Connect with us

Technology

Atomos Space’s first orbital mission is a trial by fire

Published

on

Few missions more vividly embody the maxim “space is hard” than Atomos Space’s first demonstration mission, which the corporate managed to drag back from the brink of disaster – greater than once.

This demonstration mission, called Mission-1, was launched into orbit on a SpaceX Falcon 9 rocket on March 4. The mission’s goals are extremely ambitious: the 2 spacecraft – an orbital transfer vehicle called Quark-LITE and a goal vehicle called Gluon – will ultimately reveal extremely complex maneuvers, including rendezvous, docking, orbital transfer and in-orbit refueling.

The company faced two major problems related to communication and the rotation rate of the spacecraft – and (largely) solved each problems despite massive limitations, sparse data packets, and very limited bandwidth. (So ​​limited, in reality, that the team needed to limit flight software updates to a text string of just 145 characters).

Advertisement

“It was relentless,” Atomos CEO and co-founder Vanessa Clark told TechCrunch.

William Kowalski, COO and co-founder of the corporate, agreed. “What makes it so difficult is that even in our situation, we’re trying to extrapolate the status of a very complicated system from maybe 100 bytes of data,” he said. “That’s a lot. You guess what’s causing it, knowing that a few of those guesses could lead on you down a path from which you may never get well.”

Problems began just hours after the 2 interconnected spacecraft were launched from the Falcon 9 upper stage. The deployment was nominal, and Atomos received the first signal from the spacecraft seven minutes after deployment. The mood was solemn.

But 40 minutes passed before the corporate received one other signal. Then eight hours.

Advertisement

Atomos expected data packets every jiffy.

“The worst (day) was Monday, when we took off, that evening,” Kowalski said. “It was 11 p.m. at night, it was me and the chief engineer… and we didn’t hear anything and we just think: Have we failed? Did they die? We gave it a shot, but it just didn’t work. It was really a punch in the gut.”

Mission controllers didn’t discover the basis cause until 24 to 48 hours after deployment, and did so with the assistance of one other company with on-orbit assets. After pulling some strings, they managed to speak on the phone to the chief systems engineer of the satellite communications company Iridium. The spacecraft used third-party modems that used the Iridium intersatellite link network and likewise used the Iridium constellation as relay satellites. The Atomos spacecraft was moving too fast and in direct contrast, it couldn’t perform a data “handshake” with the Iridium satellites to truly transmit the knowledge back to Earth.

Atomos engineers implemented a series of software updates that reduced duty cycles and ensured that the radios would all the time be on, even when the spacecraft was in a low-power state.

Advertisement

When engineers tried to resolve the communication problem, nevertheless, they encountered one other problem: the spacecraft was rolling at an especially high rate of 55 degrees per second (they were designed to deal with roll rates of as much as 5 degrees per second). In addition, the spacecraft slowly rotated in order that the solar panels not faced the sun. This meant it was a race against time and the spacecraft’s batteries completely depleted.

“We had two charts,” Kowalski said. “We plotted our power trend for when we predict we will probably be facing away from the sun and have (at) zero power, in addition to the sink rate. The removal rate needed to be delivered to zero before the ability dropped to zero.

The problem was exacerbated by limited communication; teams weren’t in a position to definitively confirm that anything was mistaken until the fourth day after deployment, and the spacecraft could only process recent commands between long periods of what were essentially communications blackouts.

Slowly, over the course of several days, they managed to slow the spacecraft down. The team achieved one other major victory after they were able to ascertain high-bandwidth communications, a space-to-space link on a Quark-LITE device communicating via the Inmarsat network. On Thursday, the corporate made its first attempt at establishing broadband connectivity and successfully maintained communication with the spacecraft for six minutes.

Advertisement

During this era, mission controllers received 17 times more data than since launch. As a result, mission controllers received enormous amounts of information on the state of the spacecraft. The news wasn’t all positive – certainly one of the OTV batteries was badly damaged by aggressive cycling and it appears the GPS aboard certainly one of the spacecraft needed to be reset – but these are easy fixes, Clark said.

The company plans to begin commissioning the drive system on Tuesday or Wednesday. If all goes in response to plan and engineers determine that the support system provides accuracy and aiming control, they are going to test operation without torque bars and response wheels. The company intends to deploy the spacecraft inside about a month, with all mission objectives expected to be achieved by the top of June.

