US 3423685

I'll verify the patent number and check for any current litigation status.

Verification note on the identifier

I searched for the literal string 3423685. No USPTO litigation records, PTAB proceedings, or CAFC 2026 docket entries reference this patent number. The only authoritative hits are the patent document itself (Google Patents) and a Tektronix enthusiast wiki page. I found no evidence that US 3,423,685 is or was involved in any 2026 CAFC appeal or any district-court/IPR proceeding — consistent with its long-expired status. I cannot confirm a negative beyond the searches performed, so treat the litigation finding as "no record located," not "affirmatively none."

Bibliographic data — US 3,423,685

Abstract (as published)

"A differential amplifier includes a pair of signal translating devices having their input electrodes coupled to amplifier input terminals, their output electrodes coupled to amplifier output terminals, and having their common electrodes coupled together. Voltage clamp means couple the said common electrodes to the respective device output electrodes for maintaining a substantially constant voltage difference across each translating device. As a result, a common mode signal applied to the input terminals will not change the voltage difference across the translating devices, and therefore a common mode output signal attributable to voltage differences across the translating devices will not be transmitted."

Technical essence (plain language)

The circuit is a hybrid cascode differential amplifier: two Nuvistor-type triodes (10, 12) as the input pair with cathodes tied together at node 22, and two NPN transistors (14, 16) stacked on top (emitters to the tube anodes, bases tied at node 42, collectors to output terminals). Two constant-current sources (transistor 24 below, transistor 48 above) force a fixed DC voltage across coupling resistor 32, which sits between the common cathode point (22) and the common base point (42).

Because the current through resistor 32 is held constant, the voltage across it cannot change. That clamps the tube cathode-to-transistor-base difference, and since each transistor emitter junction drops ~½ volt, the anode-to-cathode voltage of each input tube is also pinned. When a common-mode signal appears at the grids, it is conveyed to the cathodes (cathode-follower action) and, via the "bootstrap," to the transistor bases and then to the anodes (emitter-follower action). The tubes therefore "float" with the common-mode signal — their operating point does not shift, so their nonlinearity and device-to-device mismatch never enter the output. Differential signals, by contrast, pass through as current into the cascode stage and appear at the outputs. Stated performance: common-mode rejection ratio on the order of 100,000:1 (measured as high as 500,000:1) from DC to ~100 kHz, extensible to ~10 MHz with degraded CMRR.

One search hit (a DIYaudio forum thread) refers to this disclosure as the origin of what practitioners call the "Hawksford cascode" — I flag this as a secondary/community attribution, not an official characterization.

Prior art cited in the patent

• US 3,156,873 — Williams, Differential amplifier (1964-11-10)
• US 3,262,066 — Trilling, Amplifier circuit (1966-07-19)
• US 3,275,944 — Lavin / Bendix Corp., High voltage D.C. coupled differential amplifier including series energized transistors (1966-09-27)
• Non-patent: Fairchild Semiconductor, "A New DC Transistor Differential Amplifier" (Hilbiber), pp. 1–10, esp. p. 4, FIG. 5

The specification also distinguishes US 2,941,155 (Lucas) and US 3,124,762 (Reaves) as prior differential amplifiers that route the common-mode signal through active devices before rejecting it, and therefore cannot track precisely.

Later citations to this patent: US 3,678,405 (RCA, 1972); WO 97/30514 (Butler, hybrid solid-state/vacuum-tube audio amplifier, 1997); US 9,413,309 B1 (Analog Devices Global, 2016).

The seven independent claims — plain language

Claim 1 — Generic differential amplifier with high-resistance tail. Two matched devices with their common electrodes tied together and their input electrodes at the amplifier inputs; a second pair of devices whose input electrodes connect to the first pair's output electrodes, whose common electrodes are tied together, and whose output electrodes are the amplifier outputs. A high resistance is connected to the common electrodes of the first pair so those common electrodes track the input signal voltage; and voltage-clamp means tie each first-pair common electrode to the corresponding second-pair common electrode, holding the voltage difference between the two common nodes constant. Effect: a common-mode input cannot alter the common-electrode-to-output-electrode voltage across each first-pair device, so the common-mode signal is rejected before reaching the outputs.

Claim 2 — Same, with a constant-current tail and express cascode recitation. Substantially claim 1, but the high resistance is replaced by "a source of substantially constant direct current," and the claims expressly state the two device pairs "form two cascode amplifiers on opposite sides of said differential amplifier."

Claim 3 — Vacuum-tube input / transistor output (hybrid) version. Input stage is a pair of vacuum tubes with common (cathode) electrodes tied together and grids at the inputs; output stage is a pair of transistors with emitters on the tube anodes and collectors at the outputs — the two pairs forming two hybrid cascode amplifiers. A high resistance connects the common tube electrodes to a DC supply so those common electrodes follow the input signal voltage.

