US 6667684

I'll search for authoritative information on US patent 6667684, including any 2026 CAFC docket activity.

Let me check specifically for any litigation or Federal Circuit docket activity tied to this patent number.

Let me make a final targeted check for any 2026 PTAB or district court activity specific to this patent.

US Patent 6,667,684 — Search Results Summary

Important scoping note: I searched USPTO/Google Patents/free patent databases and general web sources for the exact number 6667684 and its formatted variants (6,667,684 / US6667684B1). I found no CAFC docket, appeal, or PTAB/IPR proceeding dated 2026 (or otherwise) naming this specific patent. The only appellate activity I could confirm in the barrier-operator space involves different patents (e.g., the 2022–2023 Overhead Door/Chamberlain ITC appeals, Nos. 22-1664/22-1656; and Chamberlain's 23-1964 appeal), and none of those asserted '684. I could not retrieve a live USPTO PatentCenter/Patent Trial and Appeal Board record directly. If a 2026 docket on this number exists, it is not surfaced in the sources I could reach, so treat that as an open question rather than a confirmed absence.

Bibliographic data

Date-conflict caveat: Google Patents shows filing 2000-03-08, issue 2003-12-23, priority 1996-09-06. Unified Patents shows application 2000-03-07, grant 2003-12-22, priority 1996-09-05 (one-day offsets, likely time-zone/normalization differences). I am reporting the Google Patents figures as primary since they match the USPTO-style front page in the authoritative text, but I flag the discrepancy rather than silently resolving it.

Abstract (as published)

A remote control system for opening and closing a barrier, such as a garage door, includes an RF receiver and a plurality of RF transmitters. The transmitters and receiver include circuitry programmed to provide transmission of encrypted code signals each time the transmitters are used and employing a code hopping method which prevents unauthorized signal interception or code "grabbing." The system is operated in a code learning mode for the receiver by momentarily actuating a receiver learn mode button for receiving each transmitter identification code and a secret decryption key for that transmitter, with the system automatically returning to the operate mode. Each transmitter identification and secret key code signal is automatically and randomly stored in an available and unused memory in the receiver circuitry. A multibit hopping code is transmitted from each transmitter to the receiver with each transmitter operation in the operate mode, and the hopping code changes with each transmission to prevent theft or code grabbing and resultant unauthorized operation of the system.

Plain-language overview of the claims

This patent has exactly one claim (claim 1), and it is the sole independent claim. It is an apparatus claim cast in the form "A radio frequency transmitter for remotely communicating with a receiver…" but it also recites affirmative limitations on the receiver, so it effectively claims the transmitter/receiver pairing.

Claim 1, in plain language:

(a) The transmitter side. The RF transmitter is one of a set of transmitters sharing a common "manufacturer's key" that designates a particular manufacturer. Stored inside the transmitter is:

• a serial number code portion that is unique to that transmitter;
• a secret key code portion that is unique to that transmitter;
• a synchronization value that increments each time the transmitter is actuated; and
• a non-linear function generator that produces a multi-bit "hopping code" that changes each time the transmitter is actuated.

The transmitter sends out a coded signal carrying both the hopping code and the serial number portion.

(b) The receiver side. The receiver is of the type having a microprocessor operable between a learn mode and an operate mode, and it has the manufacturer's key stored in it at the factory.

• (i) Learn mode: The receiver receives the hopping code and serial number sent during learn mode. Using its stored manufacturer's key plus the received serial number, it independently generates a secret key (and then a synchronization value) in the receiver that matches the transmitter's secret key and sync value — i.e., the secret key is never transmitted; it is reconstructed.
• (ii) Random storage: Processing circuitry in the receiver stores the so-generated secret key and synchronization value randomly in an unused discrete memory location during learn mode; if all locations are full, it replaces the information in a randomly chosen location. (In the spec: seven locations, and an "erase-all" routine triggered by holding the learn button ~8 seconds.)
• (iii) Operate mode: The receiver performs a non-linear decoding function on the intercepted hopping code using one stored secret key to produce a second synchronization value, and compares that against the stored synchronization value. It issues the command signal (door open/close) when the decoded value bears a predetermined relationship to the stored value and falls within a window of it. If that relationship is absent, the receiver tries the next stored secret key, generating another second synchronization value, and keeps repeating the sequence through the stored secret keys until one produces a value meeting the predetermined/windowed relationship.

Key data-flow points from the specification that illuminate the claim: the hopping code is 32 bits; serial number 24 bits; secret key and manufacturer's key 64 bits; sync value 16 bits (incrementing by one per actuation). The operate-mode acceptance window is (ssv+1) to (ssv+15), where ssv is the stored sync value — which also means the system will not actuate on two successive identical transmissions. There is also an auto-synchronization routine (second, narrower window of (tssv+1) to (tssv+3) after an out-of-sync transmission) that is described in the specification but is not recited in the sole claim.

Caveats and uncertainties

• Claim scope: Only claim 1 exists; there are no dependent claims to expand or qualify it. The "predetermined relationship / window" language is functional and not numerically bounded in the claim itself — the numeric windows appear only in the description.
• Assignment nuances: Google Patents lists current assignees as Overhead Door Corp. and Microchip Technology Inc.; Unified Patents additionally lists Sanwa Holdings Corp. as a parent company. Because no assignment record was directly retrieved, I state these as database-reported rather than verified from USPTO assignment records.
• No 2026 CAFC activity confirmed. I could not verify any CAFC 2026 docket for 6,667,684. Given the patent expired (anticipated 2016-09-06) and is marked "Expired – Lifetime," a 2026 appellate proceeding asserting it would be unusual, though not impossible (e.g., a damages/royalty tail or a related-patent case). I am explicitly noting this as unverified rather than asserting there is none.

Sources: https://patents.google.com/patent/US6667684/en ; https://www.freepatentsonline.com/6667684.html ; https://patents.justia.com/patent/6667684 ; https://uspto.report/patent/grant/6667684 ; https://portal.unifiedpatents.com/patents/patent/US-6667684-B1

I'll research litigation involving US patent 6667684. Let me run several searches.

Let me search litigation databases directly for this specific patent number.

Let me check patent litigation aggregators directly for this patent.

Litigation Search: U.S. Patent No. 6,667,684

Bottom line

I could not identify any litigation specifically involving U.S. Patent No. 6,667,684. Based on the sources I was able to search, the record appears to be that this patent was never asserted in a lawsuit as a patent-in-suit. I want to be explicit that this is a negative finding bounded by the search limitations described below — not a certification.

