Roche 454 GS FLX+ DNA Genome sequencer teardown

[@davidallen6421]

Oh man a 454! I used to work on sequences generated on one of these! I took a few notes as I was watching: Filters/flow control on reagents: yes those are filters, either 0.33 or 0.22 micron. Pretty standard fare for mol-biol labs. They don't pass microbes, keeping your samples free of foreign DNA. Reagent cost: very high cost reagents is standard practice in this business, it's how they make their money. Now, you _can_ make your own (they don't tell you what the reagents actually are, but the reactions taking place in the machine are pretty standard DNA copy operations, so you can take an educated guess, then mass-spec/HPLC the real ones to see if you're right, and make up your own and give it a go). Having said that, the reaction chamber slide thingy (can't remember what it's called) is a consumable (ie can't be re-used, because how do you *know* you've cleaned all the DNA from the last run off and that it won't compromise the next one?) and it probably comes as part of a reagent kit for the machine, so they've got you over a barrel there. Peltier-cooled reagent reservior lack of active temp control: not necessary. The actual enzymatic reactions taking place within the machine are controlled, but all that peltier is doing is keeping reagents chilled so they don't denature or degrade at room temp during the course of the run (pretty common for biol stuff dealing with reactions producing light). The actual temp isn't really important, most labs just use domestic fridges/freezers for this kind of storage. Labour-intensive sample-prep: Yep, that's a given. To get the most out of this machine, you need to provide it with very carefully measured samples. The DNA to be sequenced needs to be sheared into lots of strands of a particular lenth, have special joiners and the beads attached then introduced at a very well defined concentration so you don't overload the sequencer. High-speed cetrifuges for seperating DNA are everywhere in mol-biol labs, so they expect you to have access to standard equipment. Basically, you pretty much expect to employ someone full time to look after the machine. They'd feed it samples, plus do basic analysis to get the crude data to something you can begin to analyse. The plate behind the reaction chamber: Should be held at 37 deg C for the DNA polymerase reaction. That camera is really interesting, I hope you can get it up and running. Thanks for another great video Mike :)

[@cliveramsbotty6077]

Would you be so kind as to give a brief simple explanation of what the machine is used for? Thanks

[@davidallen6421]

No probs. I've already done so a little further down the comments thread. kzbin.info/www/bejne/jpLYnoimaayjedk&lc=z124jfo5nlepwjjis04cepwzgwanyjjowzc0k.1481463132881029 That link was meant to take you there, but it doesn't seem to work. Sigh. Scroll down and look for a comment by TheAmmoniacal and you'll find it in the reply thread.

[@matiastripaldi406]

Just a question. Why would you need a DNA polymerase reaction in a sequencer? Is it like a PCR reaction where you multiply the DNA fragments to make them easier to sequence?

[@topoisomerase]

@@matiastripaldi406 most DNA sequencing is done by synthesis. By that I mean you produce the complementary strand of what you're sequencing, and the DNA bases (A,C,G,T) you add to make that strand either generate a product (like PPi with 454) or include a modified base (ILMN and most others) that allows you to detect that a DNA base was added. DNA polymerase is the enzyme that adds the bases to the growing complementary DNA strand. You do this sequentially and as such read the sequence. There is often a PCR step (or other DNA amplification technique) upstream of the sequencing reaction itself as well. That PCR may or may not use the same DNA polymerase as is used in the sequencing step.

[@Tadesan]

Your shortening of biology to “biol” gives me really uncomfortable feelings. How do you pronounce your abbreviation?

[@landtechnik4k]

when a UFO crashes, we all would like to watch mike disassemble it for the meantime we watch him undo human made crazy tech

[@hene193]

And when the wife asks what did you buy this time. "Oh just an old obsolete genome sequencer" I guess everyone has their hobbies.

[@douro20]

Those heaters are from a company to the north of me in Leavenworth, KS. They are the world's largest manufacturer of thin-film heating elements and also one of the largest manufacturer of ceramic PCBs.

[@MaxKoschuh]

ceramic PCBs, yummy :-) I would love to build a low noise audio preamplifier around a MIL-spec OP-AMP on a ceramic PCB. Would be neat.

[@bborkzilla]

$500,000.00 to $250.00 in 9 years? That's extreme obsolescence.

[@turboslag]

The pace and price of progress!

[@grndkntrl]

The machine was manufactured in 2012. It's now 2016. That's only 4 years.

[@turboslag]

kyoukoku That's an eternity in electronics and high tech! How many iphones and ipads have there been in that time?

