Printed electronics

Terraformer · 2018-09-23 15:30:47

We might as well push for complete self-sufficiency, and see what the limits are...

So, printed electronics, in particular microprocessors. They're not as powerful as the ones modern chip fabs produce, by far, but they should be able to provide more power than the Apollo program had available - and hopefully enough to run the desktop chip printer, closing the loop and enabling the first desktop self-replicating industrial ecosystem.

https://www.printedelectronicsworld.com … nology-now
https://spectrum.ieee.org/tech-talk/sem … ransistors

It seems that Zinc Oxide can be suspended in an ink and used to print transistors onto a plastic backing. Zinc is relatively abundant, and simple enough to refine that we've been doing it since at least the 12th century. I don't know about the chemistry involved in forming the ink, though.

Terraformer · 2018-09-23 15:33:22

Also - http://www.3ders.org/articles/20160408- … -inks.html

SpaceNut · 2018-09-23 15:42:18

Have worked in the electronic industry and metal on metal oxides can be used to make semiconductor parts but printing them will make the devices much larger than the standard way of doing them. We have been able to print the carbon resistors and make the electrical connections as well with the technology. Though these do not handle the same amounts of current.
The metal on metal oxides do create capacitors at material change over layers. Its still going to be a period of time before this will be a main stay to making electronics.

kbd512 · 2018-09-23 19:54:45

SpaceNut,

The thin film discrete component electronics in question may be larger and consume more power than microelectronics providing equivalent function, but if the resultant systems are so simple, modular, and resistant to radiation and temperature effects that their fabrication or repair is essentially a one-and-done proposition requiring a single tool (like a specialized 3D printer), then perhaps critical systems like environmental controls and communications equipment would benefit from that type of architecture because they're comparatively simplistic to troubleshoot and more robust.

Remember how the Star Trek computers were basically a bunch of plastic semiconducting ribbons or boards that could be replaced to restore function and there was a specialized core computer system deep within the engineering section of the ship? That's what I think we need. The core functionality can be provided by specialized (and expensive) radiation-hardened system-on-chip type microelectronics, whereas all the sensors and ancillary electronics assemblies (similar to the old avionics subsystem boxes in fighter jets that all handled discrete functions, like a box to control the chaff or flare dispensers, for example) should be easily field serviceable / repairable without the requirement for a full electronics repair shop or special cleanliness requirements.

SpaceNut · 2018-09-23 20:03:35

Metal on metal oxides are easily damaged by static discharge, they are not radiation hardened, are easily damaged by repetition flexing, and would mean going back to a bus card structure to a systems computer which is not a bad thing. Just not where we are with single board computers which would be turning the clock back almost 30 years to when they were.

louis · 2018-09-24 16:50:23

Could be v. useful particularly in the early stages of colonial development when labour is such short supply.

But I don't see why a Mars colony couldn't, with the help of 3D printing of the machinery, build a sophisticated microprocessor manufacturing facility on a scaled down version of this sort of factory:

https://www.youtube.com/watch?v=qm67wbB5GmI

Terraformer wrote:

We might as well push for complete self-sufficiency, and see what the limits are...
So, printed electronics, in particular microprocessors. They're not as powerful as the ones modern chip fabs produce, by far, but they should be able to provide more power than the Apollo program had available - and hopefully enough to run the desktop chip printer, closing the loop and enabling the first desktop self-replicating industrial ecosystem.
https://www.printedelectronicsworld.com … nology-now
https://spectrum.ieee.org/tech-talk/sem … ransistors
It seems that Zinc Oxide can be suspended in an ink and used to print transistors onto a plastic backing. Zinc is relatively abundant, and simple enough to refine that we've been doing it since at least the 12th century. I don't know about the chemistry involved in forming the ink, though.

Terraformer · 2020-05-09 12:26:45

A company promising semiconductor printing

Printing of micro and nanoscale electronics and sensors (down to 20nm).

Hmm. The missing piece in the self-replicating tech stack?

tahanson43206 · 2023-08-24 08:41:31

I would like to see this topic continued and developed.

The topic of electronic reliability is top of mind for me, as one of my ancient systems has started acting flaky, and after a relatively new power supply began causing random system shutdowns due to an unreliable supply of a particular voltage, on another system.

