" rapid prototypes of complex products can now be assembled in one’s own garage—a game-changer reminiscent of the early days of personal computing."
I love 7400s and ICs just as much as everyone but I have a feeling some Primary Investigator was feeling some political weight to publish lest he perish (is it NSF/NIH grant season?).
Take the uL precision of the pipette'ing equipment they demonstrated. 2-10uL. They didn't mention the size of the micropipette they built so I can't compare the actual tolerance percentages, but to put it into perspective, those disposable pipettes you used in middle-school with a capacity of 5uL has an accuracy 1st sigma of 0.015uL.[1] If their actual testing numbers are anything like good manufacturers, you'll see just a standard Gaussian distribution where your accuracy and tolerance posts numbers so conservatively with their nominal precision just to keep their reputation stellar(2). So 2-5uL of error even if that has a full L capacity (hint: no, no it does not) and it's still going to be about as good as what you were working with in grammar school.
He mentions 3d printing out labware on ABS and I'll leave looking the MSDS on ABS left as an exercise to the reader as to why that's an absolutely horrible polymer to choose, and why borosilicate is still used in your 12 year olds science class and in clean rooms at Sigma-Aldrich. You don't want to perform any of the standard o-chem 101 caffeine acid-base solvent extractions in labware made of ABS (or almost any other plastic except for PTFE or PFA and even then.. eh).
The thing that scares me the most is this: "Luckily, freely available online resources facilitate self-learning. In addition, many makers hone their skills directly at the workbench, simply by attempting to recreate or modify existing designs."
Anyone who's tried to inherit their share of codebases has experienced that "this is complete shit, I can do it better" followed by "oh, I see, this comment from 2003 says that the new VP wanted some dumb functionality that required adding that trigger that seems to do nothing". You uncomment a line, the whole base crumbles. A common electronics project is a variable power-supply using jellybean LM317s and what not. You make a attempt to make a 'simple' modification in a power supply design (whoops my homemade transformer's 10:1 step-down was inverted!) and bad things could happen.
"even build sophisticated lab equipment from scratch for a mere fraction of what commercial alternatives cost." No. The real game changer is Ebay. Biohit's mLINE's lowest series, readily available for around $100 (at worst $150) on ebay is $150. Take a look at [3] and turn to page 13 with me to see the calibration matrix. Then flip to around 48 where the precision and accuracy numbers are posted and youre well hovering around 2-3% until you drop to the ten-millions of a liter when you see a jump. Hey, all multimeters can't have the John Fluke Co's precision ;). $100 bucks, can be thrown in the autoclave, offers numbers you can publish to Nature with. GS/MS' that used to cost a quarter of a million dollars from Agilent and HP can be picked up for 30k. Your grant money is getting pushed so much farther because of the secondary global market (Professor Smith just got his first 5 year 3 million grant, does he get a brand new unit for 250k or a 5 year old piece of kit for 30k and another post-doc for the duration?) Agilent just lost a sale and comes out with a better featured model this summer instead of next, driving costs even further down. The maker community can do great things, but open source and 3d printers aren't going to change the world - they are too imprecise for one-offs which push the industry forward and too expensive for commercial production (though industrial design engineers making ergonomic things have an awesome feedback loop going from CAD to tangible in a few hours must have revolutionized their industry). People have been sharing their own circuits amongst themselves in electronics magazines for 40 years; scientists have been collaborating since well before Leibniz via mail, 'open-sourcing' their findings amongst peers[pun not intended] in the Royal Society since the 17th century (and via patents since probably sometime around there give or take 50 years). Science advances because people like Euler, Riemann, Feynman, Tao, Perelman and thousands of other people who make micro-advancements that end up being one small grain of sand that makes the beach of their industries just need to know.
I suspect the author's motivation was to just make other scientists aware that this is even possible. By saying that you can make crappy pipettes, it's really saying that you can make much less sophisticated stuff really well.
My lab's 3D printer is constantly cranking out something to help with experiments. Any single thing seems kinda trivial - a jig that makes it so some tube doesn't have to wrap around another piece of equipment in some awkward way, a holder to keep a heater a few millimeters closer to a reaction, a grip to hold up a piece of cardboard - but in aggregate these things are making our experiments more reliable and easier to perform.