[@alerighi]

This must be declared illegal, it's pure planed obsolescence... also doing this for medical stuff, it's a lot worse, first at all, because the price is charged to the state (at least in counties with medical assistance, like all the European ones) and the money goes to a company that makes work and pay taxes in other countries, and also because the right to health it's a constitutional one, so i think it should be illegal that private companies make billions of billions at the expense of sick people and the state (and the citizen taxes)

[@jcnossen]

I don't think it's planned. The cost of the runs with all the expensive reagents (as he mentioned a run costs about 6000$) quite quickly make the machine obsolete, as there will be a new machine that's able to reduce the end user price of DNA sequencing. It would be against the patients interest if they keep using this machine too long.

[@Lactuca]

"mixing small volumes of very expensive clear fluid" That's how I describe my job and that's really all it is. Done lots of 454 sequencing over the years, never had the chance to take one apart though. Done a few qRT-PCR machines though

[@philorkill]

Thank you for this. I'm always impressed how fast this super complex pieces of kit can get obsolete.

[@frollard]

8:22 looks like the molding process where you have a negative mold, and you paint the polished smooth cavity (also acts as mold release), then either fiberglass or vacuum form or thermoset or resin cast the internals in a tumbling mold form. It's how they make boat hulls. (see the episode of how its made re: sea-doo watercraft)

[@gglovato]

keep the UPS that's a top of the line unit!

[@ChipGuy]

I think he has already given that away, according to his tweets...

[@beyonddeath123]

maybe several years ago, thats only worth 200-250$ us with good batteries

[@gglovato]

JH UPS don't really depreciate, specially APC online ones, they last forever and you only need to change batteries there's no reason for them to depreciate, at least here in my market 10+ yr old UPS still cost the same as a new one

[@beyonddeath123]

I only know as I bought one recently, companies get rid of them after lease, so you can get them cheap if you know where to go. I bought a 3000va with new batteries for 350 cdn pesos, though moving it was a bitch.

[@Mythricia1988]

Would be worth quite a lot in some countries. Like here in Norway you pay 100$ for a piece of shit 600va unit made in China. You've no right to complain about your Canadian pesos just yet! :P

[@Mythricia1988]

"Possibly a few astronomy type people wetting themselves" Damn right... Holy shit the pixel pitch. Big humongous (back-illuminated!!) buckets for light gathering. Like, holy crap. 7.4um is considered "rather large" as far as astronomy CCD's go. This one has 13.5um pixel pitch. Compare this to instruments from Santa Barbara Instrument Group (SBIG), and I think I'd safely value this camera pretty highly - you can easily pay upwards of 10'000$ for a 16mpix 7.4um cooled CCD camera, and this has larger pixels which is a pretty attractive feature. And it's cooled! So are many amateur CCDs, but the combination of massive huge pixels and very low-temp capability means the noise figures on this thing must be incredibly good under optimal conditions. Looking at the QE (Quantum efficiency, so, the ratio of incoming photons turned into readable electrons), this thing goes right up over the 90% mark at some wavelengths. For comparison, a typical color CMOS sensor used in a DSLR camera gets up to a whopping 35% or so. Damn. I don't care that I'd never be able to mount this on my 8" SC telescope (this camera is almost as large), nor power it, nor extract any images from it, but damn.

[@Keith_Ward]

Looks like the CCD on this model is 4k x 4k

[@Mythricia1988]

Yup, in the 16Mpix range, which combined with cooling and huge pixel pitch, and back-lit array, puts this camera comfortably in the "obscenely expensive" category if compared to off-the-shelf astronomy cameras. Heck it would be extremely good even in the 2k format... High resolution is not necessarily as important in astronomy - a 4k by 4k camera often ends up oversampling images on most telescopes. Meaning the optical resolution of the telescope is worse than the pixel resolution of the camera. Binning can still help make use of the excess resolution though, so if you can have high resolution AND big pixels, it's fine. Normally you have to pick one over the other.

[@WickdPerfekT]

Mythricia mate, all you need is a larger mount bahaha

[@vincei4252]

21:40 Been tearing down one of these things but I actually managed to get one that is fully working with the computer and all the software. I couldn't get the camera cover open but Mike pointed out that there is button on the computer to open the cover in a video I sent him which I totally missed. Once pressed it takes quite a while and some whirring before the cover unlocks and you can open the camera.

[@RadRider33]

Mike, in regards to the metric observation: I work at an American medical device company and all we do is metric. Much cleaner and simpler. Everyone groans when we need to use IPS...

[@altimmons]

I’m glad to see metric hardware on it, it’s so much more sensible - anyone in the sciences should have an appreciation for the metric system because it’s a sane system

[@altimmons]

I live in the US and in my technical work i typically work in metric

[@AdamWHankins]

I also live right across the street from Spectral Instruments here in Tucson. good friend of mine works for them in development

[@Konecny_M]

Aluminium can be readily nickel plated, that explains the look of the heatsink. Makes sense if any portion of it is exposed to high vacuum section (anodising is a no-go there due to the porosity, hence creating nasty virtual leak).

