Accton WiFi mini-router serial hack

Andrew Hornblow has just paid me a visit,& displayed his ~US$30 Accton MR3201A Open Mesh mini-router serial hack. This hack (c/w level shifter fitted onto a trimmed Kiwi Patch Board),readily allows PICAXE style serial feed, perhaps initially sourced via 433 MHz. Signals can then be WiFi handled & ultimately web displayed -Andrew has uploaded diverse happy user pix. Check how dry(Aust/NZ) or frozen(UK/USA) your garden is from a web cafe half a world away!

Open-Mesh now offer a more professional watch dog fitted OM1P (~US$49) as well.

I'm still getting my mind around wireless possibilities the technique offers, but we're looking at rustling things up as an Instructable or SiChip article. "Aha" moments or PICAXABLE applications welcomed- we may even use them as case studies. Any interest ? Stan

This looks fascinating. I checked out the link - lots of photos of kids doing stuff with this, and hey, if kids can get it to work, surely an old fogey like me can too?

Ok, hack a wireless mesh module to interface to a picaxe. Any chance of some more details?

I presume the mesh grows off a router (which sets the IP address), so you need a minimum of a router and two of these mesh boxes. Is it possible to demonstrate something really simple - an 08M doing a serout, and the bytes appearing on another 08M via this mesh?

Hey- interest 6 months later! Applications abound & it's perhaps time that details of the serial hack were posted. I'm presently ordering a swag,c/w universal PSU, which will work out at ~US$40 each once heady p&p is factored in. Stan.

... only because you gave it a plug in that other thread Stan.
The power of advertising huh?

Obviously advertising does work as I also sought out more info on Open Mesh this time round.

This seems to be a traditional 'hack a router' approach, of which there are many, including steliosm's proposal with BifferBoard ...

http://www.picaxeforum.co.uk/showthread.php?t=15907

There seems to be three issues with such approaches -

First the hardware which may often be expensive or only regionally available with shipping potentially adding a lot of cost. That can limit people's interest and hence why ubiquitous well known brand routers get favoured for hacking as they are over the counter purchases from the local PC store.

Second that such hacking is often less than clearly explained in what can be achieved and how it can be used. I saw Andrew's Picasa pictures last time and simply thought 'so what ? Pretty, but what does it do, how can I use it, how do I use it?'

Third, and most importantly IMO, there's often a complete lack of any detailed How To for those of us who don't have a clue. It often appears to be a Catch 22, if you know what it's about and what you are doing then well and good, if you don't then you never will. It appears there's a very steep learning curve and often little to help one up it.

All this conspires to prevent people becoming involved even if they wanted to. The best 'hacked projects' IMO are those which are treated as pseudo-commercial projects with concise and accurate "buy <this>, do <that>, then you can do <this>" instructions and ideally support forums to help with problems and people who want to take things further, add capabilities.

Really successful hacks are those which are really well explained, easy to do, and made as simple as possible to be done. Most people want an easy lead-in, want to understand what it is they are achieving and what the possibilities could be, and enough information to take that forward if they wish. Most don't have the skills or knowledge to be on the leading edge of creating something, but there will be many more who would like to use the creation.

I'm interested in connecting PICAXE via Wi-Fi or router if the price is right or the gains worthwhile but soldering a serial port on is the least of my concerns; there's a lot more to it. If someone does come up with a 'hack a router to use a PICAXE' project which addresses my concerns I'd probably buy into it and would expect others to do so as well.

The challenge, as in all hacking, is building up a community which creates and champions the hack and creates an environment in which that hack flourishes.

Hippy, that is what I'm trying to do with the OpenBridge project. The project targets 2 groups of people: The ones that want to be part of the developent and the ones that they just want to use the functionality. It clearly depends on the user which groups will join.
All the information will be available for anyone to replicate the functionality or just flash their embedded device and have a network module that just works.

Mmm-HaR ? Yes- it's indeed yet another "Hack a Router",& many (including myself) have been-there-flashed-that. For lucid insights into typical serial hacks perhaps check these LaFonera or LinkSys WRT54 conversions.

As mentioned earlier,Accton offerings caught our eye for their cheapness (as low as US$19 each without PSU),small size & "PICAXEable" Open Mesh user friendliness. We've noted that many youngsters seem happy with WiFi related projects as well- perhaps because there are fewer confusing wires & cables.

I am also interested in this concept and in Steliosm's project, and posted a somewhat detailed howto here ("Net access for picaxe with openWrt router"): http://www.picaxeforum.co.uk/showthread.php?t=14252

If Stelios is proposing to implement something like the Simplelan interface, with expansion to provide the other features he mentions, I would be interested in working on such a project. I like to work in Lua (on the router side), and so appreciate his desire to use that language (which is available just about anywhere, including on PCs) Simplelan was good, if limited, and its interface to the picaxe should provide a standard and expandable sample.

