Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

https://github.com/dmontgomery40/openframe_printer

OpenFrame M1: open-hardware monochrome LED electrophotographic printer — design package, executable engineering models, firmware, and RFQs
https://github.com/dmontgomery40/openframe_printer

Last synced: 19 days ago
JSON representation

OpenFrame M1: open-hardware monochrome LED electrophotographic printer — design package, executable engineering models, firmware, and RFQs

Awesome Lists containing this project

README 

          
# OpenFrame Printer Newbuild v2



OpenFrame M1 is a from-scratch modular printer design package. It is **not** a conversion layer, bridge appliance, or modified existing printer. The v2 package pushes the idea toward a real new machine: engine constants, HV channel targets, LED bar timing, process-cartridge geometry, fuser thermal model, paper-path timing, connector maps, interlock matrix, and a future inkjet/nozzle R&D branch.



## Primary build target



**OpenFrame M1 Rev G package, derived from the original Rev A geometry**



- Technology: monochrome dry electrophotographic printer with a stationary LED exposure bar

- Print mode: 1-bit monochrome, 600 dpi

- Speed target: 12 ppm Letter

- Active LED bar: 5120 pixels, 42.333 µm pitch, 216.747 mm active width

- Process speed: 62.0 mm/s

- OPC drum: 30.0 mm diameter

- Fuser target: 178 °C surface temp, 5.0 mm nip, 80.6 ms dwell at process speed

- Print path: IPP/AirPrint/Mopria externally, deterministic OFP1 raster/engine protocol internally

- Business/model constraint: public service docs, public module interfaces, direct replacement parts, no cartridge DRM lockout, no cloud requirement



## What changed in v2



- Locked a **Rev A architecture** around a new 30 mm OPC drum and 5120-pixel LED bar.

- Added exact connector maps for power, motors, sensors, LED bar, HV module, cartridge contacts, fuser, and UI.

- Added HV bias channel table with nominal/range/current-limit/ramp targets.

- Added fuser thermal model with generated warm-up CSV.

- Added motion timing model with jam windows and sensor positions.

- Added process cartridge cross-section, cartridge requirements, and open consumables spec.

- Added inkjet/nozzle R&D constants so the nozzle path is mapped, while keeping the first build on the easier monochrome electrophotographic path.

- Research Rev B corrected OPC exposure units to µJ/cm², fixed LED shift-clock math, added exposure-to-development delay guardrails, and separated cited research grounding from novel design hypotheses.



## What changed in Rev C



Rev C turned the research corrections into engineering engines — solved geometry, checked electrostatics, and a working calibration loop — instead of guardrail prose:



- **Solved drum station map** (`openframe_printer/station_map.py`, doc 31): exact roller-clearance math places every process station around the drum. Headline: the Rev B 50 ms exposure-to-development target is geometrically impossible on the 30 mm drum package (minimum feasible ≈ 160 ms); the binding spec moves to the OPC as a measurable contrast-hold requirement (≥90% at 240 ms).

- **Electrostatic voltage ladder** (`openframe_printer/voltage_ladder.py`, doc 32): connects PCR bias → charged surface → exposure → developer window in one checked chain. It proves **Rev A as tabled cannot print** — −720 V DC on the charge roller yields a −70 to −220 V drum, not the assumed −600 V — and proposes two concrete fixes.

- **PIDC-first calibration, implemented** (`openframe_printer/pidc_model.py`, doc 33): hypothesis H1 as running code — model, noise-tolerant fitter, latent-voltage-targeted LED pulse chooser, and a synthetic rig whose kill criteria pass/fail automatically.

- **Unit-safety layer** (`openframe_printer/units.py`): every generated JSON passes a unit-plausibility lint, so the Rev B mJ/µJ bug class now fails the build instead of waiting for a reviewer.



## What changed in Rev D



Rev D found and fixed a Rev C artifact drift: Rev C's prose and CSV retired the impossible −720 V PCR bias, but the generated HV JSON and smoke test still allowed the old value. Rev D makes the fixed HV ladder executable and adds two new quantitative engines:



- **Generated HV table is now authoritative** (`openframe_printer/hv_model.py`): `out/v2_hv_bias_channels.json` exposes only `PCR_CHARGE` Option A (`A_dc_only`, −1180 V nominal) or Option B (`B_ac_dc`, −600 V DC plus 1.7 kVpp AC). The old −720 V PCR target is now a retired value and cannot pass the smoke test.

- **HV consistency gate** (`out/v2_hv_consistency.json`, doc 35): checks the generated HV rows against `out/v2_voltage_ladder.json`. If the ladder, CSV, and generated table drift apart again, the build fails.

- **H8 developer-roller probe budget** (`openframe_printer/dev_probe.py`, doc 36): turns the developer-as-electrostatic-probe idea into a signal/noise budget. A 64×64-pixel patch gives about 6.45 nA full-scale ideal signal; a realistic first test should use 128×128 patches and a real transimpedance/current-sense DEV_MON mode. Plain HV voltage readback is not assumed good enough.

