US 6478404

I'll search for the specific patent number and any related litigation/docket records.

Let me run a couple more targeted searches for litigation/docket records.

Search results: US Patent 6,478,404

Caveat on docket search: I searched for the specific number in connection with CAFC docket activity in 2026 and with USPTO litigation records. I found no authoritative CAFC 2026 docket entry, appeal, or ITC/District Court litigation record naming US 6,478,404. The only hits returned were (a) the patent's own family documents, (b) unrelated HP patent litigations (Nokia, Sisvel/Wilus, R2 Semiconductor, Oracle/Terix), and (c) a USPTO PTAB petition paper (petition 1537806) that cites Torgerson's US 6,478,404 as background art in a third party's prosecution argument — not as an asserted patent. I cannot confirm any 2026 appellate activity; treat "no docket found" as a negative search result, not proof that none exists.

Bibliographic data (from the patent record)

Source: https://patents.google.com/patent/US6478404/en (metadata as provided); US 6,860,587 record via https://webapp1.dlib.indiana.edu/virtual_disk_library/index.cgi/[5628977](/patent/5628977)/FID979/OG/html/1292-1/us06860587-20050301.html

Abstract (verbatim)

"An ink jet printhead having three columnar arrays of drop generators that produce drops having an drop volume that enables multi-pass printing at a print resolution that is not less than 1/(2P) dpi along a reference axis, wherein P is a drop generator pitch, and three columnar arrays of FET drive circuits respectively adjacent the columnar arrays of drop generators."

Plain-language overview of the independent claims

Claim 1 — the "narrow three-color printhead" claim

An ink jet printhead built on a thin-film die, characterized by:

1. A printhead substrate made of multiple thin film layers (silicon die with resistor, metallization, passivation layers).
2. Three side-by-side columnar (vertical) arrays of drop generators running along the die's length.
3. Each column provides a different ink color (e.g., cyan/magenta/yellow) and contains at least 96 drop generators, spaced at a pitch P.
4. The three columns are packed no more than 1,060 µm apart — this is the core "narrow" limitation that keeps the die small.
5. Drop volume is sized to allow multi-pass (shingled) printing at a resolution not less than 1/(2P) dpi along the axis parallel to the die's length — i.e., the small drops let you interleave passes to double (or better) the native nozzle resolution.
6. Three columnar arrays of FET drive circuits sit on the same die, one adjacent each drop-generator column, to switch the heaters.

In short: a compact, three-color thermal ink jet die where narrow column spacing + small drop volume + integrated per-column FET drive arrays deliver high effective print resolution without a wide (expensive/fragile) die.

Claim 21 — the "energy-compensated narrow three-color printhead" claim

Contains all of the structural features of claim 1 (thin-film substrate; three side-by-side columns, different colors, ≥96 generators each, pitch P; ≤1,060 µm column separation; drop volume enabling ≥1/(2P) dpi multi-pass printing; three adjacent FET drive circuit arrays), plus three additional limitations:

1. Each drop generator includes a heater resistor with resistance of at least 100 ohms (a high-resistance heater that permits smaller drive FETs).
2. Power traces connect the drop generators and the FET drive circuits.
3. The FET drive circuits are configured to compensate for variation in the parasitic resistance presented by those power traces — i.e., the drive transistor's own on-resistance is deliberately varied along the column so that heater-plus-trace total resistance is roughly uniform, yielding uniform drop energy across the array.

In short: claim 1's narrow three-color architecture combined with high-value heater resistors and location-dependent FET on-resistance trimming to equalize delivered energy.

Dependent-claim highlights (non-exhaustive): P in the 1/300"–1/600" range (cl. 2, 22); drop volume 3–7 picoliters (cl. 3, 23); ground busses overlapping FET active regions (cl. 5, 24); on-resistance < 250,000 Ω·µm²/A (cl. 6, 25), gate oxide ≤ 800 Å (cl. 7, 26), gate length < 4 µm (cl. 8, 27); on-resistance ≤ 14 Ω or ≤ 16 Ω (cl. 9/10, 28/29); FET arrays in regions ≤ 220 µm or ≤ 350 µm wide (cl. 16/19, 34/35); substrate aspect ratio LS/WS > 2.7 with WS ≈ 4200 µm or ≈ 3400 µm (cl. 20, 36–38).

Things worth flagging (analyst observations)

• Two independents, both apparatus claims, with claim 21 adding the heater-resistance and parasitic-compensation limitations absent from claim 1. Claim 1 is the broader architectural claim.
• Antecedent-basis quirk: claims 17 and 18 recite "wherein WS is about 4200 / 3400 micrometers," but their parent claim 16 never introduces "WS." (The same pattern appears at claims 37/38 vs. 36, where 36 does recite LS and WS.) This is a potential § 112(b) issue I'm noting from the face of the claim text; I have no record of any court or PTAB addressing it.
• The continuation (US 6,860,587 B2) has a different claim 1 — it drops the drop-volume/multi-pass-resolution element and recites "at least 96 drop generators" without pitch P. So the two family members are not claim-identical.
• Cited prior art of record includes EP 0679518 A1 (Seiko Epson), US 5,713,673 (Sony), US 5,757,400 (Spectra — high-resolution matrix ink jet arrangement), US 6,325,489 B2 (Seiko Epson), and EP 0 955 166 A1 (Hewlett-Packard, thin film ink jet printhead). US 5,683,489 A also appears in the citation table but is titled "Method of producing iron carbide" — an apparently unrelated reference; I am not auto-correcting or reinterpreting that citation.
• Status: the record lists the patent as expired as of the 2021-01-30 anticipated expiration (20 years from the 2001-01-30 priority/filing date). I did not independently verify this against live USPTO Patent Center fee/maintenance records; the Google Patents listing shows maintenance-fee payments at the 4-, 8-, and 12-year intervals but no later entries.

Uncertainty statement: I have high confidence in the bibliographic data, abstract, and claim text above because they come from the full patent document supplied. I have low/no confidence regarding any 2026 CAFC appeal or litigation involving this patent — my search found nothing, and an expired patent's remaining assertion window would be narrow, so the appropriate conclusion is "no evidence found," not "confirmed none exists."

I'll search for litigation involving US patent 6478404.

Let me search more targeted litigation databases.

Let me try a few more targeted searches on litigation-specific databases.

Summary of Findings: Litigation Involving U.S. Patent No. 6,478,404

Bottom line: I found no litigation specifically involving U.S. Patent No. 6,478,404 in the searches available to me.

I ran multiple targeted searches against publicly indexed sources, including Google Patents (which aggregates litigation and family data for the patent), Unified Patents insight pages, PTAB/USPTO document repositories, and general web searches tied to the exact number "6,478,404." None returned a district court complaint, ITC investigation, PTAB proceeding, or appellate (CAFC) appeal asserting or challenging this specific patent.

