My Decentralized Reusable Covid-19 Test Development History
What you see is my prototype of hopefully 400 times reusable at home particle filter-based virus/covid19 test as part of my XPRIZE competition.
How it all started? Soon after Wuhan. In March of 2020, I got a crazy idea about how to perhaps create the first cheap reusable covid test using laser and interference change.
“If we breathe air full of median 120 nm Covid 19 particles 5.2log10ml = (168000? in 1mm²) thru filter(acting as multi-hole grating) and use cheap Laser. We should expect to get different interference pattern thus producing pretty much instant cheap and reusable test we need so bad”
This first idea (and later abandoned due to finding a cheaper way) was to project cheap laser pointer thru a filter(multi-slit interference) with holes of specific nm size and separation.
The resulting interference pattern is expected to change as the filter gets clogged and the spacing of open holes is widened. A cheap laser pointer would in theory then allow cheap 3$ measure of filter pore changes happening in nm scale.
As you can see in the simulation above. We can detect nm changes in filter changes via fast image spectrum analysis FFT DCT Or consider (most mobiles now have GPU) detect deviations via AI pre-trained on clogged filters.
Fact that this image capture and AI inference can run in-browser via tensorflow.js means we can with low friction collect valuable data like shedding rates and virus density in central DB. Just one breath through a tube attached to mobile multiple times a day.
I tried to post it on Reddit coronavirus to get feedback hopefully advice but repeatedly post was deleted by admin every time I tried. he said
“it is not a peer reviewed question”
I had no idea there is such a thing as a peer-reviewed question. I was somehow luckily been able to post and squeeze some feedback from a few virologists on virology Reddit at least. They confirmed the need for pre-filtering. But #coronavirus was where all the relevant eyes from the field were.
No asking for opinion from field possible.
I went as far as asking for help from other Reddit admins. Nope. Sharing ideas in times of crisis not possible. I posted ideas to the prestigious journal nature.com just for fun. And received an expected but still hilarious response.
Nature submission: “where is a letter from your university”
without it, the submission will not continue. ;) While I can understand. Those were special times. And I believed that stone should not be left unturned searching for a solution. And no idea dismissed before hearing. I posted on my LinkedIn or Facebook searching for support. Everybody silent.
As I become desperate. I decided to post to the whole academic society in Slovakia in the relevant photonics field. And so there go 200 mails at tuke.sk Kosice university. After a week just two responses. Both not willing to help.
When I was ready to give up completely I sent one last mail to one professor overseeing the photonics lab. As in every mail. I described the idea and asked for his opinion on whether it would work.
Curiously he didn't dismiss nor supported the idea but at least did sent me links to two labs he felt like doing something similar. One was a lab in Israel and one in New York.
But it turned out both were doing just expensive PCR variants with an additional laser. Sent mails to both anyway even to specific emails of people within those labs. But never got any response.
I contacted many organizations releasing development funds at that time. But they never responded except two that said they are sorry. I tried to register online for an EU grant. Registered login but didn't pass the part requiring an accountant secretary and to be a legal entity.
Then I tried to start the project on Kickstarter.
But Kickstarter didn't support Slovakia neither did Indiegogo or Gofundme.
And after many emails begging their support to allow mine project even though from a country not supported by them. I was repeatedly been told that
Kickstarter doesn't support projects like that
Which was weird considering the state world was in. Especially New York at that time.
I started to give up.
But then. After a few months, I noticed There is an XPrize foundation searching for a cheap fast covid test willing to support its development to 5 teams.
Yay music to my ears. Being without work now entry fee 100e was hard. But I believed in my idea strongly so. Little did I knew back then that they would want an additional 100e for each from 7 variants of an idea I had. Luckily they changed mind later.
Finally, I Joined the XPrize Rapid Covid Testing competition
Here is an explanation of the idea as presented to XPrize.
I passed to 2nd round but never got any response to this basic assumption.
I call mine approach simply as two filter test.
To explain it I would like to start with question.
If we agree that there is high count of 120nm[Zhu et al, 2020] sized particles in sputum with covid19 [ 10⁶-10¹¹ rnas/ml max values in sputum https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7224694/]
Then the obvious question is.Is there usually such large amount of particles in 150–100nm range in saliva of healthy person?
If not can we exploit this information to create fast cheap test by filtering out and detecting particles in 150–100nm range?
If so. Then I have multiple different detection approaches discussed laterBut back to basic saliva histogram assumption and whether it itself is sound.
Autumn is approaching fast and there will be a lot of people soon that are sick but due to different virus.What is usually seasonally there in 150–100nm range ? what are the virus densities are and for how long?
Are virus densities/shedding times measurably different smaller than covid19?
Can we actually isolate covid19 just by selecting proper two filters/thresholds ?In worst case if not.
