Letters

Toward the latter half of WWII, landing strips and airbases were being built all over the Pacific on small, otherwise totally isolated islands. Little contact was made with the natives to explain what was going on, and in some cases the locals were incredulous to see the new arrivals performing strange rituals on their tarmac runways and receiving airdropped “gifts” from the gods/ancestors.

When the bases were abandoned, in some cases the locals started replicating the structures and activities of the base personnel, performing marches and drills, talking into replica radios, and waving lit firebrands on the runways and waving them. Elements of Christianity were often woven into the practise in an attempt to attract the same blessings as the military staff had. These culturally contaminated pseudoreligions are called “cargo cults“, and these words have since been applied to any misapplication of a procedure to an end without understanding how it works.

It might be premature of me to describe modern Electrophoresis as it is performed in most labs as “Cargo-cultish” until I’ve properly demonstrated that I have a better alternative, but having read this paper over coffee this morning, I was really reminded how little me or my fellow lab staff really bother to understand Electrophoresis, a fundamental, essential part of our daily work with DNA and the part whose quality is usually what ends up being scrutinised on peer reviewed papers!

Ask someone which buffer to use and you’re always offered one of three options: 1x TAE (Tris-Acetate-EDTA), 1x TBE (Tris-Borate-EDTA) or 0.5 TBE for those edgy folks who’ve been doing this for a while. TBE offers better resolution when you go to look at your results, and can be run at higher temperatures, so people often favour TBE for routine use. Some people in the know will suggest TAE where you’ll need to do something to the DNA afterwards, because the Borate in TBE is high enough concentration that it can mess up DNA extractions from the Gel and enzymatic reactions later. I only learned this upon reading the aforementioned paper, and I’ve frequently lost my DNA upon gel extraction, which I now realise may be down to my selection of buffer.

As the paper intricately explains, the use of Tris buffers is a holdover from DNA electrophoresis being borrowed from protein electrophoresis. What followed after its adoption was a brief spree of experimentation, followed by a trend towards standardisation that lead to everyone adopting TAE or TBE so they could all understand one another’s protocols. The assumption was that any effort put into improving gel resolution was wasted effort, as only marginal gains would be made.

However, there is a big problem with Tris buffers that haunts anyone performing Electrophoresis and causes all manner of problems, even when it isn’t recognised to be the cause of these problems. Heat generation due to the current passing through the gel can cause the gel to soften or even melt, and worse still as the heat increases the conductivity of the gel increases also, leading to runaway heating and poor outcomes.

Problems caused directly or indirectly by heat include:

• “Smiling” bands,
• Smeared or blurred bands due to uneven softening of gel surface
• Band diffusion, forcing one to perform oversized gel extractions
• Total Gel Meltdowns
• Having to run at low voltages and missing dinner.

The reason behind this runaway reaction is the inclusion of Tris and excess Sodium from the preparation of the EDTA used in the buffer.

EDTA, they argue, isn’t even needed nowadays. Apparently they ran a gel using creek water without EDTA and had few issues; the enzymes used today have few unwanted activities under electrophoresing conditions. Or so we are told.

Tris, meanwhile, is too conductive and permits a lot of current through the Gel at a given voltage. As the gel heats, it lets even more through.

The authors experimented instead with a simple electrophoresis buffer composed of Sodium Borate at 10mMol (among other experimental buffers) and found it to be ideal on almost all fronts; Sodium Borate is cheap, can be prepared as a buffer from only one crystalline ingredient, gives excellent resolution, heats very slowly and can be used at very high voltages without issue (meaning faster gels).

If this is true, it means that Gel Electrophoresis as it is commonly performed is not only flawed but overpriced; a quick check of the price of 1L 10x TBE buffer from Bioplus shows that it costs $12.75. For 17.49/kg from the Science Company, I can make 52L of 10x Sodium Borate (1x SB Buffer is 1.907g/L Disodium Borate Decahydrate, the usual crystal form).

So the costs per litre of the two are $1.28 for TBE, and $0.04 for SB. For your investment of 4c you get better gel resolution at higher voltages (meaning less time waiting), which can apparently be more reliably purified and used downstream for enzymatic reactions such as ligation or PCR.

This is great if it’s true in my labwork at the CCRC, but it’s even better news for DIYbio folks worldwide: You can often find Sodium Borate on ebay where you’ll never see TAE/TBE, and the cost difference is pretty staggering.

More on this when I next run a gel: I have a bottle of 5x and 1x SB buffer next to me, just waiting for me to add Agarose and give it a try.

Rapid Prototyping is addictive. I knew it would be when I bought the Makerbot, but I didn’t know just how much. Since the moderate success of Dremelfuge, I’ve gone a step further into multi-part, assembled devices, and I’m proud of the result.