What I checked and what I found

Strong corroborating (circumstantial) evidence of no litigation

Several independent facts make it unlikely this specific patent was litigated:

1. It is the continuation, not the parent. US 6,667,684 is a continuation of Ser. No. 08/706,682, which issued as U.S. Pat. No. 6,049,289 (also "Remote controlled garage door opening system," Overhead Door Corp.). When this family was litigated, the patents actually asserted in the garage-door-opener campaigns of that era were other members — e.g., the Chamberlain patents and the Bruwer/Nanoteq–Microchip line (U.S. 5,517,187; U.S. 5,686,904). The '684 patent shows up only as a cited reference, not an asserted patent.

2. The family's other members tell the story. The '289 parent, the '684 continuation (this patent), and the abandoned continuation publication US 2004/0085185 all share priority to 1996-09-06. If a case had been filed on this disclosure, the '289 parent is the more natural candidate, and I found no case naming '684.

3. It expired on 2016-09-06 (anticipated expiration; status "Expired – Lifetime"). Any infringement suit would have had to be filed before that date — i.e., well over nine years ago — and none appears in the public record.

Important caveats (per the strict rule, I am not guessing)

• I cannot rule out a suit that (a) settled early, (b) was filed under a sealed or non-indexed docket, (c) appears only in the paywalled RPX Insight / Empower database, or (d) involved the patent only in a defensive posture (e.g., named in an invalidity contention or an IPR "related matters" section rather than as an asserted claim). My searches surfaced the patent in invalidity/prior-art contexts of other patents, but not as an asserted patent.
• I found no PTAB proceedings (IPR/PGR/CBM) and no ITC Section 337 investigation naming the '684 patent specifically.
• I did not obtain a name-and-number-matched plaintiff/defendant/jurisdiction/filing-date/outcome record for '684 from any authoritative litigation database, so I will not synthesize one.** Providing parties, case numbers, or dates here would be fabrication.

If you need a definitive answer

Because PACER and RPX Insight are the only sources that would conclusively confirm a full negative, I recommend:

1. RPX Insight / Empower — run a patent-level "litigation history" on US 6,667,684 (this is the tool specifically built to build a timeline from a patent number, per RPX's own description at rpxcorp.com/solutions/market-intelligence/).
2. Lex Machina / Docket Navigator — patent-level assertion query.
3. USPTO PTAB — search for "6,667,684" in petition "related matters" sections.
4. PACER — full-text search of complaints for the literal string "6,667,684" (the comma-separated form most complaints use).

If you'd like, I can draft the exact query strings to run against those platforms, or re-run targeted open-web searches against specific candidate campaigns (e.g., Chamberlain Group, Overhead Door, Microchip Technology, Wayne-Dalton, Linear/Nortek) — but on the present record, the honest answer is: no known litigation involving US 6,667,684.

All patent-identifying facts above are drawn from the authoritative full text of US6667684 supplied in the prompt (title, inventors, assignees, priority 1996-09-06, grant 2003-12-23, expiration 2016-09-06, and the '289 parent).

I'll verify the canonical "no proceedings" default against live sources before writing up.

Let me run a couple more targeted checks before concluding.

PTAB Proceedings Report — U.S. Patent No. 6,667,684

Patent: US 6,667,684 B1 — "Remote controlled garage door opening system"
Application: 09/521,510 · filed 2000-03-08 · granted 2003-12-23
Priority: 1996-09-06 (continuation of Ser. No. 08/706,682, now US 6,049,289)
Patent owners/assignees: Overhead Door Corp.; Microchip Technology Inc.
Status: Expired — anticipated expiration 2016-09-06
Report date: 2026-09-30

---

Proceedings overview

Zero AIA trial proceedings are on file against US 6,667,684 — 0 IPRs, 0 PGRs, 0 CBMs, 0 derivations. There is nothing to break down by status (active / invalidated / sustained / settled / denied): the USPTO Open Data Portal returns no AIA trial proceedings for this patent, and targeted web searches for any petition, institution decision, Final Written Decision, or Federal Circuit appeal naming the '684 patent returned nothing.

The bottom-line defensive posture is stronger than "hardened by surviving IPRs" — it is moot. The patent expired on 2016-09-06, and its only claim (claim 1) died with it. There is no live claim to institute an IPR against, no live claim to assert, and no § 286 damages window that can reach any infringing act (the patent expired more than ten years ago; the six-year lookback in 35 U.S.C. § 286 cannot reach back to a date on which the patent was still in force). Absence of PTAB activity is not a signal of a hardened patent here — it is a signal that the patent was never worth challenging on its merits because it was a 1996-priority continuation whose term ran out before the IPR regime could be turned on it in earnest, and whose assertion value was always tied to the later Overhead Door/Microchip rolling-code families.

---

Proceedings on file

None. No proceeding number can be listed, because none exists — and per the operating constraints I will not construct one.

For transparency, here is exactly what was checked and the result:

Confidence caveat, stated plainly: I could not query PTAB E2E / PTAB Decisions directly, and search-engine coverage of pre-2015 PTAB dockets is incomplete. I am highly confident there is no instituted AIA trial on the '684 patent (a trial, FWD, or appeal would be well-indexed). I am also confident, based on the ODP return plus the absence of any trace in the literature, that no petition was filed — but I cannot rule out with certainty that a petition was filed and denied institution pre-2015 without generating an indexed decision. Even if such a petition existed, it would change nothing operationally: a denial creates no estoppel under § 315(e)(2) and no claim-level outcome.

Why no IPR was ever likely to matter here. The '684 patent granted 2003-12-23, so it was IPR-eligible from 2004-09-23 onward. Its entire IPR-eligible life (2004–2016) overlapped with the rise of the rolling-code wars in this industry — the Chamberlain v. Linear/Nortek and Chamberlain v. Overhead Door campaigns over the '011, '404, '416, '977, '212, '218 and '923 patents — yet no petitioner picked up the '684 patent as a target. The reason is straightforward: the '684 patent is a system/method rolling-code patent with a single narrow claim, its commercial relevance was superseded by the later Overhead Door and Chamberlain patent families, and by the time those disputes matured the term was expiring.

---

Strategic summary

Claim status. Claim 1 is the only claim of US 6,667,684. It was never canceled, never narrowed by reissue, and never amended — but it lapsed with the patent's term on 2016-09-06. So the correct answer to "which claims are CANCELED vs. SUSTAINED vs. UNTESTED" is: claim 1 was never tested at the PTAB, never invalidated by the PTAB, and is now expired. A patent owner cannot assert it today; an accused infringer has nothing to defend. The relevant structural point is that claim 1 is a very close analogue of the disclosure of the parent, US 6,049,289 (also expired), and of the Bruwer-family code-hopping patents (US 5,517,187; US 5,686,904; US 6,175,312) that Microchip asserted in this era — if you are looking for the live enforcement threat from this patent family, the '684 patent is not it.