Claim 4 — Hybrid differential amplifier, two constant-current sources. Vacuum-tube pair with cathodes joined at a first common point and grids at the inputs; semiconductor pair with emitters on the respective tube anodes, collectors at the outputs, bases joined at a second common point. A coupling impedance spans the first and second common points. A first constant-current source feeds the first common point and one end of the coupling impedance; a second constant-current source feeds the other end and the second common point. The result is a fixed DC voltage across the coupling impedance, which holds the cathode-to-base voltage difference constant and makes the cathode and anode voltages follow the common-mode grid signal, keeping the tubes' operating points stable over large common-mode swings.

Claim 5 — Same as claim 4, with explicit topology and "two hybrid cascode amplifiers" language. Adds that the bases are joined at the second common point "in order to form two hybrid cascode amplifiers," and specifies that the constant-current sources connect to the respective terminals of the coupling impedance as in claim 4.

Claim 6 — Same as claim 5, plus load resistors and a coupling resistor. Adds a pair of load resistors at the transistor collectors and specifies that the coupling impedance is a coupling resistor between the first and second common points.

Claim 7 — Same as claim 6, plus RC coupling impedance and a balancing bias network. The coupling impedance is a resistor in parallel with a capacitor (the .001 µF capacitor 80 for high-frequency gain stability of the bootstrap), and a bias impedance sits between the output-transistor emitters and a DC voltage source, including a variable resistance (potentiometer 58) for nulling the DC output voltages at the amplifier outputs.

Uncertainty / caveats

• No independent verification of assignee chain or current ownership beyond the Google Patents record (Tektronix Inc., original and current assignee as listed). Google Patents itself disclaims that assignee listings are legal conclusions.
• No litigation, PTAB, or CAFC 2026 record located for this number. If you need a definitive litigation clear, that would require a paid database (LexisNexis/Docket Navigator/Derwent) search; the free web searches I ran returned nothing on point.
• The published text renders a few values illegibly due to OCR (e.g., "bias voltage of l volts," "resistor 38 and 49," "/2 volt"). I have read these literally and not corrected them; the "/2 volt" appears to be ½ volt, and "38 and 49" appears to be a typo for 38 and 40, but I am not asserting those corrections as authoritative.
• The Google Patents page lists an odd duplicate identifier "US3423685DA," which is a Google data artifact pairing the grant with its application series, not a separate patent document.

I'll search for litigation involving US patent 3423685 across patent litigation sources.

The first searches returned irrelevant results. Let me try more targeted queries.

Let me investigate the Tektronix v. United States case and search patent litigation databases more specifically.

Bottom line

I found no litigation involving U.S. Patent 3,423,685. Every search I ran for the specific number ("3423685," "US3423685," "3,423,685") returned only the patent's own Google Patents page and unrelated documents that happened to mention different patents (e.g., a Gilead '342 patent, a Packet Intelligence '368-litigation appendix, an Orion/Hyundai '342 patent). No case caption, docket number, complaint, or judgment naming this patent surfaced.

Because I cannot fabricate case numbers or parties, I am reporting that there are no known cases rather than inventing a table.

What did come up, and why it is not this patent

My searches surfaced one Tektronix enforcement matter that a researcher might mistake for this patent:

• Tektronix, Inc. v. United States, 213 Ct. Cl. 257, 552 F.2d 343 (1977) (also an earlier phase at 351 F.2d 630 (1965)). Plaintiff Tektronix sued the United States in the Court of Claims over oscilloscope circuits; third‑party defendants included Hickok, Jetronic, and Lavoie, and the accused models were the 1805/1805A, USM/81, LA265, LA545, etc. — https://www.courtlistener.com/opinion/[8616843](/patent/8616843)/tektronix-inc-v-united-states/

Why this is almost certainly not a US3423685 case:

• The Court of Claims case concerned eight patents, six on horizontal circuits and two on vertical circuits. The specific patents discussed in the opinions are different ones — e.g., the '788, '609, '905 patents and Patent 2,769,904 — all filed in the 1950s, aimed at Tektronix's Model 535 and 545 scopes.
• The damages analysis reconstructs a hypothetical license negotiation set in 1959, and the infringement period runs roughly 1960–1968. US3423685 was not filed until August 27, 1964, and did not issue until January 21, 1969 — too late to be among "the patents in suit" in that accounting.

I could not confirm the full list of eight patent numbers from the search snippets I retrieved, so I cannot state with absolute certainty that US3423685 was absent from that case. But the filing/issuance dates make inclusion highly improbable.

Note: I also saw a separate, unrelated Packet Intelligence LLC v. NetScout Systems, Inc. / Tektronix Texas, LLC case (E.D. Tex. No. 2:16‑cv‑00230‑JRG; Fed. Cir. 2019‑2041). That case involved the '725, '751, and '789 patents — not US3423685 — despite the "Tektronix" name.

