Hacker News new | past | comments | ask | show | jobs | submit login
Tesla Ventilator (youtube.com)
162 points by ericzawo 11 months ago | hide | past | favorite | 135 comments

Yes, this is an impressive amount of work done in a few weeks by a small team at Tesla. However, medical devices, especially for life support systems, are not a Hackathon project.

These devices breathe air into patients lungs. They need to be manufactured in clean room conditions, with medical grade parts, not plumbing parts from a big box store. They must be designed and tested to ISO specifications. If this fails, the patient can die.

I can't imagine any hospital or doctor careless enough to connect one of these to a patient, even in this current emergency.

The smarter idea is to ramp up existing proven designs, not reinvent the wheel and basically do a publicity stunt about your engineering talent.

> However, medical devices, especially for life support systems, are not a Hackathon project.

This effort they're doing really doesn't seem like they're treating it like a "Hackathon" project. They're putting real engineering resources behind it.

Whether or not they can produce something in time is still up in the air, but it seems to me Musk is attacking the problem from several different angles and hoping something works: he's purchased off the shelf BIPAP equipment and donated it (which turns out to not be as useful as hoped), he's got engineers homegrowing what appears to be a legit ventilator that may or may not get to volume production and/or certification in time, and he's helping established ventilator manufacturers with some of their manufacturing bottlenecks.

All three of these approaches are unlikely to be fruitful, but at a time when we're desperate and looking for quick innovation, it seems strange to dismiss people with the resources to try.

> This effort they're doing really doesn't seem like they're treating it like a "Hackathon" project. They're putting real engineering resources behind it.

Engineers can and do hackathon work, but real engineering work is way more about verification and validation than hacking stuff together. Products need to ensure that they will work reliably within all operational limits. You don't ensure that with an all nighter and a 3D printer.

That's one of the main reason why software development is not recognized as an engineering field. Verification and validation are the bread and butter, because if that culture doesn't exist and if these efforts don't happen then people die.

> You don't ensure that with an all nighter and a 3D printer.

Does the video linked look like the result of "an all nighter and a 3D printer" to you?

> Does the video linked look like the result of "an all nighter and a 3D printer" to you?

As you might understand from the language used and your failure to see a 3D printer, my description is what you may describe as a hyperbole.

But still, to put it in terms you might better understand, why do you think that medical devices take years to design and develop and get to the market? There are more things to it other than the old tinfoil stories.

This sounds a little like an argument that it's simply impossible to get this done any faster, so we shouldn't try.

I mean, I like to think we have clear enough standards for reliability and testing that if we manage to perform some excellent engineering and meet all of those standards in record time, and the federal government is willing to expedite the approval red tape, there's no technical reason we shouldn't be able to get it done a lot faster than what is typical without sacrificing safety.

Or put another way: I really hope the safety of our existing medical devices isn't simply based on the process being slow.

>But still, to put it in terms you might better understand, why do you think that medical devices take years to design and develop and get to the market?

Because of insane regulatory hurdles, and a refusal to adopt standards accepted by other nations' comparable regulatory agencies? See, for example, KN95 masks. Or for a non-coronavirus example, look at how slow the FDA has been to approve SLIT immunotherapy methods for allergy treatment, despite European agencies having approved them for decades. Which results in Americans seeking allergy immunotherapy to either get off-label prescriptions for SLIT therapy, or they have to get the much more expensive injections that require multiple visits a week to a doctor's office and supervision by trained (and therefore expensive) professionals.

> See, for example, KN95 masks.

IIRC, the issue with KN95 respirators is not the standard itself, but a massive problem with not-up-to-spec counterfeits being offered on the market.

Yea, it does, right down to the graph visualization that doesn't fill the screen.

Seeing as the engineers all touched their faces multiple times during the video (fidgeting with their poorly-fitted masks), I think it's safe to assume they know little to nothing about medical care, and this ventilator looks cool but will have no real world usefulness.

With hospitals in first world countries to larger or smaller extent already using 3D-printers to strengthen faltering supply lines I think the distinction is not so clear. Let a thousand flowers bloom and the doctors decide what they want to use. They can make the best tradeoffs.

And you need a rigorous process to insure they can be actually manufactured to specifications at scale. Which is why medical devices require so many layers of approval.

> If this fails, the patient can die.