The folks who set up shop on Mars are going to need reliable electronics that lasts longer than the 7 years that seems to be an average for system boards on Earth in 2023.

We have a member (marc) who was working on a PhD thesis project to develop an ultra reliable computer based upon TTL chips. The system would have been slow compared to modern integrated circuits, but in the context of Mars, it would have been repairable at the chip level, and fabrication of such primitive chips would be easier to set up on Mars than a modern high density integrated circuit manufacturing facility.

(th)

Calliban · 2023-08-25 16:25:07

I hadn't heard of printed electronics before. It could potentially be a very useful replacement for the low end electronics that are coming out of China. Those electronics are very likely to become unavailable as China collapses, which is looking increasingly likely.

On Mars, the ability to print simple electronic devices would provide a huge advantage. Simple linear controllers would useful. Temperatures controllers, sunlight tracking systems, things like that.

tahanson43206 · 2023-08-25 17:13:23

For Terraformer ... re topic ...

If you have time, please see if there might be updates available ...

It seems to me this is an important topic ... it is (or appears to be) an early step toward the StarTrek replicator, which assembles atoms into desired configurations at breath taking speed. We humans can see atom placement working all the time, in every plant and tree, but the process is languid in comparison.

A 3D printer for Integrated Circuits would be an important step forward, and I would think there would be a market for chips made that way right now.

As a reminder, NewMars member marc is working on a TTL based minicomputer designed to operate with 36 bit words, in a ** very ** robust manner. We haven't heard from marc for a while, but I assume he's been working on the system all this time.

The project involved writing an assembler, as well as designing all the hardware.

(th)

tahanson43206 · 2023-08-28 06:45:33

For Terraformer re topic ...

Here is a report on advances in the art of 3D printing of integrated circuits...

It appears that current technology can print at micron scale. The goal appears to be to print at nano scale.

https://www.nano-di.com/resources/blog/ … the-future

The capability described in this article would be helpful for a small community on Mars, where independence from Earth as soon as possible would be advantageous.

(th)

kbd512 · 2023-08-28 16:05:52

I think the idea of a simplified computing system, lasting for decades because it's not based upon the sort of microcircuitry that the universe is so unkind to (galactic cosmic rays, solar radiation, static discharge destruction over time), is greatly under-appreciated. What I view as a "future proof" fault-tolerant computing system, though, will ultimately be based upon long-lasting materials (glass or diamond or ceramic metals) and the use of very specific wavelengths of light both as input and for processing, rather than very precise electrical control.

For example, Microsoft's light-based AIM computing technology doesn't store information, it continuously processes it and reacts according to changes in light patterns. It requires very little power relative to Silicon-based electronics, it's focused on processing rather than storing data, and the volume of throughput it can process is wildly in excess of what integrated circuits with the same size and input power are capable of, because multiple wavelengths of light are not limited to a binary state. AIM doesn't store data in registers, operating more like an information sieve intended perform a particular computational task to generate usable output. AIM is very much like combining the function of the sensor and firmware of an IR-guided missile. The missile's sensor input is used for pattern matching something that represents the shape of an enemy aircraft and wavelength range of a jet engine exhaust, which is normally very different from a flare. AIM is a far more sophisticated implementation that doesn't clock-driven switching to match sensor input with database values to produce the output using more clock cycles. The processor / sensor / database are one and the same thing, not subject to voltage regulation, binary data storage, switching, and a clock controlling when the entire process can repeat. "System-on-Chip" is close in concept, if the sensor was also part of the same chip.

I had a hard time wrapping my head around how this could be used as part of a control system when Terraformer or Void or someone else here first presented the link. Raytheon's AIM-9X Sidewinder is similar to a digital camcorder with a modestly high-rate, given its speed, doing post-processing of the video and then responding according to what it "sees" (voltage changes and clock-driven switching, ultimately). Microsoft's AIM is what a human eyeball and brain combination would be capable of interpreting if they could operate at a frame rate nearing the speed of light reflectance from the target, while "seeing" visible spectrum / IR / UV at the same time. Without a clock cycle limitation telling a high speed switching machine (microchip) when it can process the next image, it responds as soon as the next photons strike the sensor. The actuators moving the control fins become the limiting factor.