Kowalski and Clark attribute a part of the startup’s success to the incontrovertible fact that it is highly vertically integrated. The team, which worked 100 hours per week within the first week after deployment, was in a position to use their in-depth knowledge of spacecraft design to resolve emerging problems.

“It was obviously very painful, but it is reminiscent of the words of the CEO of Nvidia: ‘I wish you great suffering.’ We went through it and it wasn’t great at the moment, but now that we’ve gotten there, we’re definitely better,” Clark said.

Advertisement

This article was originally published on : techcrunch.com
Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Technology

The time of the American semiconductor market in 2025.

Published

on

By

semiconductor, chips, AI, Intel,

It was already a turbulent yr for the American semiconductor industry.

The semiconductor industry plays a major role in the “AI race”, which the US appears to be determined, so it’s value being attentive to this: from the appointment of Lip-Bu Tan-which has not waste time to work, attempting to revitalize the Heritage Company-Joe Biden proposing latest rules of exports AI AI along the way that would not or can follow.

Here’s what happened since the starting of the yr.

Advertisement

Power

Last reversal

May 7: Just every week before the “Frame of Artificial Intelligence Diffusion”. According to many media, including Axios AND BloombergThe administration won’t implement restrictions when it was to start out on May 15 and as a substitute works in its own framework.

April

Anthropic doubles the support of chip export restrictions

April 30: Anthropic has doubled because of the limitation of exports of chip systems in the USA, including several corrections to artificial intelligence framework, akin to imposing further restrictions in level 2 and dedication of resources to enforcement. The NVIDIA spokesman rejected, saying: “American companies should focus on innovations and get up to the challenge, instead of telling high stories that large, heavy and sensitive electronics are somehow smuggled in” Baby bugghs “or” next to the lobsters live “.

Planned exemptions at Intel

April 22: Before connecting profits with Q1, Intel said he was planning to release over 21,000 employees. The exemptions were to enhance management, something that the general director of Lip-Bu Tan has long said that Intel must do and help in the reconstruction of the company’s engineering.

The Trump administration further limits the chip export

April 15: The NVIDIA H20 AI chip was hit with the requirement of export license, the company revealed in the SEC application. The company added that it expects for $ 5.5 billion fees related to this latest requirement in the first quarter of the tax yr in 2026. H20 is the most advanced AI Nvidia chip can still export to China in some form or fashion. TSMC and Intel reported similar expenses in the same week.

Advertisement

TechCrunch event

Berkeley, California
|.
June 5

Book now

Advertisement

NVIDIA seems to talk of further export of chips

April 9: According to reports, the general director of Nvidia, Jensen Huang, was noticed for dinner at the Mar-A-Lago Center in Donald Trump. Then, NPR was reported Huang could have the ability to save lots of AI H20 NVIDIA systems from export restrictions after investing in AI data centers in the USA

Alleged agreement between Intel and TSMC

April 3: Intel and TSMC allegedly reached a preliminary agreement on the commencement of a joint project of Chips. This joint undertaking would work in Intel devices, and TSMC would have 20% shares in the latest undertaking. Both firms refused to comment or confirm. If this contract just isn’t accomplished, this might be a good preview of potential offers in this industry.

Intel rotates from non -corporate assets, proclaims a brand new initiative

April 1: CEO Lip-BU TAN immediately worked. Just just a few weeks after joining the Intel, the company announced that it might rotate resources unrelated to the core in order that it could focus. He also said that the company would introduce latest products, including non -standard semiconductors for purchasers.

March

Intel calls the latest CEO

March 12: Intel announced that a veteran of the industry and a former board member, Lip-Bu Tan will return to the company as general director on March 18. At the time of his appointment, Tan said that Intel could be a “engine -oriented company” under his leadership.

Advertisement

February

Intel’s Ohio Chip Plant is delayed again

February 28: This yr, Intel was to start out running its first chip factory in Ohio. Instead, the company slowed down the construction of the plant for the second time in February. Now the design of semiconductors value $ 28 billion won’t end with the construction until 2030 and might even open only in 2031.