[@arcadeuk]

Wow, I love these cool biotech teardowns. That's Patreon money well spent!

[@ahelms]

I work for a custom machine integrator company in the US and we only use metric. We only use US fasteners whenever forced by the customer.

[@TheAmmoniacal]

Been studying this stuff for many years as a biochemist, but still never got to use one :( Very interesting to see one tore down to compare theory with reality! Know exactly how it works.

[@uski]

Dare to explain for non-biologists like us all ? :)

[@GoldSrc_]

Yeah dude, I want to know how that thing works.

[@davidallen6421]

It's hard to explain this in a youtube comment, but I will give it a go (note, this is way simplified): First you need to know something about DNA. It's double stranded, each strand is like a string of letters (called bases). There are four letters that make up the strands: A, T, G and C. The two strands are wound around each other in the famous double helix due to their complementary nature. Each letter has a complement, A complements T, and G complements C. So you might get a double strand like this (the 5' and 3' is mol. biol. jargon denoting the direction of the strands) : 5'-ATGCCGTAGCAT-3' 3'-TACGGCATCGTA-5' The complimentary bases of one string stick to those in the other string via hydrogen bonding, which is what holds the double helix together. It's an important property that is leveraged by the sequencer. Okay, so you do a whole genomic DNA extraction from your organism of choice. Your sample will contain millions of cells, which means you get millions of copies of the chromosomes. You take your purified chromosomal DNA and shear it up into lots of little pieces approx a few hundred bases long. Now you 'ligate' (attach) some very short pieces of DNA with a known specific sequence to the ends of all your unknown sheared-up chromosomal sample DNA (these are adapters). You then denature the double strands so they fall apart into single stranded pieces. Take the single stranded pieces and add them to a suspension containing little tiny manufactured balls with short pieces of single stranded DNA (ssDNA) complimentary to the special adapters you stuck on your sample previously. Your ssDNA will be able bind to the adapters on the balls via complementary pairing, forming short double-stranded DNA right near the ball, with the still-unknown rest of the single strand floating off it. The balls outnumber the ssDNA pieces many times over, so on average, you only get one ssDNA piece on a ball and most won't get any DNA at all. Now put these balls onto the 'nano-titre' flow cell slide (the glass thing with the strips in front of the camera in mike's video). This thing has tiny little nano-sized holes in it. The balls will fall into the holes, taking the ssDNA strands with them. Think of it kind of like a nano-sized series of honey comb cells. Now there are so many balls and so little DNA relative to the numbers, that on average, there should be only one ssDNA piece per honey cell (if there even is one in there at all). The reason they are in the holes is, we can now gently wash reagents over them and do reactions without the balls (and hence DNA) moving relative to the camera (which is very important as we'll see). Okay, now comes the important part. The actual reaction. The reaction is to take your ssDNA pieces and make a copy of them one base or letter at a time. The enzyme that does this is called DNA polymerase. DNA polymerase cannot start on a single stranded DNA strand, it MUST start from double stranded and extend the complimentary strand. It's a good thing we have a very short piece of known double stranded DNA formed from the complimentary binding of the adaptors sticking the ssDNA to the the ball, huh? ;) Confused? Yeah, me too. It's ASCII art time! 5'-ATG(*DNA pol.*) -> enzyme moves along making a new strand complimentary to the original 3'-TACGTACGATCATGCTAAG-5' 3'-TACGTACGATCATGCTAAG-5' 5'-ATGCAT(*DNA pol.*) -> 3'-TACGTACGATCATGCTAAG-5' 5'-ATGCATG(*DNA pol.*) -> 3'-TACGTACGATCATGCTAAG-5' And so on and so forth... The trick here in this machine is that the reagents they use are fluorescent. So each letter has a different fluorescent dye attached to it. We'll also only let the reaction proceed for one step at a time. You add the enzyme and the reagent for one base (say A). All the ssDNA pieces in the honey comb cells that start with a T (complimentary to A) will have an A added to them by the enzyme. After the cycle, the dye is excited and induced to fluoresce and the camera takes a photo of the slide. Then, you move on to T. Again, if there are any strands that now have an A exposed next in line, the enzyme adds the T, and you take a photograph. Then G, photo; then C, photo. This cycle keeps on repeating through all the letters over and over until you have a complete copy of the original strands and a series of photographs taken along the way. The enzyme can only add one base at a time because you only give it one base at a time to add. In order to get the actual sequence, we simply (HA!) analyse the photographs. NASA star field algorithms are used to look for changes to the dots in the pictures as the cycle progresses. Each dot point represents one DNA strand in a nano-well. As the colour of the dot changes from one cycle to the next, the colour in the photo will change from one to the next. The algorithm simply reads off the sequence using the colour as a marker. At the end, we have the complimentary sequence to the ssDNA strand that was in each well. Then the hard part is putting all the pieces of the puzzle together (assembling the sequence) to get a complete unbroken sequence of letters from one end of the chromosome to the other. I strongly recommend you have a read of the wikipedia articles on DNA and DNA sequencing. There are some good animations on youtube of the old-school sequencing methods (still in use), like the Sanger method. If you understand how that works, you will see this is just a fancy pants way of doing it the old way :) It just does it massively parallel.