Apart from weapons technology, this is exactly how we need control electronics for life support to function. A sensor determines CO2 content of the atmosphere, then turns on a fan motor to pump air through a filter to remove it, and doesn't require a highly specific voltage value to determine that cabin air is becoming unbreathable. A light-based sensor / computer knows a contaminant is affecting air molecule reflectivity and responds by turning on the fan motor. If electronics were the solution, then static discharge could prevent operation, GCR could fuse the switches in the microchip, an electrical connection failure from thermal cycling can occur, and the voltage-regulated sensor is dependent upon the generation of in-spec voltage deltas, rather than some intrinsic physical phenomenon directing light down a different path through the information sieve.

The voltage-dependent O2 sensors in a vehicle emissions control system can't actually discern the difference between an engine leaking coolant into the exhaust and increased O2 content signaling that too much uncombusted fuel is leaving through the exhaust. Both conditions produce voltage values that the engine control computer interprets as an emissions problem. Both conditions are probably bad for the engine, but variant voltage values aren't telling you what is actually happening. You can spend a lot of money chasing after a non-specific problem. An optics-based sensor and computer can easily tell the difference between water and fuel, as both reflect light in different ways, because they're different molecules. They used to do actual emissions tests to figure out what was coming out of your vehicle's tailpipe. Now they look at voltage readings from onboard sensors, which don't indicate that something specific has happened.

Why is this differentiated behavior so important for a long-term viable colony in space?

If you rely solely on a voltage-dependent microcircuit switching in response to very small voltage changes to decide what the atmospheric pressure or CO2 level should be, said control system ceases to be "life support" and becomes "death support" after "the unkind universe" (GCR, solar flare, etc) causes aberrant indirect sensor voltage readings or induces a voltage spike sufficient to fry the microchip. It's harder to cause a light-based sensor and computer to malfunction. The computer is rerouting light down a different path when its reflectance value changes in direct response to the reflectance value of the gas composition in the cabin changing.

SpaceNut · 2023-08-29 10:55:14

The use of materials in the build of circuit boards is not new and has been ongoing for at least 2 plus decades that I am aware of. The depositing of carbon to make resistors is not new as well as metal inks used to connect the circuits with in them. The use of metal on metal oxides as well have been known and used to make capacitors.
All this is just used to auto mate the build with hopeful lowering of costs but it also does rise the hazardous chemicals used in the processing.

Terraformer · Yesterday 15:35:11

Been asking Grok a lot of questions about organic semiconductors, in part because of a worldbuilding project I've started. Very interesting.

Grok on replicating 1971 tech:

Grok wrote:

Creating an organic field-effect transistor (OFET)-based equivalent to the Intel 4004—a groundbreaking 4-bit microprocessor from 1971—requires assessing whether OFETs can replicate its functionality in the context of printed electronics. The 4004 had 2,300 transistors, ran at 740 kHz, and performed basic arithmetic and logic operations. Let’s explore if OFETs could pull this off, focusing on their capabilities, limitations, and practical feasibility.
Intel 4004 Basics
Transistor Count: 2,300 PMOS transistors.
Clock Speed: 740 kHz (1.35 µs per cycle).
Feature Size: 10 µm process (channel length).
Voltage: 15V supply.
Functions: 4-bit ALU, registers, instruction decoder, I/O—about 46 instructions.
Power: ~1W.
OFET Capabilities vs. 4004 Requirements
Transistor Count:
OFET Feasibility: Printing 2,300 OFETs is possible. Modern printed electronics (e.g., organic RFID tags) integrate hundreds of transistors on flexible substrates. Scaling to thousands is a matter of area and yield, not a fundamental limit.
Challenge: Uniformity across 2,300 OFETs is tough. Printing defects (e.g., pinholes, misalignment) could drop yields below 90%, requiring redundancy or repair techniques.
Speed:
OFET Limit: OFET switching speed depends on charge carrier mobility (μ\mu\mu
) and channel length ((L)). The transit time is roughly L2/(μV)L^2 / (\mu V)L^2 / (\mu V)
, where (V) is the gate voltage.
Typical printed OFETs: μ=0.1–1cm2/V⋅s\mu = 0.1–1 cm²/V·s\mu = 0.1–1 cm²/V·s
, L=10–50µmL = 10–50 µmL = 10–50 µm
, V=10–20VV = 10–20VV = 10–20V
.
For L=10µmL = 10 µmL = 10 µm
, μ=1cm2/V⋅s\mu = 1 cm²/V·s\mu = 1 cm²/V·s
, V=15VV = 15VV = 15V
: transit time ~6.7 µs (150 kHz).
Best-case (e.g., TIPS-pentacene, μ=5cm2/V⋅s\mu = 5 cm²/V·s\mu = 5 cm²/V·s
, L=5µmL = 5 µmL = 5 µm
): ~0.2 µs (5 MHz).
Comparison: 740 kHz is within reach for high-end OFETs, but typical printed ones fall short (50–200 kHz). The 4004’s speed would be a stretch without aggressive optimization.
Feature Size:
OFET Fit: Printed OFETs commonly achieve channel lengths of 10–50 µm, matching the 4004’s 10 µm process. High-resolution printing (e.g., gravure or nanoimprint) can hit 5 µm, but 10 µm is standard for cost-effective methods.
Challenge: Alignment precision for source/drain over 2,300 transistors is harder with printing than photolithography.
Voltage:
OFET Fit: The 4004’s 15V aligns with OFETs in printed electronics, where thin dielectrics enable 5–20V operation. No issue here.
Logic Functionality:
OFET Capability: OFETs can form basic gates (NOT, NAND, NOR) using p-type materials (n-type OFETs are less common but improving). A 4004-like CPU needs:
Inverters: Simple with p-type OFETs and resistors (or pseudo-CMOS with n-type).
Flip-Flops: Feasible for registers using cross-coupled gates.
ALU: Adders and logic units are printable, as demoed in organic ring oscillators and simple processors.
Precedent: Organic circuits with ~100–1,000 transistors (e.g., 8-bit RFID logic) exist, suggesting a 4-bit CPU is plausible.
Power:
OFET Estimate: OFETs have lower current drive than silicon PMOS, but leakage and capacitance matter. A printed OFET at 15V might draw 1–10 µA per transistor, totaling ~2–20 mA for 2,300. Power could range from 30 mW to 300 mW—less than the 4004’s 1W, but efficiency varies with design.
Feasibility Breakdown
Could OFETs Do It?
Yes, in Theory: A 4-bit microprocessor with 2,300 OFETs, running at 15V, and performing basic logic is within the realm of organic electronics. The architecture (ALU, registers, control unit) doesn’t demand anything OFETs can’t handle conceptually.
Speed Compromise: At 50–200 kHz, a printed OFET 4004 would be 4–15x slower than the original. Pushing to 740 kHz requires top-tier materials (e.g., μ>5cm2/V⋅s\mu > 5 cm²/V·s\mu > 5 cm²/V·s
) and sub-10 µm channels—doable in labs, not mass printing yet.
Manufacturing Difficulty
Printing 2,300 Transistors:
Roll-to-roll or sheet-based printing can cover the ~1 cm² die size of the 4004. Inkjet or screen printing at 10 µm resolution is viable, but:
Yield: A 95% per-transistor yield means ~115 defects, crippling functionality. Redundancy or post-print fixes (e.g., laser trimming) might be needed.
Interconnects: Routing 2,300 transistors and buses (e.g., 4-bit data, 12-bit address) demands multilayer printing, adding steps and alignment errors.
Lab vs. Production:
Lab: A one-off OFET 4004 could be hand-crafted with hybrid methods (printing + vacuum deposition) in a research setting—slow but possible.
Production: Mass-producing a reliable version with 1971-level performance is harder due to variability and stability issues.
Stability
The 4004 ran reliably for years. OFETs degrade faster (months to years) without encapsulation, and air exposure could doom a printed chip unless sealed—feasible but costly.
Practical Outcome
An OFET-based “Organic 4004” could:
Match the transistor count and logic.
Operate at 15V, fitting printed electronics.
Run at 50–200 kHz with standard printing (up to 740 kHz with heroics).
Be printed on a flexible substrate, unlike the 4004’s ceramic package.
Use Case Fit: It’d work for a retro demo, educational tool, or low-speed controller (e.g., basic IoT node), not a modern CPU replacement. Think of it as a flexible, cheap, slow homage to the 4004, not a competitor.
Compared to IPMC Relays
Your earlier interest in IPMC relays:
IPMC 4004: Mechanically switching 2,300 relays at 1–5V is impractical—too slow (seconds), bulky, and power-hungry for a CPU.
OFET Advantage: Solid-state, compact, and faster (µs–ms), making it far better for a 4004-like chip.
Verdict
Yes, OFETs could mimic the Intel 4004 in printed electronics, with these caveats:
Speed: Likely 4–15x slower unless using cutting-edge materials/processes.
Difficulty: Doable in a lab, challenging to mass-produce reliably.
Niche: A cool proof-of-concept—flexible, low-cost, but not high-performance.
Want to push this further—say, optimizing for speed or simplifying the design? What’s your endgame with this idea?