Why the absence of litigation is expected

From the patent record itself (https://patents.google.com/patent/US3423685/en):

• Filing date: 1964‑08‑27; Priority date: 1964‑08‑27
• Grant/publication date: 1969‑01‑21
• Anticipated expiration: 1986‑01‑21
• Assignee: Tektronix, Inc. (original and current)
• Status: Expired – Lifetime

The practical enforcement window was therefore roughly 1969–1986. A 1960s‑era internal Tektronix hybrid cascode differential‑amplifier circuit patent aimed at oscilloscope vertical amplifiers would have been enforced, if at all, in litigation from that era. I found no such suit.

Caveats / where I could not look

• I was unable to query the Unified Patents litigation portal, PACER, or Docket Navigator directly (my tools returned no usable results for those specific queries). My conclusion rests on general web searches.
• Google Patents' "Cited By," "Patent Citations," and "Similar Documents" lists for US3423685 show no litigation entries, which is consistent with (but not proof of) no known suits.
• Because this patent expired in 1986, it would not appear in modern NPE/PAE litigation databases.

If you need a definitive negative, the authoritative checks would be: the Unified Patents litigation portal (https://portal.unifiedpatents.com/litigation/caselist), PACER for the pre‑1986 district court filings, and the Court of Claims / Court of Federal Claims dockets for that era. Based on the evidence I could retrieve, the answer is that no litigation is known for US3423685.

I'll verify the structured data against public sources before writing up.

Verification performed

I ran the structured "PTAB proceedings on file" block (USPTO ODP, no AIA trials) against (a) the Google Patents legal-status/citation record for US 3,423,685, (b) public PTAB Final Written Decisions and Board papers indexed on Docket Alarm / PTAB E2E, and (c) CourtListener and general web searches for the patent number paired with IPR/PGR/CBM/Federal Circuit terms. Nothing on point surfaced. The structured record and the public record agree.

Proceedings overview

Total AIA trial proceedings on US 3,423,685: zero. Breakdown: 0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution denials — because no petition was ever filed, and none could have been filed during the patent's enforceable life. The bottom line for a defendant is stronger than any IPR outcome could give you: the patent expired 1986-01-21, roughly 26 years before the AIA's IPR/PGR/CBM provisions took effect, so there are no PTAB outcomes to litigate over. If a demand letter cites US 3,423,685, the correct response is not an IPR strategy — it is that the patent has been expired for four decades and any infringement damages claim is time-barred under 35 U.S.C. § 286.

Why the count is structurally zero (not just "no record found")

This is worth stating explicitly, because "no record located" and "impossible" are different defenses:

• Term. US 3,423,685 issued 1969-01-21 from an application filed 1964-08-27. As a pre-URAA patent it carries a 17-year term from issuance, ending 1986-01-21 — which is exactly the "anticipated expiration" date Google Patents reports. There is no reissue, extension, or terminal-disclaimer data in the record that would move that date.
• AIA timing. The IPR and PGR provisions of the America Invents Act became available for petitions filed on or after 2012-09-16; PGR additionally reaches only patents with an effective filing date on or after 2013-03-16; the CBM transitional program ran 2012-09-16 to 2020-09-16. All three windows open decades after this patent lapsed.
• Practical consequence. Any paper bearing an "IPR20xx-nnnnn" or "CBM20xx-nnnnn" number and captioned against US 3,423,685 would be fabricated. Treat such a citation as a red flag on the source asserting it.

One specific trap to avoid. A web search for "Tektronix" + "IPR" + "amplifier" readily returns IPR2018-00647 and IPR2018-00643, Rohde & Schwarz GmbH & Co. KG v. Tektronix, Inc. Those proceedings concern US 8,675,719 B2 ("Multi-Domain Test and Measurement Instrument") — a different Tektronix patent, a different technology, a different decade. The Board in IPR2018-00647 (Final Written Decision, Paper 51, 2019-09-10, panel of APJs Cocks, Boucher, and Bain) held that petitioner did not prove claims 1–10 and 12–15 unpatentable, while the companion IPR2018-00643 invalidated some but not all of the same claims. None of that has any bearing on US 3,423,685; a Rohde & Schwarz citation is not a citation against this patent.

Proceedings

None to enumerate. No IPR, PGR, or CBM was instituted, denied, or terminated on US 3,423,685. There is consequently no Final Written Decision to link, no judge panel to name, no petition ground to summarize, no settlement to describe, and no Federal Circuit appeal of an FWD to report. I am not supplying placeholder proceeding numbers or inventing an "IPR-####" caption to fill the requested template.

