DIY Plate Reader
738 views
Skip to first unread message
Inventoriffic
Aug 29, 2009, 7:34:38 AM8/29/09
to DIYbio
Hi,
I am planning on building a microplate reader for some experiments I
have planned. I'm just going to outline my plan to see if others have
some better ideas, or see folly where I don't :). It's been many years
since I was in biotech and I never went into the hardware side before.
I am using an arduino based platform for this.
1. Proof of concept Build basic prototype spectrophotometer
- insert curvette's manually to switch between sample & controls
- basic LDR + LED combo
- no wavelength filter
2. Programmatic wavelength alteration for the basic spec
- test a laser based light source
- use a diffraction grating + servo to alter the wavelength of the
light source
At this point I will test the accuracy at a local lab.
3. Build prototype stepper motor based plate reader
- designed so that I have control over x-y position + wavelength
My major question is around finding a good diffraction grating (or
alternative wavelength modifier) that I can control easily to get the
desired output wavelength. Any ideas? All criticism is helpful :)
I am planning on building a microplate reader for some experiments I
have planned. I'm just going to outline my plan to see if others have
some better ideas, or see folly where I don't :). It's been many years
since I was in biotech and I never went into the hardware side before.
I am using an arduino based platform for this.
1. Proof of concept Build basic prototype spectrophotometer
- insert curvette's manually to switch between sample & controls
- basic LDR + LED combo
- no wavelength filter
2. Programmatic wavelength alteration for the basic spec
- test a laser based light source
- use a diffraction grating + servo to alter the wavelength of the
light source
At this point I will test the accuracy at a local lab.
3. Build prototype stepper motor based plate reader
- designed so that I have control over x-y position + wavelength
My major question is around finding a good diffraction grating (or
alternative wavelength modifier) that I can control easily to get the
desired output wavelength. Any ideas? All criticism is helpful :)
JonathanCline
Aug 29, 2009, 12:33:01 PM8/29/09
to DIYbio
On Aug 29, 6:34 am, Inventoriffic <inventorif...@gmail.com> wrote:
> 1. Proof of concept Build basic prototype spectrophotometer
> - insert curvette's manually to switch between sample & controls
> - basic LDR + LED combo
> - no wavelength filter
believe have a temperature coefficient as well (one datasheet I just
found shows +/- 20%). It is probably better to use a
phototransistor. Also, the projects I've seen use fiber optics to
route the light to the sample. You could try light pipes instead.
The focusing optics are important.
## Jonathan Cline
## jcl...@ieee.org
## Mobile: +1-805-617-0223
########################
Tito Jankowski
Aug 29, 2009, 12:34:25 PM8/29/09
to diy...@googlegroups.com
Cool, what experiments are you interested in doing after it's working?
Tito
Inventoriffic
Aug 29, 2009, 7:23:53 PM8/29/09
to DIYbio
Thanks Johnathon, that makes sense. I was also thinking of taking the
old CCD off a webcam minus it's filter as another thing to try for
detectors.
old CCD off a webcam minus it's filter as another thing to try for
detectors.
Inventoriffic
Aug 29, 2009, 8:07:08 PM8/29/09
to DIYbio
Once I have it working I have many experiments planned :)
- broad spectrum analysis of hormonal changes due to stress/food
- correlations of hormonal concentrations to simpler measures
(questionnaires etc)
- changes in hormonal state due to hypnotic suggestion (my friend is a
hypnotist)
In general there have been few [public] studies collecting
longitudinal data on these. Nearly all of the studies I have been
involved in have so much overhead and so little data actually
collected that it's no wonder the field is still in it's infancy.
- broad spectrum analysis of hormonal changes due to stress/food
- correlations of hormonal concentrations to simpler measures
(questionnaires etc)
- changes in hormonal state due to hypnotic suggestion (my friend is a
hypnotist)
In general there have been few [public] studies collecting
longitudinal data on these. Nearly all of the studies I have been
involved in have so much overhead and so little data actually
collected that it's no wonder the field is still in it's infancy.