If these patients receive no help, they die anyway. Given the choice between almost certain death or some experimental piece of hardware - give me the experimental hardware.

> I can't imagine any hospital or doctor careless enough to

> connect one of these to a patient, even in this current

> emergency.

I'm sure they will go through accelerated testing.

> The smarter idea is to ramp up existing proven designs,

> not reinvent the wheel and basically do a publicity stunt

> about your engineering talent.

Retooling has an enormous cost too (time and money) - yes the design will be proven but you won't nearly have the tools or expertise to implement them.

I'm not a Tesla fan-boy by a far shot, but I don't think they deserve this hate for what is clearly a good effort on their behalf.

The US already has a model for this though: During WW2, Ford Motor Company didn't stroke Henry Ford's ego by designing and building their own bombers -- they license built B-24s.

Yes, retooling costs. But tooling costs as well, and costs more besides.

With Elon's PR reach, I would expect that he could get at least one vent manufacturer to license a design to him for a dollar/handshake/whatever.

Elon's companies are also in parallel working with an existing manufacturer (Medtronic) to produce vents of their design (or more specifically to produce a part that is a bottleneck in their production). Putting a small percentage of resources to trying out a second model (this one) is just good risk management.

Tooling does not cost when they already have it, that is why they are focusing on ways to repurpose parts they are already producing as parts of the ventilator.

Tesla has partnered with Medtronic to increase their production capacity (https://www.massdevice.com/tesla-to-reopen-new-york-facility...). So Tesla isn't approaching this as a hackathon project, they've partnered with one of the biggest medical manufacturers who no doubt are offering their resources and expertise in this field, something Tesla obviously has none of which shows that they're trying to prove their serious about helping out

It took two years for Ford to start producing B-24s, leading credence to the time-cost of retooling.

Reference: https://en.wikipedia.org/wiki/Willow_Run

Because you said Ford, he was an anti-Semite, and his company was also part of the Nazi war machine.



There are two major differences, though. Cars are much more similar to airplanes than they are to ventilators, and the lead time for the war effort is much longer. We missed the window for rushing experimental ventilators into production. Looking at projections[1], there's about a 5-10 day window to get these into the hands of doctors. China donating 1000 ventilators to NYC is a much bigger deal.

1: https://www.bloomberg.com/graphics/2020-coronavirus-where-is...

Cars are about as similar to airplanes as airplanes are to washing machines.

Airplanes are exponentially more complicated than cars.

So discouraging to see such low quality and cynical comments at the top level here.

It's been made very clear over the weeks here on HN how Tesla and SpaceX between them, with their overlapping engineering resources (Elon, for example) have deep experience with life support systems, at a level approved by NASA even. To diss their work as a hackathon project is pretty low.

It's also been widely discussed that this crisis requires us to try some new things, and think outside of the box.

Elon donates massive quantities of masks and various types of the specific breathing devices asked for by medical facilities, and then takes it to the next level, putting his company reputation on the line, and you call it... a publicity stunt.

But when Jack Ma donated masks with a lot of fanfare, you said...

>Kudos to Jack. ... this shows selfless entrepreneurship.

Seems like a bit of a double standard.

>The smarter idea is to ramp up existing proven designs, not reinvent the wheel

Reinventing the wheel has a long history of success. But that's another story. Closer at hand, they already addressed this in the video, and why they did not take the path you thought would be smarter.

Specifically, just to reiterate what was made very clear in the first few seconds of the short video, it would impact supplies needed by existing manufacturers.

> various types of the specific breathing devices asked for by medical facilities

It's interesting that this is the spin you've decided to put on Elon donating the incorrect type of medical device to hospitals and then attempting to pass those off as ventilators. People have grown cynical because of Elon's own self-serving behavior and twitter publicity stunts evaporating any potential good faith meaning to his actions.

They specifically named those devices in their requests to him (C-PAP, BiPAP) so what you describe is not what happened at all.

The whole point of this project is not to use up supplies needed to build ventilator but to compliment the ventilators they are already building by using Tesla parts.

This is fundamentally a logistics project. Plus, they have access to Medtronic engineers, and as long as the FDA approve the devices, I have no problem with it.

> This is fundamentally a logistics project. Plus, they have access to Medtronic engineers, and as long as the FDA approve the devices, I have no problem with it.

