Long Overdue Update: This post has turned out to be one of the all-time most popular on my site, which surprises me to no end. Who'd have thought my crummy heatsink-and-tin thermal cycler would be cooler than isolating glowing bacteria or printing a 52,000g centrifuge? But, who am I to question human interest. It's not like my interests are particularly normal anyway. However, I do think this post needs updating, since people keep returning to it and asking questions.
The reason I haven't pursued this project is largely because it's being done far better by others more equipped to create a fully open-source thermal cycler. Look to http://openpcr.org to see this exciting project as it develops. As I write this, there are still 6 days left in the OpenPCR Kickstarter fundraising project, which has been wildly successful already. The more money they get through this fundraising, the more likely it is that they can include “bonus features” in the final model. If you donate $512 to the project, you can even choose to have them send you one of the first machines once they are ready! I have done so, and without doubt I'll be reviewing the machine and working with it as soon as I have it, and sharing the data on this blog.
If OpenPCR works out, it'll be the cheapest programmable PCR machine in the world. However, it'll be far more than just the cheapest; with an open-source PC-based frontend, it'll be the most customisable and future-proof PCR machine I'll have yet encountered. I thoroughly look forward to making it part of my daily lab routine if so.
Back to the original post:
So, my last post (ages ago) was all about my dreams of designing an open lab toolkit for anyone to use, professionally 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!