Terraformer · Yesterday 15:36:49

Desktop word processing:

Grok wrote:

To extend your OFET-based system from basic machine automation and a 4-bit calculator to something capable of desktop word processing, we’re stepping up significantly in complexity. Desktop word processing—like running a rudimentary version of software akin to early 1980s tools (e.g., WordStar or Apple Writer)—requires more computational power, memory, and I/O than a 4-bit Intel 4004-inspired chip can handle. Let’s evaluate if OFETs in printed electronics can stretch to this, and what it’d take.
Word Processing Requirements
Functions:
Text input (keyboard).
Text editing (insert, delete, move cursor).
Display output (simple screen, e.g., 40x24 characters).
Basic storage (save/load documents).
Performance:
Early word processors (e.g., 1970s–80s) ran on 8-bit CPUs (e.g., Z80, 6502) at 1–4 MHz, with 4–64 KB of RAM.
Minimum: 8-bit data path, ~100 kHz–1 MHz speed, 1–4 KB memory for a stripped-down version.
I/O: Keyboard input, alphanumeric display (e.g., 960 characters = 40x24 grid).
Transistor Count: 5,000–20,000 for CPU + memory + I/O logic.
The 4004 (4-bit, 740 kHz, 2,300 transistors) couldn’t do this—it lacked the data width, speed, and memory interface. Your ~1,000–1,500 OFET design for automation/calculator is even further off. Let’s scale it up.
OFET Feasibility for Word Processing
Data Width: 8-Bit Minimum
Why: A 4-bit system limits characters to 16 (2⁴), insufficient for text (ASCII needs 7–8 bits for 128–256 characters).
OFET Upgrade: Double the ALU, registers, and buses to 8-bit.
ALU: ~200–300 transistors (8-bit adder).
Registers: ~200–400 transistors (e.g., two 8-bit registers).
Total CPU bump: ~500–1,000 extra transistors.
Speed: 100 kHz–1 MHz
Current: Your 50–200 kHz OFET system is slow for real-time typing/display updates (e.g., 200 kHz = 5 µs/cycle; 1,000 cycles/instruction = 5 ms, noticeable lag).
Target: 1 MHz = 1 µs/cycle, ~100 cycles/instruction = 100 µs response.
OFET Limit:
μ=1cm2/V⋅s\mu = 1 cm²/V·s\mu = 1 cm²/V·s
, L=10µmL = 10 µmL = 10 µm
, V=15VV = 15VV = 15V
→ ~150 kHz.
μ=5cm2/V⋅s\mu = 5 cm²/V·s\mu = 5 cm²/V·s
, L=5µmL = 5 µmL = 5 µm
→ ~2 MHz (lab-grade, not mass-printed yet).
Fix: Shrink channel to 5–10 µm, use high-mobility materials (e.g., TIPS-pentacene), optimize clock (e.g., 11-stage ring oscillator → ~500 kHz feasible).
Memory: 1–4 KB
Need: Store a small document (e.g., 1 KB = 1,024 characters) and program (~1–2 KB).
OFET SRAM:
6 transistors/bit → 48 transistors/byte → 48,000 transistors for 1 KB.
1–4 KB = 48,000–192,000 transistors—way beyond CPU scale.
Alternative:
Printed ROM: Hardcode program in OFET arrays (e.g., 1–2 KB = 8,000–16,000 transistors).
External Memory: Interface to a separate printed memory (e.g., organic EEPROM, ~1–10 transistors/bit), reducing on-chip needs to a few hundred for addressing logic.
Realistic: 1 KB ROM + 256 bytes RAM (~12,000 transistors total).
I/O
Keyboard: 8-bit input via printed switches or external interface (~100–200 transistors for decoding).