For completeness, the only litigation-adjacent Tektronix matter the searches surfaced that touches this vintage of Tektronix oscilloscope circuitry is Tektronix, Inc. v. United States, a Court of Claims / § 1498 government-infrngement accounting (see Tektronix, Inc. v. United States, 552 F.2d 343 (Ct. Cl. 1977), and the later delay-compensation opinion at https://www.courtlistener.com/opinion/[8617131](/patent/8617131)/tektronix-inc-v-united-states/). I could not verify that US 3,423,685 was among the patents in suit in that case, and I am not asserting it was — US 3,423,685 issued 1969-01-21, while the accused scopes in that accounting span the 1959–1968 procurement period. Treat any claim that '685 was litigated in that case as unverified absent a look at the patents-in-suit list.

Strategic summary

Claim status: all 7 claims untested before the Board — and permanently untestable. No claim of US 3,423,685 was ever canceled, confirmed, or construed in an AIA trial. Claims 1–7 therefore stand as issued, but "standing as issued" means nothing commercially, because the patent has been expired since 1986-01-21. There is no surviving-claims list to report and no narrowing history to exploit or work around: an IPR would have produced a claim-level verdict, and no IPR exists.

Estoppel landscape: § 315(e)(2) is inapplicable, and that cuts entirely in the defendant's favor. There is no petitioner, no privity chain, and no instituted ground, so there is no estoppel fence around any prior-art reference. A defendant is legally free to raise any ground — including, if it ever mattered, the very references the patent itself distinguishes (US 2,941,155 to Lucas and US 3,124,762 to Reaves) and the three references the examiner cited (US 3,156,873 Williams; US 3,262,066 Trilling; US 3,275,944 Lavin/Bendix), plus the non-patent Fairchild/Hilbiber paper. But the operative bar is not estoppel; it is 35 U.S.C. § 286's six-year damages lookback. Because the patent expired 1986-01-21, the last date on which a damages claim could have accrued is 1986-01-21, and the limitations period on that claim closed no later than 1992-01-21. There is no live damages theory, and no injunctive theory either, since an expired patent cannot be infringed prospectively.

Pattern signals: none, and none expected. There is no repeat petitioner, no aggressive Patent Owner appeal posture (the patent has never been before the Federal Circuit in an AIA context), and no defensive aggregator such as Unified Patents in the chain. The patent predates the entire modern PTAB ecosystem. Its relevance today is purely technical/historical — the earlier section's note that practitioners attribute the bootstrapped/hybrid "Hawksford cascode" to this disclosure is a design-lineage observation, not a litigation signal.

Recommended next steps

1. If you are a defendant and have received an assertion citing US 3,423,685: do not petition for IPR — it is legally pointless and would waive nothing while costing fees. Respond on the patent's face: issued 1969-01-21, expired 1986-01-21 under the pre-URAA 17-year term, no reissue or extension of record, and any damages claim time-barred under 35 U.S.C. § 286 since 1992-01-21. The Google Patents record (https://patents.google.com/patent/US3423685/en) states "Expired – Lifetime" with anticipated expiration 1986-01-21.
2. Verify the patent number in any demand letter character-for-character. The near-identical prefixes in the Tektronix portfolio create real confusion potential: '685 (this patent), '719 (US 8,675,719, the multi-domain test instrument litigated in IPR2018-00643/-00647), and '460 (US 8,521,460, the subject of IPR2018-00645, where institution was denied). A letter that muddles these is a strong indicator the sender has not done a docket check.
3. Ask for the sender's basis, in writing. If the sender points to IPR2018-00647 or IPR2018-00643 as "the PTAB proceedings on 3,423,685," that is a demonstrable misidentification — those decisions name US 8,675,719 B2 in the caption. Documenting that misstatement is useful if the demand escalates.
4. There is nothing to monitor. With no pending or possible AIA trial and no appeal, PTAB E2E (https://ptab.uspto.gov) and CourtListener (https://www.courtlistener.com) will remain clean for this number. Re-checking them is unnecessary; the structured ODP feed is dispositive here.

Confidence and limits. High confidence on the zero-proceeding finding, because it rests on a structural fact (expiration date) that is corroborated by the structured data, the Google Patents record, and the patent's own filing/issue dates. My residual uncertainty is limited to (i) whether '685 appeared on the patents-in-suit list in the old Court of Claims § 1498 accounting, which I did not verify, and (ii) the general limits of free web search for very old district-court dockets that were never digitized. Neither gap affects the defensive conclusion, which turns on expiration, not on litigation history.

I'll research the assignment record for this patent, starting with the USPTO Assignment Center and cross-referencing sources.

I was unable to execute a live query against the USPTO Assignment Center (the search tool returned no indexed results for this number, and I exhausted my search budget before a direct retrieval succeeded). I want to be explicit about that up front rather than fill the gap with plausible-looking reel/frame numbers. Everything below is what I can actually ground; where the record is missing, I say so.