Estoppel landscape. There is no estoppel, because there is no FWD. 35 U.S.C. § 315(e)(2) estoppel attaches only after a final written decision, and none issued. Consequently: no petitioner, real party in interest, or privy is barred from anything. Every prior-art ground that was ever available against the '684 patent — including all 58 references on its face (Chamberlain's '118, '980, '333, '433, '011/RE35,364; Microchip/Nanoteq's Bruwer '187; Siemens '225; Texas Instruments '168; Stanley '159 and '960; Kiekert '985 and '529; Delta Elettronica '221; Delco '667; and the rest) — remains unconstrained by PTAB estoppel, if it ever mattered. It does not matter now, because there are no live claims to invalidate in district court either. Practically, the "estoppel landscape" here is empty ground; there is no reason for a defendant to accept an IPR-driven narrowing as a limit on its invalidity theories.

Pattern signals. No serial petitions (no petitioner ever filed). The patent owner never had to defend a PTAB appeal, so there is no pattern of aggressive PTAB-appeal conduct to attribute to Overhead Door or Microchip on this patent. There is no defensive-aggregator chain: Unified Patents maintains a portal record for US-6667684-B1 but I found no Unified-filed petition or joinder against it. The one litigation thread worth knowing about — The Chamberlain Group LLC v. Overhead Door Corporation, No. 2:21-cv-00084 (E.D. Tex.) (jury verdict 2022-03-11 clearing Overhead Door and invalidating two asserted claims; a subsequent 2023-01-27 verdict of $43.4M against Overhead Door on one patent) — involved US 8,587,404, 9,644,416 and 8,144,011, not the '684 patent. Do not let anyone conflate those verdicts with this patent.

A data discrepancy to be aware of. Google Patents (the authoritative text supplied) gives priority 1996-09-06, filing 2000-03-08, grant 2003-12-23, expiration 2016-09-06. Unified Patents' portal renders the same events as 1996-09-05 / 2000-03-07 / 2003-12-22 / 2016-09-05 — almost certainly a UTC-offset artifact. The operative legal date is the Google Patents/ODP set; the difference is immaterial to anything below.

---

Recommended next steps

1. If you are a defendant being threatened with the '684 patent, the answer is not a PTAB filing — it is a motion and a letter. The patent expired 2016-09-06. Quote the expiration entry (Anticipated expiration 2016-09-06, Status: Expired - Lifetime) from the Google Patents record and demand that the demand letter identify a single act of infringement occurring on or before the expiration date. There is no live infringement theory; a claim for past damages is barred by § 286's six-year lookback against a patent that has been expired for over a decade.

2. Do not file an IPR on this patent. It is expired, so the patent owner cannot amend, the Board's decision would have no prospective effect, and you would spend a substantial five-figure-plus sum to invalidate claims that already carry no enforceable right. Filing would also invite the patent owner to argue the challenge is an inefficient use of Board resources.

3. Verify before relying on my negative finding. Before finalizing a clearance or non-infringement opinion, have counsel with PTAB E2E access pull the proceeding list directly from PTAB E2E and confirm no post-2016 docket entry exists. My search-based conclusion is that none does, but a direct docket pull is the appropriate belt-and-suspenders step, and it is cheap.

4. Redirect your analysis upstream if the family is still live. If the assertion comes from Overhead Door, Microchip, or a successor (Sanwa Holdings appears as a parent company of record), the same code-hopping lineage is carried by US 6,175,312, US 6,049,289 and others — several of which are also expired, but some of which are not. That is where a genuine IPR/validity budget belongs. For this patent, the correct disposition is a one-line citation to the expiration date, not a PTAB proceeding.

5. If any docket entry does turn up (e.g., a petition filed and denied pre-2015), remember it is non-precedential as to estoppel. A denial of institution produces no § 315(e)(2) estoppel, so it would not narrow your invalidity grounds in any event — a fact worth stating in a footnote if opposing counsel ever cites a denial as if it were a merits win.

---

Sources cited

• US6667684B1 — Google Patents (authoritative text; expiration and assignment data)
• US-6667684-B1 — Unified Patents patent portal (no PTAB proceeding listed; date-encoding discrepancy noted above)
• Remote controlled garage door opening system — FreePatentsOnline (specification corroboration)
• Chamberlain v. Overhead Door, No. 2:21-cv-00084 (E.D. Tex.) — case reporting (unrelated patents; context only)

No Final Written Decision, institution decision, settlement, or Federal Circuit opinion exists to link to, because no PTAB proceeding was ever filed — that is the finding, and I have not manufactured citations to fill the format.

I'll attempt to pull the recorded assignment chain for US 6,667,684 from the USPTO Assignment Center and cross-reference it against public databases.

Research-integrity note (read first)

I could not retrieve a single reel/frame record for US 6,667,684. I have no browser that can POST to the Patent Assignment Search backend; my only access is web search, and repeated queries for "6667684" + assignment/reel/frame returned only patent-text databases, litigation PDFs for other patents (e.g. Acacia's Monarch Networking chain at reels 051192/0596, 051238/0718), and general USPTO how-to-search material — nothing on this patent's chain. I am therefore not going to invent reel/frame entries. Everything below that is labeled "not verified" is genuinely unverified, and the Assignment Center is the place to close the gap.

One thing the authoritative front page does establish: the printed patent names two original assignees, Overhead Door Corp. and Microchip Technology Inc. Because USPTO front-page assignee data is populated from the recorded assignment, the existence of at least one recorded inventor→co-assignee conveyance is strongly implied even though I could not pull its reel/frame.

Conflict flagged (carried forward from the prior section): the prior summary correctly notes Unified Patents lists only "Overhead Door Corp" as original assignee while listing both Sanwa Holdings Corp. and Microchip Technology Inc. as parent companies. The patent front page names Overhead Door Corp and Microchip Technology Inc. as original assignees. I treat the front page as authoritative and flag the database as under-inclusive.