Microlathe is a Makerbot/Reprap printable Lathe that uses a Dremel for rotary power. I spent a day and a half designing the first draft of it in OpenSCAD, another evening printing the parts, and the minutes I could grab over the last few evenings testing it. The result? It’s fairly hazardous, requires careful balancing, and it works just fine on wood dowelling. So on the whole, a big success!

Here’s a link to a video of me demonstrating Microlathe. It’s taken on my HTC in low light, so the quality is poor; apologies! Microlathe draft two will merit High Definition, I think.

One of the reasons I designed and made Microlathe was because I wanted a free lathe. Another reason was to contribute to a pattern of accelerating returns I’ve become aware of and excited about recently, in the sphere of rapid prototyping.

When you give someone a tool, they can use that tool (within its limitations) to make things. You might reasonably expect someone with an axe and some trees to make fuel, but they could in principal master the art of the axe and use it to create lasting works like furniture or a shelter. It’s just really really hard to do with an axe, is all. A clever artisan might use his axe to make a simple mallet, which would enhance his ability to accurately control the force of the axe blade. He might use these to produce still more useful tools and products. This is an example of accelerating returns; tools making better tools to make better tools faster.

Leap forward, and give someone a rapid prototyping machine. These tools, be they laser cutters or 3D printers, are hugely open ended, and because they use CNC control the user can take his or her time in carefully planning each parameter of the final work before beginning, and even share the result if it works with other users. The essence of “Measure twice, cut once” in carpentry and many other crafts boils down to “If you muck up one step of the way, you’ve ruined the end result”. This reliance on expert skill and patience in creative or constructive arts has probably been one of the biggest barriers to people getting involved in making until recently, and with a CNC machine, it is no longer strictly necessary. This means you don’t have to spend years mastering the art of the axe just to make your mallet, you just need to click “print”.

So with this one starting tool, you can imagine a situation where a person can design/download and assemble a plethora of relatively complex tools in short order given only the cheap feedstock needed to run the Rapid Prototyper. That’s what I’m looking forward to and trying to drive forward, because this accelerating return is going to help push innovation to new heights at a grassroots level.

A Makerbot cost me about €750, all told (Of which €110 was shipping!). A full suite of tools might cost me that much or more, easily, but if they were printable it’d cost me a few extra euro. Probably less than 20. Given enough feedstock, I can even make the structural parts of another printer, and give that to my friend, who can make a fab lab of his own, etc. etc.

Obviously the Makerbot can’t make everything. For example, it can’t make its own heater barrel, which needs to be made out of metal. That task calls for a lathe. And now, you can print one of those. And because it’s open-source and available online, if it doesn’t work right or suit you as-is, you can just improve it.

Someday I’m hoping to see a printable CNC Router and Lathe on Thingiverse, so that I’ll be able to have my computer-controlled robot fablab build me almost anything I can desire or imagine, fast. The first thing I’ll probably make when it’s ready? Another Fab Lab!

Since my last post, I’ve been a very busy person.

Dremelfuge is now available for purchase on Shapeways from my shop there. There are two versions, one with an axle for chuck-fitting machines, and another with a bore into which the cutting-disc-holder from a standard dremel can be fitted. Price varies by location, but even at the $65 price which includes shipping to Ireland plus VAT, you could buy a Dremel to match it and still come in under €100 for a functioning centrifuge. I gather the price falls to $55 for American buyers.

Here’s a video of me demonstrating Dremelfuge. I tested it with standard microcentrifuge tubes, and found that it stably spins them anywhere from 5000g to somewhere above 20000g. I say “somewhere above” because the tubes shatterd somewhere between the third speed setting and the fifth on the dremel.
The math shows that the average force on a microcentrifuge tube quickly exceeds that of the commercial centrifuges I use in the lab. They go as high as 14,000g. Dremelfuge plus a Dremel 300 can put over 50,000g on a sample. Except that’s too much for the tubes so they shatter.

One nice bonus is that it seems to be very stable on a Dremel 300; there’s little to no vibration or rattling, even with highly unbalanced loads.

So here I have it: A centrifuge attachment for drills or rotary tools which spins them with even more power than the official thing, and costs a tiny fraction of the price to make and operate. I call that a success by every metric!

Thanks to Makerbot for making this possible in the first place, and my fiancee and family for their patience.

As always, I don’t endorse use of Dremelfuge as anything but an ornament, for reasons of liability.

Update: I’ve tested Dremelfuge in my lab with E.coli cells and HL60 human suspension cells. It pellets both excellently! I’ve already shown it to spin down Miniprep columns, and the math shows it hugely exceeds the power of a standard lab centrifuge when used with a Dremel 300 (€89 in Argos and useful for just about everything else, too).