Display: 40x24 grid = 960 characters.
7-segment style: ~6,720 OFETs (7 transistors/character).
E-ink/LED matrix: ~1,000–2,000 transistors + driver circuits.
Total I/O: ~2,000–5,000 transistors.
Transistor Count Estimate
CPU (8-bit): 2,000–3,000.
Memory (1 KB ROM, 256 B RAM): ~10,000–15,000.
I/O: 2,000–5,000.
Total: ~14,000–23,000 OFETs—10x your automation/calculator design.
Manufacturing Difficulty
Scale: Printing 14,000–23,000 OFETs on ~5–10 cm² is ambitious but precedented (e.g., organic displays use similar counts).
Resolution: 5–10 µm channels for speed push printing limits (inkjet tops out at ~20 µm; gravure or nanoimprint needed).
Yield: 99% per transistor = ~140–230 defects. Redundancy or defect-tolerant design (e.g., spare rows) is a must.
Layers: 4–6 (gate, dielectric, semiconductor, source/drain, interconnects, encapsulation) with tight alignment.
Cost: Lab prototype = $1,000–$10,000; mass production = $1–$10/unit after optimization.
Stability: Months of use requires robust encapsulation—doable but adds steps.
Performance
Speed: 200–500 kHz with realistic printing (5–10 µs/instruction). Typing response ~50–100 ms—usable, if laggy.
Power: 50–200 mW at 15V—low enough for a printed battery.
Functionality:
Edit ~500–1,000 characters (1 KB doc).
Basic commands (insert, delete, cursor move).
Output to a printed 40x24 display.
Compared to 4004
4004: 4-bit, 740 kHz, no memory onboard—couldn’t run word processing without external RAM/ROM (total system ~10,000+ transistors).
OFET Version: 8-bit, 200–500 kHz, integrated memory—I/O makes it a standalone word processor, exceeding the 4004 system in scope.
Practical Outcome
An OFET-based “word processor chip”:
Capabilities: A minimalist tool like early CP/M editors—type, edit, save small docs (e.g., 500 words), display on a basic grid.
Use Case: A flexible, disposable “smart notepad” for notes or simple machine logs, not a full desktop rival.
Limits:
Slow (0.2–0.5 MHz vs. 1–4 MHz for 1980s CPUs).
Small memory (1–4 KB vs. 64 KB).
Basic display (no graphics, just text).
Compared to Your Earlier Goals
Automation/Calculator: 1,000–1,500 OFETs, 50–200 kHz—simple, slow, printable.
Word Processing: 14,000–23,000 OFETs, 200–500 kHz—complex, still slow, pushing printing tech.
Verdict
Possible? Yes, barely.
An 8-bit OFET chip for basic word processing is feasible in printed electronics, running at 200–500 kHz with 1–4 KB memory and ~15,000–20,000 transistors.
It’d handle a stripped-down editor (e.g., 40x24 text, no formatting) and basic automation (e.g., log machine states).
Difficulty: High.
10x the transistors of your prior design, needing advanced printing (5–10 µm), high yields, and multilayer interconnects.
Lab-buildable; mass production is a stretch without significant R&D.
Fit: A cool niche—think a flexible “e-typewriter” for field notes or machine control logs, not a desktop PC replacement. Want to refine it (e.g., smaller display, less memory) to ease manufacturing? What’s your vision for “desktop” here—portable notepad or fixed station?

