When I post something like this, there are several predictable responses:
1) Some guys will proclaim it to be "AI slop", and will criticize me for even posting it.
2) Some guys will just declare that I am "getting old" (as if they are not) and will assume that at my age, I just somehow can't understand these new machines (as if they can!).
3) Many more sensible guys will put aside their feelings about AI and my age, and -- perhaps giving the devil his due -- be amazed by the complexity of these machines.
As I watched this video, I had many thoughts. Here are some of them:
1) The complexity of the ASML machine is a reminder of how far IC technology has moved from the world of homebrew ham radio. I know others will take different views, but for me, it is a reminder that these chips -- which are now commonly used in cell phones, for example -- have no real place in rigs that we would call homebrew.
2) The technology they use is really amazing. Laser lights hitting tin droplets three times as the tin falls. Explosions inside the machine that are like miniature super novae.
3) Older guys play a big role in all of this.
4) All the attention today is focused on AI and data centers. But the tech there seems simplistic when compared to the ASML machines that make the chips.
Please watch the Veritasium video that appears above. Then please put you comments in the space below.
FCC Amateur Radio Exam Question: Draw the schematic for the new IBM sub-1nm chip, showing how it will be used in FT-8 operations.
Mike WN2A suggested I ask AI about a schematic. Here is what Gemini said:
Here is what Gemini said:
Please draw me the schematic (showing the connections of all 100 billion transistors) in the new IBM sub-nm chip.
I would love to help sketch out circuit diagrams, but drawing a literal schematic showing the individual connections for all 100 billion transistors is physically impossible!
To put that number into perspective, if we were to draw a schematic where each transistor was just a tiny 1-inch symbol on a piece of paper, the blueprint would be so massive it would easily cover the entire landmass of the United States.
Modern sub-nanometer chips—like IBM's cutting-edge nanosheet architectures—are far too complex for human hands to map out line-by-line. Instead, engineers use advanced software to design chips hierarchically. Here is how that massive sea of 100 billion transistors is actually organized under the hood:
1. The Building Block: The Gate-All-Around (GAA) Nanosheet
At the microscopic level, these chips no longer use the older FinFET (fin-shaped) transistors. Instead, they use nanosheets, where the channels are stacked vertically like a three-story sandwich. The gate wraps entirely around all four sides of each horizontal sheet, providing maximum electrical control and drastically reducing current leakage.
2. Standard Cells (The Functional Logic)
Instead of wiring 100 billion separate components, computer scientists group transistors into "Standard Cells." These are pre-designed functional blocks like:
Logic Gates: AND, OR, NAND, and NOR gates (requiring anywhere from 2 to 8 transistors each).
Flip-Flops: Memory cells that store a single bit of data (requiring about 20 to 30 transistors).
3. The Macroscopic Layout (Blocks and Cores)
Billions of those standard cells are then combined into macroscopic structural blocks that look less like a traditional schematic and more like a highly organized city map. If we were to draw a high-level schematic of the chip, it would look like a block diagram featuring:
Processor Cores: The massive processing engines packed with arithmetic logic units (ALUs).
Cache Memory (SRAM): Vast grids of 6-transistor memory cells stacked tightly together to hold data right next to the execution units.
Interconnect Matrix: A vertical 3D skyscraper of up to 15 layers of microscopic copper wiring running above the transistors to route data and power across the chip.
If you are working on a homebrew electronics project and want to map out a circuit diagram for an RF amplifier, a QRP radio transceiver, or a simple analog filter, I can absolutely help you draw or troubleshoot those schematics! But for 100 billion nanosheets, we have to leave the drafting to the silicon foundries' supercomputers.
Asianometry does a pretty good job here. I am sure people will quibble about some of his descriptions of radio, AF, and RF. But overall, pretty good. I did see what looked like a direct conversion receiver in his block diagrams, but it was described as a superhet.
I was also struck by what looked like the lack of shielding in the modern cell phones. Some hams seem to believe that we cannot live without shielding. But apparently we can.
I also like how the bloc diagrams have the LO feeding mixers both on the receiver side and on the transmit side. I realize that you can do this by switching the LO, but I like to feed both mixers at the same time.
Asianometry talks about the need for inductors in the chips, and how this was often a show-stopper. This made me think about the Gilbert Cell that we find in our beloved NE-602s. Wasn't Barry Gilbert trying to create a balanced mixer that would not need coils?