[@kosikko4400]

Thanks a lot buddy for explaining this process, very interesting !

[@davidallen6421]

You're welcome. I've learned so much from Mike and other knowledgeable people in the comments of his videos, that it's good to be able to return the favour and be _that guy_ that knows something about an obscure and expensive piece of equipment. Cheers :)

[@JacGoudsmit]

I can see how you misjudged the size of this thing. At the start of your video it really looks like it's maybe 24 inches (60cm) wide but at 7:06 your hands showed up and I went HOLY CRAP THAT THING IS BIG. :-)

[@mikeselectricstuff]

I did mean to put something in front for scale but forgot.

[@JacGoudsmit]

Your hands do just nicely as a size reference, thanks :-)

[@Tinker001]

Until you realize he's actually a professional basketball player & his hands are twice the size of a normal persons...

[@christopherleubner6633]

That camera has an image intensifier with a fiber taper attached to it. You could build a crazy sensitive astronomy or microscope camera with it. 🤓

[@JBaughb]

This was fantastic. I love teardowns of over engineered tech.

[@msfraser82]

The optical connector on the camera is an MT-RJ duplex multi-mode. You can get multimode patch leads which convert this to multimode ST connectors, which would connect to a standard gigabit ethernet interface (it said gigabit ethernet on the interface board silkscreen). I've only seen the MT-RJ connector used on some old fixed-configuration 100Mbit Cisco switches.

[@kd1s]

And I can tell you there's a missing component. There should be a server with heavy disk space behind the unit. Did a contract back in 2013-2014 doing the IT side of sequencing exomes. Massive amounts of storage and encrypted drives. They used Ion Proton sequencers.

[@kd1s]

Oh and yes, reagents are kept very cold/frozen. And the process you'll find has a hell of a lot in common with developing color film.

[@MrHack4never]

Hey Mike, jugding from all your medical equipment teardowns, are you planning on getting a MRI scanner for teardown?

[@Cofcos]

Having been in an MRI several times, I'm sure they must weight literal tons. And when removed from their site are already torn down into unrecognizable parts. Though, if he could manage to get invited to a teardown of an MRI machine, I'd love to see it...

[@FrozenHaxor]

Lol.

[@BenjaminEsposti]

He should get one of those 100kV+ xray transformers! X3

[@FrozenHaxor]

+Benjamin Esposti He already does.

[@Electronichub_05]

@@FrozenHaxor i need one too haha The arcs must be insane

[@MaxKoschuh]

Mike, this is wonderful. Thank you!

[@ChipGuy]

Yeah I enjoyed that bit of kit very much. Thanks Mike. Somehow those valve blocks with the SMC valves on them are kinda sexy. I'd prefer the smaller Bürkert valves though.

[@semidemiurge]

I live about 3 miles from Spectral Instruments in Tucson, AZ. Not sure if that helps at all.

[@mikeselectricstuff]

Do they have a dumpster?

[@semidemiurge]

I will definitely check tomorrow. Also, a lot of the engineers and scientists in the area are avid cyclists and I think I might have ridden with a SI employee on a recent group ride.

[@semidemiurge]

This telescope uses the 800 series. slotis.kpno.noao.edu/LOTIS/system.php

[@MaxKoschuh]

such a small world

[@ChipGuy]

28:00: Do I see a black patch of magic smoke on that board where the fan cable is attached or is that just a shadow?

[@ChipGuy]

I am afraid that the camera is actually toast and since it seems to be the most expensive part of the entire machine (worth $500.000 new) it's the reason this unit has been scrapped in the first place.

[@icesoft1]

Looks like a 'solder-plated' portion of the board there, and you're seeing the 'dark' glare off of it from deeper in the camera...

[@JWH3]

How did you infer it's toast? A small burn on one part of the board doesn't mean toast, could be repairable.

[@ChipGuy]

It was just an assumption. And yes even that damage is most likely repairable.

[@ChipGuy]

Nope I did not. It was a reflection since there is no soldermask on the board. Yay!

[@palharley]

The Detail of your videos is fantastic. I love when you connect to the SPI and dig into the the underbelly of the code.

