jmwilson

There are metal replacement toners that replace the silver in a print with either gold or platinum. This way the print can be made normally with an enlarger and silver chemistry to overcome the contact printing limitation. Toning improves the longevity of the print and provides many of the same tonal effects, depending on how long the print sits in the bath. Unfortunately, the chemicals involved are quite toxic. I remember one of the instructors at the Rayko darkroom in San Francisco running a "toning day" and it was very busy with a number of people coming in to tone their work and share the considerable material and hazmat fees for the chemicals.

"unlike silver gelatin prints that hold the chemicals in a layer on top of the paper, platinum photography uses chemicals that are soaked into the paper, so the image lies embedded in its fibers."

Fiber paper for silver prints is also a thing, and easily available from commercial sources. It requires more extensive washing compared to RC (resin coated) paper, and it tends to curl and warp after wetting. To get prints flat after drying, you need a big heated press that looks like a giant panini maker.

I do mention the NanoVNA as a measurement option that has better accuracy since it can do phase measurement. It's an amazing tool, though its lower limit for frequency is 50 kHz and I'm not sure how its dynamic range holds up there. Ideally you want the option to go very low for this measurement. Like most VNAs sold today, the NanoVNA is really designed for looking at RF and microwave circuits.

I only know two commercial VNAs in-production that are made specifically for this kind of work: the Keysight E5061B (5 Hz - 3 GHz, $50000+ when optioned out appropriately) and the Omicron Bode 100 (1 Hz - 50 MHz, I think $5k-10k).

I have data collected for 3x MLCC 10 uF + 1x 100 nF but need to make the charts and update the article. Short answer is it's basically just as good and has fewer resonance peaks. I'd also like to do the test with some tantalum 10 uFs to throw in some ESR and see if that tames the impedance spikes even more.

I wrote an article last month about testing the received wisdom about decoupling capacitors: https://jmw.name/projects/exploring-pdns/

A lot of design guidelines and advice boil down to cargo-cult rules like "sprinkle 100 nF caps everywhere," and many people don't have don't have the tools to measure PDNs and decide if that rule is actually good enough.

I made a custom PCB to emulate different configurations of capacitors of various types and layouts. With a spectrum analyzer and tracking generator, it's possible to measure and visualize several points that are important: what does "physically close" mean, the effect of DC bias on capacitors, and how does parallelizing capacitors together affect the circuit. The author of this is using an oscilloscope to look at time domain behavior, but generally I think of these things in the frequency domain.

Having first been exposed to the DFT in a very similar matter to this text, I think there is a significant advantage in first deriving it from the discrete time Fourier transform (DTFT, not to be confused with DFT, and notwithstanding the name is actually a continuous function) instead of trying to introduce it independently on its own. This is the approach taken in Oppenheim and Schafer textbook, but requires the reader have a bit more background in mathematics.

In this way, the DFT is seen as a sampling of the DTFT when the signal is convolved with a window function, and explains why the spectrum of a signal is smeared when its period is not a multiple of the transform size. This textbook says "there is no leakage when the signal being analyzed is truly periodic and we can choose N to be exactly a period, or some multiple of a period" -- actually there still is, it's just that the DFT happens to sample precisely at the nulls of those sidelobes. The sidelobes are further seen as a consequence of the window function, and explains why certain window choices have better sidelobe attenuation at the tradeoff of wider main lobe/lower frequency resolution.

With shifts and masks you know where the bits are. With bitfields, you don't because the specification leaves everything up to the compiler.

  struct foo {
    char a : 4;
    char b : 4;
  };

Is a in the high-order 4 bits, or the lower 4 bits? Both choices are allowed, so it's up to the compiler and makes the code non-portable.

Link to the cited paper, instead of a blurb written about the paper: https://onlinelibrary.wiley.com/doi/10.1002/spe.2874 https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/spe.28... (PDF)

Yep, the closest I ever got to a SGI was drooling over their product brochures as a kid. The cost of a modest Indy was about the same as a mid-range car. It's hard to grasp as a modern PC user that these workstations could handle classes of problems that contemporary PCs could not, no matter what upgrades you did. Today, it would be like comparing a PC to a TPU-based (or similar ASIC) platform for computing.

From what I've read, Oxide is making racks of servers and has no interest in workstations that an individual would use.

"Much simpler than webauthn" Is it? Trying out https://webauthn.io/, I can log in using a security key without hitting the command line and copy & pasting some base64 string.

"Private keys never leave end users' devices" How does it guarantee that, considering it saves the private key in the file system? It can be trivially copied off device from there.

The examples also appear to encourage key reuse, and since the challenge is the current time, the signed auth message can be replayed against any other site where the user set up using the same key.

The author says neither is right, but clearly prefers lon, lat ordering.

"Geographical tradition favors lat, lon. Math and software prefer lon, lat."

But why does math and software prefer lon, lat? Unlike endianness, where it appears little-endian is the better choice at the hardware level, I'm drawing a blank why one ordering is better for computation.

