"The thermal mass of the heat block also plays a key role in controllability."
This is because the old school control system is only second order. When operating at faster speeds which seem unstable, the control system has to be upgraded as well. Optimal filtering techniques in digital signal processing (for example can be done on a $1.00 dsPIC chip) with 32 taps would take care of that- a real control system and with rapid response, and internally calibrated for unique manufacturing differences. Much better than simple PID. Temp sensors would have to be embedded to get accurate readings.
Entire system cost to the end user has to be considered. A lighter mass will be cheaper to ship for example.
This is because the old school control system is only second order. When operating at faster speeds which seem unstable, the control system has to be upgraded as well. Optimal filtering techniques in digital signal processing (for example can be done on a $1.00 dsPIC chip) with 32 taps would take care of that- a real control system and with rapid response, and internally calibrated for unique manufacturing differences. Much better than simple PID. Temp sensors would have to be embedded to get accurate readings.
Entire system cost to the end user has to be considered. A lighter mass will be cheaper to ship for example.
## Jonathan Cline ## jcline@ieee.org ## Mobile: +1-805-617-0223 ########################On 2/3/16 11:11 PM, Josh Perfetto wrote:
Keep in mind that the primary purpose of a heat block is to promote thermal consistency between reactions. It may seem like a basic/old design, but it's a passive design that does an excellent job, like the wheel. The thermal mass of the heat block also plays a key role in controllability. Lack of these features was a detriment to several of the designs discussed previously in this thread.
As you've noted, increased heat block thermal mass does decrease ramp rates, and heat block design is paramount for both consistency and increasing ramp rates. Newer technology like additive manufacturing does enable new designs, but we haven't seen the costs become competitive with CNC fabrication yet, the costs of which have steadily declined in recent years.
People concerned about reducing PCR costs to the lowest possible levels should be taking a serious look at total costs including reagents. Thermodynamically the faster you want to go, the more enzyme and primer you need. I'm not saying we shouldn't work to do PCR faster, and one of my personal goals is to produce a device that can do 30-40 cycles in 1 minute. But if your goal is absolute cost minification, it's silly to be looking at ramp rates greater than 1-2 C/s.
-Josh
On Wed, Jan 27, 2016 at 6:27 PM, Jonathan Cline <jcline@ieee.org> wrote:
Curiously, none of the supposedly open designers of previous PCR thermocyclers contribute their lessons learned from their previous projects. Well it's not so curious perhaps, just greedy human nature, which even GPL attempts to exploit for it's own purposes. Not so open, those self-proclaimed open people, are they? At least not after even some meager dollar signs are seen.
The design you're thinking of was I believe the lowest tech copy of the Lava-amp or perhaps a hybrid. It wasn't the first such design of it's kind, but a well publicized one. Lava-amp relied on thermodynamic mixing of liquids due to temperature gradients only, no pumping (thus much slower protocols too, I suppose), with three temperature zones (analogy: hot, medium, cold).
One funny thing about mol bio experiments is that they are fragile in so many ways. Yields are rarely (read: never) measured by biologists so it is difficult to characterize. However some papers have noted experimental result differences depending on plastics used. Surface area may play a role in that difference (chemistry in the liquid being biased by the plastics). The last thing anyone wants during an expensive and time consuming protocol run is a drop in what I'm calling yield but could also be called quantity of viable biological product (protocols have too many already close-to-failure-mode variables, many of them unknown-unknown's). Doing several protocols just to build something in order to build something later, which then later fails, is a massive head-scratcher and frustration builder. This also in part leads towards the "let's stick with what humans have always done, to reduce changes, even though it's now an automated machine process, so let's use reaction tubes."
## Jonathan Cline ## jcline@ieee.org ## Mobile: +1-805-617-0223 ########################
0 comments:
Post a Comment