[@davidallen6421]

I'm a little surprised to see this machine was manufactured so late (2012 iirc). By that point, there were far better 2nd gen sequencing options available that produced orders of magnitude (literally) more data and deeper sequence coverage for roughly the same cost. My department purchased an Illumina machine in about 2010 that blows the specs of that 454 out of the water. Similar upfront cost, but slightly more cost per run. However, it produced so much more data per run, that you could actually put multiple genomes on the slide at the same time (making sure to use a special DNA labelling technique to tell the final data apart) and do crude 'multiplexing'. This further reduced the cost/genome sequenced. Actually, the reason might be that although 454 didn't get as good depth of coverage, the legths of the DNA reads were much longer than more modern machines (100s vs 10s of bases), making it easier to actually assemble your sequence data into a genome. This is very useful if you're sequencing something that's never been sequenced before, but somewhat useless if you've already got a genome to compare to.

[@Tadesan]

So instead of getting more useful data upfront it gives you a whole lot more data that need a lot more processing to be meaningful?

[@blank7921]

To determine if it is molded look for the ejector pins. Or molding marks things like date codes etc will stick out like a sore thumb. It is possible it is urethane cast.

[@stonent]

Tubes marked W and R, W for water (the external port) and R for Reagent. So likely flush with reagent and then a final flush with distilled water.

[@catt87]

Really looking forward to you having a play with that camera module.

[@johnkeates9434]

It is standard gigabit ethernet over optical fiber. You can use a standard fiber-to-copper adapter. The connector type and fiber type denotes a wavelength, which would be important to get right for the converter box ;-)

[@DavidMills_Physicist]

I've got a board for the camera if you wish to borrow it for a few weeks.

[@mikeselectricstuff]

I found a few on ebay - only worth getting if I can get the software - do you have that ?

[@DavidMills_Physicist]

Yep - got the software too. Sent you an email.

[@campbellmorrison8540]

The trolley is really nice. I have pulled a fair bit of lab stuff apart and Im pleased to see your hands are dirty too. Its surprising how dirty gear that is in a very good environment gets

[@andycristea]

Awesome video! Thank you! Can't wait to see the camera video!

[@Darkassassin09]

Defiantly fascinated to see what you can do with that camera when/if you get it running

[@sdshlanta]

+mikeselectricstuff as for why i is metric, the us medical industry only uses metric measurements.

[@Tadesan]

those cables just actuate the mechanism... that is beautiful

[@altimmons]

The way this would work is that you would have mixed in with the bases ACTG some tagged defective ACTGs with fluorescent probes such that copies of dna would terminate at random points - this gives you a blast and you have to use the computer algorithm to line them back up- different length sequences will come off the column at different times based on charge vs drag and the camera would like read the fluorescent probe

[@Sixta16]

Nice to see new videos coming!

[@the6p4c]

Damn, I really wanted to see that camera working. Looking forward to that next video, if it comes.

[@Spirit532]

The image sensor is most likely behind a tapered FOP. The fiber-optic output is most likely just GigE, but I'm almost certain the actual protocol is nonstandard. CameraLink over fiber is extremely rare and stupidly expensive.

[@mikeselectricstuff]

It isn't tapered -it's a 60x60mm sensor! Fibre protocol is Cypress Hotlink over fibre. Video coming shortly

[@Spirit532]

mikeselectricstuff Whoa! That's a huge sensor, x-ray direct imaging sizes! Explains the price now.

[@Mythricia1988]

60x60mm? That's downright stupid. In like a hilarious, good sort of way. I don't even know of any standard telescope mounting options for such a sensor - it's larger than the "jumbo" sized 2" optical backs on most large amateur scopes, 11" aperture and above. Practically all smaller scopes are only going to have a ~38mm diameter usable imaging circle or so.

[@Spirit532]

@100531782681373837774 Huge silicon is used in direct x-ray imaging. The most high-end~($50-200k) models have stitched sensors of over 30x40cm(but usually TFT, so not quite silicon).

[@Spirit532]

mikeselectricstuff I just found the sensor used in your camera. www.sta-inc.net/sta4150/

[@DoRC]

by the spectral instruments brochure you have yourself a 4k by 4k camera with a 27mm sq sensor. Dude that thing is incredible. I really hope you can get it to do something cool.

[@mikeselectricstuff]

Not necessarily - there are several CCD options listed

[@DoRC]

they have pictures of the different models and yours matches the top of line one!

[@DoRC]

mikeselectricstuff oh ok. the brochure made it look like the one that is in the form factor like yours is the 4k one.

[@richardparry7532]

The fibre connector is an MT-RJ. It used to be popular about 10 years ago for Ethernet connections but replaced by LC which is still the standard. You can pick up MT-RJ network cards on eBay for a few pounds, although not sure what speed that runs at. The connection could be 10, 100 or maybe 1000Mbps. I'd also be surprised if it's Ethernet though, as unlikely there is IP connectivity going on in the camera. I'd hazard a guess it has some special card that communicates. Great video Mike.

[@TheDefpom]

They would still injection mould, just using aluminum mould instead of steel.