So that’s it as far as I’m concerned: Dremelfuge is a fully functioning centrifuge. Can’t wait to see it in use on some cool projects!

I’ve had on my mind an idea for some time that I’ve wanted to try. Having a Makerbot has enabled me to experiment with mad science on a level I’ve not been able to before, so here it is: DremelFuge, a printable drill/rotary tool attachment that spins microcentrifuge tubes!

I uploaded a quickly mashed-together first draft to Thingiverse, but didn’t have a chance to print it that day as planned because I lost my laptop in town while Christmas shopping. Thankfully, I found the laptop since. Just tonight, I got an email from a friend in Washington telling me DremelFuge had been featured in Makezine, which blows my mind completely. Well, not being content to let it remain unprinted for a moment longer, I set to making it.

It was my hope that I’d be able to put it to immediate use and have something great to add right away, but unfortunately it doesn’t work just yet. However, that’s simply a matter of solving two design mistakes: Firstly, spacing the cavities further apart to increase the strength of the printed object, and secondly providing some means of actually loading the microcentrifuge tubes! Unfortunately as made, the object made no allowance for actually putting the tubes onto the rotor, which of course makes it impractical to use. I aim to fix this as quickly as possible.

Without further ado, the good news: It survived a full-speed test on the best drill we own, which tells me it should survive the rigours of actual use as well!

DremelFuge Speed Test on Powerdrill

More on this as I develop it!

History of Linux and I

I tried twice previously to switch to Linux, and for a few reasons didn’t end up having any luck.

The first time I did so was around 2005, when I was living out of home for the first time. My room was beautifully minimalist; just a double-bed, a wardrobe, a Shuttle X desktop PC and a 5.1 surround sound system. The wooden floors and old timber beam made it warm and cosy. I decided to top it all off, I should remove all the evil from my digital experience, and ditch Windows.

So, I got a new copy of SUSE Linux and installed it. It worked fairly well, but I had to battle a few nVidia Drivers (Thanks, incompetent nVidia. Just release the source, please? Your money’s on the hardware, so let us support it!), Flatscreen support in the X window system, and ultimately found that once everything was working, there was waay too much command-line use in the day-to-day of using the system. Also, lots of programs wouldn’t run, and gave no reason or even acknowledgement that they’d been clicked on. I got fed up and morosely moved back to banal, ugly, locked-down Windows. For games only of course, I did all my other business on MacOSX, a far superior system in every respect.

Then again, perhaps in 2007, I tried again with Ubuntu. Again, I had to battle nVidia (no, they don’t really move at all as a company!). Again I had issues with a few day-to-day issues, and decided it wasn’t working for me. Thankfully I discovered Linux Mint, which was a big improvement. Linux Mint is Ubuntu plus the things they should include in Ubuntu. Drivers, for one. It had issues of its own, such as a confusion arising when you used the Ubuntu repositories to install things and it developed an identity crisis, but it worked better. Ultimately I just let it slide though, because I used my Mac for everything else and that was good enough for me.

Three Times Charmed: Ubuntu Karmic on Macbook

With those two previous times under my belt, I returned to Linux when I got fed up with two things: Firstly, that Apple had become just as bad as Microsoft for mistreating their own customers and telling them “love it or leave it”. I left it. Secondly, that my Macbook was beginning to show signs of age. I’d bought it in late 2006, and it had 1GB of RAM, a dual-core 2Ghz processor (still very nice!) and only 128GB HD space. Some things were running slowly, and there was a screen-flicker issue that Apple refused to acknowledge as a design flaw, going so far as to lock and delete discussion of the issue on their forums.

It needed an upgrade, and the cheapest way to speed up a PC is to install Linux, so I said I’d give it another go. Ubuntu had just released their new system, Karmic Koala, and there was documentation on getting every single relevant feature of the Macbook running on it, as well as info on Dual-booting MacOSX with it. I figured I’d get new hardware, and once upgraded I’d install Snow Leopard and Karmic Koala together.

I got a dual-boot working on the old Harddrive to test things out, and it worked beautifully. However, when I installed the new Harddrive, Snow Leopard wouldn’t boot from the DVD. Nice one, Apple; this is your fault for using a stupid bootloading system that wasn’t ever installed on non-Apple hardware. Ubuntu, however, booted perfectly. Installed perfectly. Configured perfectly. Following the instructions as written, I found that everything worked without a hitch, and I now find myself equipped with a perfectly functioning laptop with a beautiful desktop environment, speedier than ever with 2.4GB of RAM, and sporting a more satisfying 256GB of Hard Drive space.