Important caveats and context

1. The record I could retrieve (from Google Patents) shows no litigation events. For this patent, the Google Patents page lists only the prosecution/family history — original assignee Hewlett-Packard Company, current assignee Hewlett-Packard Development Company, L.P., an anticipated expiration of 2021-01-30, and status "Expired – Lifetime." It shows no "Litigation" or "Cases" section, which is where Google Patents normally lists any assertion of the patent.

2. This is consistent with how HP operated. Later enforcement by HP on inkjet printhead/cartridge technology tended to involve other, later patents in the same family line rather than the '404 patent itself. For example:

   • HP's ITC action against Asia Pacific Microsystems (APM), MicroJet Technology, and PTC Holdings (reported in the ITC publication I retrieved) asserted patents such as 6,412,917; 6,???,053; 6,???,347; 6,???,817; and 6,???,279 — not the '404 patent.
   • The Slingshot Technologies litigation against HP (as shown in the PTAB document I retrieved) asserted patents such as 7,018,012; 7,819,498; 7,841,712; and others — again, not the '404 patent.
   • A separate Dutch proceeding (HP v. Benson, ECLI:NL:RBDHA:2018:4810) involved EP 360, not the U.S. '404 patent.

3. Searches for near-identical numbers should not be substituted. I took care to look for the exact string "6,478,404 / 6478404." Matches I encountered either (a) referred to the patent itself as a patent document, or (b) were false positives from unrelated numbers (e.g., "647 Accused Product" in a Neapco/GM driveline case; "6478404" fragments in unrelated financial data). None represent litigation on this patent.

Why I cannot definitively say "no litigation exists"

• I did not have direct access to paid/closed litigation databases. The definitive sources for assertion history are Docket Navigator, Lex Machina, Unified Patents' litigation portal (portal.unifiedpatents.com/litigation/caselist — which I was unable to query directly), and PACER. My public-web searches are not a substitute for those.
• The patent expired on 2021-01-30. Any past litigation would therefore have occurred between issuance (2002-11-12) and expiration. Most assertion activity against manufacturers of compatible inkjet cartridges/printheads (e.g., the ITC actions) used different HP patents, but I cannot fully exclude a case that simply isn't well indexed on the open web.

Recommendation for a definitive answer

To confirm whether the '404 patent was ever asserted, I would recommend running the patent number through:

• Unified Patents litigation portal (https://portal.unifiedpatents.com/litigation/caselist)
• Docket Navigator / Lex Machina (patent → district court cases)
• PACER (party name searches: "Hewlett-Packard" plus patent number in complaint text)
• USPTO Patent Center / PTAB (to confirm no IPR/PGR/CBM or reexamination)

Based on the evidence I was able to gather, the best-supported conclusion is that there is no known litigation involving U.S. Patent No. 6,478,404, but I flag that this reflects the limits of my available sources rather than a certified negative search of commercial litigation databases.

I'll verify the ODP's "no proceedings" default against public sources before writing this up.

Let me run a couple more targeted checks to make sure nothing is missed.

One search hit referenced US 6,478,404 inside a PTAB petition file — let me pin down whether that's a challenge to this patent or merely a citation of it.

Proceedings overview

Zero AIA trial proceedings are on file for US 6,478,404. The canonical USPTO Open Data Portal list in the "PTAB proceedings on file" block is empty, my independent web checks (Google Patents, PTAB E2E petition documents surfacing in search, CAFC/§ 337 reporting, IPR decision PDFs) surfaced no IPR, PGR, or CBM naming US 6,478,404 / US 09/774,811 as the challenged patent, and no Federal Circuit appeal from a PTAB decision on this patent exists. Breakdown by status therefore is 0 active / 0 invalidated / 0 sustained / 0 settled / 0 institution-denied.

The bottom line for a defendant is not either of the two buckets in the framing question. This patent has never been tested at the PTAB at all — so it is neither "hardened by surviving IPRs" nor "already dead." But the more important fact is temporal: the patent expired 2021-01-30 (20 years from its 2001-01-30 filing date; Google Patents lists legal status "Expired – Lifetime" with "Anticipated expiration 2021-01-30"). An expired patent cannot support injunctive relief, and under 35 U.S.C. § 286 damages are limited to the six years preceding the complaint. On today's date, that lookback window (back to 2020-10-01) captures only about four months of pre-expiration conduct. Your practical exposure is a shrinking damages tail, not an injunction — and there is no PTAB estoppel map to worry about because nobody has petitioned.

No proceedings to itemize

I am deliberately not generating the per-proceeding ### {PROCEEDING_NUMBER} — {Petitioner} v. {Patent Owner} blocks, because doing so would require inventing proceeding numbers, panels, and institution dates. Per your constraint, I will not fabricate. Instead, here is what the record actually contains and what I ruled out.

Near-misses I checked and rejected (do not mistake these for IPRs on this patent):

• Slingshot Printing LLC v. HP Inc., W.D. Tex. (1:20-cv-00186-ADA). US 6,478,404 shows up in that litigation record — but as a cited Hewlett-Packard printhead patent in a document table ("Torgerson et al., US 6,478,404, Ink Jet Printhead, Nov. 12, 2002"), i.e., it is referenced, not asserted or challenged. The Slingshot/HP and Slingshot/Canon PTAB activity (e.g., IPR2023-00312, IPR2023-00313, IPR2023-00313's Ex. 1026 = the Slingshot v. HP claim construction order) concerns different HP-origin patents — U.S. 7,290,864, 7,484,823, 7,594,708 — not the '404. Source: https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1550720](/patent/1550720) and https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1537806](/patent/1537806)
• "The '404 patent" hits in PTAB papers by TQ Delta and by Wound Healing Technologies (Patent Owner's Discretionary Denial briefs). These are completely unrelated patents that happen to end in 404 (TQ Delta's transceiver patents; a wound-therapy patent). A plain-text "'404 patent" search is actively misleading here — flagging it so nobody on your team repeats the mistake. See e.g. https://content-archive.fast-edgar.com/20250801/A5Z6NQ2CZ222HZ4Z2V2A2ZZ2MTPQZM2IZ272/R16.htm
• ITC Inv. No. 337-TA-1015 (HP v. Memjet, 2016) and the HP v. Memjet/Driveprint/other printhead campaigns. No '404 claim appears in the asserted-claim lists I reviewed. Source: https://www.govinfo.gov/content/pkg/FR-2016-07-01/pdf/2016-15621.pdf

Where I looked: the ODP-derived structured block (canonical), Google Patents family and legal-status data for US 6,478,404 B2 (https://patents.google.com/patent/US6478404/en), USPTO PTAB petition document repository (https://ptacts.uspto.gov), the PTAB decisions library (https://www.uspto.gov/patents/ptab/decisions), and Federal Circuit materials via CourtListener (https://www.courtlistener.com). Confidence: high that no AIA trial exists on this patent. I did not independently verify the continuation US 6,860,587 B2 (filed 2002-10-04 as a continuation of this application, same family) for PTAB activity — treat that as unverified, and it is worth one direct E2E docket check if a demand letter ever cites the '587 in addition to the '404.