[@io504]

I very much approve the Pink Floyd server :) Fantastic video mike, really enjoyed the detailed teardown

[@icesoft1]

That fiber-optic connector looks to be a standard MTRJ connection, and if the camera was also offered in Gig-Ethernet, it's likely also Gig-E over a different medium (besides the usual copper). Being this recent, it probably is actually Ethernet, plugged into a matching SFP module plugged into a fiber Ethernet card on the PC/Server.

[@mikeselectricstuff]

Looked into it some more and it has Spartan 2 FPGAs so a pretty old design, now less optimistic that it will be Ethernet.

[@icesoft1]

Looking at some of the sales literature, you are probably correct. I turned up what looks to be some of the proper interface cards (perhaps), have a look on eBay for "Spectral Instruments camera interface card", there seemed to be a few of them on there that matched the interface description of 'Either 50-pin connector or fiber-optic' (per the sales literature).

[@apathyisanartform]

Pyrosequencer. DNA pattern adheres to the beads and and bases are introduced bound to ATP. ATP is then cleaved by luciferase (lightning bug enzyme) creating a small flash of light. The camera must capture that.

[@AureliusR]

Mike, reading the PDF you linked shows the rack of tubes you have is specifically for the pre-wash stage. You would fill those tubes with the Pre-Wash Buffer after cleaning out the casette from the previous run. So that tray would not be in there during normal operation -- it does indeed come with cassettes (as they call them) which slide in the same way that tray does. Also, the back tube is chilled, as it contains the Sequencing Enzymes (whatever they are) and they have to be kept below -15C from what I understand.

[@AureliusR]

That tray you have is referred to as the Pre-Wash Cassette.

[@AureliusR]

These are the kinds of instructions you get: Wash the appropriate BDD using a soft bristle brush and Sparkleen solution. Rinse thoroughly with nanopure water and let the device air dry on a paper towel. Nanopure water?! Add 6.8 µl of Apyrase to the 20 ml supplemented Titanium Bead Buffer in the conical tube. Re-cap and label the conical tube “EB Wash”. Mix gently by inversion, and place the conical tube on ice. 4. Add 30.6 µl of Apyrase solution to the bottle of supplemented Titanium Bead Buffer. Re-cap and label the bottle “BB2”. Mix gently by inversion, and place the bottle on ice. 6.8ul? How do you even measure that much? Vortex the Enzyme Beads and the PPiase Beads and place them in a Magnetic Particle Concentrator (MPC) for at least 2 minutes for the beads to pellet. Invert the MPC several times and wait at least 2 minutes again. Carefully remove the supernatants and then remove the tubes from the MPC. Right....

[@AureliusR]

Yes, that PicoTiter plate is replaced every time. It actually gets submerged in one of the prep liquids until it's ready to be inserted, which I found intriguing.

[@mikeselectricstuff]

That cooler gets to about freezing point at room temp, but then the heatsink overtemp cutout trips so probably doesn't run quite that cold.

[@andrewmorton9683]

It's easy to measure microlitre quantities with capillary tubes - look up "tlc spotting capillary". (TLC = thin layer chromatography in this context.)

[@bearb1asting]

Really dig your work, and your teardowns.

[@stevec00ps]

(Half way through so you may have already answered this) Not sure it's standard Ethernet fibre - "Fiber optic communication; standard communication to computer by fiber optic cable to proprietary PCI or PCIe card." on www.specinst.com/Brochures%20Rev%20B/800S-camera-broch_revB.pdf

[@msylvain59]

The construction seems to be a lot more clean and tidy than the Horiba blood analyzer.

[@DavidAlanGilbert]

+mikeselectricstuff The problem with pacemakers is that magnets are used for medics to switch them in the field; so it's not an issue with interfering with the operation so much as they're intended to switch mode; www.medscape.com/viewarticle/749751_5 (so for example pacemaker users are told not to put headphones in their shirt pockets).

[@johnfranks]

AVE needs one of these to finish his bartending robot.

[@KingOfKYA]

For fiber google "MTRJ (Male) to LC multimode " the orange means it will likely be 1gbit. (the lc connector will fit most pcs/switches) with a 850nm sfp, but it may be another odd feq. Good news is all of that fairly cheap second hand:P

[@agoodm]

Most APC UPS have the ability to step up or down the voltage to keep output at an optimal level without going to battery unnecesarilly.

[@DextersTechLab]

Nice score Mike, i was watching that one for a while and would have had it but i was knee deep in Quantel shite :-P Will watch with interest tomorrow.