I haven’t even bothered looking at installing Snow Leopard now; everything I want works on this. Better, in fact! Let’s take a look at a pleasing case history:

Google Sketchup and WINE – A How-To Example

One of the things I never got working right before, which I found particularly frustrating, was WINE. The “Compatibility Layer” is supposed to make Windows programs work with Linux natively, but my previous efforts to get a response out of WINE hadn’t borne fruit. On Karmic, I just installed it from the Ubuntu Software Centre (I decided to use the latest beta version) and it was right there in the Applications Menu; a virtual C-drive you could install windows apps into. Any file ending in .exe was diverted straight to WINE as well, opening with the normal window appearance of a Linux program with the interior appearance of the native Windows application. Perfect!

I decided my first trial software would be Google Sketchup, because it’s my favourite application for Makerbot modelling. It’s really the only thing I’d need OSX for at this point, and even then it runs terribly on OSX. So, I downloaded Sketchup 7 and did a bit of Googling on how to install it in Ubuntu/WINE.

I must say at this point; the Command-line nature of Linux isn’t gone. However, it is foolproof now, whereas before it’d give you a cryptic middle-finger answer and then sit there uselessly. So far on Ubuntu I haven’t had an issue; mostly because, if something isn’t installed correctly, Ubuntu tells you immediately how to install it. Example: if you were to type “dos2unix skp_to_dxf.rb” (because someone online told you that’d solve your problems!) but “dos2unix” wasn’t installed, Ubuntu would return “type ’sudo apt-get install dos2unix’ to install”. When you type exactly that and try again, it works. So the Command line, though still around, isn’t that bad, and it’s come to the point where my mum could use Linux without ever seeing a Command line, anyway. I just see them all the time because I’m doing crazy things, I think.

Moving on. I had Sketchup 7 downloaded and an account from 2008 of a guy installing Sketchup 6 on older versions of Ubuntu/WINE before me. Reading through it, he suggested that you can expect Sketchup to crash on the first time you open it, but if you do a little registry hacking afterwards it’d run normally.

I did my homework. You do registry hacking by running ‘wine regedit’ in the Terminal, which brings up exactly the same interface as the normal Windows regedit. You’re advised to back up the two registry files that are found in “/home/yourusername/.wine” first. Do so by typing “nautilus /home/yourusername/.wine/”, which will pop up that folder, and copying them.

My first attempts to open sketchup didn’t get as far as the crash he predicted; the splashscreen that asks you to pick a template crashes first, followed by the google bug detector crashing too. A quick look at the documentation of Sketchup 7 on WINE suggested that you can turn off the Splashscreen by unchecking the “Show on startup” box, but I couldn’t do that without causing a crash, and it didn’t save. Looking again, someone had said that it was crashing due to an internet request gone wrong, but by starting it while offline this wouldn’t happen. So I did, and it didn’t. Unchecking the box, I proceeded into the program itself, which crashed as it was expected to.

Then, I followed the instructions from the 2008 account exactly, and when I reopened Sketchup it worked fine.. better, in fact, than it ever had on MacOSX!

However, it didn’t end there. I looked up the 3rd party .stl output plugin for sketchup, which is an essential tool for anyone but especially for makerbotters who use .stl for exporting files to the robot. Downloading the file, I did what the 2008 account suggested; plugins require a quick conversion between their windows form and a unix form before they’ll work right, so you just do exactly as he suggests (go to terminal, then go to your plugin, type “dos2unix yourpluginname“, install the package as Ubuntu suggests, then try again when it’s done. It works right away.). Once that was done, I just put the file into the plugins folder like one would on Windows, and tested it in sketchup; it worked right away without hassle.

I since put the properly converted Linux skp to dxf/stl plugin on the Makerbot users group, so others don’t have to go to that trouble.

In summary

I love the new Ubuntu, and I finally think I’m on Linux to stay. I’d recommend it to anyone, because although it might sound daunting from my account to a casual user, you won’t actually need to do any of this stuff if you don’t want.

Office applications, Internet, Music and Video, the whole lot, is all freely available through the handy Ubuntu Software Centre and installs with the click of a mouse. I’ve never had a bug.

There’s also loads of free games, including clones of Civilisation and Simcity 2000. Very addictive.

It also has a nice feature Windows or Mac don’t offer; *free* cloud storage built in with desktop synchronisation. So you can drop your files in a special folder and expect them to be online if you find you need them later when you’re away from your personal PC. Very handy! And, shareable with other people or between your PCs.