Indeed, they just need to fly the devices, and parts for final assembly from the Medtronic factories in China.

In the current situation, it makes zero sense to spin new assembly lines when old ones are situated in the world's biggest manufacturing supercluster, and can be realistically scaled tenfold if they commit to that.

What I read going round Chinese BBSes now is that Medtronic's "factory" in Ireland is just doing a screwdriver assembly from Chinese parts, and all that just to do the infamous "Double Irish" manoeuvre

China has banned exports of medical supplies.

Edit: Trusting them to supply medical components seems like a huge risk at this point.

Yes, as of April 1. I wrote about it below.

It is not true. China is in post-pandemic time and they even have to sell the over manufactured products to other countries.

And many people are shipping and donating masks to their friends oversea.

http://weekly.caixin.com/2020-04-04/101538661.html http://international.caixin.com/2020-04-04/101538867.html

China suppressed the outbreak, but is still seeing double digit new cases per day. It could easily see a huge outbreak in a few weeks and based on past behavior should not be trusted to continue exporting medical supplies.

Urns in Wuhan Prompt New Questions of Virus’s Toll


But how long will an FDA approval take? Are they likely to take shortcuts because it's an emergency?

I mean personally I wouldn't object to them lowering requirements temporarily because we're in a crisis situation, but I hope they won't give permanent approval for all of the hacked ventilators that a lot of companies seem to be producing right now - once the pressure is off, stop using the temporary ones until they've been pulled through the FDA wringer thoroughly.

I knew someone at Hacker News will make a comment like this. Hacker News is cynicism at its finest.

"I can't imagine any hospital or doctor careless enough to connect one of these to a patient, even in this current emergency."

This is utter BS. Imagine that you are in a situation where you are trying to ration lives by selecting who gets access to a ventilator. Won't you not use this device then too?

Please, for god's sake, appreciate and be thankful that we have such great companies in US trying to help our people out. For perspective, if you were in a developing country, like India etc., all you have is hope and no company will be stepping up (they just don't have the technical wherewithalls).

Actually, we had a (world renowned) medical company where I live annouce an emergency ventilator design, complete with functioning prototype. The design was quickly torn apart in public by doctors working on actual Covid-19 patients (the caseload is still manageable here). The device was then shelved.

If a well-known medical company can produce something that doctors say is essentially worse than doing nothing, due to the lung physiology of Covid-19 patients, then you betcha Tesla can end up doing the same.

Like others have said: start from a design that works, and scale it up. This is not a time for NIH.

Well, except there's a big issue here, which is that Tesla can really only manufacture from parts that Tesla has, unless there are clear parts availability that can be quickly ramped up and delivered to the specific factories and in the form those factories can use.

Which is why the medical community basically shrugged at Tesla's PR pronouncement about being the white knight that will save everyone with their DIY ventilators built from car parts that might arrive sometime within 18 months if they stick to the normal Tesla production schedule.

And that was before Tesla mixed up BiPAP machines with actual ventilators...

Sure. They might pull it off though. It does seem consistent with their ethos to at least try.

Do you have a link for the story? It sounds like a good case study.

Yes - of course. Tesla and GM shouldn't even try to help out because obviously they will build the wrong thing.

What BS. Take your cynicism elsewhere. You literally are the living embodiment of my point above.

No, you misinterpret what I said. They should most definitely help out - by aiding in manufacturing a known working design at bigger scale.

They are doing exactly that already, partnered with an established company with proven design to help them ramp up production by manufacturing some hard-to-come-by parts.

And then Musk decided to burn some money on top of that, trying to find a way to produce even more machines from existing parts and/or parts that he can get more easily. And people give him shit for that? Seriously?

Has their effort to produce a ventilator like that a high chance of failure? Sure. But maybe their hail marry will succeed and save lives after all. And if they fail, at least they will have tried.

The only actual problem here would be if they produced a worse-than-useless machine that kills people but is put into use anyway. This hasn't happened yet, tho, but people right now are attacking this strawman viciously anyway.

DO you realize that these are car manufacturing companies trying to figure out how they can build ventilators. To build based on existing know-how will likely be incredibly expensive and prohibitive. As such, it is reasonably to expect them build these ventilators based on the realities of their current supply chain, technology, factory capabilities etc.

It is so surprising that I have to explain this to a tech audience.