- **Transfer impedance control** (`openframe_printer/transfer_model.py`, doc 37): replaces fixed transfer-voltage faith with a current-limited paper/nip impedance sniff. Normal paper impedances run at the voltage target; extreme dry/high-impedance cases are rejected or slowed instead of silently under-transferring.

- **New falsifiable hypotheses** (doc 38): developer TIA mode as a calibrator, density-aware transfer waveform shaping, and a sacrificial non-image impedance strip for low-cost paper classification.

- Focused test suite: `python3 scripts/model_tests.py` runs alongside the generated-artifact smoke test.



## What changed in Rev E



Rev D hardened the electrostatics; Rev E covered four subsystems no revision had modeled at all — the wire protocol, the safety chain's fault tolerance, the raster screen, and the consumable mass balance:



- **OFP1 exists now** (`openframe_printer/ofp1.py`, doc 39): every revision since v1 *named* the "deterministic OFP1 protocol"; none defined it. Rev E ships the binary framing (per-frame CRC-16/CCITT, whole-page CRC, blank-line SKIP compression), a reference encoder and engine-side decoder that round-trip a page bit-exactly under arbitrary byte chunking, and proof that a flipped bit is NACKed, never printed. The transport budget is honest: a worst-case page needs 7.6 Mbit/s — 78% of the USB Full Speed bulk ceiling — so production is USB High Speed, FS is the degraded service mode.

- **Interlock single-fault analysis** (`openframe_printer/interlock_faults.py`, doc 40): exhaustively enumerates stuck-at faults with adversarial firmware. The chain as documented has **7 single-point-failure nets** — one welded cover switch, one shorted loop, or one stuck enable gate leaves HV/LED/fuser live with a door open. Rev E's dual-chain topology survives independent electrical faults; Rev F adds the missing mechanical common-cause fault.

- **EP-safe halftoning** (`openframe_printer/halftone.py`, doc 41): EP cannot stably develop an isolated 42 µm pixel, so the host screen is now an executable constraint. Rev E chose the right clustered-dot direction; Rev F fixes the partial-seed floor bug and gates zero isolated or sub-2×2 features.

- **Toner mass balance and the no-DRM gauge** (`openframe_printer/toner_budget.py`, doc 42): retires a live doc/artifact contradiction — docs claimed "about 2400 pages" per 80 g while the generated math said ~4800; neither survived. The loss-adjusted model (90% transfer efficiency, 8% hopper residual) rates ~4000 pages, sizes the cartridge waste cavity the RFQ requirement implies (≈28 cm³), and derives the pixel-count gauge constant (~11 mg per million black pixels) that replaces lockout chips.

- Focused test suite: `python3 scripts/model_tests.py` grew the suite that Rev F later extends.



## What changed in Rev F



Rev F is the correction pass on Rev E's new engines. It found one real raster bug, one safety-model blind spot, one remaining toner artifact drift, and two missing continuous-operation models:



- **Halftone floor gate** (`openframe_printer/halftone.py`, docs 41/43): Rev E's raw 2×2-seeded threshold still emitted one-pixel partial seeds below the declared 4/64 highlight floor. Rev F clips sub-floor tones and suppresses sub-2×2 connected components. `out/v2_halftone_floor_gate.json` proves the old bug and the new fix.

- **Interlock common-cause model** (`openframe_printer/interlock_faults.py`, docs 40/44): Rev E's dual-chain topology survives independent electrical stuck-at faults, but a single shared door-actuator failure can make both contacts read closed. Rev F requires topology C: dual electrical chains plus a physically diverse energy separator opened by the access door.

- **Toner artifact consistency** (`openframe_printer/toner_budget.py`, docs 42/45): Rev E removed the bad 2400-page claim but left the naive ~4820-page number in base design calcs under an unqualified key. Rev F removes that key and gates the explicit naive upper-bound label.

- **Fuser paper-load balance** (`openframe_printer/fuser_power.py`, doc 46): the warm-up curve was not a 12 ppm throughput proof. Rev F charges the fuser for paper, moisture, and toner. With the current 800 W / 0.235 °C/W assumptions, nominal 75 gsm media lacks a 10% steady heater reserve.

- **LED thermal feed-forward** (`openframe_printer/led_thermal.py`, doc 47): H9 is now executable. The raster payload predicts LED-group thermal droop and emits bounded pulse compensation before photodiode/PIDC trimming.

- **New falsifiable hypotheses** (doc 48): adaptive minimum-dot unlock, interlock transition diagnostics, fuser media inference, LED/PIDC residual fusion, and transfer-informed toner gauge correction.

- Focused test suite: `python3 scripts/model_tests.py` now runs **57 tests** alongside the generated-artifact smoke test.



## What changed in Rev G



Rev G attacks the places where Rev E/F still let a physically correct-looking package overclaim build readiness: real-time host feeding, line placement, LED optics, fuser runaway safety, missing erase/quench, HV stored-charge decay, humidity/material derating, emissions containment, and registration-edge slack.



- **OFP1 real-time spool budget** (`openframe_printer/ofp1_realtime.py`, doc 49): Rev E proved frame correctness, not that USB Full Speed plus a 32-line ring can feed a moving engine. Rev G shows the Rev E ring tolerates only ~21.9 ms; a 100 ms first-build pause target needs 93,733 bytes. USB FS at 75% of its theoretical bulk ceiling cannot sustain 12 ppm worst-case pages.