If you are a power-user though, clearly the new Ubuntu (and the new WINE) is a heady, wonderful thing. You can easily set up and use things with the Command line without feeling like an idiot, and the system finally offers feedback when things don’t work. If it’s a program you haven’t got installed, it’ll tell you how to install it. Networking is beautifully simple at last, in fact it’s easier than it ever was on Windows and it’s on-par with MacOSX. Tethering with my HTC Hero required zero setup; when you activate it on the phone, it detects instantly on the computer and connects.

Finally, if you’re a Makerbotter, it’s a viable option with the inclusion of working Sketchup, ReglicatorG, and Skeinforge. The instructions for installing RepG are clear and easy on the replicator.org website, and Skeinforge runs as it always does on Mac. Just remember to manually drag the contents of the Skeinforge zip file to a visible directory, or it’ll be hidden by default for some reason.

My satisfaction with Ubuntu 9.10 thus imparted, I’ll retire and go to bed. Many thanks to the Ubuntu crew, to the developers of WINE, and to “lewiswadsworth” of SketchUcation.com!

Well, since my last post the Makerbot has been busy!

Busy breaking and requiring troubleshooting and maintenance. But that’s behind us now; I’ve a lot more knowledge of Makerbotting, and a much tougher Plastruder as a result. Many, many thanks to Mark Adams for the replacement part he printed and sent to me!

It Broke

So, I’ll keep it brief. The Makerbot is prone to a failure that occurs when plastic melts too high up in the barrel; this leads to increased friction, but also to some molten plastic being shoved up into the insulator by displacement as the filament forces its way down. This molten ABS hardens in the insulator, causing a tight blockage. The result is that printing requires steadily higher printing temperatures to melt the blockage upstream, which can lead to two potential failures:

1) Your insulator overheats, softens and the heater barrel is forced out by the hefty forces applied to the filament inside. Very bad news, as this usually de-threads the insulator and demands a replacement. Building a plastruder using an M6 nut to take the forces instead of the insulator seems to be a reliable preventative measure.

2) The diverted forces instead cause your insulator retainer (or worse, the lasercut parts above it) to shatter violently.

In my case, it was the retainer piece, a circular piece of acrylic (as given in the kit) that is apparently there to keep the insulator from wobbling away from the filament’s entry point and also to shatter instead of the more important parts above. Thankfully, a printable replacement is available on Thingiverse.

I fixed it and upgraded it

While I awaited the donated replacement part (did I mention I love this community? Fellow Makerbotters rock!), I ordered and installed a proper internal power supply, so the Makerbot is prettier and more mobile. I also was donated a little hose clamp from my Mechanic up the road, which some recommend as both a brace and a heatsink for the insulator. Finally, I rebuilt the plastruder so the circuit board was out-of-the-way, which makes it easier to see what I’m doing.

The end result:

And it prints!

I used a delightfully geeky item for my test print: A Clip-On Reading LED for glasses. It printed quite beautifully, but didn’t fit my glasses (which are usually pretty small-framed). So, I drilled two holes, threaded a tiewrap through and used it as a wedge to fit smaller frames. I plan to put a red LED in there so I can use it for nighttime navigation without losing my nightsight! I just need a battery now.

I spent my weekend doing a lot of geeky stuff. The two high achievers were: 1) Getting my iPhone to have always-on internet despite not having a Data Plan (kinda), and 2) Getting my Makerbot working at last, and making a few test prints.

Firstly, I’ll reveal my ingenious hack for the iPhone. Actually, it’s pretty simple, and possibly already done elsewhere. Anyway.

Personal Area Network Hack

If you have a mobile broadband modem for your computer, and you’re using a Mac, follow these instructions and you’ll have yourself a personal Wifi network for use with your smartphone, book reader or other-laptop in less than a minute, with only one free install.

1) Download InsomniaX for your mac. Because Apple can be idiotic sometimes quite often about permitting Users to do what they like with the hardware they fairly purchase, there’s no inbuilt feature to prevent the laptop from sleeping when you close the lid. This app allows you to toggle on/off this feature. Necessary to go wandering around with your hacked wifi PAN.

2) Fire up your Mobile Broadband, ideally using a USB extension cable so the Modem isn’t poking out the side of your computer so much.

3) Go to your Wifi bar at the top of the screen, click it, and choose “Create Network…”. Using default settings (Channel 11/Automatic) and a name of your choice should do. No spaces here, please.

4) Go to your System Preferences from the Apple Menu in the top right. Choose “Sharing”. From the list of options on the left, choose “Internet Sharing”. In the dropdown menu for “Share your connection from:” choose your Broadband connection/modem/device, then select Airport in the menu below. And optionally Ethernet or whatever if you need other options. Then simply check the box next to “Internet Sharing” on the left, and click “Start” when it asks you if you’re sure.