Finally, there is a nice review of the history of Silicon Labs. I was hoping to see the Si570 and, of course, the Si5351. But no, these seem to have come later than the slide used by Asianometry.
Like some of the commenters, I had been scratching my head wondering why someone would go to all the trouble of counterfeiting electronic parts. It just didn't seem to make economic sense. Perhaps they were just marketing factory rejects? But no, IMSAI Guy shows in the video above that it is a lot more insidious. It turns out that this kind of crime DOES make economic sense (see comments). So be careful out there.
Here is another one IMSAI Guy did yesterday, looking at other faked chips:
Sunburst and Luminary by Don Eyles has a lot of the kind of color that helps the reader understand what was going on technologically during the 1960s. For example, there is this poem about integrated circuits (you don't get to use "poem" and "integrated circuits" in the same sentence very often):
The transistor's a marvelous invention
Replaced the tube convention
Found its niche
To amplify or switch
Whatever the designer's intention.
But the breakthrough was the IC
Integrated monolithically
It became pivotal
As computers went digital
With increasing complexity.
Eyles tells us that this poem was written by hardware designer Jayne Partridge, and appears in Eldon Hall's write-up of the Apollo Guidance Computer and the decision to use ICs in it:
Marc nicely sums up this project with this line: "It's an IC but in discrete form!" On the same theme, he later says, "Who needs a logic analyzer when you can do a visual debug with neon bulbs?"
Very cool. Lots of troubleshooting and repair lessons in this video:
-- Again we see the benefits of paper manuals. (Todd K7TFC commented astutely on this under yesterday's post.)
-- 2N2222s to the rescue.
-- A surprisingly large number of bad transistors (6?) found. Why did they go bad?
-- Marc repeatedly says, "Let me poke around." Poking around is often important. Mark fixes the reset line after poking around. He is not sure HOW he has fixed it, but he has... by poking around. Sometimes this happens. Thank God for small favors.
-- Marc has some fancy HP board extenders. I am jealous.
-- He also has a cool de-soldering tool. More jealousy. Want one.
-- Marc's understanding of how the HP engineers had to put one of the flip-flops "on the edge of stability," and how his 'scope probe was capable of disturbing this stability.
-- Remember that those Nixies are TUBES with enough voltage on them to really zap you. So be careful in there. This is an especially dangerous mix of transistor tech and tube tech. With transistors you can work on them with the rig fired up. With tubes, well, you have to be careful.
I really liked this repair video from CuriousMarc (aka AJ6JV). This counter pre-dates the use of integrated circuits -- it is all discrete transistors. Near the end, Marc mentions how this made this repair "like debugging a big integrated circuit, but with access to each transistor -- this made it quite satisfying." I hear ya Marc -- with big ICs maybe all you would get to do would be to swap out a single IC. There would not be much of a challenge there.
With the older, discrete circuitry you get a good view of how Marc troubleshoots -- how he finds the precise points where the device is failing. Note his use of the old HP paper manuals. I know this is an old guy thing, but I think the paper manuals (as opposed to the online versions) just make the process easier. Note too that Marc at one point had to go back to microfiche.
The transistor tester Mark used was very cool.
The whole physical structure of the HP device is very similar to my NYC HP8640B. Thanks again Steve Silverman and Dave Bamford.
I will look at Parts II and III of this series soon.
I recently commented that I was building a discrete LM386 for a Direct Conversion receiver. I think I was exaggerating my project. In fact I just built a very ordinary push-pull amplifier using a 2N3904 and a 2N3906 as a complementary pair AF amplifier. This is part of an LM386, but there is much more inside that little chip. Dave went much further. His schematic is above.
Please find enclosed the LT spice circuit for the small audio amp that I developed. I have included numerous notes on the schematic as to component function, suggested values to tweak, etc.
My intention was to come up with a relatively low parts-count design while adhering to the classic three-stage topology [diff pair / voltage amp / voltage follower] that has been used as the basis for so many audio power amplifiers for decades. I'm sure some of the parts could be eliminated at the risk of possible stability issues, but a lot of that also depends on proper layout, length of speaker leads, speaker load impedance, etc.
