.

Do you have an idea how much of bandwidth a MIPS cpu like that can push?

The chip is actually made for routers: it has 802.11n "MAC/BBP", 5-port 10/100 switch, etc. http://www.anz.ru/files/mediatek/MT7620_Datasheet.pdf I wonder how "Hardware NAT with IPv6 and 2 Gbps wired speed" works if the switch is only 100 Mbits...

I guess if what you want out of this is really a router, you are supposed to either build the rest of the hardware yourself or wait for yourself, or wait for the "motherboard" module for this.

I looked into a lot of these boards a while back and most are like this one: crippled with 100mbps Ethernet. Sad to see that yes, this one is too. The Ethernet module is far slower than what the CPU can push.

Doesn't destroy it, but it's certainly a black mark. Are 1gig chipsets that much more expensive than 100mbps these days?

It's a chip for very low-end routers: 2.4GHz only, no gigabit switch (although apparently it can be equipped with two gigabit ports), and a CPU that would struggle to do 50Mbps WAN-LAN with firewall and QoS. It's only redeeming quality is that it apparently has a PCIe lane, but the CPU is still going to have trouble feeding whatever NIC you connect to that.

I do, thanks.Hmm, let's see if I can give you the information without breaking a confidentiality agreement. Ah, here is a discussion on the openwrt mailing list: https://lists.openwrt.org/pipermail/openwrt-devel/2013-Augus... As you can see, they reported 90mb/s (they also quote 710, but I think that is the switch doing that). And that's on a mips74k.

I know it's made for routers, but most likely not for doing a lot in software. As you quote, it does hardware NAT. There's a reason they put that in.

There isn't a fundamental reason why the mips couldn't do the job, but running stock linux it's going to be thrashing its L1 cache, and it doesn't have an L2. Also, next time you play with one of those,check the hardware counters for how many instructions/sec it actually manages to process running flat out :-)

90Mb/s at least? Makes sense that they won't go for the gigabit Ethernet. Any idea how that compares to ARM with the same frequency? I'd say about the same. That's not bad, right?

I would guess about the same; I didn't have two similar ones to compare. The mips might be a bit slower though. I had the sense that the limitation was really L1 thrashing of code space, so the ARM being a bit denser might help. It would also help if the SOC had a smaller latency to memory as well - don't know what this one has.

Bear in mind that it's probably a packet rate rather than a bit rate limitation. That really bit us in our application because we were trying to run many lines of voip, which means lots of small packets. - YMMV.

Good point. I remember working on MIPS M4k CPUs many years ago which could almost service a single 1Gbps link (at 100% CPU though). Without a gigabit link on the other side fast WLAN is pointless though. It was just too tempting to see a little board for networking. I agree though, networking is not the primary market.

Expensive enough to largely be impractical.

From what I've seen, open source development that involves hardware typically involves reverse engineering specifications and/or emulation. Neither is practical for state-of-the-art.

Not to mention, MIMO is just a "wee" bit tougher than simpler schemes.

  802.11ac WiFi, 5GHz, preferrably supporting a 4x4 MIMO

Retail price just went up to $300. Four dual-band wifi radios are expensive.

If I could have a router or even just an AP with complete hardware documentation in public and 100% open source firmware for it that would be an ok price. With the advertised $25 of course greed kicked in as well. ;-)

I'm curious to know why you want to "avoid the 2.4GHz network". Can you please explain why?

I can't speak for rffn, but many people want to avoid using the 2.4GHz spectrum due to congestion.

In the 2.4GHz band there is a limited amount of space available, and there are now many wireless routers in typical residential neighbourhoods all competing for the same space. Many other devices such as cordless phones, baby monitors, and wireless speakers also use 2.4GHz and interfere with wifi signals. Most microwaves generate significant amounts of noise in the 2.4 GHz band when in use.

In the 5GHz band there is much more space available, and there are many fewer competing devices and sources of interference. The penalty of 5GHz is slightly less range, and more expensive wifi equipment.

Exactly. Where I live I see about 20 networks in 2.4GHz, not even talking about Bluetooth, 802.15.x and other 2.4GHz users. Over the latter WLAN should have priority though if the other devices are standard conforming. In 5GHz I can still only see one or two other networks.

Scrolling on this site w/ a mac is incredibly frustrating. It keeps resetting to the top of the page.

Linux too.

This webpage doesn't seem to interact well with Mac's touchpad scrolling...

It doesn't interact well with any scrolling

Is it open hardware if the ISA of the processor is not free (in both senses)? I think MIPS (which this uses) is free, but I see a lot of open hardware where I doubt the ISA, much less the architecture, is free.

