Smart Cycler qPCR Machine Construction

Recently I’ve been talking to a few people about at-home molecular testing for SARS-CoV-2 that is to say companies lie Lucia, Detect, and Cue. The general impression I have of these tests is that there was a small but healthy market during the pandemic, which has dried up now the pandemic is maybe? kind of? over? In particular Detect have withdrawn their test citing lack of demand and discussion I’ve had confirm this general impression.

These discussions led me to qPCR machines, and the observation that while the cheapest qPCR machines are around $5000. The BOM cost is probably much less than <<$100 per “plex” (target), and the COGS could be similar to Lucira’s disposable device if sufficiently optimized.

So, today we’re going to pull apart an old SmartCycler qPCR machine (similar to the one I’ve been playing with in my “lab”).

In the image above, you can see that the SmartCycler is composed of 16 identical independent modules. Each of these contains a complete 4 channel qPCR instrument, with the chassis housing the backplane and PC interface.

The SmartCycler uses Cepheids custom qPCR tubes, these tubes are sandwiched between two PCB with resistive heating elements and a temperature sensor. There’s no heat block here. Cepheid use very thin walled tubes to enable fast thermal cycling, and a heat block would likely cause issues with this. The platform works with hot-start chemistries, so they run above room temperature and use a cooling fan to reduce the temperature when cycling.

It’s possible that the temperature sensors are fancier RTD sensors, in which case they cost ~$1, but the whole heating assembly most likely costs <$3.

Looking at the rest of the qPCR module (called the iCORE) you can see the cooling fan to the left and the optical sub-modules which couple against the tube:

The PCBs make it clear that VSE manufactured these boards, and I suspect were involved in the design process too:

The detection module (shown above) uses TSL257 “light-to-voltage” converters. These are cheap integrated photo-diode/amplifier sensors which cost about $2 in low volume. The optical components most like form most of the BOM cost. The lenses are all plastic, which should cost almost nothing but each detection wavelength has a dichroic mirror, and three filters. This seems like over-kill and I’m not exactly clear why they use this approach.

Filters are ~8x4mm. These are probably the hardest components to get hold of cheaply in low volume. Edmund Optics will charge you something like a $160 for suitable filters in the smallest size they have. My limited experience buying optical filters from Shenzhen suggest you should be able to get filters like these for a few dollars. But even if you cut up Edmund filters, you’d still only be looking at ~$10.

The use of multiple filters perhaps suggests that they use a combination of lower spec (non-OD4) filters to reduce costs.

On the emission side, we see a similar optical system mated with an array of LEDs:

Again, these likely cost at most a few dollars. And there’s nothing here that suggests to me a BOM of more than $100, though manufacturing costs may exceed this given that they sold this as a medical device in relatively small quantities. But overall, the fundamental BOM cost of a qPCR machine is very low…