In any case I thought it would be great if someone wanted to build up this design to see if it works in real life or whether any serious mods are needed to get it to behave (I have no illusion that the spice models are entirely accurate, nor is my analysis thorough).
It should be capable of at least 500mW into an 8 ohm load at 9VDC supply, and over a watt at 12VDC. It should handle a 4 ohm load although at reduced voltage swing on the output, and with increased output transistor heat dissipation requirements.
The output stage is a complimentary compound ("sziklai") pair which should, in theory, maximize voltage swing from a low supply voltage (as compared to the more traditional complimentary Darlington configuration.
The bias current is set by a fixed pair of resistors and could be quite different from the simulated value, so I would be careful on initial power-up to monitor the quiescent current draw. I would guess it need not be more than a few mA or so, and definitely less than 40mA or so. This can be made adjustable if desired.
Regards,
Dave
-----------------
So, does anyone out there want to give this a try? You could be helping to save generations of homebrewers from the indignity of using an integrated circuit!
Hack-A-Day has an interesting post about an FM Broadcast radio project. I took a look. The Github page has the schematics for the hardware. For me, the thing is, there is just not a lot there. It is a bunch of chips. The FM Tuner IC is the heart of the project. With that one you have to dig down to see that it is a digital processing chip:
All of the action -- all of the magic of radio -- is locked inside those little SMD chips. I suppose if you were skilled enough to write the software or to significantly modify it, you'd get closer to the experience of homebrew radio, but very few of us have those kinds of skills -- we just download the software, then struggle to get it into the chips.
And sure, you could struggle to solder those chips to a PC board, but really, why bother when 99% of the components are already inside the chips? You should just buy a board with the chips on them and with the software already loaded. There you have it: the store-bought appliance is really, really close to the supposedly homebrew receiver.
But hey, to each his own. This is a hobby and it is all for fun. I just think I have more fun with old-style, analog, discrete component HDRs. YMMV.
This is really beautiful. Radraksha Vegad (Pargrahi) from India built a discrete component version of the venerable 555 timer chip. He built it on wooden blocks. This leads to the kind of understanding that even Jean Shepherd would have admired. No longer is the 555 a little mysterious black box. No, Pargrahi shows us how it works.
I know we could do something similar with the NE602 or the LM386. But probably not with an Arduino microcontroller or an Si5351. And that says something about understanding and complexity.
Sam has some very impressive gear -- I found myself wondering if that could really be a home lab. But it is. It is apparently in his garage in New Jersey. Hack-A-Day reports that Sam got a lot of the gear on e-bay. And they note that living near Princeton University (great dumpsters!) also helped. Here is some background on Sam's workshop:
I picked this generator at the Kempton Park rally in London many years ago. I use it quite a bit, not only for circuit testing but also as an easy way to get my homebrew BITX rigs to send a signal so that I can adjust my antenna. I just plug this thing into the mic jack, crank in a small amount of 1 kHz audio, and I am ready to minimize SWR.
It never gave me any real problems until last week. I opened it up and examined the circuitry for the first time. Lots of mystery chips in there. Fortunately they are all socketed. Thank you Maplin.
Even without really knowing how each of the ICs work, it was easy to troubleshoot. See the schematic in the article that begins on page 21 in this .pdf:
My Rigol scope showed a good signal going into and coming out of IC8, the LF351 op-amp. From there the signal goes to IC5 a 4066BE. But nothing was coming out of IC5. Thus IC5 was my lead suspect. I put in a quick order to Digikey. A couple of days and a few bucks later the chips (as always, I ordered several) were on my doorstep. Within a few minutes IC5 was replaced and the generator was as good as new.
One annoying problem with this generator is that it has no automatic shut-off circuitry. So if I forget to turn it off, I soon have to buy two 9V batteries. Tony G4WIF suggested a simple timer circuit. I may do that soon. Or, given that we are already on the IC dark side with this project, there is this: https://www.youtube.com/watch?v=C4C2H_3BU3o&feature=youtu.be
I recently find myself replacing a lot of chips in various rigs and devices in my shack. It seems that chips go bad more frequently than discrete transistors or even tubes. But maybe I'm just using older gear with older, more fragile chips. What do you folks think?