Will such test doable at home still help reduce strain on classical but non scalable tests by testing thousands instead of millions?particle sizes of common seasonally present viruses:
rhinovirus 20– 30 nm avg 25 nm
influenza 80–120 nm avg 100 nm
— — — — — — — — — — — — — — — — — -
covid 19 60–145 nm avg 125 nm [Zhu et al, 2020]
mpv 150–200 nm avg 170 nm
rsv 150–250 nm avg 200 nm
piv2,3 150–250 nm avg 200 nm
The laser variant was cool and all but isn't there a cheaper simpler way?. So I was considering the led diode luminance variant. But will luminance drop be measurable without an expensive avalanche diode?
With a smaller concentration area probably yes but then flow resistance would skyrocket and common dust and contaminants will become a problem.
Or one focusing on the change of electrical properties of such clogged filters. After all virus shells are lipids. A lot of lipids.
Capacitance change? Resistance Change? That could be cheap and flow drop is not issue either.
Or colored ph paper in this concentrated 1mm² hole. Where drop of fluid breaks virus lipid shells releasing a lot of acidic RNA thus changing ph and color?
Or spectrum variant requiring just sunlight and grating or cheap prism. needs mobile phone again.
I even envisioned this
fun crazy weird mechanical contrapiction ;D
(as a proof of concept)
The idea of the last one was that you use two n95 filters (from respirators you have around? I know stupid but this is proof of concept only) which in reality have larger like 1000nm holes or more but are statically charged from the factory.
Discharge static charge on one for example with isopropyl. You breathe in thru pair of foams and large 150nm particle filters and two funnels concentrating filtered air to two small spots on two filters.
The surface area /resistance of charged one capturing particles should in theory keep rising thus moving a needle?
But manufacturing billion of lasers or gratings or ph strips or using precious n95 filters worried me supply chain-wise and were unrealistic in such large quantities at once.
So in the end I abandoned all variants above and focused only on the simplest cheapest filter only variants
Variants measuring the drop in flow (air or fluid) thru a small clogged area on the filter. After all. If they would work why make things more complex than they need to be. The advantage of instant results was sacrificed thou.
Air filter variant(I definitely should have focused on this one instead of pressure variant).
Air balloon(to solve high filter resistance and need for big volume) filled by breath time measurement variant. Reusable mechanical filter moving in x/y to next 1mm spot after each use.
High virus density and flow concentrated to single 1mm² area of 100nm filter should be in theory clogged fast. For the air/breath test idea was to let a people fill the balloon with air and measure how long discharge would last over two filters.
One with a large area blocking larger particles than 120nm and concentrating resulting flow to single spot of a 100nm filter.
Reusability is a big advantage. We need just a small 1mm area of large filter. And just move to new unexposed part after each use. Estimated 400 uses were for the smallest cheapest 23mm filters. sold for 0.24euro
large particle large area filter will not clog that fast or can be cheap washable plastic or commercial nylon or pc 150nm filter preceded with dense replaceable foam.
But the density of virus particles in-breath is lower than in saliva. Smaller area bigger balloon?
Semifinals 1st Fluid filter variant.
Any soda bottle with 2L of 150nm pre-filtered water and diluted saliva installed on top. 150nm filter below red gasket. 50-100nm between blue parts.
The idea is to measure time 2L or smaller? of fluid will take to flow thru. A reusable variant is easy to do the same way as above Just different bottom with a mechanical filter mover but I had no time.
Now I know. why not use electronics but just gravity? Think about reusability and scalability. Billions of tests means tons of precious metals or batteries.
In theory, we could theoretically 3d print the whole test including filters by introducing 1:1000 gearbox to common cheap 400$ UV resin SLA printer to reach 50nm z-axis layer resolution.
Then the only material for the whole test would be UV resin 1L like 30$ where 30ml for 1 test pretty much free.
Beauty is. Manufacturing would become decentralized and due to how SLA printers work you can print 400 tests at once in 4h.
But that's all just bunch of theory and what-ifs and I am a bit skeptical about filter/resin/printer quality and reproducibility. This resin filter one is perhaps too crazy I guess ;)
In a rush to get some measurements, I modified the air variant to compressor. The fatal mistake in retrospect.
Semifinals 2nd compressor filter variant.
This pressure rise variant was designed just for XPrize judges. Since I needed to report measurements somehow and make its basic idea as clear as possible.
But since I could not afford a precise manometer(500e) and XPrize didn't lift a finger for us I improvised. I changed the balloon variant to compressor one since it had psi(with horrible precision but at least something) measurement built-in.
I was struggling with pressure leaks due to a bad seal design.The whole compressor turned out stupid idea considering how delicate the Pragopor filter was and I bet it was forcing particles thru/breaking them too?.
Lapse of judgment under pressure I guess. On the other end, vacuum filtering is common with dense filters?