Strategic summary

Claim status: every claim of US 6,478,404 is UNTESTED. Claims 1–20 and 21–38 (the two independent claims and their dependents) have never been before the PTAB. There are no canceled claims, no sustained claims, and no adverse PTAB record to cite. That cuts both ways for a defendant: you cannot point to a Final Written Decision and say "claim 1 is dead," but the corollary is that the patent owner also has no PTAB win to brandish, and no institution decision or FWD has ever validated these claims against a skilled prior-art challenge. The nearest thing to a validity determination is prosecution: the file history shows the application was allowed over the art cited on its face (EP 0679518, US 5,713,673, US 5,757,400, US 6,325,489, EP 0955166 among six references of record). Likely prior-art candidates for a § 102/§ 103 attack are precisely those references plus the HP Journal articles and HP's own earlier printhead patents (US 4,719,477; US 5,317,346; US 5,469,199) that the specification incorporates by reference — a § 325(d) risk (same or substantially the same art already before the Office), so any petition should lead with art that is not of record and explain why the examiner materially erred.

Estoppel landscape: essentially blank — in your favor. Because no one has ever petitioned, 35 U.S.C. § 315(e)(2) estoppel binds no one here. There are no privity chains to trace, no "reasonably could have raised" bar, and no petitioner-side estoppel shutting down grounds in a district court. Every prior-art ground a defendant can develop is available. The real gates are not estoppel but procedure and economics: (1) § 315(b) — if you (or a privy/real party in interest) have been served with a complaint asserting the '404, your petition is due within one year of service, and that clock is unforgiving; (2) § 325(d) — the Office has already considered the references on the face of the patent, so you must show material examiner error under Advanced Bionics; and (3) the expiration date — with the patent expired since 2021-01-30, an IPR bought today invalidates nothing prospectively. Its only value is knocking out the claims that support the residual § 286 damages window, and that window shrinks every day.

Pattern signals: none. This is a single-assignment Hewlett-Packard-origin patent (original assignee Hewlett-Packard Company; current assignee Hewlett-Packard Development Co. LP per the ODP record) that has never been the subject of a serial-petitioner campaign, a defensive-aggregator challenge (no Unified Patents or RPX filing appears in any of the materials I reviewed), or PTAB appeal activity. There is no litigation-driven IPR cluster around it. It is a 2001-filing, 2002-issuance printhead patent whose 20-year term ran out in 2021 — which is the most likely explanation for why no IPR was ever filed: by the time printhead assertion activity heated up in the Slingshot/HP/Canon disputes, the patent's useful life was nearly or entirely over. That is the correct read of the empty docket, not "the patent is so strong nobody dared challenge it."

Recommended next steps

1. Treat "no PTAB activity" as the finding, but do not treat it as a defense. The absence of an FWD means you have no ready-made invalidity record and cannot invoke § 315(e)(2) against anyone. Do not represent to a court (or to your client) that claims have been canceled — they have not.
2. Lead with the expiration date and § 286 in any response to a demand. US 6,478,404 expired 2021-01-30 per its Google Patents legal-status entry (https://patents.google.com/patent/US6478404/en). No injunction is available; damages are capped by the six-year lookback under 35 U.S.C. § 286. As of 2026-10-01, a complaint would reach back only to roughly 2020-10-01, leaving a ~4-month pre-expiration damages window that will reach zero. Demand a pre-suit damages analysis with claim-by-claim pre-expiration product mapping before you consider any payment.
3. If you do file a petition, watch the § 315(b) clock from the date of service and prepare a § 325(d) Advanced Bionics rebuttal up front, because most of the obvious art here (EP 0679518, US 5,713,673, US 5,757,400, US 6,325,489, EP 0955166, plus the HP Journal articles) is already of record or incorporated into the spec. Note the institutional posture: an expired patent can still be reviewed, but the Board has in other cases proceeded with claim construction under Phillips rather than the AIA's broadest-reasonable-interpretation standard for expired claims — confirm current practice before you rely on either construction, as this point is fact- and panel-dependent.
4. Verify the continuation separately. US 6,860,587 B2 (filed 2002-10-04) is a continuation in the same family and presumably carries the same 2001-01-30 priority and a 2021 expiration, but I did not confirm its own docket. If a demand letter cites the '587 alongside the '404, pull both dockets directly from PTAB E2E (https://ptacts.uspto.gov) rather than relying on the '404 record.
5. Assign confidence honestly in your memo: high confidence of zero AIA proceedings on the '404; medium confidence regarding the family continuation; and note explicitly that "'404 patent" text searches return unrelated TQ Delta and wound-therapy patents.

Primary links: patent record — https://patents.google.com/patent/US6478404/en · PTAB E2E petition/decision repository — https://ptacts.uspto.gov · PTAB decisions — https://www.uspto.gov/patents/ptab/decisions · Federal Circuit opinions — https://www.courtlistener.com

I'll research the ownership chain for this patent, starting with the assignment records.

Let me verify the specific reel/frame records and check for any later transfers.

Ownership Analysis — U.S. Patent No. 6,478,404 ("Ink jet printhead")

Source note (read first): I could not execute the dynamic query on USPTO Assignment Center itself; the reel/frame data below is drawn from the USPTO assignment records as mirrored in the patent's Google Patents "Legal Events" table (https://patents.google.com/patent/[US6478404](/patent/US6478404)/en) and cross-checked against USPTO Official Gazette / assignment documents surfaced elsewhere. Where a field (notably Correspondent) is not visible in the sources I reached, I say so rather than filling it in. Verify at https://assignmentcenter.uspto.gov/ (search patent number 6478404).

Inventors

All five named inventors, with the residence listed on the issued patent (US 6,478,404 B2, Nov. 12, 2002):

Employer basis: All five residences cluster around Corvallis, Oregon, site of HP's inkjet printhead design/manufacturing operation; all five assigned to Hewlett-Packard Company by the April 2001 assignment (reel 011702/0788); and the patent's face names Hewlett-Packard Company as assignee. That is conclusive as to assignee, and strongly probative as to employment — but the assignment document itself does not state "employee," so treat "HP employee" as an inference from residence + assignment, not a recorded fact.