[@OneBiOzZ]

i have to say it looks like no fibre ethernet i have seen before, but i have seen them use all sorts of connectors internally inside fibre switches to route internally including just off the shelf SPDIF fibre so who knows it could be just a high speed serial interface like LVDS but you might want to try looking inside to see what chip is driving it and if there is a space for ethernet magnetics you might be able to solder on (such as gb or 10gb)

[@stonent]

Looks like this could be the camera controller: www.ebay.com/itm/Spectral-Instruments-PCBA-SCH-FO-Gigabit-I-O-3097-Rev-G-Camera-Interface-Card-/291956437009?hash=item43f9f5b011:g:kXsAAOSwo4pYFJgf It uses 850nm fiber, so it may be possible to use a standard 850nm fiber to ethernet media converter box to get into the camera. The camera brochure indicates an SDK is available for the camera upon request.

[@michaelaj4495]

stonent nice find man

[@Mulletsrokkify]

That's not a standard ethernet fibre connector Mike. The usual standard LC or the older SC and ST look very different. That's an MTRJ connector, probably quite easy to get an LC to MTRJ fibre somewhere

[@rkan2]

There are few US-ebay listings which show the IBM-1U server and one of them also shows it running Red Hat with Roche software and three Firefox windows, which seem to have been opened at start..: :P

[@RhythmGamer]

That fiber connector is MTRJ, not really that common but its for ethernet comm

[@you-dont-know-me]

How do you obtain data from the serial port? I want to learn that!

[@AureliusR]

Dunno if you've learned in the 6 months since -- but it's pretty simple. RS-232 (aka serial) has a very distinct pattern. Whenever you see a 4-pin header near a processor, or near the edge of the board, there's a decent chance it'll be serial. RS-232 voltage levels are +/-15V, but often microcontrollers send the same signals with TTL levels (0/5V or 0/3.3V) instead. In either case, you just get a serial-to-USB dongle of some kind (there's about a billion different varieties floating around on eBay, or you can shell out for a nicer one from a supplier -- Adafruit have a good one) and connect it up. And voila, you'll get serial comms. Sometimes it'll be actually connected to a DE-9 port, in which case it probably *is* RS-232, and in that case you can (you guessed it) get an RS-232 to USB converter. The only difference between the two types of dongle is the RS-232 type has a MAX232 or similar chip that converts between the two voltage signalling levels. Let me know if you have any other questions.

[@ChipGuy]

The housing looks like a vaccum low pressure mould. Something like this www.citim.de/de/Polyamidguss-Vakuumguss

[@MaxKoschuh]

very interesting, never heared of that before

[@ChipGuy]

There is a smaller company that also does room temperature vaccum mould (Raumtemperatur Vakuumguss) into silicone forms: www.mfd-modelle.de/ They can include premade metal brackets, brass thread inserts and other stuff into the mold. Here is one example: www.ibpmt.com/hdc_e/adoc.asp?hdc_e/hdc_overview.htm This housing is produced with that vaccum mould technology

[@Morbuto]

Could be, but it's pretty expensive for a piece that big, especially since only a few companies have the vacuum chambers large enough for parts like this. Another option is RIM (reaction injection molding).

[@Graham_Langley]

+Morbuto RIM/PU was my thinking too.

[@SionynJones]

Thanks Mike

[@joe72205]

Pacemakers (at one time) used a reed swtich to change operating mode (e.g. to high-activity). I don't think the magnetic field affects the actual operation of the device.

[@heathwellsNZ]

Many thanks for such an interesting video as always.

[@DoRC]

what kind of search terms do you use to to find things like this on eBay?

[@proluxelectronics7419]

Looks like Ethernet connection running "SI Image software suite" for config,

[@90SecondsofAviation]

Mike, might we know how You find these things in ebay ? specific categories or what ?

[@markhodgson2348]

Computer waste or wee recycling ♻️

[@90SecondsofAviation]

@@markhodgson2348 yo man that was 4 years ago

[@ejayerik]

Incredible teardown.

[@mbainrot]

the cooler does have a temperature sensor on it

[@nRADRUS]

but another side of radiator ?

[@tbbw]

They could be smart enuff to have figured out that if they feed X ammount of Amps into those peltiers X ammount of units of heat gets moved. Unless the role of that thing is to just keep the liquid frozen when it's not used and they could more or less run them on almost full blast until it was time to unfreeze it.

[@JasperJanssen]

The s/n on the camera is 1119, which rather puts an upper bound in the number of DNA sequencers made.

[@docdaneeka3424]

Lower bound, surely.

[@JasperJanssen]

No. The serial number would probably be numbered through for all the other variants of the camera that didn't go into DNA sequencers. Obviously it doesn't tell you how many were made after this one, but it does tell you that no more than that many were made up to 2012.

[@docdaneeka3424]

Then by that reasoning it doesn't tell you anything about how many DNA sequencers were made. I don't see how seeing serial number x tells you there are no other machines out there with serial number greater than x? I could go on ebay and buy the same machine with serial number 1120? How do you get that '1119' was the last one made in 2012?