@ lbenson : I overlooked your earlier posting but certainly agree with your comments on OpenWrt and is part of why I threw in the towel as it seemed impossible to go anywhere fast. I'm almost an absolute newbie with Linux, especially with embedded Linux, doubly so on routers, but could go far if I could just get over the early hurdle and reach the epiphany moment.

Lua I glanced at but didn't study it ( too many languages in the universe ) but it looks like it would be useful if it allows programmable serial access which seems the one thing missing for most embedded Linux I've touched; my NAS supports Perl and PL2303 but no serial modules installed. Basically it's having the 'this is all you need' stuff as with PICAXE which would move the idea on.

The advantage of Open Mesh over BifferBoard is it's less than a quarter the cost, two-thirds including shipping one-off to UK, and includes Wi-Fi and a neat case. If I found one in Maplin's for £10 I'd buy one so perhaps we ( 'us' not Rev-Ed ) need to talk to manuka to see what better shipping rates are on offer.

Hippy the OpenMesh router seems like a very cheap alternative to bifferboard, but you will still have to 'hack it' and make it work the way you need it to work. Plus, you need to hack it and also keep in mind that you will have to have a 'control interface' in order to configure the wireless stuff (ESSID, WEP/WPA/WPA2 keys, etc.) as well. So, the way I see it, it will work after you manage to hack it, flash it and configure it.

I'm more into creating a device that will transparently network-enable your project. I saw in this forum more projects using a simplelan module rather than using a hacked router. This is actually telling me that a transparent network device will help more people network-enabling their projects and also create new projects that are networked enabled.

The way I see it, is that we should be focusing for any new project on usability, fun and interaction rather than figuring out how our project will send a 'Hello World' out to the Internet.

The price is an issue of course. I've seen similar 'easy networking' approaches which would require a lot of developing and testing hours of work. So, they end up not being cheap after all.

@lbenson: You are welcomed to help me out with the project. I have the embedded linux part covered, but I do need help writing Lua code.

I agree with Hippy's comments here. I'd like to build useful wireless net-able data connections - and even though I'm an amateur radio operator and experimenter there appears to be one heck of a learning curve to much of the web-wireless world.

I would very much appreciate seeing a website that described it all clearly, and showed me exactly what to do with what (and why) in order to get such a system up and running. A site that defined terminology - showed close-up pictures - etc.

I seem to keep bringing up my friend, Max in these discussions. His website shows his projects in great detail - photos - schematics - theory of operation - software - including everything that might help someone do what he did, as well as provide them with a solid base to do more.

The sites of some of the PICAXE forum contributors do so as well. It's simply the old adage, "Anything worth doing is worth doing well." We forget that old saying far too often these days, and wind up with a newer adage that applies far too often, "Garbage in - garbage out."

I sure wish everyone had the time and energy to be as thorough as these sites. It would turn the web from merely a useful tool into a veritable heaven for all hobbyists.

So much for the rant. Like I said - I agree with Hippy.

@John West,

What, then, is Max's website?

Basically any router which has usb and can run openWrt (bifferboard has recently returned to openWrt as default recommended OS) can interact through usb serial with a picaxe--and of course, connect to the internet. Some can do so using WIFI, tho wireless client is problematic on openWrt machines (someone please show me how).

OpenWrt-capable devices include at a minimum, Asus wl-520gU (today, $33.27 at Newegg), Asus wl-500gPv2, Buffalo wzr-hp-g300nh, TP-LINK TL-WR1043, Bifferboard, NSLU2, Sheevaplug, Dockstar (the latter 3 can run fuller linux OSs like Debian). None of these are hard to flash.

Linux/unix is notoriously complex, and on these machines you use the command line (openWrt does support a web page setup interface, LUCI). Nevertheless, it requires only a tiny subset of linux to enable these devices to mediate between a picaxe and the internet. I am myself a linux newbee, tho less so than I had once hoped to be. I'll try to provide a detailed howto on setting up a WL-520gU, as an addition to the previous howto I posted.

MPep said:

@John West,

What, then, is Max's website?

Click to expand...

maxmcarter.com

I worked with him on a few of the projects on his site - notably the twist-modulation technique and the laser communications project. I also wrote the article on Longwave xmitting antennas and one on converting Syntor-X 2-way radios to the amateur radio bands. My stuff is probably the least coherent and least detailed on his site.

The whole point being that some rare website creators work hard and smart and do a beautiful job of defining and describing a project and showing others how to put it all together and make it work, giving their site true value, (we call it "sweat equity" in the states) - while most folks just slap incomplete junk up on a website and waste enormous amounts of other web viewer's time.

I've spent hour after hour going to so many crappy pseudo-tech websites looking for bits of useful information, I could barf. Gotta be a better way. Whatever methodology the prominent search engines use for listing web sites just plain sucks.