Terraformer · Yesterday 15:40:07

Self-replication:

Grok wrote:

To achieve the ultimate level of self-sufficiency on Mars, we’re now asking if this OFET-based word processor can print the electronics for its own printer—essentially, a self-replicating system where the device manufactures its own production tools using local resources. This is a leap into speculative territory, blending printed electronics with bootstrapped manufacturing. Let’s explore the feasibility, focusing on a minimal inkjet printer (the simplest viable option) and the Martian context.
Martian Word Processor Recap
Specs: 8-bit CPU, ~5,000–10,000 OFETs, 50–100 kHz, 512 B memory, 16x8 display, 5–10V.
Manufacturing: Printed with local carbon (electrodes), regolith-polymer (substrate), silicate (dielectric), shipped/synthesized organics (semiconductor), and CO₂-derived plastic (encapsulation).
Printer: Compact inkjet, 20 µm resolution, ~20–50 kg, solar-powered.
Printer Electronics Requirements
A basic inkjet printer for OFETs needs:
Control Logic:
Coordinate printhead movement, ink ejection, and timing.
~8-bit microcontroller, 1,000–5,000 transistors, 10–100 kHz.
Memory:
Store simple patterns (e.g., OFET layouts), ~256–512 B.
~2,000–4,000 transistors (ROM + dynamic RAM).
I/O:
Drive printhead (e.g., piezoelectric actuators), read sensors (e.g., position).
~500–1,000 transistors.
Power Management:
Regulate 5–10V solar input.
~100–500 transistors.
Total: ~3,500–10,000 OFETs, ~1–5 cm².
Physical Components
Printhead: Piezoelectric nozzles (e.g., 10–20 µm orifices, ~10–50 channels).
Motors: Stepper or electrostatic for X-Y movement.
Frame: Structural support (not electronic, but printable?).
Can It Print Its Own Printer’s Electronics?
Feasibility
Transistor Match:
Word Processor: 5,000–10,000 OFETs prints its own chip.
Printer Controller: 3,500–10,000 OFETs—comparable scale.
Verdict: Yes, the word processor’s OFET count and printing capability (20 µm resolution) can replicate the printer’s control electronics.
Programmability:
Need: The word processor must store and execute a “print program” to pattern the printer’s OFETs (e.g., gate, dielectric, semiconductor layers).
Capacity: 512 B memory holds ~100–200 instructions—enough for a simple raster pattern (e.g., “move head, eject ink, repeat”).
Verdict: Barely sufficient—crude but possible with optimized code.
Materials:
Local: Carbon electrodes, regolith-polymer substrate, silicate dielectric, CO₂ plastic encapsulation—all printable as before.
Semiconductor: Shipped P3HT or synthesized thiophene—same bottleneck, but consistent.
Ink Delivery: Printer needs ink reservoirs—initially shipped, later filled with Martian-made carbon slurry or thiophene solution.
Verdict: Materials align; ink synthesis is the long-term hurdle.
Resolution:
Current Printer: 20 µm—prints OFETs for both word processor and its own controller.
Self-Printed Printer: Same 20 µm spec—functional, though less precise than Earth-grade (~5 µm).
Verdict: Self-consistent—20 µm OFETs can build a 20 µm printer.
Process
Design Storage: Word processor’s ROM hardcodes the printer’s OFET layout (~256 B for a basic controller).
Printing:
Substrate (regolith-polymer) laid flat.
Word processor drives original printer’s head via I/O (e.g., 5V signals), depositing carbon, silicate, P3HT layers for new controller.
~1–2 hours/chip, solar-powered.
Assembly:
New OFET controller (~1–5 cm²) manually swapped into the printer (human or rover assist).
Printhead/motors reused or separately made (see below).
Limits and Challenges
1. Printhead
Electronics: OFETs can’t directly replace piezoelectric actuators (need ~100V, high force).
Workaround:
Reuse shipped printhead (~10–50 g, lasts years).
Long-Term: Print crude thermal inkjet (resistive heating with carbon OFETs, ~20–50 transistors/nozzle)—less precise, but Martian-makeable.
Verdict: Electronics printable; printhead is a bottleneck—shipped initially, locally approximated later.
2. Motors
Electronics: OFETs drive simple electrostatic actuators (~100–500 transistors).