Note on a sixth name: The same Corvallis team appears on the sibling filing US 6,523,935 ("Narrow ink jet printhead," filed the same day, Jan. 30, 2001) with a sixth inventor, Simon Dodd, who is not on the '404. The '404 is therefore a sub-set of a larger same-day, same-team family (also US 6,422,676, US 6,398,347, US 6,488,363 cluster).

Departure pattern: No evidence of any inventor departing within 12 months — the assignment executed 2001-03-27 to 2001-03-30 (≈2 months post-filing), all to HP, with no subsequent re-assignment of the patent away from HP ever. Nothing in the record supports a fire-sale reading.

Original assignee

• Entity on the issued patent: Hewlett-Packard Company, Palo Alto, CA (Delaware corporation).
• Product embodying the claims: Yes. The claims cover a three-columnar-array thin-film thermal inkjet printhead; this is the architecture HP shipped in its tri-color inkjet printhead/cartridge line (the spec ties the printhead to HP's "pen"/print-cartridge products and to the printer of FIG. 11). The commercial embodiment is HP's consumer inkjet printhead hardware, not a paper asset.
• Primary line of business at filing: Computer hardware, peripherals, and imaging/printing — with inkjet printheads as a flagship franchise.
• Current status of the original assignee: Superseded by corporate reorganization, not bankruptcy. Hewlett-Packard Company (HP) survived; in 2015 HP split into HP Inc. (PC/printing) and Hewlett-Packard Enterprise. The printing IP was held through Hewlett-Packard Development Company, L.P. (HPDC), the LP that is the current assignee here (see timeline). HPDC remains a live HP entity; the patent expired for fee purposes 2021-01-30 (Google Patents status: "Expired – Lifetime").

Assignment timeline

Two recorded conveyances touch title; the rest of the events are maintenance-fee (FPAY) and grant-status entries, not assignments.

• 2001-03-27 → 2001-03-30 (executed) / recorded 2001-04-09 — Reel 011702/0788

  • Conveyance: Assignment (Assignment of Assignors' Interest)
  • Assignor: Joseph M. Torgerson; Robert N. K. Browning; Mark H. MacKenzie; and others (record lists three by name; "and others" covers Miller and Bakkom)
  • Assignee: Hewlett-Packard Company, Palo Alto, CA
  • Correspondent: Not visible in the sources I reached. Google Patents' legal-events mirror does not expose the correspondent field. For context — but not verified against this reel — HP's historical assignment correspondent of record was its captive "Hewlett-Packard Company, Intellectual Property Administration, P.O. Box 272400, Fort Collins, CO 80527-2400" department. This is an in-house corporate IP department, not a third-party NPE filer.
  • Context: Initial inventor→corporate assignment (standard employment/IPAD practice; not a sale or transfer of a business).

• 2005-01-11 (executed/effective) / recorded 2005-01-11 — Reel 015583/0106

  • Conveyance: Assignment
  • Assignor: Hewlett-Packard Company
  • Assignee: Hewlett-Packard Development Company, L.P. (HPDC), Houston, TX (Texas LP; general partner HPO Holdings, LLC)
  • Correspondent: Not visible in the sources I reached. Analogous HP→HPDC recordings elsewhere carry HP's own in-house correspondent (e.g., a companion HP→HPDC recording at trademark reel 2689/0939 lists correspondent "Hewlett-Packard Company, 3000 Hanover Street, MS 1051, Palo Alto, CA 94304"). Again: in-house, not an outside NPE firm. Cite this as analogous, not as this reel's field.
  • Context: Internal corporate reorganization — migration of HP's printing/imaging IP into the HPDC holding LP that came out of the HP/Compaq-era restructuring. Not a sale; same ultimate corporate family.

• 2006-05-12 / 2010-05-12 / 2014-04-28 — FPAY (years 4, 8, 12). Maintenance-fee payments by the then-owner. Not assignments.

• 2021-01-30 — Anticipated expiration (status: Expired – Lifetime). No further title events.

Continuation family (for completeness, not a transfer): The continuation US 10/265,073 issued as US 6,860,587 B2 ("Ink jet printhead") directly in the name of Hewlett-Packard Development Company, L.P. (filed 2002-10-04), i.e., the family already sat with HPDC by 2002–2003.

Timeline diagram

timeline
    title Ownership of US 6478404
    2001 : Filed by Hewlett-Packard Company
         : Inventors assign to HP Company
    2002 : Patent issued
    2005 : Assigned to HP Development Co LP
    2021 : Patent expired

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT. Both recorded transfers (reel 011702/0788, 2001-04-09; reel 015583/0106, 2005-01-11) run between operating entities: inventors → Hewlett-Packard Company, then HP Company → Hewlett-Packard Development Company, L.P. HPDC is a holding LP, but it is HP's captive IP holder for a business that ships the covered printheads — not a licensing-only shell, and the transfer was intra-family, not a divestiture.

2. Known asserter in the chain — NOT PRESENT. No assignee at any point matches a public NPE list (Acacia, Marathon, IV, Wi-LAN/Mosaid/Conversant, Pendrell, etc.). The only NPE in this story sits on the other side of the table: HP's own patent intermittently turns up as an evidentiary exhibit in HP Inc. v. Slingshot Printing LLC, IPR2020-01086 (petition filed 2020-06-15), where Slingshot Printing LLC (a Fortress-backed NPE) is the target, not the owner. Owning-side NPE signal: absent.

3. Repeat correspondent across the chain — UNCLEAR / no finding. I could not retrieve the correspondent field for either reel from the sources available. What is visible is consistent with HP's in-house IP Administration department handling both recordings; a corporate department appearing twice is not the "same NPE lawyer runs unrelated shells" pattern the signal is designed to catch. No NPE-attorney recurrence finding can be made on this record.

4. Cascading transfers — NOT PRESENT. Exactly one ownership change post-invention (HP→HPDC, 2005), 4 years after filing, in a single step. No chained LLCs, no shared-correspondent clustering, no <24-month cascade.

5. Pre-litigation transfer — NOT PRESENT. No infringement suit asserting the '404 was located in the searches available (consistent with the earlier litigation summary). With no suit, there is no "transfer 6 months before filing" event to point to.

6. Bankruptcy fire-sale — NOT PRESENT. Neither HP Company nor HPDC ever filed for bankruptcy; this patent was never part of a Kodak/Nortel/Polaroid-style asset sale. It died by fee-expiry, not by liquidation.

7. Privateering — NOT PRESENT. The opposite is closer to true: HP used the '404 as documentary support in an IPR against an NPE. There is no evidence HP handed this patent to an entity to assert on its behalf (no SEC/8-K disclosure, no assignment to a litigating vehicle).