How is starting over from scratch less expensive?

Given the time to explore cheaper parts, manufacturing processes, etc., starting over can always result in a cheaper device.

> can always result in a cheaper device.

  ^^^^^^^^^^ oxymoron alert

Agree with you, but just one small nit pick is that Mahindra (Indian car manufacturer) is now making ventilators too, and Anand Mahindra (CEO) has been exchanging tweets with Elon.

> If this fails, the patient can die.

In the current situation, I doubt anybody will give any regard to your sentiment. Though, I myself will.

> The smarter idea is to ramp up existing proven designs, not reinvent the wheel and basically do a publicity stunt about your engineering talent.

This is what China did. Ctrl-C, Ctrl-V, and knockoff ventilators began to fly off assembly lines in less than a month.

P.S. In other news, A VERY BIG THING that no news seem to be reporting. China has effectively banned nearly every medical export as of April 1 through enforcement of impossible to comply paperwork:


A man I know, who works at a medical clothing, and bedding factory, says that even such basic things that should not require any certification are now banned for export.

> In the current situation, I doubt anybody will give any regard to your sentiment. Though, I myself will.

I think the thought process for a lot of people will go more along the lines of "If you were given the choice of not having a ventilator (because they ran out) and dying, or using an unproven Tesla ventilator and maybe surviving, what would you do?".

That's assuming that Tesla (or other companies working on this) actually ends up manufacturing and supplying these things, and this is not just a publicity stunt.

I'm given to understand that Chinese traditional medicine is effective as well, and with their reported success in controlling the epidemic, we should look into those techniques as well.

How much can we count you in for?

I'm not so sure about your conclusion of banning exports.

Your link at SCMP says "companies must demonstrate to the Chinese customs agency that their products meet the relevant standards of the countries to which there goods are being exported."

Isn't this also a good idea to ensure quality?

> Your link at SCMP says "companies must demonstrate to the Chinese customs agency that their products meet the relevant standards of the countries to which there goods are being exported."

> Isn't this also a good idea to ensure quality?

May be, but the motivation for such regulation coming out of nowhere, and subsequent closing of reasonable exceptions, is not good at all.

It effectively reduced the number of eligible exporters from thousands to single digits per province, mostly completely irrelevant to fighting the epidemic.

This was regulation was made with 100% deliberate intent to be impossible to comply in order to shut down exports, without overtly announcing that they do that.

As of now, pretty much every medical export out of China is being smuggled, and bribed through the customs.

>but the motivation for such regulation coming out of nowhere,

I would guess (I have no info and I am just a random dude) that is because the bad press caused by reports of bad tests imported from China.

It's not coming out of nowhere. We've already seen hundreds of thousands of defective face masks, and hundreds of thousands of under-performing tests, being shipped: https://www.euronews.com/2020/03/29/netherlands-recalls-hund...

The motivation appears to come from recent news articles about how Chinese tests don’t work, or their respirators are faulty.

It turned out it was because non-certified Chinese companies were exporting these products, so the government cracked down.

From the article:

> of the 102 Chinese companies with CE accreditation, just 21 were licensed to sell their products within China

It's not clear to me what they're trying to do here. Why did they make this new law that prevents 80% of the accredited companies from exporting medical supplies?

> It's not clear

I think it's 100% clear.

> Why did they make this new law that prevents 80% of the accredited companies from exporting medical supplies?

Your questions already has an answer: that was made with that exact intent.

As I wrote above, even things that should not require any medical certification as such, like medical clothing and bedding, are now being turned around at customs. And those few random companies having full certifications, are turned around at customs too!

Sorry, I should have been more clear and concise. Why do they want to limit exports from certified companies?

> Why do they want to limit exports from certified companies?

Put some thought to it. It's China, the government here doesn't need to resort to any legal tricks to crack down on anything, especially now.

The sentiment that exports are being banned for fear of faulty tests being sent to export customers? One must be joking. That's, again, China, a country not having any problem with that.

On other hand, exploiting this convenient pretext to deny the West medical supplies, under a guise of an iron wall adherence to the process is unassailable from the West, when the West always poked China for its inability to control exports.

On other hand, the Chinese red cross (which is a 100% state ran organisation here,) has no problem buying, and exporting exactly the same medical supplies, and equipment.