Simon Quellen Field
Aug 29, 2009, 10:50:55 AM8/29/09
to diy...@googlegroups.com, invent...@gmail.com
Try using multiple LEDs.
An infrared, a red, an orange, a yellow, a green, a blue, a violet,
and an ultraviolet. That makes 8 LEDs, perfect for an IO port.
Switch them on one at a time and record 8 different bands.
An infrared, a red, an orange, a yellow, a green, a blue, a violet,
and an ultraviolet. That makes 8 LEDs, perfect for an IO port.
Switch them on one at a time and record 8 different bands.
Inventoriffic
Aug 30, 2009, 8:46:06 PM8/30/09
to DIYbio
Yeah, I thought about that, but then you are limited to the wavelength
of the LED's you select.
After playing in the weekend I have come up with something that looks
like it should work for a full spectrum spectrophotometer.
1. Use a full-spectrum light source
2. Fiber optics to focus the light - thanks to Jonathon
3. A cd. Angle of reflection is positioned by a servo
4. Another fiber optic transfer (essentially selecting a single
wavelength of light)
5. Sample illumination
6. Absorbance detection
This should give me the ability (with calibration) to have a
wavelength specific spectrophotometer.
The problem with this is I would like it auto-calibrated as the servo
will likely have drift.
I have been thinking of using an experiment like:
http://www.practicalphysics.org/go/Experiment_124.html
With a diffraction grating such as:
http://www.rainbowsymphonystore.com/difgratfilsh.html
The 1504 model has 1000 lines/mm which should give me reasonable
resolution.
Any thoughts?
On Aug 30, 2:50 am, Simon Quellen Field <sfi...@scitoys.com> wrote:
> Try using multiple LEDs.
> An infrared, a red, an orange, a yellow, a green, a blue, a violet,
> and an ultraviolet. That makes 8 LEDs, perfect for an IO port.
>
> Switch them on one at a time and record 8 different bands.
>
of the LED's you select.
After playing in the weekend I have come up with something that looks
like it should work for a full spectrum spectrophotometer.
1. Use a full-spectrum light source
2. Fiber optics to focus the light - thanks to Jonathon
3. A cd. Angle of reflection is positioned by a servo
4. Another fiber optic transfer (essentially selecting a single
wavelength of light)
5. Sample illumination
6. Absorbance detection
This should give me the ability (with calibration) to have a
wavelength specific spectrophotometer.
The problem with this is I would like it auto-calibrated as the servo
will likely have drift.
I have been thinking of using an experiment like:
http://www.practicalphysics.org/go/Experiment_124.html
With a diffraction grating such as:
http://www.rainbowsymphonystore.com/difgratfilsh.html
The 1504 model has 1000 lines/mm which should give me reasonable
resolution.
Any thoughts?
On Aug 30, 2:50 am, Simon Quellen Field <sfi...@scitoys.com> wrote:
> Try using multiple LEDs.
> An infrared, a red, an orange, a yellow, a green, a blue, a violet,
> and an ultraviolet. That makes 8 LEDs, perfect for an IO port.
>
> Switch them on one at a time and record 8 different bands.
>
Inventoriffic
Aug 31, 2009, 7:12:45 AM8/31/09
to DIYbio
I read my post again and I wasn't very clear...
- I am using the cd as a cheap diffraction grating to split the white
light into it's constituent parts. I could equally use a prism I
guess.
- I angle the cd so that the correct wavelength is passing through the
fiber optics that illuminate the sample
Hopefully that's clearer.
- I am using the cd as a cheap diffraction grating to split the white
light into it's constituent parts. I could equally use a prism I
guess.
- I angle the cd so that the correct wavelength is passing through the
fiber optics that illuminate the sample
Hopefully that's clearer.
Simon Quellen Field
Aug 31, 2009, 12:19:27 AM8/31/09
to diy...@googlegroups.com
You didn't mention why you thought you needed a full spectrum for a plate reader.
I would think 8 colors is overkill.
NASA doesn't use more than about 4 or 5 colors for most earth imaging.