Senators call for more chip export restrictions

February 3: US Senators, including Elizabeth Warren (D-Mass) and Josh Hawley (R-MO), wrote a letter to the Secretary for Trade for Howard Lutnicka Calling Trump’s administration for further restriction Export of the AI ​​system. A letter addressed especially to AI H20 NVIDIA systems, which were used during training of the R1 Deepseek “reasoning” model.

January

Deepseek releases its open model “reasoning”

January 27: The Chinese startup Ai Deepseek caused quite mixing in the Silicon Valley when he released the open version of his model “reasoning” R1. Although this just isn’t a special message of semiconductors, the alarm in the AI ​​industry and Deepseek semiconductors meant that it still has an impact on the chip industry.

Order Joe Biden on the export of chip

January 13: Only the incumbent week remained, former President Joe Biden proposed extensive export restrictions on AI systems made by the USA. The order has created a 3 -level structure that determined what number of American systems will be exported to every country. According to this proposal, level 1 countries didn’t face any restrictions; Countries 2 level 2 had a chip purchase limit for the first time; And level 3 countries received additional restrictions.

Advertisement

Dario Amodei from anthropics weighs the limitations of chip exports

January 6: Co -founder of anthropics and general director, Dario Amodei, co -author of opinions The Wall Street Journal Supporting existing control controls of the AI ​​system and indicating them as the reason why the Chinese market of artificial intelligence was in the US. “He also called on the incoming President Donald Trump to impose further restrictions and shutting the gaps that allowed AI to get these tokens in China.

(Tagstranslate) Intel

This article was originally published on : techcrunch.com
Continue Reading

Technology

One of the long -term VC Elona Muska suits his former employer after alleged dismissal

Published

on

By

Concept illustration depicting messy litigation with an illustrated gavel on a multicolored background

Josh Raffaella, who has deep roots as an investor of the Silicon Valley and was supported by many firms Elon Musk, suits his former employer, massive trillion dollars Aum Brookfield Asset Management, reports the New York Times.

A major part of Raffaella’s criticism concerns how Brookfield covered losses related to the pandemic of real estate and claims that the company released him after submitting the criticism of informants at SEC. His lawsuit gives allegations akin to fraud and bribe, while Brookfield deny all offenses rapidly, said The Times.

In February, Brookfield quietly closed the Venture Capital unit run by Raffaella and threw some assets on one other unit, Bloomberg reported at the moment. One of Raffaella’s complaints in the lawsuit is that Brookfield didn’t buy so many shares in firms belonging to musk because he provided the possibility of purchase.

Advertisement

Raffaella had shopping transactions in Musk, akin to SpaceX, XAI and a boring company, claims the claim. Bloomberg announced that his Brookfield fund was an awesome supporter of Twitter’s takeover by Musk.

The lawsuit is a really public battle of Raffaella, who previously worked as a partner at VC, known at the time as a drapeer Fisher Jurvetson. (Today it’s a set of funds.) In DFJ Brookfield, it has helped this company spend money on Musk, akin to Solarcity (acquired by Tesla), Spacex and Tesla.

(Tagstranslate) Brookfield

This article was originally published on : techcrunch.com
Advertisement
Continue Reading

Technology

The AI ​​X Dream Machine library project is intended to help in underestimated innovators

Published

on

By

Dream machine, Van Jones, Google, Ai

Through the project of the AI ​​X library, the CNN commentator, activist and entrepreneurs Van Jones and co -chairman of the projects Inayah Bashir and Steven Pargett be sure that the Black and Brown communities have free, practical opportunities to learn the force of artificial intelligence.


Libraries have long been secure for knowledge, access and community. Now they grow to be starts in innovation. Through the CNN commentator, the activist and entrepreneur Van Jones and co -chairing Project Inayah Bashir and Steven Pargett be sure that black and brown communities have free, practical possibilities of learning the strength of artificial intelligence.

Launched by the trio Dream Machine Innovation Lab in cooperation with Google, AI X Library Project already has an impact on cities corresponding to Brooklyn and Atlanta, with plans to extend on Miami, Detroit and Los Angeles.