It's too bad there is not yet a good "quality index" that can be applied to websites. If we could stop having to wade through all the blatant crap, the intrinsic value of the Internet would skyrocket.

It's too bad there is not yet a good "quality index" that can be applied to websites. If we could stop having to wade through all the blatant crap, the intrinsic value of the Internet would skyrocket.

Click to expand...

Too true John, too true.

With every respect to web publishers (myself included),it's fair to say that "site quality" may be very prone to enormous personal & cultural interpretation. It's often near impossible to get folks (even within the same country/gender/age group etc) to agree on what constitutes powerful style/design/message. Some reach for the IR remote at a mere glimpse of such recent gorgeous visuals as "Toy Story 3" & "Avatar"-special effects in the latter were done by Weta Digital here in NZ as it happens.

Let's not forget today's sophisticated web surfers may justifiably have suspicions about motives behind gold plated sites. Plan & simple messages may be more lucid overall than bells & whistles. Hence the Australian "Silicon Chip" monthly has a multi decades duration "Serviceman's Log" column that makes for outstanding reading- it's apparently the first article most regulars read.

Style interpretation is perhaps akin to musical appreciation - most of us flower power oldies can't stand even the thought of today's offerings, and vice versa -Frank Sinatra's songs are painful torment to most teenagers...

But back on the "Open Mesh" topic. Their P&P is via Federal Express & it transpires- shock- the UK is not on their preferred country shipping list. P&P quotes to NZ vary between US$40 & $70, but seem almost unrelated to what's actually ordered. Hence a single US$19 Accton mini router incurs almost the same p&p as a swag of routers c/w power supplies! The trick looks to be bulk purchase & share out locally after arrival- any Kiwis fancy joining my latest 6 pak order? Stan.

Just discovered Open Mesh web site dollar prices are USD not NZD nor AUD, so that pushes costs up slightly, about the same as BifferBoard for a one-off. Multiple unit orders can spread shipping cost but needs someone to take management of that on.

Another aspect since manuka mentioned LaFonera and lbenson the NSLU2; hardware longevity and availability. BifferBoard wins on that as a purpose built product, where repurposed or hacked kit designed for something else can drastically change or cease to exist overnight. This can particularly be the case where a product is sold at minimal cost with the business making its profit from service charges which a hacked product won't ever deliver.

The two obvious consequences are that much of the effort in creating and documenting a hack can go to waste and, while there's a perfectly good hack, there's 'nothing' to use it on.

Re longevity, the NSLU2, for example, has been out of production for years, yet it remains a useful device. I have 4 of them running, including one controlling a picaxe through usb serial for pan and tilt for a usb webcam. If the NSLU2 dies, I won't have to replace it with another--any device which has two usb ports and supports openWrt will do. For that matter, any programmable linux device with two usb ports will do.

Hardware always changes, but if the old software still runs, you may have the longevity you require. I still have a dos grep.com program from 1988. It's probably been on a dozen different PCs. I use it almost daily to find tidbits from various text files of notes I have taken over the years. Likewise with the picaxe. One might have a program written for an 08. If your last 08 gives up its smoke, and you can't find another, an 08M will do, or a 40X2.

I definitely agree with all that in principle; I'm using similar pre-Windows software today but the caveat is "if the software runs" and you can find a compatible platform. That there are various builds of firmware required for different hardware, the need for specific files to get particular platforms running once flashed, works on this but not on that incompatibilities and all that suggests it's not always as plain-sailing as it could be.

The 08 to 40X2 analogy is interesting; you need to change to using new pin names, may run into TTL v ST input problems if not careful and max chip current capabilities on others, and if the code relies on the 08's 16-step ADC that may need software changing. Not always as simple as take your code and run it on new hardware. As long as someone can achieve what's needed it's not a problem.

Look at all that effort which went into making iPlayer streaming to various non-supported systems possible which has all come to nothing with a change at the BBC's end.

Even Microsoft had to have an XP Mode for Windows 7 as software which should just run sometimes doesn't. I think we've all seen the problems which can arise with USB drivers.

I'm not knocking the idea, but as I noted I've got a NAS running Perl and PL2303 drivers, but no one has been able to show me what I need to do to be able to communicate with the serial port. Nor has anyone been able to tell me how I can assign two distinct IP addressess to the single LAN port. All should be possible, apparently, but for me a promising start hit a complete brick wall.

As long as those brick walls can be avoided then everything will be great.

Hippy, that is exactly the problems that I'm looking into resolving. Obviously, it's not easy to support any kind of linux embedded devices running various versions of OpenWRT.
I'm taking out the hardware from the picture and I introduce a replacement hardware that anyone will be able to use for as a networking module. The end user will never have to know what it's inside the hardware or what version of software it's running.

I've moved my NAS discussions to the Sandbox ...

http://www.picaxeforum.co.uk/showthread.php?t=15921