8. Defensive aggregator (anti-NPE) — NOT PRESENT. The chain does not terminate at RPX, AST, LOT, Unified, or OIN. It terminates with HPDC, an operating-company affiliate, and expired 2021-01-30.

Verdict

Operating-company assertion.

Justification: the complete assignment chain for US 6,478,404 consists of exactly two recorded conveyances — inventors → Hewlett-Packard Company at reel 011702/0788 (executed 2001-03-27/30, recorded 2001-04-09) and Hewlett-Packard Company → Hewlett-Packard Development Company, L.P. at reel 015583/0106 (executed/recorded 2005-01-11) — and no assignment to any NPE or shell entity appears at any point. The current assignee is an operating-company affiliate whose products practice the claimed printhead, so the ownership sits squarely in the "operating company" category; I flag explicitly that no suit asserting this specific patent was located, so "assertion" here reflects operating-company ownership of an asserted-capable asset rather than evidence of enforcement of this particular patent. The patent's later appearance as an exhibit in HP Inc. v. Slingshot Printing LLC (IPR2020-01086) places it in a defensive posture against an NPE — the inverse of a troll signal.

Verify at the USPTO Assignment Center: https://assignmentcenter.uspto.gov/ (and the legacy index at https://assignment.uspto.gov/patent/index.html).

Contradiction / gap flag for the analyst: My litigation summary states no litigation touches the '404. A search result for IPR2020-01086, HP Inc. v. Slingshot Printing LLC shows a Docket Alarm docket entry titled "Exhibit-1004 — US Patent No. 6,478,404 to Torgerson et al." (petition dated 2020-06-15). That is not a contradiction of "no litigation on the '404" — it means the '404 was used as an exhibit (documentary/prior-art evidence) inside a proceeding whose challenged patent is a different, Slingshot-owned patent. Recorded because it is relevant to the ownership narrative (HP wielding its own patent against an NPE), but it is not an assertion of the '404 and does not alter the ownership verdict.

I'll search for the specific patent and its cited prior art references to verify details and assess § 102 relevance.

Let me verify the specific references and the discrepancy flagged earlier regarding US 5,683,489.

Prior-Art Analysis — U.S. Patent No. 6,478,404 ("Ink Jet Printhead")

Note on tool budget: I confirmed the identity of the '404 patent and retrieved substantial detail on the four ink-jet references (US 5,757,400; US 6,325,489; EP 0 679 518 A1; EP 0 955 166 A1). My verification searches for US 5,713,673 and for the US 5,683,489 identity conflict were cut off by the tool step limit. Where I could not verify against a live document, I say so and rely on the face-of-patent citation data supplied in the authoritative full text, which is the controlling source per the operating rules.

1. USPTO record for the number as searched

The number 6,478,404 resolves to a single U.S. patent — no substitute was returned:

The "References Cited" list consists of six references (four U.S./two EP). Those six form the scope of this analysis. I also list the three references incorporated into the specification by name (see §3).

2. The six cited references — one by one

(A) EP 0 679 518 A1 — Seiko Epson Corporation

• Full citation: EP 0 679 518 A1, "Print head, especially for use in a serial printer."
• Dates: priority/foreign filing 1994‑03‑02; published 1995‑11‑02.
• Description: An ink‑jet (serial) print head that executes an interlaced printing scheme. The head has N dot‑forming elements arrayed in the sub‑scan direction at a spacing K times the dot‑line pitch (K and N mutually prime); the sub‑scan is adjusted to N times the dot‑line pitch, and the dots have a diameter at least 1.4× the dot‑line pitch. A hybrid control switch selects minute‑feed printing for the head/foot areas and interlaced printing for the body, eliminating "white‑line" streak defects caused by paper‑feed error at densities above ~720 dpi.
• § 102 posture: Qualifies under § 102(b) (published >1 yr before the 2001‑01‑30 filing).
• Potential anticipation: Does not anticipate claims 1 or 21. It is directed to the printing method/pitch geometry for raising effective resolution by interleaving passes — i.e., it speaks to the "multi‑pass printing … not less than 1/(2P) dpi" element (claim 1) and to pitch P (claims 2/22) — but it is silent on a three‑color columnar drop‑generator architecture, the ≤ 1060 µm column separation, thin‑film FET drive arrays, ≥ 100 Ω heater resistors, or parasitic‑resistance compensation. It is § 103 art, not § 102 art, against the independents.

(B) US 5,683,489 — ⚠️ record conflict (flagged)

• Full citation (per the authoritative Google Patents citation table in the supplied full text): US 5,683,489 A, Shoji Hayashi, "Method of producing iron carbide," filed 1995‑01‑20, issued 1997‑11‑04; marked * cited by examiner.
• Conflicting identification: The printed front page of the '404 patent (as reproduced in a litigation exhibit I retrieved) lists "5,683,489 A * — 11/1997 — Miyazaki et al. — 347/41." Class 347/41 is an ink‑jet classification, which is inconsistent with an iron‑carbide chemistry patent.
• Rule applied: The supplied full patent text lists this number as "Method of producing iron carbide" (Hayashi). I am not auto‑correcting it. But the two data sources disagree on the title/assignee for the same number, so I report the conflict rather than resolve it.
• § 102 posture: Qualifies under § 102(b) on date (issued 1997‑11‑04) regardless of which title is correct.
• Potential anticipation:
  • If the reference is genuinely "Method of producing iron carbide" (Hayashi), it is non‑analogous art and anticipates nothing in a bubble‑jet printhead claim.
  • If the reference is really the Miyazaki et al. ink‑jet document (class 347/41), it would be a multi‑color/ink‑jet arrangement reference and could be relevant to the "different color" / columnar‑array element of claim 1 — but the 347/41 subclass alone does not establish that it discloses the ≤ 1060 µm spacing, drop‑volume/multi‑pass element, or FET drive arrays. On the record available, no claim is anticipated.
• Cross‑reference: This is the same anomaly flagged in the previously generated Patent Summary; I confirm it is unresolved and cannot be treated as clean § 102 art.

(C) US 5,713,673 A — Sony Corporation

• Full citation: US 5,713,673 A, Nemoto et al., "Recording head, recording device and recording method."
• Dates: filed 1994‑03‑11; issued 1998‑02‑03. (Front‑page classification ~400/120.02.)
• Description: A recording‑head / recording‑device / recording‑method disclosure. The 400/120 subclass is the typewriter‑/serial‑printer recording‑head area (thermal/transfer‑type recording head), not a thermal bubble‑jet thin‑film printhead.
• § 102 posture: § 102(b) (issued >1 yr before filing).
• Potential anticipation: None of claims 1 or 21. The reference is not directed to an integrated thin‑film ink‑jet die with three color columnar drop‑generator arrays and FET drive circuitry. At most it is background on multi‑element recording‑head arrays. (Caveat: I could not complete a live verification of this reference's full text.)