Generally in the "maker" community open hardware is based on layout and software level openeness. So board layouts are shared even if the internals are not.

This is what allows adruino to be open as they share layout files and software, but the Atmel AVR micro-controller they use does not have an open design.

It does get confusing since there are open hardware projects that go down to silicon, but for the most part that level of design is just beneath the concern of the people who want to internet enable their fridge or whatever. Just like most computers contain closed source code in the form of controllers and all but the most extreme open source types just don't care that they have a closed source USB controller.

Well, my beef with it is that the proprietary blobs should be interchangeable, and that requires an open ISA, the silicone can be as proprietary as it wants, as long as the interface is open.

That would make any attempt at creating such a device very hard, if not futile, and not economical.

The ISA is just a small part of what's required. All the other interfaces are just as important such as pinouts, memory mappings, memory interface, bus support and so on.

There's very few SoCs from different vendors that are interchangeable even among the ones that uses an open ISA. This requires you to either build everyhing from scratch including the CPU, or to use more generic and perhaps interchangeable CPU or microcontrollers and rather push the special components to other chips instead of having those components on a SoC - which gains you nothing.

Meaning the value of using an open ISA is miniscule as long as the code is available and easily translated to another ISA by a compiler.

The problem I often see with "open" devices like that is that they typically work with exactly one Linux kernel binary that shipped with the device. They usually don't work with upstream kernels.

Even if you can get original kernel source, it might not be trivial to get it to compile. Porting the required changes to a newer kernel once a security vulnerability was discovered can be very time consuming.

my favorite is the banana pi: gigabit, sata etc

http://www.lemaker.org/

http://www.linuxx.eu/2014/08/banana-pi-raspberry-pi-upgraded...

uhm I checked if this could be a good solution for mediaplayer usage, but it seems not: http://openelec.tv/forum/124-raspberry-pi/71202-banana-pi

What are the power specs? For battery powered applications like the shown mini-quadcopter you really need to know the watts as well as weight. Admittedly, WiFi has a very spiky power consumption profile.

Current consumption: 0.13A - same module, wifi off 0.17A - same module, wifi enabled 0.14A - dock (RJ45 included) + wifi off 0.20A - dock (RJ45 included) + wifi enabled

according to one of WRTnode from http://openrouter.info/forum/viewtopic.php?p=36058

and WRTnode is about 20g in weight.

Looks like a Carambola2 (19€). http://8devices.com/carambola-2

Same amount of RAM and flash, but different SOC.

There's several similar boards; off-hand I can think of the VoCore, the AsiaRF AWM002, and a couple of unbranded Chinese ones.

This looks really awesome. I've been contemplating a KickStarter project for a Linux networking device for a while, and this may well make it possible at an attractive per-device price point. Everything else was either too pricey or not quite powerful enough.

Previous submission that sank: https://news.ycombinator.com/item?id=8181321.

This proves one thing: you need to keep trying.

or get a "flashy" website that enrages people enough to upvote rant

Can anyone comment on the read/write performance of SPI flash? Seems like it would be similar in performance to SD card, but without the wear leveling?

AFAIK, SD cards do not have wear leveling, SSD drives do.

Many SD cards have wear leveling built in, but most of the algorithms and CPUs used for the controllers are quite crap at wear leveling. If you start looking at more industrial targeted SD cards you'll find quite decent wear leveling.

Most of them do some kind of wear leveling, but they usually use cheaper flash and less capable controllers and the controller algorithms are mostly oriented around making them usable for sequential writes as you see with a digital camera, not for making a good drop-in replacement for rotating disks.

SPI flash is usually NOR flash (not NAND like you'd find in an SD card or SSD). NOR flash is usually SLC and will take 100k writes/erases and is quite simple to access (as compared to MLC NAND which only usually takes 3k erases and is a pain to access comparatively due to things like ECC requirement).

SPI flash generally will do on the order of 0.5 MB/s erases, slightly faster writes, and somewhere in the realm of 4 to 12 MB/s reads, give or take. It's not high performance but it's simple, cheap, and low pin-count. The host controller is responsible for wear leveling, erasing each needed sector/block, and all writes, as SPI flash is unmanaged (compared to SD cards which are managed memory).

The best mips based wireless router dev board I've ever seen! and I've added an OLED screen on it, works like a charm ;-)

An excellent price point with wifi included.

hardware is very good

The board is very good

good hardware

I always find it hard to take things seriously when they are written in poor english. "Will coming soon" makes me worry about the quality of the rest of the project, maybe more than I should.

FWIW, I have ordered some parts from Seeedstudio (including a Xadow board) and was quite satisfied with the service.

Seeed isn't a typo, to save anyone else checking.