You can tell chlorophyll from green paper with RGB and an infrared channel.
Just looking at the levels of R, G, and B gives you 16 million colors on your computer monitor.
With 8 bits of resolution on your A to D converter, and three LEDs, you can easily
split the spectrum into 16 million pieces, which ought to be enough. Sodium
yellow will excite the green and the red LEDs more than the blue one, and you
can get the frequency pretty darn close with 24 bits of data.
Inventoriffic
Aug 31, 2009, 5:35:42 PM8/31/09
to DIYbio
Hi Simon,
On Aug 31, 4:19 pm, Simon Quellen Field <sfi...@scitoys.com> wrote:
> You didn't mention why you thought you needed a full spectrum for a plate
> reader.
Many protocols require measuring absorbance at specific frequencies,
e.g. 490nm.
> I would think 8 colors is overkill.
> NASA doesn't use more than about 4 or 5 colors for most earth imaging.
> You can tell chlorophyll from green paper with RGB and an infrared channel.
> Just looking at the levels of R, G, and B gives you 16 million colors on
> your computer monitor.
> With 8 bits of resolution on your A to D converter, and three LEDs, you can
> easily
> split the spectrum into 16 million pieces, which ought to be enough. Sodium
> yellow will excite the green and the red LEDs more than the blue one, and
> you
> can get the frequency pretty darn close with 24 bits of data.
I'm confused - does this give you specific frequencies somehow?
On Aug 31, 4:19 pm, Simon Quellen Field <sfi...@scitoys.com> wrote:
> You didn't mention why you thought you needed a full spectrum for a plate
> reader.
e.g. 490nm.
> I would think 8 colors is overkill.
> NASA doesn't use more than about 4 or 5 colors for most earth imaging.
> You can tell chlorophyll from green paper with RGB and an infrared channel.
> Just looking at the levels of R, G, and B gives you 16 million colors on
> your computer monitor.
> With 8 bits of resolution on your A to D converter, and three LEDs, you can
> easily
> split the spectrum into 16 million pieces, which ought to be enough. Sodium
> yellow will excite the green and the red LEDs more than the blue one, and
> you
> can get the frequency pretty darn close with 24 bits of data.
Simon Quellen Field
Aug 31, 2009, 6:10:06 PM8/31/09
to diy...@googlegroups.com
You calibrate it with a mercury spectrum from a fluorescent light.
In principle all you need are two receivers -- one at each end of the spectrum.
Suppose each receiver has a Gaussian response to light (or a response
curve the computer can transform to Gaussian). You take an 8 bit reading
of each receiver. You add the two to get the total light available, and then
divide each reading by the total. This gives you two 'distances' from the
actual frequency. Starting with a few known frequencies (from the mercury
spectrum) you calibrate and get an 8 bit frequency result.
Suppose one LED was infrared at 1,000 nm, and the other was UV at 400 nm.
You have a 600 nm spread, divided by 256, for better than 3 nm accuracy.
Adding more LEDs with different color response gives you more resolution,
and better accuracy, as well as reducing the effects of any light with a frequency
outside the range of interest.
And you have no moving parts.
It gets better. With a microprocessor and 8 LEDs, you can set up the circuitry
so that one of the LEDs is lit up (an output) while the others are being read as
inputs. Cycle through all 8 LEDs this way, and you have lots of data to play with,
since each emitter is a different color, and each receiver has a different frequency
response. Lots of redundancy, and lots of resolution (not quite the 256 bits you
might have thought, but way more than you need).
In principle all you need are two receivers -- one at each end of the spectrum.
Suppose each receiver has a Gaussian response to light (or a response
curve the computer can transform to Gaussian). You take an 8 bit reading
of each receiver. You add the two to get the total light available, and then
divide each reading by the total. This gives you two 'distances' from the
actual frequency. Starting with a few known frequencies (from the mercury
spectrum) you calibrate and get an 8 bit frequency result.
Suppose one LED was infrared at 1,000 nm, and the other was UV at 400 nm.
You have a 600 nm spread, divided by 256, for better than 3 nm accuracy.