We also prototype things that will ultimately be made in the machine shop. It takes about a month to get any feedback due to their backlog, so being able to print the design in plastic and just make sure it fits and is practical to use can save months.
I love 7400s and ICs just as much as everyone but I have a feeling some Primary Investigator was feeling some political weight to publish lest he perish (is it NSF/NIH grant season?). Take the uL precision of the pipette'ing equipment they demonstrated. 2-10uL. They didn't mention the size of the micropipette they built so I can't compare the actual tolerance percentages, but to put it into perspective, those disposable pipettes you used in middle-school with a capacity of 5uL has an accuracy 1st sigma of 0.015uL.[1] If their actual testing numbers are anything like good manufacturers, you'll see just a standard Gaussian distribution where your accuracy and tolerance posts numbers so conservatively with their nominal precision just to keep their reputation stellar(2). So 2-5uL of error even if that has a full L capacity (hint: no, no it does not) and it's still going to be about as good as what you were working with in grammar school.
He mentions 3d printing out labware on ABS and I'll leave looking the MSDS on ABS left as an exercise to the reader as to why that's an absolutely horrible polymer to choose, and why borosilicate is still used in your 12 year olds science class and in clean rooms at Sigma-Aldrich. You don't want to perform any of the standard o-chem 101 caffeine acid-base solvent extractions in labware made of ABS (or almost any other plastic except for PTFE or PFA and even then.. eh).
The thing that scares me the most is this: "Luckily, freely available online resources facilitate self-learning. In addition, many makers hone their skills directly at the workbench, simply by attempting to recreate or modify existing designs." Anyone who's tried to inherit their share of codebases has experienced that "this is complete shit, I can do it better" followed by "oh, I see, this comment from 2003 says that the new VP wanted some dumb functionality that required adding that trigger that seems to do nothing". You uncomment a line, the whole base crumbles. A common electronics project is a variable power-supply using jellybean LM317s and what not. You make a attempt to make a 'simple' modification in a power supply design (whoops my homemade transformer's 10:1 step-down was inverted!) and bad things could happen.
"even build sophisticated lab equipment from scratch for a mere fraction of what commercial alternatives cost." No. The real game changer is Ebay. Biohit's mLINE's lowest series, readily available for around $100 (at worst $150) on ebay is $150. Take a look at [3] and turn to page 13 with me to see the calibration matrix. Then flip to around 48 where the precision and accuracy numbers are posted and youre well hovering around 2-3% until you drop to the ten-millions of a liter when you see a jump. Hey, all multimeters can't have the John Fluke Co's precision ;). $100 bucks, can be thrown in the autoclave, offers numbers you can publish to Nature with. GS/MS' that used to cost a quarter of a million dollars from Agilent and HP can be picked up for 30k. Your grant money is getting pushed so much farther because of the secondary global market (Professor Smith just got his first 5 year 3 million grant, does he get a brand new unit for 250k or a 5 year old piece of kit for 30k and another post-doc for the duration?) Agilent just lost a sale and comes out with a better featured model this summer instead of next, driving costs even further down. The maker community can do great things, but open source and 3d printers aren't going to change the world - they are too imprecise for one-offs which push the industry forward and too expensive for commercial production (though industrial design engineers making ergonomic things have an awesome feedback loop going from CAD to tangible in a few hours must have revolutionized their industry). People have been sharing their own circuits amongst themselves in electronics magazines for 40 years; scientists have been collaborating since well before Leibniz via mail, 'open-sourcing' their findings amongst peers[pun not intended] in the Royal Society since the 17th century (and via patents since probably sometime around there give or take 50 years). Science advances because people like Euler, Riemann, Feynman, Tao, Perelman and thousands of other people who make micro-advancements that end up being one small grain of sand that makes the beach of their industries just need to know.
3d printers sure are fun though.
[1] http://www.hecht-assistent.de/en/product-field/blood-test-ur... [2] I'm pretty sure that Nichicon underrates their electrolytic caps, I know Vishay does that with their precision resistors, and Panasonic definitely does with their 18650 green LiPos. [3] Datasheet, page http://www.biohit.com/resource/files/media/manuals/lh/mline/...