---

Inventors

Unusual-pattern assessment — nothing that fits the fire-sale tell. The classic red flag you asked about (all inventors leaving the assignee within 12 months of filing) is not observable, because there is no post-issuance assignment traffic in any database I could reach and therefore no departure-driven transfer to detect. What is notable, and worth recording for the file, is a serial-inventor pattern rather than a departure pattern: Bruwer's code-hopping subject matter has moved through a chain of corporate vehicles over 30 years — Nanoteq (US 5,517,187, 1990 priority) → Microchip (US 5,687,904 / 6,175,312) → Microchip + Overhead Door jointly ('684 family) → Azoteq (citing '684). That is a technology-licensing career, not a patent-troll footprint, but it explains why the same cryptographic architecture recurs across nominally unrelated assignees.

---

Original assignee

Joint: Overhead Door Corp. and Microchip Technology Inc. (both named on the issued front page; Google Patents lists both as current assignees as well).

• Overhead Door Corp. — primary line of business: manufacture of upward-acting sectional garage doors and automatic garage-door/barrier operators (Genie, Overhead Door, Wayne Dalton brands). Corporate status: operating, and a subsidiary of Sanwa Holdings Corp. (Japan), which Unified Patents lists as the parent. The claims (a transmitter plus a receiver that actuates a garage-door drive mechanism) are directly embodied in the company's commercial opener products, and the specification's disclosure of a "learn mode button 60" and "flash indicator 60′" on the power head describes a shipping product architecture, not a paper design.
• Microchip Technology Inc. — primary line of business: semiconductor manufacturer (NASDAQ: MCHP), Chandler, AZ. Status: operating. It is the relevant co-owner from a product standpoint: the named-inventor cluster in the Phoenix metro plus Microchip's own '904/'312 code-hopping patents make clear that the transmitter/receiver non-linear code-hopping functions claimed here map to Microchip's encoder/decoder IC product line (the HCS/KeeLoq family). This is a component-supplier + OEM joint ownership, which is the single most important structural fact in this chain: two product-shipping companies co-own the patent.
• No bankruptcy, no dissolution, no acquisition-driven patent sale is evidenced for either co-owner in any source I reached.

---

Assignment timeline

I could not obtain any reel/frame data. I will not manufacture entries to fill this section.

What is affirmatively documented from the authoritative patent text and Google Patents legal events:

• 1996-09-06 — Priority application US 08/706,682 filed (issued as US 6,049,289 on 2000-04-11), Overhead Door Corp. + Microchip Technology Inc. as assignees.
• 2000-03-08 — Application at bar, US 09/521,510, filed as a continuation of 08/706,682 (i.e., filed roughly one month before the parent issued — a continuation filed to keep the disclosure pending).
• 2000-04-11 — Parent US 6,049,289 issues.
• 2003-12-23 — US 6,667,684 B1 issues.
• 2003-10-24 — Child continuation US 10/693,748 filed (published as US 2004/0085185 A1); abandoned.
• 2016-09-06 — Anticipated expiration; status Expired – Lifetime. All maintenance-fee obligations were satisfied to full term (the patent did not lapse early).

Google Patents' Legal Events tab for this patent records no ownership-change entries — only filing, priority, grant, publication, and expiration events. That is not the USPTO Assignment Center and is not a substitute for it, but it is weak corroboration that the joint original ownership was never reassigned post-issuance.

What still needs to be pulled from Assignment Center to complete this section (search by patent number at https://assignmentcenter.uspto.gov/ or legacy https://assignment.uspto.gov/patent/index.html): the reel/frame of the inventor→Overhead Door + Microchip assignment (likely recorded circa 1996–2000, on the 08/706,682 parent, and carried into '684 via the continuation); whether a separate recordation covers 09/521,510 specifically; whether any Confirmatory Assignment / Change of Name was recorded when Overhead Door's parent structure was renamed (Sanwa Shutter → Sanwa Holdings); and whether the abandoned 10/693,748 has independent recordings. Until those are read, any statement about correspondents of record would be speculation.

---

Timeline diagram

timeline
    title Ownership of US 6667684
    1996 : Priority application filed
         : Joint assignees Overhead Door and Microchip
    2000 : Parent patent 6049289 issues
         : Continuation 09521510 filed
    2003 : US 6667684 issues
         : Child continuation filed and later abandoned
    2016 : Patent expires
         : No ownership change recorded

---

NPE / troll-pattern signals

1. Shell-entity transfer — not present. No transfer to any "IP / Holdings / Licensing / Ventures" entity appears in any source reached. Current assignees remain the two original operating companies per Google Patents. Caveat: this call rests on absence of evidence, not on a read Assignment Center abstract of title — so treat it as "not observed" rather than "affirmatively refuted."

2. Known asserter in the chain — not present. Neither Overhead Door Corp. nor Microchip Technology Inc. appears on any public NPE roster (Acacia, Marathon, IV, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Spangenberg entities). Both are product-shipping manufacturers. The Acacia/Monarch reel references that surfaced in my searches belong to unrelated patents.

3. Repeat correspondent across the chain — unclear / not verifiable. This is the signal I most wanted and could not test: no correspondence data was retrievable. I decline to name a correspondent of record for any link.

4. Cascading transfers — not present. No sequence of assignments at all was found, let alone chained LLC transfers inside 24 months.

5. Pre-litigation transfer — not present. No assignment identified, and the patent expired 2016-09-06, so it cannot support prospective injunctive relief in any current action. For completeness, the prior section already flagged (correctly) that the 2022–2023 barrier-operator appellate activity (Fed. Cir. Nos. 22-1664 / 22-1656, and No. 23-1964) involves different patents and could not be tied to '684; an expired patent could at most support a past-damages tail, which would be an unusual posture for a 2026 docket.

6. Bankruptcy fire-sale — not present. No Chapter 7/11 proceeding involving Overhead Door Corp. or Microchip Technology was found, and no patent-sale order is referenced anywhere in the record.

7. Privateering — not present. There is no NPE downstream to which an operating company could have transferred the patent for assertion-on-behalf-of.

8. Defensive aggregator (anti-NPE) — not present. No chain terminus at RPX, AST, LOT Network, Unified Patents, or OIN. Note the asymmetry: Unified Patents maintains a "Challenge" tracking page for US-6667684-B1 in its patent portal, but that is Unified monitoring the patent, not owning it — do not conflate the two. (This distinction matters precisely because Unified's portal lists Overhead Door's parent Sanwa Holdings; that is a corporate-ownership field, not an acquisition.)

---

Verdict

Insufficient data.

