Home / Space-class CPU: How do you send more computing power into space?

Space-class CPU: How do you send more computing power into space?

  Mars is beckoning.

Mars is beckoning.


Phobos-Grunt, perhaps the most ambitious deep space mission ever attempted by Russia, crashed into the sea in early 2012. The spaceship would land on the battered Mars moon Phobos, collect ground samples and bring them back to Earth . Instead, it landed helplessly in the Low Earth Orbit (LEO) for a few weeks as its on-board computer crashed just before it could fire the engines to send the spaceship en route to Mars.

In the subsequent report, Russian authorities blamed heavy charged particles in galactic cosmic rays that hit SRAM chips and led to a lock-in, a chip failure caused by excessive current passing through. To handle this fix, two processors working on the Phobos-Grunt TsVM22 computer initiated a reboot. After restarting the probe then went to a safe position and waited for instructions from the ground control. Unfortunately, these instructions never came up.

Antennas intended for communication would be fully operational at the cruise stage in Phobos-Grunt, after the spaceship left LEO. But no one planned that a failure would prevent the probe from reaching this stage. After the particle fence, Phobos-Grunt ended up in a strange deadlock. Firing of engines on board would trigger the distribution of antennas. At the same time, engines could only be fired with a command from ground control. However, this command could not come through because antennas were not deployed. In this way, a computer failure killed a mission that was several decades in the making. This was partly due to some monitoring by the team at NPO Lavochkin, a primary developer of the Phobos-Grunt probe. In development, in short, it was easier to count the things that worked on their computer than to count the things that didn't. However, every little mistake they made became a serious reminder that designing space-class computers is bloody hard. A mistake and billions of dollars fall into flames.

All involved had simply underestimated the challenge of performing computer operations in space.

Why so slow?

Curiosity, everyone's favorite Mars rover, works with two BAE RAD750 processors clocked up to 200MHz. It has 256 MB RAM and 2 GB SSD. As we approach 2020, the RAD750 stands as the current state-of-the-art single-core space quality processor. That's the best we can send with deep space missions today.

Compared to all the smartphones we carry in our pockets, the RAD750's performance is simply pathetic. The design is based on the PowerPC 750, a processor that IBM and Motorola introduced in late 1

997 to compete with Intel's Pentium II. This means that perhaps the most technically advanced space hardware up there is fully capable of running the original Starcraft (the one released in 1998, mind you) without hiccups, but something more computationally demanding would prove problematic. You may forget to play Crysis on Mars.

Still, the price tag of RAD750 is about $ 200k. Why not just throw an iPhone there and call it someday? In terms of performance, iPhones are whole generations before the RAD750 and cost only $ 1,000 apiece, which remains much less than $ 200,000. In retrospect, that's pretty much what the Phobos-Grunt team was trying to achieve. They tried to increase performance and lower costs, but they ended up in corners.

The SRAM chip in Phobos-Grunt, which was hit by a heavily charged particle, went by the name of WS512K32V20G24M. It was well known in the space industry because T.E. 2005 Page and J.M. Benedetto had tested these chips in a particle accelerator at Brookhaven National Laboratory to see how they perform when exposed to radiation. The researchers described chips as "extremely" vulnerable, and lock-ups with one incident also occurred at the minimal linear energy transfer found at Brookhaven. This was not a surprising result, remember, because WS512K32V20G24M chips have never been thought of or tested for space. They have been designed for aircraft, military-grade aircraft for that matter. But still they were easier to get and cheaper than real space memories, so the Russians involved with Phobos-Grunt went for them regardless.

"The discovery of the different types of radiation found in the space environment was among the most important turning points in the history of space electronics, along with an understanding of how this radiation affects electronics, and the evolution of curing and mitigation technologies," said Dr. Tyler Lovelly, a researcher at US Air Force Research Laboratory of this radiation are cosmic rays, solar particle events and belts of protons and electrons that circle at the edge of the Earth's magnetic field, known as the Van Allen belts, particles that hit the Earth's atmosphere are about 89% protons, 9% alpha particles, 1 % heavier nuclei and 1% solitary electrons They can reach energies up to 10 ^ 19 eV Using chips that are not qualified for space in a probe that intends to travel through deep space for several years asked for a disaster to happen. the work reported Krasnaya Zvezda a Russian military newspaper, at that time 62% of the microchips using s on Phobos-Grunt were not qualified for spaceflight. The probe design was driven by 62% of a "let's throw in an iPhone" mindset.

Radiation becomes a thing

Today, radiation is one of the most important factors that designers take into account when building space-class computers. But it has not always been so. The first computer reached space aboard a Gemini spaceship since the 1960s. The machine had to undergo more than a hundred different tests to obtain flight clearance. The engineers checked how it worked when exposed to vibration, vacuum, extreme temperatures and so on. However, none of these testicles covered radiation exposure. Still, the Gemini on board managed to work pretty well – no problem at all. This was the case because the Gemini computer on board was too large to fail. Literally . Its 19.5 kB peeking memory was housed in a 700-cubic box weighing 26 kilos. The whole computer weighed 58.98 pounds.

 First orbital meeting: Gemini VI holds the station after using the built-in computer to operate to a position near Gemini VII. "src =" https://cdn.arstechnica.net/wp-content/uploads/2019/11/p11.jpg "width =" 387 "height =" 386

First orbit: Gemini VI holds station after using its on-board computer to operate to position near Gemini VII.


In general, to calculate processor technology has always been done primarily by reducing the functional sizes and increasing the clock frequency. We just made transistors smaller and smaller moved from 240nm, 65nm, 14nm, to as low as the 7nm design we have in modern smartphones. The smaller the transistor, the lower the voltage required to turn it on and off. That is why older processors with larger functional sizes were mostly not affected by radiation – or, not affected by so-called single event upsets (SEU), to be specific. The voltage created by particle scattering was too low to really affect the operation of sufficiently large computers. However, when space-turned people moved down by size of function to pack more transistors on a chip, the particle-generated voltages became more than enough to cause problems.

Another thing engineers and developers usually do to improve CPUs is to clock them higher. The Intel 386SX that ran the so-called "glass cockpit" in space buses was clocked at about 20MHz. Modern processors can run as high as 5 GHz in short bursts. A clock rate determines how many processing cycles a processor can go through in a given time. The problem with radiation is that a particle malfunction can destroy data stored in a CPU memory (such as L1 or L2 cache) only for an extremely short moment called a lock window. This means that every second there is a limited number of opportunities for a charged particle to damage. In low-clocked processors such as 386SX this number was relatively low. However, as clock speeds increased, the number of lock windows per second also increased, making processors more vulnerable to radiation. Therefore, radiation-cured processors are almost always clocked far lower than their commercial counterparts. The main reason that space CPUs are developing at such a slow rate is that virtually every conceivable way to make them faster also makes them more fragile.

Fortunately, there are ways around this issue.

Source link