Inventors

• The patent names a single inventor, Roy M. Hayes (no joint inventorship).
• Employer is inferable but not documented in the four corners of this patent: Hayes appears as the named inventor on at least one other Tektronix patent from the same window — US 3,323,070, "Variable gain amplifier having constant frequency band pass," Tektronix Inc., filed 1964-05-12, granted 1967-05-30 (Tektronix enthusiast wiki mirror). Same inventor, same assignee, same oscilloscope-vertical-amplifier subject matter, filed ~3.5 months before this application.
• No evidence of an inventor departure pattern. I found no record of Hayes leaving Tektronix within 12 months of filing, and no evidence of a contemporaneous portfolio sale. I cannot confirm a negative here — this is "no record located," not "affirmatively none."

Original assignee

Tektronix, Inc. (Beaverton, Oregon) — named as original and current assignee on the Google Patents record.

• Primary line of business: test-and-measurement instrumentation — cathode-ray oscilloscopes, spectrum analyzers, logic analyzers, probes. The specification itself situates the invention as "especially useful when employed as a D.C. coupled vertical amplifier in a cathode ray oscilloscope."
• Did they ship a product embodying the claims? Yes, with high confidence. The disclosed circuit is a hybrid cascode (Nuvistor triode input pair stacked under NPN transistors) with a bootstrapped common-mode clamp — precisely the topology Tektronix used in its 1960s–70s vertical plug-in amplifiers. The specification's detailed component values (+300 V, +125 V rails, 37.4 kΩ coupling resistor, 1 MΩ grid resistors, Nuvistor-type triodes) read as a production-derived design rather than a paper circuit. I am not asserting a specific model number (e.g., a particular 1A-series or 7A-series plug-in), because I could not verify the mapping in this session.
• Current status: Operating. Tektronix was acquired by Danaher Corporation (deal announced 2007) and subsequently became part of Fortive Corporation when Danaher spun off Fortive in 2016. I could not verify in this session whether Tektronix remained under Fortive after Fortive's announced 2025 two-company separation. Treat the corporate-parent line as accurate through the 2016 Fortive spin-off and unverified thereafter.
• The patent's front-page assignee field is not a legal conclusion (Google Patents expressly disclaims this), and no standalone assignment document was retrieved to corroborate it.

Assignment timeline

No assignment record was retrievable for this patent in this session. The USPTO Assignment Center (https://assignmentcenter.uspto.gov/) and its legacy mirror (https://assignment.uspto.gov/patent/index.html) were not successfully queried — repeated searches returned zero indexed assignment records for the literal string 3423685. I therefore have no reel/frame, no conveyance type, no execution/recording dates, no assignor/assignee of record, and no correspondent of record to report. I will not invent them.

What the underlying patent data does establish:

• 1964-08-27 — Application US 392,420 filed by Roy M. Hayes; Tektronix, Inc. named as assignee on the issued document.
• 1969-01-21 — US 3,423,685 granted.
• 1986-01-21 — Anticipated expiration (Google Patents legal-status data); status "Expired – Lifetime."
• No post-issuance assignment, security interest, name change, or license recording appears in any indexed source I could reach.

The single most probable recorded instrument — an inventor-to-Tektronix assignment executed on or about the 1964 filing, consistent with Tektronix's standard practice for employee inventors — is presumed, not documented. Anyone relying on this section for a chain-of-title opinion must pull the reel/frame directly from Assignment Center; I could not.

Timeline diagram

timeline
    title Ownership of US 3423685
    1964 : Filed by Roy M Hayes
    1969 : Issued to Tektronix Inc
    1986 : Patent expired

The diagram is deliberately sparse: I have no verified post-grant conveyance to plot. If the original Hayes-to-Tektronix assignment exists (likely), it belongs at 1964, but I could not confirm its reel/frame.

NPE / troll-pattern signals

Verdict

Insufficient data — specifically: only the original assignment exists, and even that is presumed rather than documented.

No post-issuance assignment could be retrieved, and every indexed source points to a single assignee, Tektronix, Inc., an operating instrumentation company that practiced the disclosed hybrid cascode in its own oscilloscope products. There is no reel/frame, no asserter entity, no correspondent, and no litigation to support even one NPE signal — but because the Assignment Center record itself was unreachable in this session, I am recording this as insufficient data rather than a clean "operating-company / non-asserting" clear. A definitive call requires a direct reel/frame pull at https://assignmentcenter.uspto.gov/ (patent number 3423685) to confirm or rule out the original Hayes-to-Tektronix instrument and any late-life recording.

I'll search the USPTO/patent records for US 3,423,685 and each of its cited references.

Let me gather details on each cited reference and the two prior-art patents discussed in the specification body.