(D) US 5,757,400 A — Spectra, Inc. (Hoisington)

• Full citation: US 5,757,400 A, "High resolution matrix ink jet arrangement" (also published as EP 0 819 063 B1 / WO 97/28000).
• Dates: filed 1996‑02‑01; issued 1998‑05‑26.
• Description: An ink‑jet head whose jets are arranged as a matrix (hexagonal) array producing images spaced at ≥ 1200 lines/inch. Modular arrays sit in an ink‑chamber plate with pumping chambers fed by ink‑supply ducts, a piezoelectric actuator per chamber, and an orifice plate. It expressly discusses drop volumes vs. resolution — e.g. ~4 picoliters at 1200 dpi and ~0.5 pL at 2400 dpi — and multi‑color operation (black, yellow, magenta, cyan) applied to the same pixel location over successive scans.
• § 102 posture: § 102(b) (issued >1 yr before filing).
• Potential anticipation:
  • Claim 3 / 23 (drop volume 3–7 pL): This reference discloses drop volumes within (≈4 pL) and around that range, though as a piezoelectric system.
  • Claim 1 / 21 ("drop volume that enables multi‑pass printing at ≥ 1/(2P) dpi"): Directly relevant — it teaches small drops combined with multi‑scan/high‑resolution imaging.
  • It does not anticipate the independents, because its jets are piezoelectric matrix arrays, not thin‑film heater‑resistor drop generators with adjacent columnar FET drive‑circuit arrays, and it does not disclose the ≤ 1060 µm three‑color column spacing.
  • Assessment: the strongest § 103 combination reference on the drop‑volume/high‑resolution element.

(E) US 6,325,489 B2 — Seiko Epson Corporation (Endo)

• Full citation: US 6,325,489 B2, "Ink‑jet printing apparatus," Endo.
• Dates: filed 1997‑02‑04; issued 2001‑12‑04; marked * cited by examiner.
• Description: A serial ink‑jet apparatus with multiple nozzle‑opening arrays that independently eject black, cyan, magenta and yellow (and, in a further embodiment, light cyan/light magenta), with the arrays shifted by two print‑pitches relative to one another in the sub‑scan direction, and a constant paper‑feed of five print‑pitches per scan, so dots printed in one scan dry before adjacent differently colored dots are printed — preventing color bleed and banding.
• § 102 posture: Because the patent issued after the applicant's 2001‑01‑30 filing date, it is prior art under pre‑AIA § 102(e) (U.S. patent granted on an application filed 1997‑02‑04, i.e., before the applicant's date). It is not § 102(a)/(b) art.
• Potential anticipation:
  • Relevant to claim 1/21 elements (b)/(c) (multiple side‑by‑side color arrays) and (e) (multi‑pass/interlaced printing to build resolution), and to pitch P (claims 2/22).
  • Does not anticipate claims 1 or 21: no thin‑film bubble‑jet die, no columnar FET drive‑circuit arrays, no ≤ 1060 µm column separation, no ≥ 100 Ω heater resistor, no parasitic‑resistance‑compensating FETs.
  • § 103 combination reference on the multi‑color array / interlacing elements.

(F) EP 0 955 166 A1 — Hewlett‑Packard Company ✅ (likely the closest structural art)

• Full citation: EP 0 955 166 A1, "Thin film ink jet printhead," Hewlett‑Packard Company.
• Dates: filed/priority 1998‑04‑29; published 1999‑11‑10.
• Description: A thin‑film (bubble‑jet) printhead of the same generic architecture the '404 patent describes as conventional: a silicon thin‑film substructure carrying thin‑film heater resistors, an ink barrier layer, an orifice plate, ink feed slots, and on‑die drive circuitry organized into primitives with primitive‑select and address lines (this reference family also treats ground‑bus/transistor‑active‑region layout, as reflected in sibling HP documents I retrieved).
• § 102 posture: § 102(b) (published >1 yr before the 2001‑01‑30 filing).
• Potential anticipation — the closest candidate:
  • Likely discloses claim 1 elements (a) (thin‑film printhead substrate) and (f) (on‑die FET drive circuitry), and is directly relevant to claim 5 (ground busses overlapping FET active regions) and to claim 21's power traces.
  • However, to anticipate claim 1 or 21 it must disclose all elements in a single reference, including the three side‑by‑side color columns separated by ≤ 1060 µm, the drop‑volume/multi‑pass‑resolution limitation, and (for claim 21) the ≥ 100 Ω heater plus parasitic‑resistance compensation. On the record I retrieved, that combined disclosure is not established.
  • Assessment: the most relevant single reference structurally (thin‑film die + integrated drive circuitry), and the leading § 103 primary reference; not confirmed as a § 102 anticipation of the independents.

3. Three additional references incorporated into the specification (not on the citation list)

The '404 specification expressly incorporates these by reference and they therefore form part of its disclosure context; they are also the "Family Cites Families" entries:

These are § 102(b) art, but each is a narrow, foundational teaching and none anticipates the claims; they inform § 103.

4. Bottom line — most relevant prior art and § 102 conclusions

Ranked by relevance to claims 1/21:

1. EP 0 955 166 A1 (HP) — closest structurally (thin‑film die, heater resistors, on‑die drive circuitry). Relevant to claims 1, 5, 21 elements; leading § 103 primary reference; not a confirmed § 102 anticipation.
2. US 5,757,400 (Spectra/Hoisington) — closest on the drop‑volume / multi‑pass‑high‑resolution element; relevant to claims 1 and 3 (and 21/23).
3. US 6,325,489 B2 (Seiko Epson) — multi‑color nozzle arrays + interlacing; relevant to claims 1/21 (color arrays, multi‑pass); § 102(e) art.
4. EP 0 679 518 A1 (Seiko Epson) — interlaced printing / pitch geometry; relevant to claims 1 and 2 element on multi‑pass resolution and pitch.
5. US 5,713,673 (Sony) — general recording‑head art; minimal relevance.
6. US 5,683,489 — identity conflict (iron‑carbide vs. Miyazaki et al. 347/41); if the former, non‑analogous; if the latter, only tangentially relevant.

§ 102 conclusion (honest): On the citation record, I do not find any single one of the six references that anticipates independent claim 1 or independent claim 21 under § 102. The claimed combination — three color columnar drop‑generator arrays spaced ≤ 1060 µm, a drop volume enabling ≥ 1/(2P) dpi multi‑pass printing, and three adjacent columnar FET drive‑circuit arrays (plus, in claim 21, ≥ 100 Ω heaters and parasitic‑resistance‑compensated FETs) — is not met in its entirety by any of them. The realistic exposure is § 103, with EP 0 955 166 A1 and US 5,757,400 as the natural combination core (thin‑film printhead + integrated drive circuitry combined with small drop volume / interlaced high‑resolution printing), optionally in view of US 6,325,489 or EP 0 679 518 A1.