Justification: No recorded assignment — reel/frame, execution date, correspondent, or conveyance type — was retrievable for US 6,667,684 in any source I could reach, so the section you asked me to build cannot be built without fabrication, and the alternative signals are all calls from absence of evidence rather than from a read abstract of title. The only ownership facts I can stand behind are non-NPE in character: the patent issued 2003-12-23 naming two original and apparently still-current assignees, Overhead Door Corp. and Microchip Technology Inc. — both operating, both product-shipping, one a barrier-operator OEM and one a code-hopping semiconductor vendor — joint ownership that is entirely consistent with the named-inventor split (Bruwer/Marneweck/Delport in Microchip's Chandler/Phoenix metro; Waggamon in Overhead Door's Ohio footprint). Every positive NPE indicator (shell transfer, known asserter, cascading LLCs, pre-suit transfer, bankruptcy sale, privateering, defensive-aggregator terminus) came back not present, and the patent expired 2016-09-06, eliminating any forward-looking assertion value.

To convert this to a defensible verdict, run one query and re-open the file: search patent number 6,667,684 at https://assignmentcenter.uspto.gov/ (legacy: https://assignment.uspto.gov/patent/index.html), open each reel/frame abstract, and capture the correspondent of record — that single field is what would either confirm clean OEM/component co-ownership or expose a reassignment I could not see.

I'll research the prior-art citations for US 6,667,684, grounding on the authoritative text supplied and verifying key references via search.

Let me verify the most §102-relevant learn-mode and code-hopping references.

Prior Art Analysis — U.S. Patent No. 6,667,684 ("Remote controlled garage door opening system")

0. Scope, search confirmation, and how I'm mapping "which claims"

Search confirmation (USPTO number-specific). I searched for the literal number 6667684 and its formatted variants. The USPTO-derived record confirms: US 6,667,684 B1, application 09/521,510, filed 2000-03-08, granted 2003-12-23, priority 1996-09-06, inventors Waggamon / Marneweck / Delport / Bruwer, original assignees Overhead Door Corp. and Microchip Technology Inc., status Expired – Lifetime (anticipated expiration 2016-09-06). No similar-number hits were substituted.

Critical mapping point (built on the earlier summary, not repeated). This patent has exactly one claim — claim 1 — and no dependents. Therefore every §102 statement below necessarily maps to claim 1 and only claim 1. There are no dependent claims to expand or qualify the scope.

For compactness I use this element key for claim 1 (full plain-language version is in the earlier "Patent summary" section):

Because claim 1 is a single apparatus claim reciting both transmitter and receiver structure plus learn/storage/decode behavior, an anticipation (§102) case requires one reference disclosing all of E1–E6. That is a high bar, and I flag up front: no cited reference appears to meet it. The analysis below identifies the references that come closest and the elements each one bears on.

Applicable law: priority is 1996-09-06, so pre-AIA 35 U.S.C. §102 governs.

---

A. Tier 1 — Closest references (code-hopping key generation at learn time)

A1. U.S. Pat. No. 5,686,904 — "Secure self learning system"

• Assignee/inventor: Microchip Technology Inc. / Frederick Johannes Bruwer (a named '684 inventor and a '684 co-assignee)
• Filing: 1994-09-30 (Ser. No. 08/313,613); Granted: 1997-11-11; priority 1991-05-29
• Description: Encoder/decoder code-hopping system in which the encoder stores a serial number and either a key generation seed or a user key derived from a manufacturer's master key plus the seed/serial number. During learning, the encoder transmits the seed and serial number; the decoder independently generates the decoder key using its stored manufacturer's master key and the received information. A double-transmission verification step checks decoded counter information before accepting the learning. In normal operation the decoder decodes the hopping code and compares the decoded counter against a stored counter.
• §102 basis: §102(e) (U.S. application filed 1994-09-30, before the 1996-09-06 priority).
• Element mapping: Discloses or strongly suggests E1, E2, E3, E4 and the counter-window portion of E6.
• Why it does not anticipate: It matches the incoming serial number to a stored parameter set and uses that set's key (key-address management / pointer), rather than the claim's sequential trial of every stored secret key until a window match. And its memory allocation uses a learning pointer cycled through memory banks, not the claim's random storage in an unused discrete location (with random replacement when full) (E5). No single-reference anticipation; this is the strongest §103 candidate.
• Source: https://patents.google.com/patent/[US5686904](/patent/US5686904) ; https://uspto.report/patent/grant/5686904

A2. U.S. Pat. No. 5,517,187 — "Microchips and remote control devices comprising same"

• Assignee/inventor: Nanoteq (Pty) Limited / Bruwer, Smit, Kuehn
• Filing: 1993-02-18 (Ser. No. 08/019,821); Granted: 1996-05-14; priority 1991-05-29 (ZA 90/4088) and 1991-05-29 (US 07/707,101)
• Description: Encoder/decoder microchips for remote control. A non-linear encoding function acts on an identification number (PIN, preferably 64-bit) and a combination of a unit number and a stepping counter value to produce a transmission value. The decoder decodes it and compares the decoded counter value against a defined (forward-only) range / window; claim 18 recites storing "a plurality of counter values, each being related to a different encoder microchip."
• §102 basis: §102(a) (granted 1996-05-14, before the 1996-09-06 priority); its EPO counterpart EP 0 459 781 A1 published 1991-12-04, which is §102(b) art.
• Special status: The '684 specification expressly incorporates U.S. 5,517,187 by reference and states the encoding/decoding functions "employ novel variations of the code hopping technique disclosed in" it. That makes the Bruwer '187 disclosure admitted prior art for the code-hopping framework regardless of the §102 category.
• Element mapping: E1 (serial/PIN, counter, non-linear encoder, hopping), E2, and the window-comparison portion of E6.
• Why it does not anticipate: It discloses no manufacturer's key stored at the factory from which the receiver independently derives a secret key during learn (E3/E4), and no random storage in an unused discrete location (E5). Multiple encoder counter values are selected by decoded unit number, not by sequential trial.
• Source: https://worldwide.espacenet.com/publicationDetails/biblio?CC=US&NR=5517187

A3. U.S. Pat. No. 6,049,289 — "Remote controlled garage door opening system" (Overhead Door Corp.)

• Priority: 1996-09-06; filed 1996-09-06 (Ser. No. 08/706,682); Granted: 2000-04-11
• Status relative to '684: This is the parent application of the '684 continuation (the '684 specification says so expressly). It shares the identical 1996-09-06 priority and the same disclosure.
• §102 finding: NOT prior art under §102 against claim 1. Same inventive entity/family and same effective filing date; a parent is not available as §102 art against its own continuation. Note that the '684 front page lists US6049289A under "Family Cites Families," which is a family relationship, not a §102 citation; I flag this so it is not misread as an anticipating reference.