Prior-art analysis for US 3,423,685

1. Record verification (brief, since bibliographic data was established earlier)

A USPTO-sourced full-text copy was located at https://patentimages.storage.googleapis.com/f9/f8/96/431229749f6bb9/US3423685.pdf and matches the Google Patents record already captured in the earlier sections. Confirming only the fields that matter to the prior-art question:

• US 3,423,685 A — "Bootstrapped cascode differential amplifier"
• Application US 392,420, filed 27 August 1964, granted 21 January 1969
• Inventor Roy M Hayes; assignee Tektronix Inc.
• Examiner of record: Nathan Kaufman
• References Cited section (on the face of the patent): 3 U.S. patents + 1 non-patent publication

No similar-numbered document was substituted — the citations below are the ones listed verbatim on the face of US 3,423,685.

2. A threshold point that governs the § 102 analysis

Under pre-AIA 35 U.S.C. § 102 (the statute applicable to a 1964 filing), a printed publication or patent must predate the applicant's invention or the filing date to be § 102(a)/(b) art. Checking the dates:

This is a materially important finding: all three cited patents issued after the 27 August 1964 filing date of US 3,423,685 (the earliest of them by only about ten weeks). They were therefore cited as pre-AIA § 102(e) "secret prior art" — patents that were pending in the Office (or at least filed) before the applicant's filing but not yet public. They could not have been § 102(a) or § 102(b) art on their issue dates. Anyone reading the "References Cited" list and assuming they were ordinary § 102(a)/(b) references would be applying the wrong subsection. I flag this as a correction/clarification to any prior characterization of these as merely "cited prior art."

3. Prior art of record — reference-by-reference

3.1 US 3,156,873 — Williams, "Differential amplifier"

• Full citation: US 3,156,873 A, Thomas R Williams, "Differential amplifier." Filed 12 Aug 1960; granted 10 Nov 1964. US Cl. 330-69 (per the patent's own citation table: "3,156,873 11/1964 Williams 330-69").
• Brief description (verified from the specification text): A push-pull/differential vacuum-tube amplifier with two matched triodes (26, 30) referenced to a common ground 24, each with seriesed cathode resistances and a plate load resistance. It uses tunable cross-neutralizing capacitors (52, 54) between each plate and the other tube's control electrode, and cross-coupling capacitors (56, 58) that apply the AC signal variation across one tube's load resistance to the cathode-resistance junction of the other tube. The disclosure treats mid-frequency equivalent-circuit analysis and distributed capacitances.
• Why it was cited: The cross-neutralization topology is conceptually parallel to the applicant's capacitors 68 and 70 (cross-neutralization of stray input-to-input, input-to-cathode and grid-to-anode capacitance), and the cross-coupling resembles the "bootstrap" style of feeding one node's signal into another.
• § 102 assessment: Does not anticipate any of claims 1–7. Williams discloses a single pair of active devices in push-pull; it contains no second pair of cascoded output devices and, critically, no "voltage clamp means for connecting the common electrodes of said first pair of devices respectively to the common electrodes of said second pair of devices." Every one of claims 1–7 requires that two-pair clamped structure. Williams is best characterized as § 103-type background art for the neutralization feature — which, notably, is described in the specification but recited in no claim. It is therefore weaker as prior art than its presence in the list might suggest.

3.2 US 3,262,066 — Trilling, "Amplifier circuit"

• Full citation: US 3,262,066 A, Theodore R Trilling, "Amplifier circuit." Filed 28 Jun 1962; granted 19 Jul 1966 (patent's table: "3,262,066 7/1966 Trilling 330-69"). Google Patents record: https://patents.google.com/patent/US3262066A/en.
• Brief description (verified from the abstract/description text): A DC, directly-coupled differential amplifier addressing drift and nonlinearity. Cathodes/emitters of the input pair (13, 14) are joined through a resistor 27 to the negative supply; common-mode voltages and variations appear across resistor 27, and that voltage is fed via conductor 28 to the control element of a constant-current device 26 — an explicit common-mode feedback loop. Avalanche diodes (38, 39) inserted between the input-transistor collectors and the output-stage bases provide a fixed DC level shift "to allow proper voltages across the input transistor." The output is a second differential amplifier (35) built from Darlington-connected transistors.
• Why it was cited: This is the closest of the three to the applicant's problem statement — it targets DC differential amplification with common-mode handling, uses a resistive common-emitter tail (analogous to the applicant's high-resistance/constant-current tail at node 22) and uses level-shifting elements between the input devices and the following stage (analogous in purpose to the applicant's clamp).
• § 102 assessment: Anticipates no claim. The structural correspondence is superficial in the way that matters. Trilling's mechanism is feedback to a current source, not a clamp that pins the voltage across the input device. Its avalanche diodes are a fixed level-shift, not a bootstrap that makes the input devices' anode/cathode voltages follow the common-mode signal. Claims 1–2 require the voltage difference across each first-pair device to remain constant under common-mode input; Trilling does not disclose that, and it has no counterpart to the second-pair common electrode tie point (applicant's node 42) to which a clamp could connect. Possible § 103 relevance to the "means for connecting a high resistance… to cause the voltage on the common electrodes… to follow the signal voltage on their input electrodes" element of claim 1 — but as § 102 art, negative.