Outstanding verifications I could not complete (be aware before relying on this):

• Full-text of US 5,713,673 (Sony) — relevance assessed from classification/date only.
• Full-text of EP 0 955 166 A1 — assessed from the family/sibling HP disclosures; I did not confirm whether it discloses the three‑color, ≤ 1060 µm, drop‑volume limitations.
• The US 5,683,489 identity conflict is unresolved; do not treat it as clean prior art either way.
• I did not perform a full third‑party/obviousness‑type prior‑art search beyond the patent's own citation list; the "Similar Documents" entries (e.g., US 6,523,935 "Narrow ink jet printhead," US 6,423,676 "Compact ink jet printhead," US 6,398,347 "Energy balanced ink jet printhead," US 6,305,053 "…ground bus that overlaps transistor active regions") are same‑family/contemporaneous HP patents, not prior art to the '404 patent's 2001‑01‑30 filing, but they are worth noting because they show the surrounding claim landscape.

I'll ground the obviousness analysis by pulling details on the prior-art references of record before mapping them to the claims.

Key finding: PTAB petition papers treat "Torgerson" (this patent family) as prior art against later HP/Lexmark patents. Let me pull the two remaining of-record references and the EP family claim text.

Obviousness Analysis — U.S. Patent No. 6,478,404 ("Ink jet printhead")

Note on scope and date. This analysis builds on the previously generated sections (bibliographic data, claim text, family, and the negative litigation search). The current date instruction on this task is April 26, 2026. The '404 patent expired (anticipated expiration 2021-01-30, per the record), so any § 103 analysis is necessarily retrospective — relevant for hypothetical challenge, prosecution-history study, or family/validity comparison — but I proceed as requested.

One data-integrity caveat up front: the "prior art" I was told to use (§ "Prior Art section of this page") is the set of 6 citations of record plus the 5 family cites and 1 non-patent citation. Several are foreign-language or whose full text I did not retrieve. I flag per-reference confidence below and do not fabricate disclosures where I only have a title/abstract.

1. Legal framework and level of ordinary skill

Under Graham v. John Deere Co., 383 U.S. 1 (1966), I examine (a) scope/content of the prior art, (b) differences from the claims, (c) level of ordinary skill, and (d) secondary considerations. Under KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007), a combination of known elements is obvious where there was a known problem, a design need or market pressure, and the combination was predictable to a skilled artisan; and the reference need not be from the same field if the field of endeavor is the same or the problem is one the artisan would look to.

Level of ordinary skill (POSITA): A person with a bachelor's degree in electrical, mechanical, or materials engineering (or equivalent) and roughly 2–4 years of experience in thermal ink-jet (TIJ) printhead design, including thin-film die fabrication, CMOS/FET heater-drive circuits, addressing/primitive schemes, drop-volume control, and multi-pass/shingled printing. This is a mature, incremental art — the "predictable" combination posture of KSR applies with force.

The invention date is 2001-01-30; a reference must predate that to be prior art.

2. Scope and content of the prior art of record

Not in the citation-of-record list, but surfaced during this search and highly relevant (analogous art): US 6,309,053 B1 — Hewlett-Packard Co., "Ink jet printhead having a ground bus that overlaps transistor active regions," filed 2000-07-24, issued 2001-10-30 (EP counterpart EP 1 303 412 B1). Its specification is substantially co-extensive with the '404's FET-drive discussion and it expressly claims "FET circuits … configured to compensate for variation in a parasitic resistance presented by said power traces" (US 6,309,053 cl. 6/13/21) — the exact limitation of '404 claim 21/11–15. Its filing date (2000-07-24) precedes the '404's filing date (2001-01-30).

3. Claim 1 — element-by-element mapping and primary combination

Claim 1 requires: (i) thin-film-layer substrate; (ii) three side-by-side columnar drop-generator arrays along a longitudinal extent; (iii) each a different color, ≥96 generators at pitch P; (iv) columns ≤ 1060 µm apart; (v) drop volume enabling multi-pass printing at ≥ 1/(2P) dpi; (vi) three columnar arrays of FET drive circuits, one adjacent each generator array.

Combination 1 (primary — renders claim 1 obvious):

EP 0 955 166 A1 (HP) in view of US 5,757,400 (Spectra) and EP 0 679 518 A1 (Seiko Epson), optionally with US 4,719,477 / US 5,317,346 (HP) and US 6,309,053 (HP).

• (i) thin-film substrate → EP 0 955 166 A1 ("plurality of thin film layers"); US 4,719,477; US 5,317,346.
• (ii)–(iii) three multi-color columnar arrays, ≥96 at pitch P → US 5,757,400 (matrix arrays, multiple colors per array, small drops); EP 0 955 166 A1 (CMYK multi-printhead color architecture). A "columnar array" of ≥96 TIJ nozzles was routine (EP 0 679 518, US 6,309,053).
• (iv) ≤ 1060 µm column spacing → this is the "narrow die" objective; a result-effective design choice to minimize silicon area, expressly motivated by the compactness/cost concerns recited across this art (e.g., US 6,309,053 background: "increased substrate size and/or substrate fragility … need for an ink jet printhead that is compact").
• (v) drop volume enabling ≥ 1/(2P) multi-pass resolution → EP 0 679 518 A1 supplies the multi-pass/interlace concept ("high density print independently of the distance between the ink jets"), and EP 0 955 166 A1 supplies the enabling small drop volume (2–4 pL; ~3 pL). Together they teach exactly the claimed relationship: small drops + shingled passes → effective resolution finer than nozzle pitch.
• (vi) three adjacent columnar FET drive arrays → US 6,309,053 claims a columnar array of FET circuits adjacent the drop generators, and the HP thin-film art generally integrates drive FETs on-die.

Motivation to combine (KSR):

1. Common field & common problem. All references are ink-jet printers addressing the same triad of concerns: print quality/resolution, compactness/silicon cost, and uniform drop energy. Combination of such art is the paradigm KSR case.
2. Common assignee for the primary HP references. EP 0 955 166 A1 and US 6,309,053 share the HP inventive environment; an artisan would naturally consult HP's own thin-film printhead work.
3. Express teaching/suggestion. EP 0 679 518 A1 expressly teaches that interlace/multi-pass yields high density independent of jet spacing — supplying both the motivation and the mechanism for the "≥ 1/(2P) dpi" element without shrinking nozzle pitch. US 5,757,400 expressly links smaller drops → higher resolution → multiple colors in one compact head.
4. Predictable results. Increasing nozzle count, packing columns, lowering drop volume, and integrating FETs were each well-understood, independently variable levers with predictable effects.