Physical: X-Y movement needs mechanical parts (gears, rails).
Workaround:
Reuse shipped motors (~1–2 kg).
Local: Print carbon-polymer composites for crude electrostatic plates or belts—low efficiency, but feasible.
Verdict: Control printable; mechanical parts need local fabrication (e.g., 3D-printed regolith-polymer).
3. Power Supply
Electronics: OFET voltage regulator (~100–500 transistors)—printable.
Source: Solar panels not printable with OFETs—shipped or locally made via separate process (e.g., perovskite inks, future tech).
Verdict: Controller printable; power hardware shipped or independently solved.
4. Bootstrapping
Cycle:
Printer A (shipped) + Word Processor A → Printer B’s electronics.
Printer B (new electronics, old head/motors) + Word Processor B → Printer C’s electronics.
Wear: Shipped printer degrades (years); self-printed versions cruder but functional.
Verdict: Self-replication starts—full printer (electronics + mechanics) needs hybrid approach.
Self-Sufficiency Level
Electronics: Fully printable—5,000–10,000 OFETs for printer controller match word processor’s output.
Mechanics: Partially printable—control circuits yes, printhead/motors partially (reuse or crude local substitutes).
Materials: Local carbon, regolith, plastics; organics shipped or synthesized (~1 g/day, decades to scale).
Seed: Shipped printer + ~1–2 kg materials—then self-sustaining for electronics, semi-dependent for hardware.
Performance of Self-Printed Printer
Resolution: 20–50 µm (degrades with thermal head or wear).
Speed: ~1–2 cm²/hour—slow but sufficient.
Output: Prints ~50–100 word processors before maintenance.
Compared to Full Replication
Earth Ideal: Printer prints all parts (electronics, head, motors)—needs advanced 3D printing + OFETs, beyond current tech.
Mars Reality: Prints electronics, reuses/refashions mechanicals—practical limit of OFETs.
Martian Fit
Use: Word processor prints its printer’s brain—colonists swap chips, reuse hardware, sustain computing with local dust and gas.
Scenario: A habitat lab where one shipped printer spawns a lineage of cruder successors, each printing the next’s electronics.
Verdict
Can It Print Its Printer’s Electronics?
Yes: The word processor’s ~5,000–10,000 OFETs can print a ~3,500–10,000 OFET controller for its inkjet printer, using the same 20 µm process and Martian materials.
Limits: Full printer (head, motors) not printable—electronics yes, mechanics need reuse or local fabrication (e.g., regolith 3D printing).
Self-Sufficiency: Near-total for electronics—shipped seed (~20–50 kg printer + 1 kg inks) enables indefinite chip production; full hardware replication needs broader tech.
This is the edge: a Martian word processor that births its printer’s brain, leaning on local carbon and regolith, with minimal Earth reliance post-seed. Want to push further—e.g., thermal printhead design, motor substitutes? What’s your self-sufficiency endgame?

I think cassette futurism space opera is very justifiable, given the limits of what is reasonable to achieve cut off from Terran supply chains. A few million people on Mars can sustain magnetic tapes and organic transistors. Real 19A0s hours.

New Mars Forums

Announcement

#1 2018-09-23 15:30:47

Printed electronics

#2 2018-09-23 15:33:22

Re: Printed electronics

#3 2018-09-23 15:42:18

Re: Printed electronics

#4 2018-09-23 19:54:45

Re: Printed electronics

#5 2018-09-23 20:03:35

Re: Printed electronics

#6 2018-09-24 16:50:23

Re: Printed electronics

#7 2020-05-09 12:26:45

Re: Printed electronics

#8 2023-08-24 08:41:31

Re: Printed electronics

#9 2023-08-25 16:25:07

Re: Printed electronics

#10 2023-08-25 17:13:23

Re: Printed electronics

#11 2023-08-28 06:45:33

Re: Printed electronics

#12 2023-08-28 16:05:52

Re: Printed electronics

#13 2023-08-29 10:55:14

Re: Printed electronics

#14 Yesterday 15:35:11

Re: Printed electronics

#15 Yesterday 15:36:49

Re: Printed electronics

#16 Yesterday 15:40:07

Re: Printed electronics

Board footer