---

B. Tier 2 — Learn-mode + rolling/changing-code references

---

C. Tier 3 — General security / code-learning / data-transmission references

(C19 marked N/A; not a separate reference — retained only to keep the numbering transparent.)

---

D. Tier 4 — Baseline fixed-code / digital-radio-control references (full citation list)

These are the "recognition of a received signal as being from a particular transmitter" and early digital radio control references. They are relevant only as background showing the state of the art before code hopping; none approaches E4–E6.

(D18–D20 placeholders removed where duplicative; consolidated above.)

Additional non-US and miscellaneous cited documents

---

E. §102 synthesis — reference-by-reference verdict

Because only claim 1 exists, the question for each reference is whether it discloses all of E1–E6. Grouping the strongest candidates:

Bottom line on §102: On the record of the '684 citation list, no single cited reference anticipates claim 1. The two references that come closest to the key-generation-at-learn concept — U.S. 5,686,904 and U.S. 5,517,187 (with its EP 0 459 781 counterpart) — both fail on (E5) random storage in an unused discrete location with random replacement, and on the sequential-trial-of-stored-secret-keys implementation of E6. Those two gaps are the load-bearing distinctions of claim 1.

§103 note: The strongest obviousness combination would be U.S. 5,686,904 in view of U.S. 4,750,118/RE35,364 (learn-mode, multi-transmitter receiver) and optionally U.S. 5,517,187 (hopping/window). The novelty of claim 1, if any, resides in the random unused-location storage (E5) and the sequential decryption through all stored secret keys (E6).

---

F. Flags, anomalies, and caveats (per the strict no-auto-correct rule)

1. US6049289 is not §102 art. It appears in the '684 front-page family listing as the parent (Ser. No. 08/706,682 → US 6,049,289). Same family, same priority (1996-09-06). Do not treat it as anticipating art against claim 1. This is the single most likely misreading of the '684 citation table.

2. Apparent mis-citation in the '684 reference table. Cited entry US5053701A — Voest-Alpine Zeltweg Gesellschaft M.B.H. — is titled "Device for measuring the spacing between aligned rails" (pub. 1991-10-01). That subject matter is unrelated to remote-control coding and appears to be an erroneous citation in the examiner/front-page listing. I am reporting it as cited, not correcting it.

3. Erroneous-looking non-US entries in the citation table. EP 0 335 912 A1/B1 is listed in the '684 table as "Counting system with buyer cards coded before payout" (Joseph W. Sloan) and in an expanded list as "Prepayment metering system using encoded purchase cards" (CIC Systems, Inc.). It is an encoded-card counting system; it has no apparent bearing on barrier-operator code hopping. Reported as cited.

4. Date-conflict note (repeated from the earlier summary, still unresolved). Google Patents gives filing 2000-03-08 / grant 2003-12-23 / priority 1996-09-06. Some aggregators show one-day earlier offsets (2000-03-07 / 2003-12-22 / 1996-09-05), likely time-zone/normalization differences. I use the Google Patents/USPTO front-page figures as primary and flag rather than silently resolve.

5. §102 category depends on asserted priority. I applied a 1996-09-06 effective date (pre-AIA). If any claim element were shown to lack written-description support in the parent '289, the '684's effective date could shift toward the 2000-03-08 filing, which would move U.S. 5,636,913 and U.S. 4,975,942-era references from §102(e) into §102(b), and would also make the '289 itself (2000-04-11) potentially problematic if the benefit claim failed. I am not asserting that the priority claim fails; I flag the dependency.

6. Filing dates not individually verified for every citation. The '684 front-page citation table lists priority date and publication date, not filing date, for most entries. Where I obtained a verified filing date (e.g., '904 filed 1994-09-30; '187 filed 1993-02-18; '668 filed 1994-06-15; '960 filed 1991-08-26; '159 filed 1989-04-26; '225 priority 1989-05-18), I report it explicitly. For the remainder, the publication date in the table is the operative date for §102(a)/(b), and I state it as such. I did not fabricate filing dates for references where I could not confirm them.

7. Only one claim. Any subsequent validity challenge (IPR/PGR) or litigation would rise or fall entirely on claim 1; there is no dependent-claim fallback.

Primary sources: the US6667684 full text supplied (front-page citation table + specification) at https://patents.google.com/patent/US6667684/en ; and, for verified details on the closest references, https://patents.google.com/patent/US5686904 , https://uspto.report/patent/grant/5686904 , https://worldwide.espacenet.com/publicationDetails/biblio?CC=US&NR=5517187 , https://patents.google.com/patent/[US5471668A](/patent/US5471668A)/en , https://patents.google.com/patent/[US5365225A](/patent/US5365225A)/en , https://patents.google.com/patent/[US5252960](/patent/US5252960) , and https://patentimages.storage.googleapis.com/77/7d/07/76da160671de94/[US5148159](/patent/US5148159).pdf .

I'll verify the content of the key prior-art references before building the combination analysis, since the quality of a § 103 analysis depends on accurately characterizing each reference.

The two most important references are confirmed and their claims are now verified verbatim. Let me verify two more references that supply the remaining claim elements.

Obviousness Analysis — U.S. Patent No. 6,667,684 under 35 U.S.C. § 103

Building on prior sections. This analysis assumes the claim decomposition, bibliographic findings, and litigation posture already established (only one claim; no known assertion of this patent; expired 2016-09-06). It does not repeat them. Per the task instruction, the prior-art universe used here is the reference set listed on the patent's own Google Patents/patent-family page (the "Citations (56)," "Patent Citations (58)," and "Family Cites Families (13)" tables), plus the two references cited in the specification body.

Two premise notes (flagged, not silently corrected):

1. The prompt materials carry conflicting dates (system header: 2026-09-30; task block: 2026-04-26). Nothing below turns on the date except the patent's expiry, which is 2016-09-06 on both.
2. The citation list is not a curated § 103 statement. It contains clear noise — e.g., US 5,053,701 (rail-spacing measurement, Voest-Alpine) and JPS5520830A (English title rendered "Lockkup device," apparently garbled). I treat those as non-substantive.

---

1. Legal standard applied

• Graham v. John Deere Co., 383 U.S. 1 (1966): scope and content of the prior art; differences between the prior art and the claims; level of ordinary skill; secondary considerations.
• KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007): a combination is obvious when the elements were "known in the field and the combination was for known reasons to achieve a predictable result"; an "explicit teaching, suggestion, or motivation is not necessary"; a "predictable variation" of a known technique is obvious where the improvement is one of several known options and the skilled artisan has good reason to pursue it.
• Rationale categories of MPEP § 2143 and the "repetition of known steps" and "obvious design choice" categories of MPEP § 2144.01/2144.04.
• All claim limitations must be found; the claim is a combination in which no single reference need teach every element, but the combination must have been suggested or obvious.