4. Claim 21 — energy-compensated variant

Claim 21 = claim 1 + (a) heater resistor ≥ 100 Ω; (b) power traces; (c) FETs "configured to compensate for a variation in a parasitic resistance presented by said power traces."

Combination 2:

Combination 1 + US 6,309,053 B1 / EP 1 303 412 B1 (HP).

• (a) ≥ 100 Ω heater — a routine design choice; higher-resistance heaters permit smaller drive FETs (the '404 specification itself states this). No unexpected result.
• (b) power traces — inherent to any on-die primitive-select/ground architecture (US 6,309,053; US 5,317,346).
• (c) parasitic-resistance compensation — directly and expressly disclosed in US 6,309,053 (claims 6, 13, 21: "FET circuits are respectively configured to compensate for variation in a parasitic resistance presented by said power traces"), including the drain-region-trimming implementation.

Motivation: The references share the identical problem statement — non-uniform delivered energy to heaters along a column due to trace IR drop — and US 6,309,053 supplies the solution (location-dependent FET on-resistance). An artisan seeking energy uniformity would combine them with a reasonable expectation of success.

Caveat on US 6,309,053: it is not in the '404's citations of record (it appears only as a "Similar Document"). Treat it as analogous prior art whose date (filed 2000-07-24) precedes the '404's filing (2001-01-30). Whether it qualifies as "another" under pre-AIA § 102(a)/(e) depends on inventor overlap, which the retrieved record does not resolve. If the inventive entities are identical, the analysis shifts to double-patenting (note the family record shows a terminal disclaimer associated with the continuation US 6,860,587 B2), not § 103.

5. Dependent-claim mapping (representative)

6. Where the obviousness case is weakest (be candid)

1. The numeric thresholds are not literally disclosed. "≤ 1060 µm," "≥ 96 generators," "1/(2P) dpi," "LS/WS > 2.7," "WS ≈ 3400/4200 µm," "≤ 220/350 µm" — I found no retrieved reference reciting these exact values. Obviousness here rests on optimization of a recognized result-effective variable (silicon area / resolution), which is a legitimate but fact-dependent KSR argument that requires evidence of what the art recognized as the variable and of routine optimization.
2. "Ground bus overlapping active regions" (claim 5/24) and parasitic compensation depend heavily on the co-pending HP sibling (US 6,309,053 / EP 1 303 412) — not a citation of record here. If that reference is unavailable (inventorship/same-entity problem), these limitations need a different source.
3. EP 0 955 166 A1 is HP's own prior work with a closely related specification; a challenger would still need to bridge from its SiC/kogation focus to the specific three-column, narrow-die geometry.
4. Reference-text gaps: I did not retrieve full text for US 6,325,489 B2 or JP 3305182 B2; I have not verified whether they supply additional claim elements. Do not treat them as fully accounted for.
5. US 5,683,489 ("Method of producing iron carbide") is, on its face, unrelated to ink-jet printheads. I flag it exactly as the record presents it and do not reinterpret or auto-correct it. It contributes nothing discernible to this analysis.
6. US 5,713,673 (Sony) is thermal-transfer/dye-vaporization, not TIJ, and is only marginally analogous (multi-color recording head, different physics).

7. Secondary considerations

No evidence of unexpected results, long-felt need, commercial success nexus, or praise is available to me in the record, and the previously generated litigation section found no assertion or PTAB proceeding against the '404. The family record does show the '404 lineage was commercially deployed (it is the "narrow multi-color" HP cartridge family, and later US 9,815,290 / 9,987,852 "Tri-color ink cartridge" patents cite it), which could support an argument of industry adoption — but I have no nexus evidence connecting such success to the specific claim elements (narrow column spacing, drop volume, on-resistance trimming).

8. Bottom line

• Claim 1 would likely be obvious over EP 0 955 166 A1 + US 5,757,400 + EP 0 679 518 A1 (with HP thin-film foundational patents US 4,719,477 / US 5,317,346), because each element is either expressly taught or a predictable design choice, and the references supply the express motivation (interlace → resolution independent of nozzle spacing; small drops → high-resolution multi-color).
• Claim 21 and its dependent compensation claims (11–15, 30–33) would likely be obvious by adding US 6,309,053 / EP 1 303 412, which discloses the precise "FET compensating for power-trace parasitic resistance" limitation, subject to the inventorship/"another" question.
• The strongest validity defense for the '404 would be non-obviousness of the specific numeric envelope (≤ 1060 µm; ≥ 96 nozzles; 1/(2P); WS ≈ 3400/4200 µm; LS/WS > 2.7) and of the claimed combination as a whole, arguing no reference or combination taught or suggested these particular limits and that they produced an unexpected compactness/yield benefit. That is a defensible but evidence-dependent position.

Consistency check with prior sections: No contradiction. The earlier-flagged § 112(b) antecedent issue (claims 17/18 vs. 16) and the differing claim 1 of continuation US 6,860,587 B2 are unaffected by this § 103 analysis.

Sources

• Full patent text and metadata (authoritative): https://patents.google.com/patent/[US6478404](/patent/US6478404)/en
• EP 0 955 166 A1 (HP, thin film ink jet printhead): https://patents.google.com/patent/EP0955166
• US 5,757,400 (Spectra, high resolution matrix ink jet arrangement): https://patents.google.com/patent/US5757400 and https://www.freepatentsonline.com/[5757400](/patent/5757400).html
• EP 0 679 518 A1 (Seiko Epson, print head for serial printer): https://patents.google.com/patent/EP0679518A1/en
• US 5,713,673 (Sony, recording head/device/method): https://patents.google.com/patent/US5713673
• US 6,309,053 B1 (HP, ground bus overlapping transistor active regions): https://patents.justia.com/patent/[6309053](/patent/6309053) ; EP counterpart EP 1 303 412 B1
• EP 1 309 452 B1 (the '404's EP family member, same claims): http://patentimages.storage.googleapis.com/90/2a/af/ebd67ed7e9990f/EP1309452B1.pdf

Uncertainty statement: High confidence in the claim text, the EP 0 955 166 / US 5,757,400 / EP 0 679 518 / US 6,309,053 characterizations (directly quoted from retrieved text). Low confidence as to the full disclosures of US 6,325,489 B2 and JP 3305182 B2 (not retrieved), and as to whether US 6,309,053 qualifies as "another" for § 102 purposes (inventorship not verified). I found no reference expressly reciting the numeric thresholds of claims 1/20/36–38, so those limitations rest on an optimization argument, not literal disclosure.