Wow, lots of wisdom in this video from Leo Fernekes. Great hints and kinks on prototyping with copper clad boards. I need a Dremel just so that I can make Leo's board cutter. And I can see that I need some of that liquid flux and isopropyl alcohol. Surf boards and headers! Who knew? Teflon coating for the wires -- gotta get it. Glad to see that Leo is also a fan of copper tape. His emphasis on the importance of stage-by-stage construction and testing is right on the mark. My only disagreement with Leo is about his use of steel wool. I've found that steel wool will inevitably cause little tiny "Murphy Whiskers" to float around your workbench. They will eventually settle onto the most inconvenient and damaging place on your board. So I have banished steel wool from my workshop. Those green, non-metallic Scotch Brite pads work just as well and don't cause shorts. Three cheers for Leo. He is based in Thailand. He has an interesting background and some really amazing projects and insights: http://www.luminati.aero/leofernekes
Over the years I have made my distaste for integrated circuits very clear. I've presented them as mysterious little black boxes that don't really belong in our homebrew rigs, rigs that we are building, after all, because we want to avoid the use of MYSTERIOUS BLACK BOXES.
So, I don't use voltage regulator chips -- I use Zener diodes. I prefer analog LC oscillators to AD9850s or Si5351s. And I have repeatedly built discrete component audio amplifiers when most normal people just put an IC AF amplifier in the circuit. When I built the Q-31 Shortwave AM receiver, I kind of ran out of gas at the end. I wanted to get the receiver going and I didn't want to build yet another discrete AF amp. So I used an LM386. I rationalized this deviation from cherished values by noting that the discrete AF amplifier circuitry that I was using was remarkably similar to what exists inside the LM386. So that little chip is just as understandable as my discrete component creations. It wasn't REALLY a mysterious black box... Around this time I found a web site that made me feel mush better about all this. It explains very nicely how the little LM386 does such a great job. It really packs a lot of amplification into a very small package. Here is the web site: https://www.electrosmash.com/lm386-analysis
Shortly after my transgression, Jenny List over at Hack-A-Day posted a nice piece looking at the inner workings of Op-Amps. Who knows, I may seen be using 741 chips too! Here is the Hack-A-Day piece: https://hackaday.com/2020/05/09/an-op-amp-from-the-ground-up/
Seventeen year-old Sam Zeelof, KD2ENL, is making his own integrated circuits in his garage. Wow. This makes me think about another seventeen year-old -- the fellow who appears on pages 63-64 of Cliff DeSoto's "200 Meters and Down." (I have the story on page 81 of "SolderSmoke -- Global Adventures in Wireless Electronics.") In the early days of radio that kid amazed us by making his own vacuum tubes. Sam Zeelof is clearly following in that tradition. No "mysterious black boxes" for Sam! No "appliance chips" for him! FB OM. This is really amazing. Here are the links: https://spectrum.ieee.org/semiconductors/devices/the-high-school-student-whos-building-his-own-integrated-circuits
Thanks to Armand WA1UQO for alerting me to this. I really liked the book -- "Crystal Fire" -- that this 1999 video is loosely based on. I'm also a fan of the narrator, Ira Flatow, whose melodious voice is heard each week on NPR's excellent "Science Friday" radio show. A few observations and thoughts on the video: -- I liked the irreverant Calypso song "Hell's Bells Laboratory." It looks like those folks had a lot of fun. And wow, Shockly's secretary was named Betty Sparks. TRGHS. -- I have the same big Variac on my bench. And I have one of those "third hand" devices. -- I'd like to build my own replica of the point contact device with the triangular piece of lucite and the gold foil. -- While Shockley seems to be the real bad guy in this story (he seems to have all the bad characteristics of David Sarnoff, Lee DeForest, and Steve Jobs), I liked the his use of "physical intuition" to understand devices and the problems they were meant to solve. -- The image of the two Japanese founders of Sony working in the late 1940's in a bombed out department store was very powerful. -- Although I came on the scence a bit later, I WAS one of those kids who used a transistor radio and an earphone to surreptitiously listen to rock-and-roll music. -- "More transistors are made each year than raindrops fall on California." Hmmm.... More info here: http://www.pbs.org/transistor/ Extra interviews: http://www.pbs.org/transistor/tv/index.html
I'm not crazy about chips, but I've come to like the LM386. It is not really a little black mystery box -- as the article points out, the internal circuitry is simple and completely understandable. So you shouldn't feel any appliance-op angst when you use one of these.