At the beginning of this, Musk said that he didn't feel like new companies (including his) could get production going in meaningful quantities by the time that they would be required. I see nothing from Tesla that indicates that has changed, and I think that largely agrees with the point that you are making.

Your last point is important, I think. Notably, Tesla & SpaceX have been working with Medtronic to provide solenoids that are apparently a bottleneck in Medtronic's production process. It wouldn't surprise me to hear that others are also working on supplies for existing production lines to help with more rapidly ramping production. I think this is more likely to be useful.

How much of all of that clean room conditions is truly necessary? I ask not to be combative or to just support Tesla, but I know the US has a habit of only accepting the absolute best medical equipment, whereas hospitals make do with good but not great equipment and are able to save lives and save money. There is a risk, but there is also the issue of peopling dying because of no ventilator. So what's better? Having no ventilator, or having an one that gets the job done but isn't made with "medical grade" equipment.

If Tesla were to just build existing ventilator designs, like Medtronic's Open Ventilator, Tesla wouldn't be getting any criticism from anyone. That would be a truly useful thing to do in this situation. For example, nobody is criticizing Tesla's partnership with Medtronic to build the solenoids that are supply-constraining Medtronic's Western production of ventilators.

But Tesla wasting time and money on a new ventilator that won't go into production during this crisis is not praise-worthy, especially considering that what they have still needs weeks of work to make it usable in a clinical setting even if they wanted to start production immediately. It's just like the submarine they built for the Thai cave crisis. Lots of overthinking and self-praise going into something that ultimately couldn't be used by the people it was intended for.

If a doctor believes it is 99% likely that a patient will die without a ventilator and none are available, then why wouldn't they try something like the one from Tesla?

A choice between almost certain death and possible life isn't really a choice.

For starters, if they had to make that choice they'd just use one of the many CPAP or BiPAP machines already available in the hospitals. Even though this would render the machine single-patient only (since they generally can't be sanitized to acceptable levels for re-use without being partially dissembled) we don't have a shortage of those devices.

Welcome to Pascal's Wager.

This is overly dismissive and clearly just a “negative for the sake of being negative” comment. I’m not a Tesla fan, but these arguments don’t really hold water and are very insufficient to just dismiss the whole thing out of hand.

Without these sorts of things, patients are dying.

This is not business as usual. We are already using various hacks to overcome the shortages, any improvement is worthwhile in the very short term.

Sorry to be the bad new bear, but most people die even with them. Go look at the respirator death rates.

As an anaesthesiologist, I’m fully aware of the published death rates. You will see that they vary from 50 to 95%. The former appears achievable.

Three weeks ago, "hackers" here were posting their ventilator "designs" based on Arduino ducktaped to a CPAP machine and getting hundreds of upvotes. Now Tesla cooperating with actual medical equipment manufacturers is not legit enough?

When you have nothing, this is something. Stick to software.

> The smarter idea is to ramp up existing proven designs, not reinvent the wheel and basically do a publicity stunt about your engineering talent.

Those are not mutually exclusive. Yes, it would be great if we could produce more ventilators, but Tesla has no impact on "real" ventilator production either way.

This is obviously a PR piece. I see nothing wrong with a company trying to improve their brand's perception by trying to save lives.

This is an entirely self-serving PR move. Anyone at Tesla knows that getting government approval on any sort of ventilator they construct will require at least 6 months, probably 12, of rigorous testing. This isn't doing anything to truly solve this crisis.

Partnering or giving resources to companies that already have approved ventilator designs would by far be more helpful.

You are basically saying "do no harm"

If it becomes a battlefield medicine situation, what should happen?

I remember reading that during WW2 when there was no penicillin, nurses would inject milk I believe into patients, hoping it would stimulate an immune response and save the patient. Seems kind of risky, what are the ethics?

What should we do if we run out of ventilators?

This is not work from scratch.

Tesla partnered with Medtronic https://www.medtronic.com/us-en/index.html It seems that they also work as parts supplier for Medtronic.

>I can't imagine any hospital or doctor careless enough to connect one of these to a patient, even in this current emergency.

Well, you can choose death, I'll choose life.

An important constrain they mention is that they are intentionally avoiding current parts to not block supply chains of existing ventilator builders

I agree. But one comment worth considering: does it take longer to retool an assembly line to make a new verified part, or to come up with a new design based on existing verified parts?

