Tesla’s Dojo Is An Interesting CPU Design

What do you get when you cross a modern super-scalar out-of-order CPU core with more traditional microcontroller aspects such as no virtual memory, no memory cache, and no DDR or PCIe controllers? You get the Tesla Dojo, which Chips and Cheese recently did a deep dive on.

It starts with a comparison to the IBM Cell processors. The Cell of the mid-2000s featured something called the SPE (Synergistic Processing Elements). They were smaller cores focused on vector processing or other specialized types of workloads. They didn’t access the main memory and had to be given tasks by the fully featured CPU. Dojo has 1.25MB of SRAM that it can use as working memory with five ports, but it has no cache or virtual memory. It uses DMA to get the information it needs via a mesh system. The front end pulls RISC-V-like (heavily MIPS-inspired) instructions into a small instruction cache and decodes eight instructions per cycle. Continue reading “Tesla’s Dojo Is An Interesting CPU Design”

Hackaday Links Column Banner

Hackaday Links: September 4, 2022

Say what you will about Tesla, but there’s little doubt that the electric vehicle maker inspires a certain degree of fanaticism in owners. We’re used to the ones who can’t stop going on about neck-snapping acceleration and a sci-fi interior. But the ones we didn’t see coming are those who feel their cars are so bad that they need to stage a hunger strike to get the attention of Tesla. The strike is being organized by a group of Tesla owners in Norway, who on their website enumerate a long list of grievances, including design defects, manufacturing issues, quality control problems, and customer service complaints. It’s not clear how many people are in the group, although we assume at least 18, as that’s the number of Tesla cars they used to spell out “HELP” in a parking lot. It’s also not clear how or even if the group is really off their feed, or if this is just a stunt to get the attention of Tesla honcho and notorious social media gadfly Elon Musk.

Continue reading “Hackaday Links: September 4, 2022”

Hackaday Links Column Banner

Hackaday Links: July 31, 2022

Don’t look up! As of the time of this writing, there’s a decent chance that a Chinese Long March 5B booster has already completed its uncontrolled return to Earth, hopefully safely. The reentry prediction was continually tweaked over the last week or so, until the consensus closed in on 30 Jul 2022 at 17:08 UTC, give or take an hour either way. That two-hour window makes for a LOT of uncertainty about where the 25-ton piece of space debris will end up. Given the last prediction by The Aerospace Corporation, the likely surface paths cover a lot of open ocean, with only parts of Mexico and South America potentially in the crosshairs, along with parts of Indonesia. It’s expected that most of the material in the massive booster will burn up in the atmosphere, but with the size of the thing, even 20% making it to the ground could be catastrophic, as it nearly was in 2020.

[Update: US Space Command confirms that the booster splashed down in the Indian Ocean region at 16:45 UTC. No word yet on how much debris survived, or if any populated areas were impacted.]

Good news, everyone — thanks to 3D printing, we now know the maximum height of a dive into water that the average human can perform without injury. And it’s surprisingly small — 8 meters for head first, 12 meters if you break the water with your hands first, and 15 meters feet first. Bear in mind this is for the average person; the record for surviving a foot-first dive is almost 60 meters, but that was by a trained diver. Researchers from Cornell came up with these numbers by printing models of human divers in various poses, fitting them with accelerometers, and comparing the readings they got with known figures for deceleration injuries. There was no mention of the maximum survivable belly flop, but based on first-hand anecdotal experience, we’d say it’s not much more than a meter.

Humans have done a lot of spacefaring in the last sixty years or so, but almost all of it has been either in low Earth orbit or as flybys of our neighbors in the Sol system. Sure we’ve landed plenty of probes, but mostly on the Moon, Mars, and a few lucky asteroids. And Venus, which is sometimes easy to forget. We were reminded of that fact by this cool video of the 1982 Soviet landing of Venera 14, one of only a few attempts to land on our so-called sister planet. The video shows the few photographs Venera 14 managed to take before being destroyed by the heat and pressure on Venus, but the real treat is the sound recording the probe managed to make. Venera 14 captured the sounds of its own operations on the Venusian surface, including what sounds like a pneumatic drill being used to sample the regolith. It also captured, as the narrator put it, “the gentle blow of the Venusian wind” — as gentle as ultra-dense carbon dioxide hot enough to melt lead can be, anyway.