- **Motion registration budget** (`openframe_printer/motion_registration.py`, doc 50): nominal process speed is not line placement. A 0.5% open-loop speed error stretches Letter by 33 lines. Rev G requires encoder-slaved LED line timing and timestamped registration edges.

- **LED optical MTF gate** (`openframe_printer/optical_mtf.py`, doc 51): 5120 emitters and correct shift timing do not prove 600 dpi latent contrast. Rev G requires MTF >= 0.35 at 12 lp/mm at the OPC plane, or equivalent measured spot FWHM <= 45 µm with <=15% one-pixel crosstalk.

- **Fuser thermal safety model** (`openframe_printer/fuser_safety.py`, doc 52): Rev F's power balance was not a runaway proof. Rev G proves firmware-only heat control has three single-fault runaway paths; independent thermostat plus one-shot thermal fuse has zero modeled single-fault violations.

- **Erase/quench ghost budget** (`openframe_printer/erase_model.py`, doc 54): the prior process stack had no explicit post-cleaning erase station. Rev G fails that state, sizes a 0.75 µJ/cm² erase dose, and predicts the ghost-repeat search distance at one drum circumference: 94.25 mm.

- **HV stored-charge bleed-down** (`openframe_printer/hv_discharge.py`, doc 55): interlocks remove enables, but they do not automatically discharge output capacitance. Rev G adds redundant bleeder sizing and gates normal <60 V / single-fault <120 V after 2 s.

- **Humidity/material derating** (`openframe_printer/environment_model.py`, doc 56): 80% RH plain paper is no longer treated as nominal. The generated model holds quality printing for a PIDC/developer/transfer sweep when toner charge falls to ~83% of the 50% RH reference.

- **Emissions containment budget** (`openframe_printer/emissions_model.py`, doc 57): a fan filter alone fails the modeled high-emitter case; the fuser/output path must be captured and bench-tested instead of left as an open plume.

- **Registration and LED-edge slack** (`openframe_printer/registration_budget.py`, doc 58): 5120 pixels only gives 0.373 mm lateral slack per side on Letter. Rev G either moves the LED requirement to 5184 active pixels or forces reduced printable width / ±0.25 mm registration proof.

- **New falsifiable hypotheses** (doc 53): clutch-start spool proof, encoder-derived drum health map, cheap slanted-edge LED MTF acceptance, cutoff-trip telemetry, transport-mode aware ppm scheduling, ghost-based erase health, bleed-decay self-test, humidity-driven calibration policy, output-tray emissions capture, and edge-slack-aware print-width negotiation.

- Focused test suite: `python3 scripts/model_tests.py` now runs **78 tests** alongside the generated-artifact smoke test.



## Important safety boundary



This is an engineering design package, not a certified appliance. The first physical rig must be a cold paper-motion rig. HV, fuser heat, and LED output are hardware-gated and default off in the firmware model. The package includes safety interlocks because a real printer contains mains voltage, hot surfaces, and high-voltage bias rails.



## Run the design generator



```bash

cd "$HOME/Downloads" && rm -rf openframe_printer_newbuild_v2 && unzip -o openframe_printer_newbuild_v2_research_revG.zip && cd openframe_printer_newbuild_v2 && python3 -m venv .venv && . .venv/bin/activate && python -m pip install --upgrade pip && python -m openframe_printer.demo && python -m openframe_printer.design_report && python scripts/smoke_test.py && python scripts/model_tests.py

```



Generated outputs land in `out/`:



```text

out/openframe_m1_page.pbm

out/openframe_m1_job_plan.json

out/v2_design_report.md

out/v2_design_calcs.json

out/v2_exposure_summary.json

out/v2_motion_events.json

out/v2_hv_bias_channels.json

out/v2_hv_consistency.json

out/v2_fuser_summary.json

out/v2_fuser_sim.csv

out/v2_led_group_map.csv

out/v2_interlock_matrix.csv

out/v2_openframe_m1_cross_section.svg

out/v2_process_cartridge.svg

out/v2_nozzle_math.json

out/v2_ep_physics_summary.json

out/v2_station_map.json

out/v2_station_map.csv

out/v2_voltage_ladder.json

out/v2_pidc_calibration_demo.json

out/v2_dev_probe_budget.json

out/v2_transfer_impedance_plan.json

out/v2_ofp1_transport_budget.json

out/v2_interlock_fault_analysis.json

out/v2_halftone_printability.json

out/v2_halftone_floor_gate.json

out/v2_toner_mass_balance.json

out/v2_toner_artifact_consistency.json

out/v2_fuser_power_balance.json

out/v2_led_thermal_feedforward.json

out/v2_ofp1_realtime_spool.json

out/v2_motion_registration_budget.json

out/v2_optical_mtf_budget.json

out/v2_fuser_thermal_safety.json

out/v2_erase_ghost_budget.json

out/v2_hv_discharge_bleed.json

out/v2_environment_derating.json

out/v2_emissions_containment.json

out/v2_registration_edge_budget.json

```