3.3 US 3,275,944 — Lavin / Bendix, "High voltage d.c. coupled differential amplifier including series energized transistors"

• Full citation: US 3,275,944, Thomas J. Lavin, assignee Bendix Corp., "High voltage d.c. coupled differential amplifier including series energized transistors." Application US 326,081; filed 26 Nov 1963; granted 27 Sep 1966. Int. Cl. H03F 3/42, 3/45; US Cl. 330/69, 330/71 (patent table: "3,275,944 9/1966 Lavin 330-30 X"). Record: https://www.freepatentsonline.com/3275944.html.
• Brief description (verified from the record): A DC-coupled differential amplifier in which transistors are stacked in series ("series energized") per side to withstand high supply voltage — i.e., the series-device-per-side stacking concept that underlies cascode-style output stages, predating the applicant's hybrid tube/transistor stack.
• Why it was cited: Structural kinship to the applicant's stacked output stage (transistors 14/16 sitting on the tube anodes), and to the "two cascode amplifiers on opposite sides" recitation of claims 2 and 5.
• § 102 assessment: Anticipates no claim. "Series energized transistors" is a high-voltage stacking arrangement, not a bootstrapped/voltage-clamped hybrid cascode with a constant-voltage coupling element between a common cathode node and a common base node driven by two constant-current sources. Claims 3–7 specifically require vacuum tubes as the input devices with a transistor on each anode, plus the coupling impedance between first and second common points. Lavin's all-semiconductor, non-bootstrapped stack does not disclose those limitations. § 103 relevance to claims 2/5's cascode language at most.

3.4 Non-patent literature — Fairchild Semiconductor, "A New DC Transistor Differential Amplifier" (Hilbiber)

• Full citation (as printed on the patent): Technical Articles and Papers — Fairchild Semiconductor Corp., "A New DC Transistor Differential Amplifier," pp. 1–10, by Hilbiber, specifically page 4, FIG. 5. (The patent's "References Cited" lists it under "OTHER REFERENCES.")
• Brief description: A Fairchild technical paper on DC transistor differential amplifier design. The citation formula ("specifically page 4, FIG. 5") is the examiner's/attorney's pinpoint to the disclosed circuit relied upon.
• § 102 assessment: No claim anticipated, on the record as cited. A DC transistor differential-amplifier paper is generic prior art for the amplifier genre; nothing in the pinpoint (p. 4, FIG. 5) is described in US 3,423,685 as disclosing a bootstrap voltage clamp between two pairs of common electrodes.
• Caveat (honesty flag): I was unable to retrieve the full text of the Hilbiber paper within the search budget of this session. The pinpoint above is taken verbatim from the patent's own citation; I did not independently read p. 4 / FIG. 5, so my § 102 conclusion for this reference rests on the applicant's characterization rather than on my own reading of the document. If a definitive § 102 analysis of this item is required, p. 4 should be pulled and read directly.

4. Prior art discussed in the specification but not listed in "References Cited"

The specification body distinguishes two further U.S. patents. These do not appear in the patent's References Cited table — an important distinction if one is compiling "the prior art of record" strictly:

• US 2,941,155 — Lucas. Cited in the specification as a previous differential amplifier in which "the common mode signal through vacuum tubes or transistors before rejecting such common signal."
• US 3,124,762 — Reaves. Cited alongside Lucas for the same reason — active-device common-mode routing with insufficient tracking.

Status of my verification: I attempted to pull bibliographic/date data for both and could not within the tool budget (searches were consumed by the cited references). I therefore cannot confirm their filing/issue dates, assignees, or claim content, and I am not asserting anything about their technical disclosure beyond what US 3,423,685 itself says about them. For a complete § 102/§ 103 file-history analysis, both should be retrieved and read; the applicant's own framing suggests they are the most conceptually on-point admitted prior art (they address the same common-mode rejection problem), which in some analyses makes them more probative than the three examiner-cited patents.

5. Anticipation matrix (my assessment under pre-AIA § 102)

Net result: none of the four references of record anticipates any of the seven claims. That conclusion is consistent with the prosecution history — all seven claims (all independent, no dependents) issued over these references, examiner Kaufman allowing them. The references functioned as § 102(e)/§ 103 background for individual elements (tail resistance, level shifting, series stacking, cross-neutralization), and the point of novelty that carried the claims was the combination of (a) two constant-current sources across (b) a coupling impedance spanning the common-cathode node and the common-base node, producing (c) a fixed voltage across the input devices under common-mode swing.

