DIY thermal cycler

This post is way long overdue. In the Fall of 2008, (or was it 2007?) I taught a small module in the ABE 221 course dealing with automation in Biological Engineering. My hope was to create a home-made thermal cycler for less than $1000. Of course many people in the DIYBio community have constructed thermal cyclers for way less, and truly, you probably don't need to pay more than $300-400 (perhaps even less) to build a reasonably good quality one. 

Anyway - here's what we built. We had the following goals:

1. Use a Peltier module to provide heating and cooling of the sample.
2. Have a large dynamic temperature range (4C - 99C).
3. Use a USB connection to program and control the thermal cycler.
4. Use a language like Python to program the instrument
5. Use off-the-shelf parts as much as possible
6. Make it look pretty, since this was going to be entered into the Engineering Open House competition.

First - let's take a look at the outside. 

Notice the obligatory warning labels. Gotta have those. Ok - so we decided to build the enclosure in clear polycarbonate. A bit pricey, but the good folks at Illini Plastics cut us a good deal. You can see the LabJack U3, a USB data acquisition and control module at the heart of the IlliniCycler. The ugly circuit board at the top is the H-bridge driver described below. 

Here's another view with the lid open. Notice we had a circuit board screwed to the inside of the lid. Once you get past the wires, you will notice the metal box at the bottom. This is the 24V 13A switch mode power supply made by Meanwell (Model SP-320-24)

The thermal block is visible to the left. You can also see two small computer fans glued to the walls for extracting the copious amounts of heat generated by the circuit. The power supply also has a fan for cooling. 

Here's a closeup of the thermal block with its lid open. The lid was designed using CAD and machined by one of our students, who did a fantastic job on it. The top part of the lid is machined out of a single block, with a cavity cut in to provide room to accommodate the PCR tubes. A small angle is screwed to the lid and separated using insluating washers to provide for a relatively cool handle to work the lid. 

The bottom block of the lid is also machined from a solid block. It has four holes drilled to receive the PCR tubes and a couple trenches on either side to house the temperature sensors (we used LM35s, which provide a 0.1V/C ramp and operate between 0 and 150C.)

The thermal block is mounted on a computer chip heatsink that has a built-in fan. There is a Peltier module between the thermal block and the heatsink. Thermal compound on both sides, and JB Weld to hold everything in place. The entire assembly is screwed to the bottom of the box and raised using standoffs so that the fan has plenty of room to circulate air. 

If you are heating the block, the fan is kept off. If you want to extract heat from the block, the fan and the heatsink help dissipate the heat from the (lower) hot side of the Peltier element. The maximum temperature difference between the hot and cold sides is specified to be 76C. If you are operating at room temperature (25C) you should be able to reach 97C or so required for dehybridizing the DNA. 

To drive the Peltier element, we need an H-bridge configuration that allows the current direction to be reversed in the Peltier module for heating and cooling. You can either use an electromagnetic system (relays) to build an H-bridge, or you can use a solid state system composed of IGBTs or MOSFETs to carry the large (~12 A) currents involved.

It is rather difficult to find H-bridges that can drive more than 2A. We had to build our own. I've built several H-bridges since, and this one was a rather primitive design. We used the 'smart' MOSFETs produced by ST Microelectronics. These mosfets can drive dozens of amps and can be driven by a TTL (5V) signal. The high side device was VNP920 (Obsolete now I believe) and the low side device was VN20N07 (Also obsolete). Both these devices have been replaced with higher current carrying products in the same series. 

High current design is not trivial and should not be attempted if by someone with no experience in designing high current circuits. The PCB tracks have to be really robust to handle the current and heat, the ICs have to be attached to properly sized heat sinks and there is always the danger of burning your finger. Make sure the parts of the circuit that get really hot do not come in contact with the rest of  your equipment or your person.

More details on the construction, the team that built it, the parts cost etc. are provided in the complete report here. The students who built it won a couple different prizes at internal competitions. They have graduated and are probably making more money than I am. 

But but but... did it work?

That depends on how you define 'work'. The system basically *can* work, if we wrote the proper control system. The way we left it, the Peltier module was heating up to about 89C, without getting up to about 97C. This was because we never really tuned our PID controller. Also, we never figured out a clever way to determine how to accurately measure holding times in the elongation cycle. 

In short, the project is still incomplete - we've never demonstrated an actual thermal cycle. The hardware works, but the software doesn't. If you are interested in working with me to produce a good control algorithm maybe we could get it to work like a thermal cycler... perhaps do more aggressive tests on it. My sense is that the circuit board probably needs revising. I hope someone will step up to the challenge. 

That said, the possibilities are endless for this device. In theory, you should be able to program things like touchdown PCR or other exotic variants. You should be able to use this for things like isothermal amplification, or Gibson assembly as well. What's more, one could use this module as part of an automation suite, or put the thermal cycler on the web and control it using a browser or an iPhone... Or better yet - put it on the Cloud. Everything's on the Cloud these days!

Comments

Post a comment

Comments will be moderated prior to becoming publicly visible.