My money is on the former and they'd be better retooling some of the line _if_ there was an off-the-shelf set of designs to work from.

That said, there's room for all sorts of approaches. I just don't see this one reasonably panning out or being used.

Personally if I were gasping for air I'd take my chances with something built from a garden hose, duct tape, and parts from Home Depot.

For more background info on what is needed for a ventilator and why it isn't that simple: https://news.ycombinator.com/item?id=22779665

The Tesla prototype seems much closer to a medical ventilator than most of the overly simplified models shown in that video. Working Medtronic is likely a large advantage here.

Adjusting air temperature and humidity seems like the biggest missing piece.

The schematic on the whiteboard had a humidifier component, fwiw.

This is a very good point. I incorporated the fact that they have access to Medtronic engineers into my post, or at least they have Elon Musk who can talk to their engineers.

Indeed, all what I saw in last few weeks on other side of the Pacific were not advancing past what Chines factories without any prior health device engineering experience were able to come with in the first month.

To add to that, most of those factories don't even have real engineers with 6 years long education.

I think this is the best proof to the immense potency of Ctrl-C & Ctrl-V approach to doing things. America is loosing a lot by shying away from it.

That video's author says he worked for Medtronic. Well... coincidental or not: Medtronic bought up Covidien, who have been alleged to intentionally sabotage a contract with the US government for a national ventilator stockpile back in 2012: https://www.nytimes.com/2020/03/29/business/coronavirus-us-v...

This is not Tesla, but Seat has already done it : https://www.seat.com/company/news/company/from-making-cars-t...

Less techie, more down to earth, they used wipe motors.

That's great. I hope wiper motors can take that kind of continuous actuation -- it's not really what they're designed for. I'm sure they'll stress test the hell out of it.

Driving for hours in heavy rain results in pretty continuous actuation of wiper motors. I'm far from a mechanic but I can't say I've ever heard of a wiper motor failing either. I'm sure they have real data on it instead of this handwaving, but as an outsider I would be surprised if it was an issue.

Not a comment on the ventilator made from a wiper motor, but I have had one fail, and had to have it replaced for something like $255.

All mechanical parts will fail (the antikythera device isn't still operational, but that's a pretty long-lived device!), it's just a matter of when and why.

That's actually really impressive. I didn't realize any of these efforts were that far along. It will be interesting to see how long certification takes.

It was certified 3 days ago and today they have started deliveries to hospitals: https://www.abc.es/motor/reportajes/abci-correos-comienza-re...

Thank you for sharing, super cool stuff.

Tesla didn't do this from scratch as people here seem to assume.

Tesla partnered with Medtronic https://www.medtronic.com/us-en/index.html It seems that they also work as parts supplier for Medtronic.

Are ventilators even helping? I see numbers like a 30% survival rate in some places, but they don't mention if that means "aren't dead yet" or recovered and removed from the ventilator. I definitely read about people who have successfully been removed and are recovering, but that's all anecdotal.

The information we need is:

1. Percentage of patients who needed ventilation, got ventilation in good time, and recovered vs died.

2. Percentage of patients who needed ventilation, didn't get ventilation due to lack of resources, recovered vs died.

Then we'll know how much ventilators are helping.

I understand that prior to this pandemic, typical survival rates for patients who required invasive ventilation was in the region of 30%. 70% die even with ventilation.

I've also heard that of Covid-19 patients that require ventilation, 80% die even with ventilation. So it has a slightly lower survival rate than other illnesses.

If you need ventilation and don't get it the mortality is really close to 100%.

The 30% survival rate means that 30% of patients are saved by ventilators. People who aren't dying without a ventilator aren't put on a ventilator, which is an extremely invasive medical procedure.

I'll be interested to see how well they can scale this up. Production functions are always dependent on the most scare component. Hopefully Tesla can leverage their already existing supplier base to easily source everything needed for this design.

I'm interested in this too. In the video, they said they're attempting to make this using as many car parts as possible, to prevent themselves from taking supply away from other ventilator manufacturers. That's a noble goal, as well as it helps them leverage their existing supply chain.

I'm curious what the mixing chamber is used for in an electric vehicle. Is it part of the climate control system?