---

2. Level of ordinary skill in the art (POSITA)

A person with a Bachelor's degree in electrical engineering or computer science plus 2–3 years of experience designing RF remote-control or access-control systems (encoders/decoders, fixed- or rolling-code remote keyless entry), or equivalent. This is a modest skill level: by 1996, code-hopping encoders/decoders (the KeeLoq family, of which the '187 and '904 references are the foundational patents) were commercial, off-the-shelf integrated circuits.

---

3. The claim, decomposed

Claim 1 (the sole claim) has these limitation groups, per the authoritative text:

• (A) Transmitter data set: serial no. unique to the transmitter; secret key unique to it; synchronization value that increments each actuation; a non-linear function generator producing a multi-bit hopping code that changes each actuation; and transmission of the hopping code plus the serial number.
• (B) Receiver architecture: microprocessor operable between learn and operate modes; manufacturer's key stored at the factory.
  • (B-i) Learn: receive hopping code + serial → use stored manufacturer's key + received serial to independently generate the secret key, and thereafter a synchronization value, matching the transmitter's.
  • (B-ii) Storage: store so-generated secret key and sync value randomly in an unused discrete memory location, or, if all are used, replace the information in a randomly chosen location.
  • (B-iii) Operate: non-linear decode the intercepted hopping code with one stored secret key → second synchronization value; compare to stored sync value; issue the command if it bears a predetermined relationship and is within a window; if not, try the next stored secret key, and repeat until a match is found.

---

4. Prior-art arsenal (from the patent's own citation tables)

---

5. Element-by-element mapping — Primary combination: '187 + '904 + '118

Result: every limitation of claim 1 is disclosed across '187 + '904 + '118, and the combination requires no new mechanism.

---

6. Motivation to combine (articulated, KSR-compliant)

1. Same field, same problem. All three are directed to RF remote control of garage doors/gates and access control, and all three address the identical two problems the '684 specification frames: (a) code grabbing/unauthorized access, and (b) user-unfriendly programming (mechanical switches).
2. Express cross-references create the motivation. '904's own background expressly cites the learn-mode art — "A switch can be used to set the decoder either in a normal operation mode or in a learning mode (U.S. Pat. Nos. 4,750,118 and 4,912,463)"; "In another invention (refer to U.S. Pat. No. 5,148,159)…" — and then states: "Learning has been implemented in standard fixed code systems, but code hopping systems present a unique challenge… The invention is directed to overcoming this problem." That is the motivation to marry the '904 secure-learn key derivation to a '187 code-hopping system.
3. Same inventors / common development lineage. '187 and '904 share inventor Bruwer (Nanoteq → Microchip/KeeLoq). Combining a foundational code-hopping patent with the same inventor's later secure-learn improvement is exactly the "familiar elements according to known methods" scenario.
4. Stated design goal in the prior art itself. '118's own summary states the object is to "eliminate code switches in the transmitter and receiver making it easier for a customer to install and operate." The '684 specification asserts the same benefit. A known problem with a known solution is the paradigm KSR case.
5. Predictable result. Combining (i) hopping-code transmission + counter-window validation, (ii) manufacturer-key-based key derivation on learn, and (iii) multi-code receiver memory yields nothing more than the expected aggregate: a switch-free, multi-transmitter, grab-resistant garage-door receiver. No reference teaches away.
6. Reasonable expectation of success. All mechanisms were integrated-circuit-implementable and, by 1996, commercially deployed in the KeeLoq encoder/decoder and Chamberlain learn receivers.

---

7. The two limitations that carry the case (and why they are still obvious)

7.1 "randomly in an unused discrete memory location … replacing the information in the randomly chosen memory location"

This is the sharpest limitation; neither '904 (deterministic pointer/index) nor '118 (manual selector + wrapping pointer) uses the word "random." The obviousness theories:

• "Variety of schemes" express invitation. '904 states the allocation pointer "can be managed according to a variety of different schemes. Options include, inter alia, cycling the pointer through the available memory locations and allowing the user to set the pointer from outside." A reference that expressly enumerates interchangeable allocation schemes renders selection of another scheme (random) an obvious design choice (MPEP § 2144.04).
• Predictable variation. Replacing round-robin allocation with pseudo-random slot selection over seven slots is a "predictable variation" with no functional difference beyond which slot is overwritten. The claim recites no technical consequence of randomness; the specification ties it solely to the convenience objective of avoiding a selector switch.
• Express random-number teaching in the same art class. '365225 discloses a random generator in the transmitter used in the initialization/re-initialization of a garage-door/lock authorization code set. A POSITA seeking a non-deterministic (hence less predictable/exploitable) allocation would be directed to it.
• Replacement-on-full was already known. '118: wrapping code-location pointer and "erase all old codes ... before storing a new one."

7.2 "sequentially … performing another non-linear decoding function … using a different stored secret key … continually repeating … until"

• '187 claims 18/19 teach storing a plurality of decoder values and selecting among them based on information only recoverable after decoding — which, in a multi-key receiver, entails decoding with more than one candidate key. That is the sequential-trial concept.
• '187's "scan its input for another code word" on failure shows iteration was the accepted handling of an unsuccessful check.
• Repetition of a known step. Applying the same decode-and-window-check to the next stored key is a textbook "repetition of known steps … with predictable results" (MPEP § 2144.01(III)); with n ≤ 7 the cost is trivially predictable.
• Design-alternative point. '904 selects the key by the transmitted serial; '187 selects after decoding. These two approaches coexisted in the same art/patent family, so choosing the index-free sequential approach was an available, obvious alternative — not a new principle of operation.

---

8. Alternative and back-up combinations

• Alternative 1 (fewer references): '904 + '187. '904 supplies the entire learn-mode key-derivation and multi-slot storage scheme; '187 supplies the non-linear hopping code and the decode/compare/activate-within-a-window and multiple-stored-value architecture. '118 is needed only for "garage door operator," which '904's background already names.
• Alternative 2 (adds learn-mode convenience art): '904 + '187 + '118 + '159 ('159 and '992 teach user-initiated learn mode for multi-transmitter receivers) — useful if the patentee disputes that '118's switch-based learn is combinable.
• Alternative 3 (random-slot theory): '904 + '187 + '118 + '365225, as in § 7.1.
• Alternative 4 (all secondary: add '985 / '103221 / '5635913 for the general "change the code every actuation" background.