Adding more LEDs with different color response gives you more resolution,
and better accuracy, as well as reducing the effects of any light with a frequency
outside the range of interest.
And you have no moving parts.
It gets better. With a microprocessor and 8 LEDs, you can set up the circuitry
so that one of the LEDs is lit up (an output) while the others are being read as
inputs. Cycle through all 8 LEDs this way, and you have lots of data to play with,
since each emitter is a different color, and each receiver has a different frequency
response. Lots of redundancy, and lots of resolution (not quite the 256 bits you
might have thought, but way more than you need).
Inventoriffic
Sep 1, 2009, 7:24:50 AM9/1/09
to DIYbio
Hi Simon,
That *sounds* great if it will work, much simpler on the hardware
side.
It seems your method assumes that a sample that requiring illumination
of 490nm light (e.g. an assay) will still absorb light at both 400nm
and 1000nm, with the absorbance being some function of the distance
from both samples. The more different illumination sources the better
if this is the case.
But - I don't know if this is a good assumption or not. From what I
understood this type of analysis needs specific wavelengths.
Actually, if I am understanding you correctly, multiple led's won't
work:
http://www.chm.davidson.edu/vce/Spectrophotometry/AbsorbanceSpectrum.html
Am I understanding correctly?
On Sep 1, 10:10 am, Simon Quellen Field <sfi...@scitoys.com> wrote:
> You calibrate it with a mercury spectrum from a fluorescent light.
> s
That *sounds* great if it will work, much simpler on the hardware
side.
It seems your method assumes that a sample that requiring illumination
of 490nm light (e.g. an assay) will still absorb light at both 400nm
and 1000nm, with the absorbance being some function of the distance
from both samples. The more different illumination sources the better
if this is the case.
But - I don't know if this is a good assumption or not. From what I
understood this type of analysis needs specific wavelengths.
Actually, if I am understanding you correctly, multiple led's won't
work:
http://www.chm.davidson.edu/vce/Spectrophotometry/AbsorbanceSpectrum.html
Am I understanding correctly?
On Sep 1, 10:10 am, Simon Quellen Field <sfi...@scitoys.com> wrote:
> You calibrate it with a mercury spectrum from a fluorescent light.
Inventoriffic
Sep 1, 2009, 11:28:52 AM9/1/09
to DIYbio
:) Looks like I'm reinventing the wheel somewhat, but provides me with
better confidence that the idea has legs.
http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6THP-4T9M60F-3&_user=10&_rdoc=1&_fmt=&_orig=search&_sort=d&_docanchor=&view=c&_searchStrId=997287443&_rerunOrigin=google&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=5b957f2606188dcd352a9c129e1b5e8e
better confidence that the idea has legs.
http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6THP-4T9M60F-3&_user=10&_rdoc=1&_fmt=&_orig=search&_sort=d&_docanchor=&view=c&_searchStrId=997287443&_rerunOrigin=google&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=5b957f2606188dcd352a9c129e1b5e8e
> That *sounds* great if it will work, much simpler on the hardware
> side.
> It seems your method assumes that a sample that requiring illumination
> of 490nm light (e.g. an assay) will still absorb light at both 400nm
> and 1000nm, with the absorbance being some function of the distance
> from both samples. The more different illumination sources the better
> if this is the case.
> But - I don't know if this is a good assumption or not. From what I
> understood this type of analysis needs specific wavelengths.
>
> Actually, if I am understanding you correctly, multiple led's won't
> work:http://www.chm.davidson.edu/vce/Spectrophotometry/AbsorbanceSpectrum....
> side.
> It seems your method assumes that a sample that requiring illumination
> of 490nm light (e.g. an assay) will still absorb light at both 400nm
> and 1000nm, with the absorbance being some function of the distance
> from both samples. The more different illumination sources the better
> if this is the case.
> But - I don't know if this is a good assumption or not. From what I
> understood this type of analysis needs specific wavelengths.
>
> Actually, if I am understanding you correctly, multiple led's won't
Inventoriffic
Sep 1, 2009, 11:39:27 AM9/1/09
to DIYbio
One further update:
http://assets0.pubget.com/pdf/19064105.pdf
The diagram in Fig 1 is essentially exactly the same as I prototyped
above, but with more simplicity and elegance! :)
On Sep 2, 3:28 am, Inventoriffic <inventorif...@gmail.com> wrote:
> :) Looks like I'm reinventing the wheel somewhat, but provides me with
> better confidence that the idea has legs.
>
> http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6THP-4T9M60...
http://assets0.pubget.com/pdf/19064105.pdf
The diagram in Fig 1 is essentially exactly the same as I prototyped
above, but with more simplicity and elegance! :)
On Sep 2, 3:28 am, Inventoriffic <inventorif...@gmail.com> wrote:
> :) Looks like I'm reinventing the wheel somewhat, but provides me with
> better confidence that the idea has legs.
>
Eugen Leitl
Sep 1, 2009, 11:50:06 AM9/1/09
to diy...@googlegroups.com
On Tue, Sep 01, 2009 at 08:39:27AM -0700, Inventoriffic wrote:
>
> One further update:
> http://assets0.pubget.com/pdf/19064105.pdf
> The diagram in Fig 1 is essentially exactly the same as I prototyped
> above, but with more simplicity and elegance! :)
Since you're not responding to private mail, I have to
>
> One further update:
> http://assets0.pubget.com/pdf/19064105.pdf
> The diagram in Fig 1 is essentially exactly the same as I prototyped
> above, but with more simplicity and elegance! :)
send it on-list.
Please do not top-post, please trim your
replies and cite correctly (message unchanged
below). Thank you.
Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
______________________________________________________________
ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A 7779 75B0 2443 8B29 F6BE
tuur
Sep 1, 2009, 7:21:03 AM9/1/09
to DIYbio
A few months ago, I developed a proposal with some friends for an
interaction-design workshop / festival. We proposed an installation
and visualisations based on spectrometry of urine, with an open source
arduino-based spectrometer.
Have a look at the project proposal here:
http://openwetware.org/wiki/Citizen_Science/The_Urinomics_Project
on the resources page (http://openwetware.org/wiki/Citizen_Science/
The_Urinomics_Project/Resources) , there's some links to earlier cheap
spectrometer projects, make sure to check the paper on the SLIM
spectrometer and the specifications for the TSL230R Light to Frequency
Converter.
keep us posted on progress and experiments!
Tuur
interaction-design workshop / festival. We proposed an installation
and visualisations based on spectrometry of urine, with an open source
arduino-based spectrometer.
Have a look at the project proposal here:
http://openwetware.org/wiki/Citizen_Science/The_Urinomics_Project
on the resources page (http://openwetware.org/wiki/Citizen_Science/
The_Urinomics_Project/Resources) , there's some links to earlier cheap
spectrometer projects, make sure to check the paper on the SLIM
spectrometer and the specifications for the TSL230R Light to Frequency
Converter.
keep us posted on progress and experiments!
Tuur
Mackenzie Cowell
Sep 1, 2009, 2:52:49 PM9/1/09
to diy...@googlegroups.com
Tuur, thanks! Let's build some solid-state specs!
This recent paper discusses the use of a variety of sensors for detecting LED light in photometer applications, including phototransisters (PTs), photodiodes (PDs), light dependent resistors (LDRs), and reverse-biased LEDs.
A quote:
The novel use of an LED as both light source and detector for analytical applications has been developed by Lau et al. [74, 75]. The emitter LED is forward biased while the detector LED is reverse biased. Instead of measuring the photocurrent directly as performed by Mims III [82] and Berry [84], a simple timer circuit is used to measure the time taken for the photocurrent generated by the emitter LED to discharge the detector LED from 5 V (logic 1) to 1.7 V (logic 0) to give digital output directly without using an A/D converter or operation amplifier. This method achieves excellent sensitivity in comparison to the method of operating an LED as a photodiode (i.e. to measure directly photocurrent) [81].
Absorbance Based Light Emitting Diode Optical Sensors and Sensing Devices
MO'Toole, D Diamond - Sensors, 2008
Mac
Inventoriffic
Sep 1, 2009, 10:41:33 PM9/1/09
to DIYbio
Tuur,
Great links - thanks!
How far did you get with the project? The TSL230R seems really nice.
Great links - thanks!
How far did you get with the project? The TSL230R seems really nice.
Abizar Lakdawalla
Jul 17, 2015, 8:51:04 AM7/17/15
to diy...@googlegroups.com
A simpler approach is to use a android or apple tablet as an illuminator (many apps that provide bright RGBW colors, I use a free app called "Lightbox"). Place the plate on top of the tablet and then take a picture through an app, such as "ELISA Plate Reader" on Android. The App will then give you values for a 96 well plate. Enzo also has an ELISA reader app. You can also determine HSV values through most photo processing software such as GIMP or JImage. Jiamge already has a module for quantitation of grid images such as 96 well plates. For the photo softwares you just need to take an image through a regular camera and import into the software.
let me know if you need more details.
Thanks
Abizar
Sebastian S Cocioba
Jul 17, 2015, 3:36:10 PM7/17/15
to diy...@googlegroups.com
If you use a photodiode as your sensor, remember its only as good as the power you feed it. You will need a very clean and stable voltage source (5v 3.3v, etc) so the ADC has a static reference. No matter how many bits of ADC accuracy you get, if you have a noisy power rail your signal to noise ratio will drop like a brick. There are many chips on the market that have integrated systems called transimpedence amplifiers but due to the miniaturization of the very high impedance gain resistor inside the chip itself Iv've found varied degrees of sensitivity across the same batch and identical part number chips. You can either spin your own PCB with a chopper amplifier that has "zero-drift" and ultra-low noise or compensate for the variation in sensitivity.
Ive built a few spectrophotometer prototypes for an upcoming publication with my research partner this fall (PLoS, nothing fancy) and have been doing some heavy reading on photodiode theory and practically own every kind of Light-to-X sensor where x is the modality (frequency, voltage, analog, digital, etc) from Mouser. There is no perfect sensor for the job. Also most of the light sensors, due to their chemistry, are biased toward the red/infrared side of the spectrum. Most white LEDs don't have a good full spectral range. Go for xenon lamps if you can, just be sure to cool them or they'll melt your enclosure. Another thing is if you are reading multiple sensors at once, arduino may not be the best. We are switching to the Parallax Propeller system since it has 8 completely independent cpu cores at 80Mhz each. I don't know why more people don't use these little beasts in their projects.
Two books I'd recommend if you happen to go the photodiode route are:
-Photodetection and Measurement by Johnson
-Photodoide Amplifiers by Graeme
Also two very very good intro electronics books:
-The Art of Electronics third edition by Horowitz and Hill
-Practical Electronics for Inventors
Unfortunately I cant share my designs until we publish this fall, which will be open source, but I can offer you my two cents on the matter of spectrophotometery. Been wanting to make a plate reader and have some ideas as well but the need for it is not present in my current research goals. We can exchange notes later if you'd like. Good luck!
Sebastian S. Cocioba
CEO & Founder
New York Botanics, LLC
--
-- You received this message because you are subscribed to the Google Groups DIYbio group. To post to this group, send email to diy...@googlegroups.com. To unsubscribe from this group, send email to diybio+un...@googlegroups.com. For more options, visit this group at https://groups.google.com/d/forum/diybio?hl=en
Learn more at www.diybio.org
---
You received this message because you are subscribed to the Google Groups "DIYbio" group.
To unsubscribe from this group and stop receiving emails from it, send an email to diybio+un...@googlegroups.com.
To post to this group, send email to diy...@googlegroups.com.
Visit this group at http://groups.google.com/group/diybio.
To view this discussion on the web visit https://groups.google.com/d/msgid/diybio/d2fc1b2e-fd59-45a2-8f27-de0cea289dd5%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Reply all
Reply to author
Forward
0 new messages