Advertisement

“We don’t just want to consume technology – we want to create, shape and run it,” said Jones during an interview with Black company. „Myślę, że AI jest potencjalnie najbliższe odszkodowania, jakie kiedykolwiek dostaniemy, ponieważ po raz pierwszy od 400 lat mamy dosłownie równy dostęp i równe szanse na coś transformacyjnego. To, o czym polega nasz projekt, jest wrażliwa na możliwość i upewniając się, że jesteśmy w rozmowie o równym dostępie do dobrej rzeczy. Jeśli chodzi o edukację, jeśli chodzi o edukację, jeśli chodzi o edukację, jeśli chodzi o edukację, jeśli chodzi o edukację, jeśli It is about education when it comes to education when it comes to education when it comes to education.

Instead of focusing only on threats of artificial intelligence, corresponding to algorithmic prejudice, Jones, Pargett and Bashir, they challenge themselves to concentrate on the chances. “Often black people are very sensitive to threat and we quickly go into conversations about equal protection against bad things,” said Jones. “This is important. But our project is to be sensitive to possibilities-insulting that we are also in a conversation about equal access to good things.”

Libraries are a perfect entry point for such a democratized technological education. “The library has always been a technology center or anything else,” Bashir noted. “In the case of insufficient or underestimated people, the library is a place where they touch new technology for the first time. When the computers came out, when the internet began, people went to libraries to learn.”

During each AI X library workshop, participants can pick from topics corresponding to marketing with artificial intelligence, writing with AI-Holności, intelligent monitors and the long run of labor. Importantly, local librarians help select which topics might be taught on the idea of the needs of their patrons.

Advertisement

“We were looking for a way to transfer this conversation to real communities,” said Jones. “A lot of online and fancy conferences are happening, but not much at the neighborhood level. Most districts still have a library. This is a trusted place. We decided to go there.”

The project attracted participants on the age of seven and old as 70, creating an intergenerational learning environment. “We develop at least three new business ideas in every workshop,” said Bashir. “In Atlanta we had a mother who runs a collective at home. She said that they already use artificial intelligence to plan the curriculum – and even used artificial intelligence to find our event.”

Other participants are a waste management employee in Miami, bringing their two sons to study next to him, and grandparents examined AI tools for the primary time. “It is amazing that families learn together,” said Bashir. “Excience exists – creativity is there.”

Changing considering, not only skill sets

Advertisement

In contrast to the digital division in the 90s, where access to software and hardware created entry barriers, Jones believes that the most important obstacle with which our communities with which they stand today is the way in which of considering. “We have no problem with the equipment because everyone has a smartphone,” said Jones. “We have no problem with the software because it is so much free now. We have a problem with” wet software “. Our brains do not process artificial intelligence as something that is for black people. It must change. “

Instead of perceiving artificial intelligence as a perfect, almighty tool, Jones encourages people to give it some thought as “a bit stupid, but free, very fast interns”. He explained: “Nothing is perfect – you need to check human work and you need to check the work AI. But do you prefer to improve something that you will get in two weeks or improve something that you will get in two minutes?”

Marlon Avery repeated this sentiment, reminding the participants that everybody is involved with AI for the primary time at their very own pace. “It’s okay to hurry,” Avery said. “AI is like a treadmill – you can walk, jogging or sprint depending on what you are ready for. It is important to get on it.”

Building liberatorial innovations

Advertisement

Ultimately, the project of the AI ​​X library concerns something greater than just learning latest tools – it is concerning the idea of ​​a greater future.

“Liberatorial innovation is not only creating new technologies and new systems – it is about creating new freedoms,” explained Bashir. “Our brains, our communities, even our definitions of what it means to be human.

Jones emphasized the urgency of quite a lot of participation in technology. “When one small group tries to design a civilization for everyone, it doesn’t work well,” he said. “We now have a chance to build something more human, more simply – but only when everyone takes place at the table.”

When Dream Machine tries to scale the project of the AI ​​X library outside its pilot cities, one thing is clear: the long run of innovation is not only happening in the Silicon Valley – it is built in neighborly libraries, one workshop without delay.

Advertisement


This article was originally published on : www.blackenterprise.com
Continue Reading
Advertisement

OUR NEWSLETTER

Subscribe Us To Receive Our Latest News Directly In Your Inbox!

We don’t spam! Read our privacy policy for more info.

Trending