5) Connect your mobile device to your new network, and let it self-assign IP and settings. It should be able to connect once it’s adjusted. Now, you should be able to close the lid on your laptop without interrupting this networking, and put the laptop + modem-on-cable into your bag like so:

My Macbook with Huawei modem on a USB cable; the blue light indicates 3G/HSPA.

Proceed to enjoy your new personal Wifi network! It’ll last as long as your battery does, which for my macbook is pretty formiddable. I imagine if you do something similar with a decent netbook you’d get a good 5hrs out of this with the screen/HDD disabled.

Something to watch out for: make sure your computer’s heat vents are pointing up/out of the bag, or it might overheat. With modern processors, this usually means it just shuts down without warning or saving anything, but it’s best not to let it get to this stage in case it leads to damage/deterioration.

Don’t come crying to me if your bag catches fire, either.

Makerbot #147 Online

In other news, on the same day I got my makerbot working after ages building and troubleshooting it! I had some great help from the good folk of the makerbot google group.

First thing was to troubleshoot the settings for the Plastruder. Here, it's making a raft successfully, though it later failed on the infill for the model because the heat setting was too low.

First build was a standard 608 Pulley, the same model used to make the pulleys that ship with the machine itself. They’re a good test, a simple model, and convenient to have as spares in case yours somehow break. If the shipping of 30+ pulleys works out favourably for you, the Makerbot crew are even offering $1 per pulley printed for them, because they’re having trouble producing enough for their kits! Very cool idea.

Left: Freshly printed pulley. Right, the one that shipped with the 'bot.

As expected, once the machine was online I had a few little teething issues that largely were due to software issues.

My initial problem (which was a long hangup) was with the ceramic insulator you’re supposed to put around the barrel of the extruder. Turns out it can allow the plastic to melt way too much, such that it blocks the plastruder. Your extruder cog-drive then chews a notch in the filament and can’t push it further, so the whole thing effectively stops. Removing the ceramic fixed the problem.

The next problem was with the temperatures being too low when the machine tried to fill in the layers of a model, so a similar build failure would occur halfway into the first layer. Some settings fixed in skeinforge solved this; all your temperatures should be at least the melting temperature of the plastic.

Finally, although the ceramic wrap caused much grief, its absence means the machine is highly unreliable near draughts or gusts of cold air. The barrel can cool below the melting temperature pretty quickly, and you get a build fail.

With those problems solved, I’ve so far printed two pulleys, a mis-sized beltclip of my design, and half of a belt-clip insert for hanging pippettes from one’s belt through said belt-clip. I am proud of my geeky weekender. Much was achieved!

My only hangup; someone with knowledge of Python would be much loved if they would rewrite Skeinforge to have a more self-explanatory interface. I mean, the barebones box-and-button interface is fine; what the hell do all those settings mean? Many thanks nonetheless to the Reprap crew for making the tool. It’s evil, but very useful.

It’s getting extra close to the new Lisbon Treaty vote, and although I’m encouraged to see fewer people falling for the crap, I’m encountering whole new “no” phenomena that sadden me.

Chief among these, and one which few outside of Ireland understand (for the most fortunate of reasons), is the Belligerent No. In Ireland, we’re particularly fond of biting off our noses to spite our faces; just because we feel like we’re being asked to go one way, we’ll instead go another without thinking about which way is best for us.

This stems from a psyche we pick up in school learning all the one-sided propaganda about Ireland’s fight for independence. To the Irish, if anyone *not* Irish asks for something, it’s because they’re after our Lucky Charms and our Independence.

As far as this perspective goes, it should be clear to anyone looking at it objectively that it’s daft. The Europeans want us to vote Yes because it’s good for them…but what’s good for them is good for us, because we’re Europeans, too. When we’re asked to do something for Europe, we should remember we’re doing it for ourselves as well.

The other notion floating around, and a sillier one at that, is the “No because we said No” vote. This is based on the argument that, because we already voted no, nobody should ever again ask us our opinion on the matter. Words like “Bullying” and “Forcing” are thrown around, as if we’re somehow obliged to vote one way or another to stop the madness.

Er..not quite. We’re being asked again because it is the opinion of virtually everyone that the first vote wasn’t fair and balanced; nobody adequately understood the issues at hand, and some incredibly rich campaigners on the No side spent some foreign money on expensive posters everywhere that convinced everyone it was a Lucky Charms scenario.

Now that we understand it a little better, and now that we’ve had some clarifications on what it means to Ireland, we’re being given another chance to vote so that our actual opinions on the matter can be had. If we vote no, it’s over, a done thing.