But 5h before the deadline I was finally able to measure the difference between clogged and unclogged filters. Reported values. And passed to 2nd round
Grande Finale and Judges
We were required to submit a video. So in that video, I tried to demonstrate a reusable variant on video. You can see it on top of the article.
I had like two days left so I cobbled it together given that most of the time was spent on iterating prototype with slow and later broken 3d printers
I did not publish all versions for scrutiny due to time. Even those on protocols.io are not properly tested due to my inability to find help with obtaining filters but posting was mandatory so that's why no measurements are there.
Mine apologies.
I did pass to 2nd round of semifinalists along with 219 other teams. But I didn't pass to 20 finalists.
Strangely variants requiring magnetic nanoparticles or gold substrate passed. Weird. Anyway. What I am most sad about. Is that I never got any help advice or filters or any feedback. Things I was looking forward to the most to move forward faster.
Still, I wish one day we will all meet and have at least a photo together.
This is the official end of the Article.
Please don't read further it's a bit depressive.
Brutal Postmortem
What went wrong. And Why.
I desperately needed filtering membranes Thermo Phisher Scientific was presented on slides as our official XPrize support partner.
“We are here for you we have connections should you need anything you are our heroes”
I remember in the first zoom meeting. I begged XPrize for filter help many times. When I asked to at least get contact with our official supply partner Thermo.
“Sorry cant help”
No contact was given. But no contact with Thermo meant no filters which meant no testing and no measurements.
The nitrocellulose filters I had were cheap 24euro for 100pcs but too fragile. When I bought them. I was also told I snatched the last pack in Czech republic for the next 3 months.
I was hoping to get hands on ideally polycarbonate or nylon ones. They are no problem to mass manufacture just understocked due to unexpected demand spike. I sent a lot of emails begging for help in all Slovak and Czech warehouses.
By some miracle, I was able to contact some high bra from phisher scientific on LinkedIn. Explained I am in XPrize and they are the official partner
And indeed a secretary from Czech warehouse called. Kudos to him. I told her the filters I need are in the Austria warehouse 30km from me. But just told me she is from the Czech warehouse cant help. I was finally able to call
Austria Phisher warehouse.
The only warehouse around that has pc filters I needed.
But they denied my registration even though I mentioned XPrize partnership. 100pcs 700$ filter bulk prices we nightmare. But hope was to find a sponsor later.
Meanwhile. Since I spent months working on prototypes when in XPrize comp. In my contractor daily job I was essentially replaced and my source of income disappeared.
So the whole self-funding adventure pretty much financially ended. I now have 2 broken 3d printers, plenty of filament and resin. One useless US C Corporation without a bank account (but hey at least IRS is now breathing on my neck next month) and
8000 euro is gone.
So that's how spectacularly I failed.
But If I learned anything.
Basically you are always on your own nobody will help you. So have some savings before you start ;)
It's hard to stay motivated.
But my brain still somehow reserves like 20% for at least some hope. ;)
The Crowdfunding Debacle.
Because Kickstarter Indiegogo or GoFundMe for some weird reason doesn't support Slovakia while neighboring Austria or Germany are no problem.
To get crowdfunding Indiegogo page advised incorporating US company.
IRS and Uncle Sam
So I used Stripe Atlas that's like 500 bucks. 6 months passed and never received EIN number. In the end, I got a tip on the atlas forum and got EIN after an expensive 2h wait on the IRS phone line.
But at least I finally got EIN to be finally able to ask for the only bank willing to help startups without visiting the US. I.e SiliconValey Bank. In March I payed my first Delaware franchising tax another 450 bucks. What's worse. Now I have like 4 days to fill the most complex tax form on the planet 1120 plus a bunch of other forms as a foreign owner. Fine for not filling 5472 is 25000$ by IRS and fine for not filling BE-13 is up to 45000$. O boy. Meantime I am paying 100$ year agent that did given me neither EIN or IRS PIN. Nir did it helped me fix incorrect Delaware stock record being 100m instead of 10M
I definitely have no money to pay for US tax advice 1200 a month a pop even though the company has zero assets. So I filled 7004 form to delay filing to october and hopefully I will somehow cobble that form together.
But guess what.
SVB too?
To finish all the great news. I received a message from SVB that my US company bank account was denied because … Wait for it … Because I am from Slovakia. Great. What is this company and all of the financial fines risk now good for without a bank account and ability to get funding?. Oh boy, what I got myself into. My last hope is Mercury bank fingers crossed hopefully they will not dismiss me with the same … Because I am from Slovakia and not Germany or Austria.
Still, If I can get my hands on some filters and mainly people willing to help with prototype development or testing or give advice.
Feeling that this test can in theory maybe save lives and I am doing nothing and days just fly is killing me…
Oh god, what I would give for a good experienced mentor or advice. Sorry for so much rant and incoherent rambling.
;D
Protocols.io
Test reproduction steps along with 3d printable files