Thanks for pointing out, we corrected it. :)

If I can share a suggestion: the phrase "Born to be proud" comes off as a bit arrogant and doesn't sound natural to a native English speaker. If you're proud of the product, or you're trying to convey the idea that it was created with good intentions, some different wording might get your point across better. Something like "a better open platform" or "born to be open" [even though that also doesn't sound natural either, unless you know what 'open hardware' is].

Here i've cleaned up some of the copy from your sales page:

"WRTnode is a high performance, low power, small, inexpensive development board intended to run OpenWRT OS. It also easily ports open source software based on Linux.

Today, more and more devices are connecting to The Internet of Things. The WRTnode is a small node which can connect to the internet all by itself and perform complex tasks, like detecting a cat running by, or identifying your speech and twittering it, or checking your email and speaking it to you, or learning how your room is laid out and finding garbage to sweep while streaming video to you over the internet. [..] Open Source hardware for OpenWrt [..] [remove the line "smart machines' heart"] complete set of high performance I/O [..] Local dns uses i.wrtno.de, wrtnode.lan, and openwrt.lan [..] default WRTnode ip 192.168.8.1 [..] WRTnode additional features (source at github.com/WRTnode):"

  ..."Born to be proud" [...] doesn't sound natural to a native English speaker...

Have to agree. It's not idiomatic English, so it really fails when given such prominence on the site.

It also has echoes of classic 1970s rock, which doesn't help (born to be wild, born to run).

Made with pride is the most literal translation. I think the intent though is Born to be free however there is no good translation of that given the many meanings of free that even native speakers cannot understand ( https://en.wikipedia.org/wiki/Libre )

It looks good now. Next time I will try your support link instead, but I am glad you saw the comment and got it cleaned up.

In the future it shouldn't be too hard to find someone willing to support and a good open source project with some proof-reading. I may not be the best candidate, but it would be a good way for someone to get started helping the project, and could make a big difference if you are targeting a larger audience.

Thank you very much, It's very kind of you all if there's anyone like chrisBob mentioned who willing to help us with our poor English, pls mail me: noel@wrtnode.com, we could pay for it. :)

Next time you translate your website to Chinese let us know how it goes. Personally I appreciate that they have made the site accessible to me as an English speaker and am fine reading around any grammatical shortcomings.

Thank you.

When I was taking Chinese classes in university, people would sometimes poke fun at the eccentric or badly accented English. I would as them how their Chinese sounds to their native-speaking instructors.

It is strange how the reaction to that was always one of confusion or dismissal.

I bet if we could read their Chinese website, the copy would be perfect.

Some manager who probably doesn't speak much English has tasked a random person on their staff who went to college in the US or UK with writing their English copy, without realizing the difference between "this person can communicate effectively 1:1 in English" and "this person can write formal English like a native speaker".

In other words, a task is being supervised by someone who doesn't understand the task. Just like every other company we deal with. As native (or at least highly fluent) speakers, we just happen to be competent to critique it.

Yes, the Chinese website (http://cn.wrtnode.com/) is perfectly fine. (Chinese is my second language.)

However, I would not immediately characterise the copy as written by a US/UK college grad though - it feels like it is written by someone who learnt English on the mainland.

I don't really know why companies in various parts of East Asia don't hire/outsource to someone who actually is highly fluent in English to write/edit their copy. Sometimes I suspect it is simply due to cost, other times I suspect they just don't realise that their fluency is not quite.. 'international English' level

(No offence to anyone who runs/writes for East Asian company websites.)

Absolutely. It is ungeileivable.

For my coworkers based in Taiwan who are somewhat fluent, they mostly don't recognize it, and really can't.

Their limitations are effectively hidden from them. Odd things still come out, but in informal contexts it's not worth correcting, and there's little opportunity for formal critique. (There's also the issue of "face" that complicates matters and drives me insane. It's easier with people who have spent a few years in the US, though.)

If there's no (near-)native speaker around to check, nobody's going to notice until it goes out to the world, and then there are a half billion people ready to jump on it.

If they recognized it, I don't think cost would be a factor. I could review/correct a site like this in an hour (or less). I might spend a day on the mini-manuals that come with some gadgets.

They don't need better translation. They need a cultural liaison. You can speak English at a very high level but without deep cultural involvement, you won't know that "get out of the city" and "get out of town" have very different meanings.

When the peasants are revolting its very different from when the food is revolting!

"I always find it hard to take things seriously when they are written in poor english."

You realize English should be capitalized, right?

The words realise and capitalise should be spelt with an 's', not a 'z'.

AFAIK in British English, yes, in American English, no.

Reply on news.ycombinator.com