(2020)

The halcyon days when one could just order a few STM32 chips...

Title should be clarified to match the article. Not everyone associated with the cult has perished. Several ex-members are alive and some remained close to the cult in a way that's hard to describe, and took on the maintenance of the website.

HBO's series, Heaven's Gate: The Cult of Cults, is highly recommended. It is a fascinating, tragic, and ultimately humanizing look at the cult, through the eyes of observers and several ex-members.

It didn't used to be this way. Tape was once affordable for individuals; ca. 2000 Travan was marketed to consumers and DDS wasn't too much of a stretch for prosumers. But they're all gone now and only LTO is still standing. Given the small, cornered market segment that tape now serves, there is no incentive to bring the price down. I'd like the option of being able to manage my own backups locally instead of using cloud storage, but the economics of Glacier remain too good until the storage size is several times what I could foresee needing.

As someone who once believed this, this statement is simply not true. Experimentation in ham radio is largely about the physical layer, with maybe some layer 2-4 work when it comes to building low-scale communications infrastructure. Encryption is not implemented at this level, and the limitations on encrypted content do not impede research into new digital modes or actual radio science.

I think it is worth asking why modern digital radio standards (e.g. P25, DMR, LTE, 5G) were largely developed outside the ham community. My survey level understanding is they came from the communication needs of their users and are not a natural evolution of something that hams would build, with or without encryption. When I was researching DMR, there's a lot of stuff that is unnecessarily clunky when considering the use case of a ham QSO. Conversely, the business users of DMR want more structure than what's provided with amateur-designed digital modes. With 5G, I think there are some real technical and capital barriers, i.e., very few amateurs have access to the grade of test equipment needed to engineer and experiment with systems at those frequencies.

Most Nixie tube drivers can't hurt you, unless they are directly rectifying the mains. DC-DC designs that start with typical low voltage can give you a quick jolt but cannot sustain their rated voltage or current if you short them.

I designed my own Nixie tube driver last year, and documented it on Hackaday: https://hackaday.io/project/170962-nixie-tube-170-v-flyback-.... I think I had more fun building the power converter than with playing with Nixie tubes.

Unlike the NCH6300HV, it's an open source hardware design with complete design notes and files for understanding how a HV power converter works.

An earlier article has a bit more description of the complaint: https://archive.md/67DiB

For those curious about how this works, the company has to show it tried and couldn't hire a US resident, but they don't have to try very hard... so the company posts a brief job opening in a local newspaper. When nobody responds, because it's the newspaper, tada! Proof there was no American-based talent capable of filling the job.

> Unless they plan to release a billion of these, sweep them up, and then have interns/grad students probe each one to extract the collected data.

After reading the article and watching the video, I think that's what they are aiming for. The coil can couple data and power through near-field interaction, so they'll release a large number, hope to recover a sample of them, and extract the data through field or lab instruments. I could see this being useful for recording environmental data and forensic tracing, but they look closer to RFID tags than a computing platform.

Strong opsec that the supporting documents are actual photos of a computer screen from the visible moire (or somehow altered to look that way).

After numerous leaks, Facebook's internal security team became very good at identifying leakers. The person responsible for this 2016 post was identified within hours and terminated the next day: https://www.buzzfeednews.com/article/blakemontgomery/mark-zu.... The leaker was easily identified by the names of friends liking the post (that and part of their name was visible).

Facebook-issued laptops are filled with spyware, monitoring everything down to the system call level, and practically every access to internal systems is logged at a fine level. The only way to exfiltrate data with plausible deniability would be to photograph the screen with an individually owned device. The fact that you searched for the internal wiki page and viewed it are nothing, but that you shortly invoked the keyboard shortcut for a screen capture, then inserted a USB drive, and copied a file ("Screen shot ____.png" even!) to it (all logged) ... congratulations, you're caught.

It basically means order sufficient parts before starting layout or even before finalizing the schematic. It's not just microcontrollers that are scarce, it's also "dumb" ICs like power management, discrete logic, A/D interface, etc.

Somebody already made one - Hoptroff, a London-based watchmaker produced a pair of wristwatches based on the CSAC a few years ago. Don't bother searching for a price or where to buy; they've since pivoted from making watches to being a time-as-a-service software company.

IMO it was a bit of gimmick. Apart from being a little too chunky for a wristwatch, just because it's an "atomic clock" doesn't mean it's "atomic powered". IIRC the CSAC uses about 1/8 W which is really pushing it for a low-power device like a watch. It might lose "one second per millennium" as advertised but good luck keeping it continuously operating anywhere near a thousand years.

It's a good question and something I'm looking into as I consider the idea of selling them. It's something that has to be approached in the context of other changes (like BOM optimization) for larger manufacturing.

I've studied SMPS design and am not totally oblivious to conducted and radiated EMI issues. I took several considerations in the project like analyzing current loops and taking near-field measurements on a spectrum analyzer. However, what ultimately matters is what happens in the testing chamber.