It said “Continental” as the manufacturer on it. If I had to guess, I would say it was the reservoir for their adjustable height air suspension.

Ah yes, that makes a lot of sense. It looks quite beefy for climate control.

This is pretty excellent. It is one of the first ventilator solutions that actually implements appropriate air sensors, a PEEP function and air flow monitoring. Hats off to the tesla engineers for throwing this together so quickly. It's pretty neat that so many car parts could be reused. I can imagine they are finalizing the software interface and optimizing the machining process to reduce the cut time on the custom manifold production. Again, awesome job guys!

Are there certification requirements for newly designed ventilators? If so, how long would that take?

Years, decades? It really depends on a lot. But on average, 12 years. FDA can work fast, and likely will here. But even for something 'simple' it can take a while. Medical devices are really complicated and take a lot of testing. It's not like NASA or something where you're only blowing up John Glenn. We're dealing with a million dead babies/grandmothers/new brides/etc, potentially.


This might be one of those scenerios similar to what's currently happening with PPE - where we just have to use uncertified equipment because we have nothing that is certified left.

Like many others have pointed out, if the option is this uncertified machine or certain death - which would you choose?

Actual proper testing and certification would likely take years - and we need them within the next few weeks.

Who pays the liability when the ventilator backs up and blows the patient's lung apart?

If you're the patient, and the doctor says "I'm sorry but we have no more ventilators. You can elect to choose this experimental device and sign away liability as a last resort. Otherwise you are going to almost certainly die."

How would you respond? If anything I'd be appalled if I was the patient, and was basically just left to die because of all the certification red tape.

>If anything I'd be appalled if I was the patient, and was basically just left to die because of all the certification red tape.

AFAIK that's what already happens now with highly experimental drugs. You can't just give terminal patients whatever you want with the excuse "they were going to die anyways".

I’m sure Trump’s FDA would grant an exemption.

I'm sure any FDA would grant an exemption.

This is not how saving lives work

The main criticism I see here is that without prolonged testing it's very possible for a piece of equipment like this to do harm to a person.

On the other side I see people saying that for those that can't get a ventilator and are going to die anyway, why not use it.

I'm curious if that category of patients exists though. With Covid are doctors able to make that call at any point? I've heard lot of Hail Mary treatments criticized because unlike in the movies there are many ailments where there's not really a clear point where someone's not coming back.

Anyone know if it's a common scenario with this where a doctor knows someone is def going to die but there is still time left to do meaningful treatment?

The FDA would take years to approve a new ventilator that was an old one with a new paint color. A new design like this would probably be approved for use in US hospitals around 2025 if they're lucky.

Might be useful sooner in a country with a more functional administrative state.

There are many reasons for long testing on medical supplies. While there might be some laxing of them for a "hail mary pass" in the situation we are currently in, we shouldn't claim that equipment shouldn't have validation.

Not sure what to think of this.

Any people working with real medical ventilators care to comment on this?

Free PR nothing else (and nothing wrong with that). Those engineers didn't call up the marketing/pr dept and ask for a professional video crew to show up - "cause we just want to share an update with the world".

It’s more than that. Tesla engineers themselves probably are itching to help. Check out GE’s workers getting mad that the company wasn’t helping to make ventilators. People join companies like Tesla and SpaceX because they want to be challenged, have a bias for action (of course, everything still needs to be qualified), and make the world a better place. In a global pandemic that can be helped by addressing a shortage of manufactured stuff, engineers aren’t going to want to just sit on their hands if there’s a chance their efforts could be helpful.

It’s not just corporate PR. It’s not just Elon Musk. It’s also these workers wanting to help the world.

That's pretty much what Trump and his team say too. It's all PR to me, until they are standing next to their product, in use, saving lives.

That would be an unwise decision. Doctors can be replaced with other doctors, but if an engineer working on the ventilator gets sick, and possibly infects the whole team, they are hard to replace, as other engineers don't have the experience and expertise that they got in the last couple of weeks (and maybe years).

Not a doctor, but I did a crash course in ventilators recently and it has the main features - inspiration/expiration pressure, volume, rate, oxygen %.

UK gov also put out a document with the MVP for a ventilator, and the Tesla one looks like it will cover most of the therapeutic features.

The one thing I would worry about is software bugs in such a rushed project.