Continue reading “Hackaday Links: July 31, 2022”

Hackaday Links Column Banner

Hackaday Links: June 12, 2022

“Don’t worry, that’ll buff right out.” Alarming news this week as the James Webb Space Telescope team announced that a meteoroid had hit the space observatory’s massive primary mirror. While far from unexpected, the strike on mirror segment C3 (the sixth mirror from the top going clockwise, roughly in the “south southeast” position) that occurred back in late May was larger than any of the simulations or test strikes performed on Earth prior to launch. It was also not part of any known meteoroid storm in the telescope’s orbit; if it had been, controllers would have been able to maneuver the spacecraft to protect the gold-plated beryllium segments. The rogue space rock apparently did enough damage to be noticeable in the data coming back from the telescope and to require adjustment to the position of the mirror segment. While it certainly won’t be the last time this happens, it would have been nice to see one picture from Webb before it started accumulating hits.

Continue reading “Hackaday Links: June 12, 2022”

Wireless Power: Here? Now?

Outside of very small applications, Nikola Tesla’s ideas about transmitting serious power without wires have not been very practical. Sure, we can draw microwatts from radio signals in the air and if you’re willing to get your phone in just the right spot you can charge it. But having power sent to your laptop anywhere in your home is still a pipe dream. Sending power from a generating station to a dozen homes without wire is even more fantastic. Or is it? [Paul Jaffe] of the Naval Research Laboratory thinks it isn’t fantastic at all and he explains why in a post on IEEE Spectrum.

Historically, there have been attempts to move lots of power around wirelessly. IN 1975, researchers sent power across a lab using microwaves at 50% efficiency. They were actually making the case for beaming energy down from solar power satellites. According to [Jaffe] the secret is to go beyond even microwaves. A 2019 demonstration by the Navy conveyed 400 watts over 300 meters using a laser. Using a tightly confined beam on a single coherent wavelength allows for very efficient photovoltaic cells that can far outstrip the kind we are used to that accept a mix of solar lighting.

Wait. The Navy. High-powered laser beams. Uh oh, right? According to [Jaffe], it is all a factor of how dense the energy in the beam is along with the actual wavelengths involved. The 400 watt beam, for example, was in a virtual enclosure that could sense any object approaching the main beam and cut power.

Keep in mind, 400 watts isn’t enough to power a hair dryer. Besides, point-to-point transmission with a laser is fine for sending power to a far-flung community, but not great for keeping your laptop charged no matter where you leave it.

Still, this sounds like exciting work and while it might not be Tesla’s exact vision, it sounds like laser transmission might be closer than it seemed just a few years ago. We’ve seen similar systems that employ safety sensors, but they are all relatively low power. We still want to know what’s going on in Milford, Texas, though.

EV Charging Connectors Come In Many Shapes And Sizes

Electric vehicles are now commonplace on our roads, and charging infrastructure is being built out across the world to serve them. It’s the electric equivalent of the gas station, and soon enough, they’re going to be everywhere.

However, it raises an interesting problem. Gas pumps simply pour a liquid into a hole, and have been largely standardized for quite some time. That’s not quite the case in the world of EV chargers, so let’s dive in and check out the current state of play.

AC, DC, Fast, or Slow?

Since becoming more mainstream over the past decade or so, EV technology has undergone rapid development. With most EVs still somewhat limited in range, automakers have developed ever-faster charging vehicles over the years to improve practicality. This has come through improvements to batteries, controller hardware, and software. Charging tech has evolved to the point where the latest EVs can now add hundreds of miles of range in under 20 minutes.

However, charging EVs at this pace requires huge amounts of power. Thus, automakers and industry groups have worked to develop new charging standards that can deliver high current to top vehicle batteries off as quickly as possible.

As a guide, a typical home outlet in the US can deliver 1.8 kW of power. It would take an excruciating 48 hours or more to charge a modern EV from a home socket like this.

In contrast, modern EV charge ports can carry anywhere from 2 kW up to 350 kW in some cases, and require highly specialized connectors to do so. Various standards have come about over the years as automakers look to pump more electricity into a vehicle at greater speed. Let’s take a look at the most common options out in the wild today. Continue reading “EV Charging Connectors Come In Many Shapes And Sizes”

The State Of Play In Solid State Batteries

Electric vehicles are slowly but surely snatching market share from their combustion-engined forbearers. However, range and charging speed remain major sticking points for customers, and are a prime selling point for any modern EV. Battery technology is front and center when it comes to improving these numbers.

Solid-state batteries could mark a step-change in performance in these areas, and the race to get them to market is starting to heat up. Let’s take a look at the current state of play.

Continue reading “The State Of Play In Solid State Batteries”