Key to this whole argument about “Bullying” is the failure to realise that ultimately, we’re the voters and we have all the power here. So what if they’re asking us to vote again? It’s a democracy, we can vote all day if we like! It’s not even costing you anything, because everyone is entitled to an hour off for free from work just to go voting. If you’re quick about it, you get a free lunch-break with your leftover minutes. Hell, I’d love it if they kept asking. I could keep offering my opinion on the vote, and get muchos coffee après-democracy.

So, Lisbon 2. Same as the one before, which means it’s exactly as innocuous. No Militarisation, Taxation, Abortion, Euthanasia, Enslavement, Loss of Liberty, Forced Incarceration, Destruction of Rights, or any other stupidly sensationalist topics Cóir, Sinn Féin or Libertas would want you to associate it with.

Just some housekeeping that’s a decade overdue. Read up on the independently verified stuff on the Referendum Commission’s website (often mistaken for a “Yes” site because it tells the truth), or look it all up on Wikipedia and check the references, making sure they’re not political propaganda sites or personal webpages. It’s all there right in front of you; make your own decision.

So, my last post (ages ago) was all about my dreams of designing an open lab toolkit for anyone to use, proffessionally or amateur. I haven’t been idle in the meantime! That said, I also haven’t had enough time to burn on this project, so it’s only just getting started. I’ve decided to start calling it OpenThermo, to get across the notion that it’s going to be Open Source and it’s, er, thermo.

To cut the long story short, I spent some time fiddling with thermistors to get a temperature sensing system working, only to discover that A) Thermistors are $#1+, and B) There’s an IC called an LM35 that accurately senses temperature and gives a clean, linear output based on the celcius value. That sped the development up a lot! I only had to get me one of these fine LM35s.. thanks to ebay, it arrived last friday.

The other area that I wasn’t looking forward to sorting out was the Peltier circuit; since the peltier heater/cooler units operate at a fairly high wattage, the Arduino wouldn’t be able to drive them on its own at all. Therefore, I had to sort out a way of getting high power through the peltier, under the control of the arduino, without burning out the latter by accident. Probably elementary to an electronics guru, but I’m not one of those. I discovered that, again, there are some wonderful ICs to do the job for me; in this case, the L293D motor controller chip apparently handles an isolated power source and high-power load while accepting low-voltage logic input. In other words, perfect.

I’ve therefore just started work on making the actual thermal cycler, finally. Shown below is my first setup for testing the L293D:

First setup of my homebrew thermocycler. Shown is a 12v battery bank, an L293D motor driver chip, my arduino, and a hacked-together peltier thermocycler rig using thermal tape, a heatsink and a little enzyme tin from the lab.

Quite the jury-rig. However, it seems to work. I would be testing it more now, rather than writing about it, except that I noticed a funny smell about 20s into the testing (a simple on/off/reverse/on/off script I’ll share later) and started looking for the source. I actually burned my finger slightly off the L293D chip! So.. Doesn’t like loads THAT high, it seems. Thankfully I come pre-armed with a solution: According to LadyAda on her Arduino Motor Shield instructions, you can just stack the chips to add twice the current capacity. I can’t do this just now, in part because I have more pressing business in the lab. That, and I’ll probably add a crude heatsink while I’m at it; definitely a DIY job for at-home time and not the lab.

Suffice to say I think it was working before I stopped it, but at the wattage I had it going it may have been about to burn out the chip.

Next task is to test out the LM35 chip, which should be elementary. I think I shall love the LM35, and want to make all sorts of funny things with it. At least the odds of my setting one on fire are far less than with the motor chip!

For a while preceding buying my own Arduino, I spent my time looking at all the cool projects people have used them for on the Make blog, Fashioning Technology and the Arduino wiki itself. Although I love all the projects that showcase the artistic use of electronics, and I’m impressed with the more utilitarian uses also, I see great untapped potential in the Arduino as a replacement technology for certain niches where equipment is prohibitively expensive.

One such niche I am intimately aware of is Biotechnology, particularly molecular biotechnology. For example, I work daily with a machine called a thermal cycler. These machines perform a seemingly simple duty; they accurately vary the temperature of a metal casing in which you place little plastic tubes containing your reaction mixes for molecular biology. You set them up to cycle through a variety of temperatures so as to help enzymes work at optimal efficiency, break the double-helix of DNA with heat, permit only certain molecules to adhere to one another during a reaction, or to inactivate enzymes through heat-treating. They are most often used for the Polymerase Chain Reaction, a cornerstone of modern genetics, and so they are often called PCR machines.

The odd thing about these machines is the cost. A thermal cycler can cost over €2,000 for a used robotic version, or as low as €200 for an ancient, damaged one. These machines almost universally have a poor interface, too; few can connect easily to a computer, and none are open-source.