I thought about that, but I wanted to ground the display in the natural cycle of the day. Once you introduce wall-clock time into the design, there are at least two problems that need solving:

1) what's the timezone? 2) what about DST?

The first is quite complex and requires knowing more than just lat/lon. Dates of DST also do change over time according to policy. The only solution I saw for these would to go full IoT so it could talk to NTP servers or have a control interface, and that would ruin the plug-it-in-and-it-just-works minimalism I was going for.

I actually used a USB connector on the previous Nixie-tube based version. It's harder since there aren't as many panel-mount options for the USB connector. You have to design the circuit board around abutting the enclosure wall. This makes sense in an injection molded or unibody construction, but less so with a wood box like this.

You're also potentially limited by the USB spec on current draw. There are ways around this, like ICs that will detect if you're connected to a charger or a data port. The barrel connector is also quite ubiquitous and I designed the power entry to work across a wide range of DC power adapter voltages.

It's not surprising at all. Anyone who worked at Google or Facebook during the decade or so ca. 2005-2015 is aware of the degree to which these companies presented themselves as a totally new and different workplace than the typical corporate employer. Facebook even had a little red handbook that stated on its cover, "Facebook was not originally created to be a company." With on-site banking, laundry, car-washes, dentistry, hobby/interest clubs, etc. they provide a mixed, almost collegiate environment that blended the workplace with a community. Many people who worked at these places fell in love with their jobs, the companies made it easy to. To some degree, anyone who was super cynical about this would select themselves out of the hiring pool.

The writer joined out of college, and perhaps was a previous intern. Again, anyone familiar with the internship program at big tech companies knows how they aim to provide not just a mentored professional experience, but a social one including events with other interns and employees.

> That's impossible to achieve across all latitudes and longitudes, and it's totally arbitrary and has zero benefit.

> Solar noon matters to precisely nobody. No one is looking at the sky to tell time.

That's why the OP said "about when the sun is highest" not "exactly when the sun is highest". And it matters if you like the idea of midday being at, well, the middle of the day. Deliberately adding another ~4% error on top of whatever is there because of arbitrary time zone boundaries, and things begin to add up and eventually you end up like Spain.

> People want sunlight after work. They don't want to be depressed because they finish the day and it's dark outside.

To state the obvious, changing the clocks doesn't change the amount of daylight available. As another commenter said, if we've learned anything from the pandemic, maybe it's that we can be flexible about work hours instead of warping our sense of time and the rhythm of the day because managers and office workers can't fathom the idea of seasonal work hours.

Right, but this situation also happens in traditional color printing as well. Ink can't be varied in intensity, so images are halftoned/dithered. Halftone screening is done at a different angle per channel to avoid moire, so there is a mix of true overprinting and places where dots of distinct color are next to each other.

> ACEP achieves a full color gamut, including all eight primary colors

Last I checked there are three primary colors, and what is a full color gamut? There are several standard choices of gamut.

What confuses me about color e-ink products is that it is a reflective display, yet none of them use a reflective color model (CMY) - all of them talk about RGB pigmented subpixels. Either what I learned about color reproduction for print media is wrong, or these devices are probably incapable of showing a large number of colors. Given the washed-out appearance of several photos of real products, I'm leaning toward the latter conclusion. I'd love to see magazine-quality color in a reflective display some day, but they don't seem to be starting from the right fundamentals. Has anyone benchmarked the color reproduction of these products?

ARM licenses and royalties are not free. The royalties are a small but significant cost of a chip (1-2%) and licenses on the IP are a large R&D cost for the designer.

China is obviously interested in RISC-V as their access to the US market is threatened, and they could be a driver of wider use.

> I don't quite follow. I think you just mean reactance at higher frequencies (since resistance is DC), and total impedance is then DCR + frequency-dependent reactance.

No, he's correct. Ferrites are a specifically-designed lossy inductor over a targeted frequency range. There is more than just DC resistance + a linearly-increasing reactance with frequency. Just like dielectrics exhibit loss and can be characterized by complex permittivity and a loss tangent, magnetic materials are also lossy and have a complex permeability. Here is a ferrite I used on the data lines for a GPS module:

https://www.murata.com/en-eu/products/productdetail?partno=B...

Look at the impedance chart at the end and you'll see the resistive component is basically zero at DC and starts to increase in the HF range, peaking close to 1 GHz. The reactance of it is also highly nonlinear.

> Hasn't that problem mostly been solved already with switch mode power supplies that are pretty small.

Perhaps you are thinking of "wall wart" power supplies that used line-frequency transformers in their rectifier design. These have been replaced and miniaturized by using offline switched-mode power supplies and point-of-load converters, but the magnetic components remain the largest components in these designs.

> With most personal electronics being battery powered these days I'm curious where you see a real need for further reduction in the size of power converters.

It's now common for modern SoC devices to require multiple voltage rails. It may require 3.3 V, 1.35 V, 1.8 V on separate pins and each requires a point-of-load converter for efficiency, and each of these take up board space.