[@TheCaitlinlopez]

Where do you get this? the gel and electricity that this use cooling and heating are very expensive

[@stonent]

Possible camera brochure: www.photonlines.co.uk/downloads/800s-camera-broch_reva.pdf "Back illuminated CCDs along with industry-leading low read noise creates the perfect camera for low-light level applications such as astronomy, bioluminescence and plate reading for drug discovery."

[@pepper669]

Hope there wasn't any nasty stuff remaining in there.

[@bluephreakr]

50:55 was worth the wait.

[@mrfrog8502]

01:00 BioTech squishy since stuff 😂

[@NanoCottage]

Yes that fibre optic connector is standard, it's an MTRJ. They were very common on switches/routers around 15 years ago, hardly ever used today. A cable to adapt this to a more common format such as LC/SC/ST is easily sourced. I'd expect it to operate at 850nm, possibly stamped on the transciever in the camera? Can you see any visible light from the fibre connector when the camera is on? Apparently the interface spec is "AIA Camera Link" edt.com/downloads/pcidv_family-3.pdf Some frame grabbers here edt.com/product/pci-dv-fox/

[@KeenanTims]

Definitely not a standard connector used in fibre Ethernet applications. It looks like an 'MT-RJ' style connector on the camera side, not common in Ethernet but could be. This pdf: www.specinst.com/Products/800s_brochure.pdf mentions a 'proprietary PCI card'.

[@commonmogoreanu7135]

Great video! Thank you Mike!

[@marcus_w0]

I'd kill to get one of these for under a thousand bucks! What do I have to search? "DNA Sequencer" didn't return any hits in ebay ;)

[@mikeselectricstuff]

Try Genome sequencer or Roche 454. Plenty in USA atm but Ebay search omits international search results from sellers who ask international bidders to contact them for shipping. Search via geo-ship.com to bypass this filter. Shipping cost will be huge unless you can get them to remove the UPS and frame, but make sure you get the PC (or at least the PCI card) and camera PSU.

[@ArtesianFalma]

It looks small to begin with, then you realise it's the size of a plane

[@MaxKoschuh]

exactly my thoughts. I was surprised when I saw the first size comparison with Mike's hand.

[@JohnWatkinsUK]

With the large sleeve, I wondered if mike had shrunk.

[@NickAlexeev]

The large plastic parts could have been made by urethane casting. Then polished after they have been removed from molds.

[@TheModerGuy]

39:34 chime triggers an epileptic episode in mike

[@redtails]

I love your teardowns. In comparison to other teardown youtubers which I follow, you have by far the most interesting items and you also seem to possess the most knowledge even beyond electronics. (it's also such a waste for this machine to be replaced after just a couple of years, though I can sort of understand if the reagents are super expensive)

[@jayc2469]

The case _sounds_ like GRP

[@davemould4638]

Low production run injection moulds can be made pretty cheaply these days. There are several companies offering a complete service whereby you upload a CAD file of the intended finished plastic part to them and they make aluminium A and B moulds using a CNC machine fed essentially from the CAD file with minimum human interaction needed, and then they can produce a few thousand parts. Your design is analysed by their computer and you get feedaback a day or less after uploading your design to highlight any errors that would make it unmanufacturable such as insufficient draft angles for the chosen plastic or features that are too narrow & deep to machine, and also warnings about likely sink dimples or flow lines etc. An engineer will also discuss the positions they plan to place the gates and parting lines to ensure that it will not cause any problems. I used a company called "Protolabs" in the UK but there are many others.

[@alext9067]

Is that a Frankenstein machine?

[@DatBlueHusky]

50:52 "ribbed for pleasure" lol

[@liamharrison3]

its weird, you dont get an idea of the size until he talks about the surface finish

[@pcuser80]

Linux as os?

[@markg735]

Unlikely. From the serial boot log I saw OSIdleCtrMax. That's a dead giveaway that it's Micrium's µC/OS.

[@Fan119]

378 views and someone has already disliked it, they are subscribed and get notifications on new videos just to dislike.

[@lorenzo42p]

could you sequence the dna of the star child skull?! love your vids. we might never get to see these things if it weren't for you.

[@SkuldChan42]

As an American fwiw - the only people who don't use metric here is the government. In most industries - its all metric.

[@ChristophFretter]

Not metric in general, but metric for fluidics is rather uncommon, even here in europe a lot of piping stuff is imperial.

[@Soundfactory24]

I still remember the time 40 Years ago when I took a mechanical kitchen alarm, which is, of course, much more complicated than this cheap device ;-)

[@DextersTechLab]

That camera looks very trick, i hope you can get it working in some way. It would be a terrible shame for something like that not to be re-purposed.

[@JGunlimited]

Checked on Ebay. I think your videos bumped the price quite a bit ;-)

[@monelfunkawitz3966]

I think I went to high school with Genome Sequencer.