This is an incubator, guys.
What organism do you know that maintains its temperature to 0.1C?
Why would you think that was important for anything you are going to incubate?
Let's suppose that you need to incubate something for 10 hours (600 minutes).
Suppose you set one incubator for 310 kelvin, and another for 310.1.
Thus one is to get 186,000 degree minutes, and the other gets 186,060 degree
minutes.
You open them both, and immediately the air temperature changes by 11 degrees
to 26, the temperature of the Petri dishes full of agar.
It takes both incubators 20 minutes to settle down to their set temperature.
You have an average of 5 degrees below the set point during those 20 minutes.
That is 100 degree minutes off already, which is more than the 60 degree minutes
by which the two incubators have been set.
But more to the point, suppose we are incubating E. coli.
Let's look at some of the protocols in use. One says to incubate at 37 Celsius for
18 to 24 hours. Another of them says to expect a 2-fold loss of transformation for
every 15 minutes that a 1 hour incubation at 37 Celsius. A third calls for incubating
at 37 Celsius for 30 minutes, then incubating 'at room temperature' for 10 minutes.
It does not look to me like anyone is calling for one tenth of a degree of accuracy in
temperature. But more to the point, given that how long the cultures are incubating
is so variable, being off by a couple degrees does not seem like it is going to make
any difference.
If it does make a difference, is temperature the right thing to be measuring?
If the reason we incubate at 37 for an hour is to get a certain density of critters
per milliliter, why not incubate until the optical density is some particular value?
Maybe we shouldn't bother with electronics at all.
Run some hot water from the tap and adjust the temperature to 37.
Get a 5 gallon insulated beer cooler and fill it with water from that tap.
Put your Petri dish in a zip-lock bag, set it in the water, and close the cooler lid.
Take it out an hour later. The water temp is still 37.
But that takes all the fun out of building the device. So let's use the device to
do sous vide cooking, or to incubate chicken eggs (which will hatch in 21 days
even if you are off by 10 degrees). Your sous vide pork chop will come out great
even if you are off by 5 degrees.
In other words, the temperature sensor in the microcontroller is already more
sensitive and accurate than you need.
-----
Get a free science project every week! "http://scitoys.com/newsletter.html"On Mon, Apr 23, 2012 at 2:20 PM, John Griessen <john@industromatic.com> wrote:
On 04/23/2012 03:28 PM, Cathal Garvey wrote:It's the physics behind all the temp sensors, the differences are in the implementation details.
Most other forms of temperature readout
that I've encountered are a bit batch-variable, and I wouldn't be
surprised if the same is true of something cool but hack-ey like your
onboard diode idea
Why would you pay more for a method or mechanism if you can't get any better results
than with the low cost method? You don't just always pay more, you shop for a low accuracy need for low dollars
and hi accuracy for hi dollars. Precision, or fine grained resolution, is inexpensive,
and you can transfer the accuracy from another temp probe to your machine and get repeatable results
without a temp standard built in.
This particular MSP430, for example, MSP430G2230IDR, has a nicely engineered diode temp sensor
inside, that can be switched in to one of its ADC channels to read the temperature of the chip.
The diode will have a small batch to batch random variation, but will always repeat very closely
as temperature differences happen. There is a temperature compensated volt reference for the
ADC so volts measured are truly accurate. Temperature repeatability of .01 deg C is probably
possible, although that's not accuracy, and would only agree that closely on long equilibration
soak times after a temperature change.
All you need is to put the chip in the stirred air you want to heat and you're accurate
to a very reasonable precision, and you could get better accuracy by calibration.
Repeatability and precision are easy to get with silicon microcontrollers.
That one I mention costs $46/qty 100, or 46 cents each and has four channels of 10 bit analog converter.
Is 0.1 degree C inherent reference standard accuracy necessary for incubation? No...
repeatability is nice, an easy cal feature is nice... Your code can have a mode where
when you press a calibrate button, it changes its look up tables to use what your external
fancy temp sensor reference standard says. The code could take input as up/down buttons
to adjust the temperature lookup table while controlling temp at 35 deg C. You would put the tip
of your hi res. platinum temp probe in the incubator for the calibration for a few minutes for each
up/down command until stable at 35.0 deg C. Then it will repeat, and be a transfer standard
for the cal thermometer for many months or years with as good accuracy and precision
as the cal thermometer even though it does not have an internal temp standard that
accurate. It does need a very stable, repeatable volt standard inside or on board, and
to make no changes to the chips used to keep its cal.
John
--
You received this message because you are subscribed to the Google Groups "DIYbio" group.
To post to this group, send email to diybio@googlegroups.com.
To unsubscribe from this group, send email to diybio+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/diybio?hl=en.
You received this message because you are subscribed to the Google Groups "DIYbio" group.
To post to this group, send email to diybio@googlegroups.com.
To unsubscribe from this group, send email to diybio+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/diybio?hl=en.
0 comments:
Post a Comment