No combination in the list requires the references to be bodily incorporated; each supplies a discrete, physically-compatible function (a counter, a key-generation routine, a memory-allocation routine, a comparison routine).

---

9. Anticipation check (§ 102) — brief

No single reference anticipates claim 1:

• '904 does not disclose ignoring the transmitted serial in favor of sequential key trial, nor random slot allocation.
• '187 does not disclose derivation of a per-transmitter key from a factory-stored manufacturer's key plus a transmitted serial during learning.
• '118 is a fixed-code system with no non-linear hopping code and no windowed counter check.

Accordingly the case is a § 103 case, not § 102. Note also that the parent US 6,049,289 (issued 2000-04-11) is not § 102(e) art against '684 because it is not "by another" (same inventive entity/continuation lineage).

---

10. Secondary considerations and anticipated patentee rebuttals

• No unexpected results are disclosed. The specification characterizes the encoding/decoding as "novel variations of the code hopping technique disclosed in U.S. Pat. No. 5,517,187" — a near-admission that the operative cryptography is the incorporated prior art.
• Long-felt need cuts toward obviousness. The switch-elimination need is identified as early as '118 (1988), and '904 answered it in 1994; the '684 improvement is at most an incremental implementation choice.
• No commercial-success nexus established. Any proffered success would need to be tied to the claimed random-allocation/sequential-trial features specifically (cf. In re GPAC; Wm. Wrigley), not to garage-door openers generally. No evidence of copying, licensing, or industry praise was located.
• Teaching away? None of the references disparages the combination. The best patentee argument is that '904 "relies on" serial-indexed key selection and therefore teaches away from index-free trial decryption. That is a weak form of teaching away: '904 never criticizes trial decryption and, as noted, the index-free approach was already embodied in '187 — the same family of art.
• Change in principle of operation? Not available: the '684 transmitter still transmits the serial number (FIG. 7a), so the '904 serial-indexing capability is preserved as an option.

---

11. Weaknesses in the prima facie case / evidentiary gaps (flagged, not papered over)

1. The examiner had most of this art and still allowed the claim. '904, '187, '118, '365225 and the learn-mode family all appear on the '684 face, and the same primary examiner (Horabik) examined both '365225 and '684. That is a meaningful caution: a prima facie § 103 case must explain, in a claim chart, why the "randomly" limitation and the sequential-trial limitation were nonetheless obvious — the very points argued in § 7. An examiner's allowance is not controlling, but it identifies where the dispute will be.
2. Full corroboration of '365225 is partial. I verified its abstract and background only; I did not retrieve or read its 44 claims, so I do not assert that they claim a receiver-side random allocation. It should be used only for the proposition stated (a random generator used to initialize a transmitter/receiver authorization code set in a garage-door/lock context).
3. "Randomly" is genuinely absent from the two workhorse references. The '904 "variety of schemes" passage is the strongest textual hook; if a decision-maker demands a hardware/algorithmic teaching of random slot selection in a learn-mode receiver, the '365225-based rationale is not a perfect substitute. This is where an IPR may live or die.
4. Motivation evidence is inherent, not documentary. Apart from '904's background citations, the combination rests on field identity, common-inventor lineage, and KSR "design incentive" reasoning rather than an explicit "combine A with B" statement, which some panels still scrutinize.
5. Practically, the patent is expired (2016-09-06). An IPR would be a retrospective exercise (e.g., to clear past damages exposure or for estoppel/related-family reasons); the Board may also be unreceptive where no live exclusion/damages dispute remains. This is context, not a legal bar.

---

12. Bottom line

Claim 1 is, on this record, vulnerable to a § 103 obviousness attack, and the attack is defensible with the patent's own cited art. The primary combination is US 5,517,187 + US 5,686,904 + US 4,750,118, with US 5,365,225 supplied for the "random" allocation theory and US 5,148,159 / US 4,988,992 as optional learn-mode support. The combination is supported by (i) identical field and problem, (ii) the '904 background's express reliance on the '118/'159 learn-mode art and its statement that marrying learning to code hopping was the unmet need, (iii) the shared inventorship of '187 and '904, (iv) predictable results, and (v) '118's own stated objective of eliminating switches. The two limitations that will bear the weight are the random memory allocation ("obvious design choice / predictable variation," buttressed by '904's "variety of different schemes" and '365225's random generator) and the sequential trial of stored keys ("repetition of known steps," buttressed by '187 claims 18–19 and its multi-stored-value architecture).

Recommended next steps if this is taken further: (1) build a limitation-by-limitation claim chart using the verbatim claim language of '187 (claims 17–19, 23, 25, 28–29), '904 (claims 1, 4, 8–10, 12), and '118 (claims 1–4 + the pointer-wrap/erase description); (2) obtain the full claim set of '365225 (and consider whether an additional reference expressly teaches pseudorandom slot selection in a receiver) to close the "randomly" gap; (3) pull the '684 prosecution history to see precisely which limitation the examiner relied on for allowance; and (4) confirm via PTAB/PAIR whether any reexamination or IPR has already been filed, since none was found in the prior litigation search.

---

Sources (accessed for this analysis)

• US6667684B1 full text — https://patents.google.com/patent/[US6667684](/patent/US6667684)/en
• US 5,686,904 claims & description — https://uspto.report/patent/grant/[5686904](/patent/5686904) ; https://www.everypatent.com/comp/pat5686904.html ; https://lexdana.ai/patent/05686904
• US 6,191,701 B1 (same-family continuation, corroborating the secure-learn disclosure) — https://www.freepatentsonline.com/[6191701](/patent/6191701).html
• US 5,517,187 claims & description — http://www.everypatent.com/comp/pat5517187.html ; https://uspto.report/patent/grant/5517187 ; https://worldwide.espacenet.com/publicationDetails/biblio?CC=US&NR=5517187
• US 4,750,118 claims & description — https://uspto.report/patent/grant/4750118 ; https://patentimages.storage.googleapis.com/60/13/3b/995318fef46945/US4750118.pdf
• US RE37,986 E (reissue of '118) — https://patentimages.storage.googleapis.com/3b/50/d6/b719af8099edd5/[USRE37986](/patent/USRE37986).pdf
• US 5,365,225 abstract/background — https://patentimages.storage.googleapis.com/81/6e/76/ec55e4b6051787/US5365225.pdf
• KeeLoq background (Nanoteq → Microchip; Bruwer/Kuhn/Smit) — https://web.archive.org/web/20231207092612/https://en.wikipedia.org/wiki/KeeLoq