What’s going on here? Am I missing some deep secret of thermal cyclers that should make them so expensive and limited? So far as I know, a modern thermal cycler technically only uses one basic technology; a Peltier thermoelectric device. These little semiconductor/ceramic plates get hot on one side and cool on the other when supplied with electricity, with the orientation depending on the polarity of the voltage. In short, the work like an electric heat pump, heating or cooling one side on demand. And, they cost less than €10 on ebay with free postage.

In principal all that’s needed to make one, then, is a Peltier unit, a temperature sensor (60c in Maplin), a little breadboarding ( < €20) and an Arduino (€20). Connect it to a PC, optimise your code, and you have a highly flexible Thermal cycler for less than €50.

This ties into something of a broader interest for me, that of someday having my own lab wherein to do my research. I think it’s just part of my personality that I feel compelled to do stuff for myself, and tend to lose interest in projects not my own. I look forward to working on nutty genetics projects that I fancy might, just might, have the potential to help improve the world..or at least look nice growing in a pot by the window.

Setting up a Genetics lab is the big obstacle here; because you need such specialised equipment and reagents, it’s not a simple matter of building a shed and ordering from a catalog. (Update: I started a shop on Shapeways where I am already selling some cheap lab equipment: LabsFromFabs) Although a growing community of people are taking an interest in homebrew biology, their ability to do so will be constrained by the price of setting up and maintaining a working lab.

Thankfully, this is slowly changing. Much as electronic technology used to be a very exclusive field until prices plummeted and homebrew electronics became an affordable hobby, certain powers that be are working to bring Biology labs into the home for enthusiasts and clubs. Update: See the LavaAmp for one such project that is already near the market!

One such effort is OpenWetWare, a wiki project that aims to build a directory of Protocols, Tutorials and Essays to teach the knowledge and practices of Biology online, for free. Most homebrew clubs I’ve encountered maintain a page on OpenWetWare to showcase their work or share their experiences.

Another, potentially even more revolutionary example is the Biobricks Foundation, which aims to standardise genetic material to enable swift, easy construction of complex genetic tools and systems. Although they suffer from a terrible website design and a general lack of beginner-friendly information or support, the system they are using is seemingly reliable and robust. It should require no more specialised equipment than four enzymes and a special breed of e.coli to start making things that glow, smell like bananas or release special enzymes in response to something. Technically.

Best of all, existing genetic “parts” (read; the bits of whole genes such as the coding sequence, promoter, terminator, etc) can be modified so that they are themselves new “Biobrick” parts, and can be added to the Biobricks Registry of Standard Parts. If someone wants your part, they can ask you to mail some over to them!

Finally, and perhaps most importantly, those interested in genetics at home will be excited to hear that I firmly believe a service will be available within this decade where you can simply order a gene from scratch and have it delivered already transformed into e.coli (at least!). Mail order GMOs.

I say this because you can go right now over to Mr. Gene and order up to 3kb worth of DNA to be synthesised from scratch at a competitive price (€0.32 per nucleotide), which is enough to contain a whole functioning bacterial gene. Not only that, but their web-tools will help you easily add in all the special sequences you want, while excluding those you don’t, while maintaining the coding sequence and optimising it by species. That’s pretty much all the hard work.

The price is still expensive; although €960 should cover 3kb at 35c/n, it can cost a bit more for a useful gene to be made due to the relative difficulties in synthesising odd or difficult sequences. However, I have learned (through my own travails working in a genetics lab) that producing and stably assembling a gene by yourself is sometimes so time consuming (months!) and expensive that you’d be better off just paying the extra money and getting a guaranteed sequence in 15 days.

It is a small step from being able to order a complete gene and the e.coli separately, and transforming them on your benchtop, to having a company do all of this for you in advance and delivering the finished, genetically modified product by courier.

It would also be a small thing to order a modified strain of another bacterium, known as Agrobacterium tumefaciens. The significant difference here is, Agrobacterium is known for its ability to transfer DNA into plants, and is routinely used as the most simplistic means of creating GMO plants. So, mail order transformed Agrobacterium = Mail order modified plants.

I eagerly look forward to this day. For the moment though, and coming back to the beginning of this post, I aim to create an open, free and publicly available design for at least one piece of specialised equipment. With this and other similar projects, I’m hoping lab equipment stops being so damned exclusive and comes quickly to the Maplin Catalog. Update: See LabsFromFabs, my Shapeways shop where I’m already selling some lab equipment!

And if someone else reading this has the tech savvy to help (I’m an electronics n00b), please do!

Edit: The price of gene synthesis wasn’t €0.35, it was $0.35. Much cheaper than I quoted.