Tesla (and SpaceX) has experience with life critical code (running vehicle powertrains independent of the rest of the vehicle subsystems; please save Autopilot critiques for another thread).

A fully validated ventilator is ideal, but if your other option is death, a ventilator that is "less than fully validated" might be preferable.

EDIT: You're right to acknowledge there might be bugs, it's important to quantify risk. You decision based on the risk(s) raised.

This is good, and I'm glad they're doing it.

They've clearly got involvement with existing ventilator manufacturers.

Are they talking to doctors in ITUs? Here's a recent Facebook thread from people involved in London's new hospital: https://www.facebook.com/permalink.php?id=103930280957826&st...

It shows that ventilation of covid-19 patients is not like ventilation of other patients.

I'll paste part of it here because I know some can't visit Facebook.

It's by Daniel Martin OBE, Macintosh Professor of Anaesthesia, Intensive Care Lead for High Consequence Infectious Diseases.

---everything below this point is a direct quote from Daniel Martin---


- Early high PEEP is probably not the right strategy and may be harmful. This is not ARDS in the early phase of the illness.

- Avoid spontaneous ventilation early in ICU admission as also may be harmful.

- There is clear microvascular thrombosis happening in the pulmonary circulation, which leads to an increased dead space.

- Also some evidence of early pulmonary fibrosis reported from Italy, possibly oxygen related, possibly inflammation related.

- Not many patients have reached extubation yet in London, re-intubation seems to be common. I highlighted our experiences of airway swelling / stridor / reintubation.

- Brompton are seeing wedge infarcts in the lungs on imaging, along with pulmonary thrombosis without DVT.

- Proning is essential and should be done early. Don’t just do it once. Threshold for many centres is a PF ratio of 13, but all agreed, do it even earlier.

- Early on in the disease, the benefit of proning lasts < 4 hours when turned back to supine, as the disease progresses into a more ARDS type picture, the effect is more long lasting.

- Many centres using inhaled nitric oxide and prostacyclin with good effect. Tachyphylaxis with NO after 4-5 days.

- Generally people are using humidified circuits with HMEs.

- A very interesting thing they are doing at Georges is cohorting by phase of disease i.e. early, late, extubation / trachy. It involves more moving of patients but helps each team to focus on things more easily.

- Leak test before extubation is crucial, others are also seeing airway swelling.

- Wait longer than usual before extubating, high reintubation rates reported. Do not extubatne if inflam markers still high.

My conclusions from this are:

- Less aggressive PEEP strategy at the beginning of the disease and go straight for proning.

- Thromboembolic disease is prevalent, look for it. No one is sure about whether we should anti-coagulate everyone, this is probably too risky.

- An extubation protocol is needed immediately.

- We should consider using inhaled prostacyclin again (like we previously did) as it seems to be working early in the disease.

Extremely curious the outcome... I hope they can get some folks in the medical field to provide quickly some feedback.

Tesla is working with Medtronic (per Medtronic's CEO on CNBC Mar 25th) to rapidly iterate on their designs.

Is this like the kid-size submarine or does it have a chance of being mass-produced?

Definitely a PR stunt like the "submarine". See this comment: https://news.ycombinator.com/item?id=22794211

I just took possession of a Model 3 about 2 weeks ago. Feel like I underpaid for it when seeing how some of the residuals are being used now.

I just love the environment in the video! It looks like they're all in college experimenting! Loving the rowdy shirts and leather jackets!

This looks promising. They're being smart to use their own supply chain as much as possible, but obviously things will get interesting when it comes to getting clearance to deploy the device. Even if the design can't be used in it's entirety, they may be able to fill part of the supply chain for other manufacturers.

One complaint - everyone keeps touching their damned masks. Keep your booger picker away from your face.

It's because their masks keep dropping when speaking. It's surprisingly hard to fit the mask in a way that you do not feel the urge to adjust it every now and then.

Hopefully they get these right instead of ordering C-PAP machines [1]

[1] https://www.forbes.com/sites/billroberson/2020/04/04/health-...

I see this everywhere. I also saw where they literally listed what machines they needed, and Tesla brought the exact machines they wanted, including B-PAP and C-PAP machines.

Some hospitals converted those C-PAP machines into ventilators using these modifications: https://engineering.berkeley.edu/news/2020/04/turning-sleep-...

Applications are open for YC Summer 2021

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact