h a l f b a k e r yPoof of concept
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
This is a suggestion for a possible hack.
BluRay players are astonishing pieces of engineering -
they
focus a laser beam down to less than 500nm, and have
the
transport mechanism to align with the spiral of data on
the
disc at a pitch of only about 300nm. All in all, it's quite
fantastically
precise, yet BluRay players are dirt cheap
because of mass production.
So, here's the proposed hack to convert a BluRay player
into a high-resolution scanning microscope.
The hard part will be producing a sample-carrying
dummy
disc. It will need to be a CD-sized disc of glass, only
0.1mm thick. Fortunately, such glass exists - it is widely
used for microscope cover-slips. I am pretty sure that
the
glassmakers would be able to supply this is CD-sized
discs.
Next, take a regular CD and punch a bunch of holes in it,
maybe 1cm in diameter. Then bond the glass disc to the
holed CD. The CD provides structural support to the
glass
disc. Samples are placed in the holes, in direct contact
with the glass.
Now comes the electronic part. You need to be able to
access the read-head drive mechanism (can't be that
hard
to hack), and to collect the raw signal coming from the
read head.
By accessing the raw signal from the read head, what
you
have is an optical scanner with a resolution of better
than
500nm, which should be able to image the samples
placed
in the holes in the CD (through the 0.1mm glass layer).
Add some fancy software and -
gadulka! - a high-resolution scanning optical microscope.
I haven't yet figured out if the optics in a BluRay player
are
confocal or not, but I believe they are (since BluRay
discs
have more than one data layer). If so, and if the Z-axis
control can also be hacked, then you actually have a
scanning confocal microscope, capable of generating 3D
images of samples.
One drawback is that you'd only be working at a single
wavelength (unless you did a lot more hacking); but even
so, a high-resolution scanning confocal microscope built
from a cheap BluRay player would be pretty amazing.
Someone has already...
https://hackaday.io...g-laser-microscope/
...been there, done that. [MaxwellBuchanan, Oct 14 2016]
Metalens works in the visible spectrum, sees smaller than a wavelength of light High efficiency ultra-thin planar lens could replace heavy, bulky lenses in smart phones, cameras and telescopes
https://www.seas.ha...wavelength-of-light
[beanangel, Oct 19 2016]
Blu Ray Lens Control
http://www.nt.ntnu..../ecc03/pdfs/456.pdf
[bs0u0155, Oct 19 2016]
[link]
|
| |
perhaps you could get the right depth of field with a microtome slice of something, then polish it down to the right focal plane for the CD |
|
| |
Do you mean put a whole slice in the machine,
instead of the glass-windowed disc? Possible, I guess
- but the windowed disc would be more versatile. |
|
| |
One limit might be centrifugal forces at normal disc-
playing speeds; however, there'd be no real need to
spin the disc that fast, and in any case fixed samples
wouldn't be a problem. |
|
| |
Oh arsely arses. This has been done <link>. |
|
| |
Yeah, but you get credit for having thought of it second. |
|
| |
Story of my life, [norm], story of my life. |
|
| |
Hmm, so I disappeared down the rabbit hole of Blu Ray
optics. So could this work, could you make a microscope
by just having a special disc and getting access to the
output of the photodiode? |
|
| |
Well, maybe sort of a bit. The drive has rudimentary
focus control for the disc, it only cares about pit Vs no
pit representing 0&1 so working out which is 0 and 1 is
relatively easy and focusing on the plane of the disc is
also easy. You could retain this system so that it auto
focuses on the plane of the disc. Problem here is that
you would have to get your sample EXACTLY in that
plane. At this point the holder becomes impossible to
make practically. You'd need a much better microscope
to make sure it lined up OK. |
|
| |
This is where you start chucking things away and
replacing them with real microscope components, like
the guys on Hackaday did. The first thing they chucked
was the whole disc spinning and head movement
system. Replaced with a "high precision x-y-stage" I
don't know which one they have but they're all in the
thousands of $. This means you can move the sample in
a
lovely X Y sweep whereas the disc system would have
created a series of arcs. |
|
| |
Once they have control of the X Y movement,
mechanism of checking focus then they're off to the
races. Except they're not. What they've made is a
polarized reflective sort-of-confocal microscope. They
can scan across their sample and assign values to pixels
based on the reflected light received by the
photodiode. To separate the excitation light from the
reflected light they have a polarizing beam splitter. So
they shine light of one polarity and look for another. |
|
| |
This creates problems, edges and angles really mess
with linearly polarized light. So is there no reflection?
Or has the shape of your sample messed with your
polarization? They thought of that though, that's what
the 1/4 wave plate is there for, changes the linearly
polarized light to circularly polarized light. The sample
reflects the circularly polarized light but in the opposite
handedness the plate changes that back to linear and
you are good. |
|
| |
The problem with the sample creating artifacts remains
however. What if your sample does not uniformly
reflect 405nm light? You will never know the difference
between something that doesn't reflect back along the
incident light path and something that isn't there. That
and reflection is great for mirrored discs, and stuff
made of metal. Not much else reflects considerable
light at those scales. Cells for example are essentially
transparent. |
|
| |
Next problem is with moving the sample about. I
assume they are not stepping the sample while it scans
past the lens, but instead just binning the output by
units of time. Commercial confocals instead scan the
excitation laser accross the back of the objective using
unbelievably precise mirrors. The stage isn't going to be
able to match that precision. The advantage though, is
that you can get away with a really dreadful lens.
Which is what they have. A commercial microscope
essentially has the whole of the sample in focus, and
the sample remains fixed relative to the objective
during the scanning. This means you need an incredible
bit of engineering to maintain an even field of view
despite the fact that the light path to the center of the
sample and the edges is different. That's where a lot of
money goes. The Blu Ray lens just has to focus one pixel
at a time. |
|
| |
It is a viable strategy, as long as you have a very, very
good stage. You could use this to make a fluorescence
microscope too, simply removing the polarizing
elements and replacing with dichroics. |
|
| |
metalens material should be mentioned [link] this is like a velvet of waveguides rather than a refractive optical element. much thinner, with higher resolution at the visible, and presumably UV |
|
| |
//Problem here is that you would have to get your
sample EXACTLY in that plane.// |
|
| |
But... the BluRay head has a voice-coil that
constantly adjusts the head height to keep the
correct layer of the disc exactly in focus. For
example, the first layer is directly underneath the
100µm (0.1mm) coating of the disc face. I don't
see why the same system would not keep the focus
100µm behind the face of the coverglass we're
using - i.e. focussed perfectly on the specimen. |
|
| |
In other words, the player is already designed to
image (by scanning) sub-micron pits with sub-
micron spacing, in a plane which is a precise
distance from the surface of the disc - and to do so
even though the disc itself may be slightly warped
and non-flat. |
|
| |
Points such as polarization artefacts and sample
reflectivity are well made. But I do like your idea
of making it a fluorescence 'scope instead - much
more versatile. A good many dyes are excited
around 400nm, as are my favourite things - Q-dots.
you'd want swappable emission filters, though. |
|
| |
//But... the BluRay head has a voice-coil that
constantly adjusts the head height to keep the correct
layer of the disc exactly in focus// |
|
| |
I suspect, and read, that this took a while to get right. I
expect the reflectance values of the surface are fairly
tightly controlled. As you move the lens toward the disc,
you should get increasing values to a peak, then
decreasing. So you can get a ball park range before the
disc starts spinning. Then it's simply moving the lens if
the value starts to fall off. Move it down a tiny bit and
it falls further, then you need the opposite direction. If
the disc is standary blu-ray surface then BOOM! A
neonatal cardiomyocyte, there might be issues with
that algorithm. |
|
| |
// keep the focus 100µm behind the face of the
coverglass we're using// |
|
| |
The working distances actually work out OK, you need
#0 or #1 thickness, but that's easy. I suspect there will
be a problem with the glass though. Your light is then
travelling an awkward path, Air-plastic-air-glass-
sample. Bit of a Zig-zaggy mess of refractive indicies
like the spoon in the half-full glass of water. It gets
worse if you have an aqueous sample... this is why
many confocals use water immersion lenses, they're not
as good as oil in terms of sheer NA values, but the
aberrations are kept to a minimum. (I'll bet the lens
plastic and the disc plastic on the blu ray match)
Interestingly, I don't think this matters too much,
because you are viewing such a restricted volume, any
aberrations wont have the same effect. I'll have to have
a think about that. |
|
| |
//dyes are excited around 400nm, as are my favourite
things - Q-dots. you'd want swappable emission filters// |
|
| |
Well, it starts to get tougher here. Individual
fluorophores aren't as bright as say, a mirror. This is one
of the reasons the laser diodes are so conspicuously
powerful for a domestic device. You need a bit of brute
force if you're going to fire your light through such a
crappy lens and lots of other optical elements.
Especially when you're dealing with a low efficiency
photodiode. Your 10mW of laser is losing say 30% in the
collimators and other organizing elements, so 7mW. say
another 30% in the lens and getting to the sample, say
5mW, then your fluorophore will yield 50% at best,
2.5mW, Your light will be emitted in a full sphere, but
you're 0.8 NA lens will only cover say 20% of that. Then
Back through your optics, and filters, say 50% gone, to
slam into your detector which is about 25% efficient.
About 15uW, which is actually workable. With QDots it
is OK, with anything else, you lose efficiency which is
amplified down the chain, worse, almost everything
else will bleach with that power. |
|
| |
Milliwatt schmilliwatt. A BluRay player has about
half a watt of laser. And bleaching scmeaching too.
Qdots all the way. |
|
| |
Next week: how to fashion a fully-functional heart-
lung machine from a rotary-dial telephone, a
disposable camera and a Corby trouser press. |
|
| |