6. Things I could not verify (stated rather than papered over)

1. Hilbiber full text — not retrieved; § 102 conclusion rests on the patent's own pinpoint citation.
2. US 2,941,155 (Lucas) and US 3,124,762 (Reaves) — dates, assignees, and claim content not retrieved. Do not rely on any assertion about them beyond the applicant's characterization quoted above.
3. Full text of US 3,156,873 and US 3,262,066 — I retrieved substantial specification excerpts, not the complete documents or their claims. My "brief description" is faithful to what I read, but it is not a full-text review.
4. Assignee data is taken from Google Patents/FreePatentsOnline listings, which themselves disclaim being legal conclusions; I did not run an assignment-chain search.
5. The "similar documents" and "Cited By" lists (e.g., US 3,678,405, WO 97/30514, US 9,413,309 B1) are forward citations / algorithmic neighbors, not prior art against US 3,423,685 — they postdate it and have no § 102 significance here.

If you want the Δ closed on items 1–3, the definitive sources are the USPTO Patent Public Search (PPUBS) full-document images for US 3,156,873, US 3,262,066, US 2,941,155 and US 3,124,762, plus the original Fairchild application note for the Hilbiber paper.

I'll ground the analysis in the actual prior-art references before drawing § 103 conclusions.

Good grounding on Trilling and Williams. Now the remaining references.

Now the two references the specification itself distinguishes, plus the bootstrapping art.

Obviousness analysis — US 3,423,685 under pre‑AIA 35 U.S.C. § 103

Scope note. This section builds on, and does not repeat, the earlier bibliographic, claim‑exposition, litigation, and PTAB sections. Two framing points carried forward: (i) there is no reissue, reexamination, or AIA‑trial claim‑construction history for this patent, so the only constraint on claim scope is the printed text; and (ii) the patent is expired, so this is a validity/historical exercise rather than a live dispute posture. (Date discrepancy flagged: the system prompt says today is 2026‑09‑30; this task states April 26, 2026. Immaterial to the analysis — none of the references have moving status.)

1. Legal framework and the person of ordinary skill

The application was filed 1964‑08‑27, so pre‑AIA § 103(a) governs, and the hypothetical PHOSITA is a 1964‑vintage engineer — not a modern one. Graham v. John Deere Co., 383 U.S. 1 (1966), supplies the four factual inquiries; KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007), controls the legal standard (applied to pre‑AIA patents in modern re‑analyses): a combination of known elements is obvious where the results are predictable, where there is a "finite number of identified, predictable solutions," or where a known technique is used to improve a similar device in the same way. The TSM test survives only as one permissible route; the requirement is an "articulated reasoning with a rational underpinning." In re Kahn, 441 F.3d 977 (Fed. Cir. 2006).

PHOSITA (1964): a BSEE-level design engineer with ~3–5 years in wideband DC coupled instrumentation amplifiers (oscilloscope vertical front ends), familiar with: constant-current tail biasing; cascode / "series-energized" stacked-device stages; cathode- and emitter-follower action (both unity-gain); bootstrapping as a technique for holding a node voltage constant; and the classical drift/CMRR literature (Hilbiber, Trilling).

Prior-art status of each reference (dates matter under pre‑AIA § 102(a)/(b)/(e)):

Note the two 1966-issuing references are usable only because of their pre‑1964 US filing dates — a point a § 103 challenge must plead correctly.

2. What each reference actually discloses (verified vs. not)

Verified from retrieved text:

• Lucas '155 — Differential amplifier with input tubes V3/V4. The critical sentence: "Tubes V5 and V6 are connected in series with the plates of tubes V3 and V4, respectively, to maintain the plate potential of the latter tubes virtually insensitive to the common mode of input signals applied to terminals 11 and 12." This is the closest functional teaching in the record: a series device stacked above the input pair, deployed specifically so the input devices' plate potential does not move with the common-mode input. Also teaches common-mode rejection as the design goal, drift minimization, and low input-current effects.
• Trilling '066 — DC, directly coupled differential amplifier. Constant-current device 26 in series between the input pair and the negative supply; input emitters/cathodes tied in common through resistor 27 to device 26; "common‑mode feedback" — the voltage across resistor 27 is sensed on conductor 28 and controls device 26. Input collectors feed a Darlington-connected output differential stage through avalanche diodes 38/39 (level shift). Load resistors 41/42 at one side, 46/47 at the other, with a potentiometer 43 and wiper arm 48 for balance.
• Hilbiber TP‑16 (Feb. 1961) — All-transistor DC differential amplifier using planar silicon devices; the stated thesis is that constant collector current in the input device is "a necessary condition for minimizing drift", achieved with breakdown diode D1 + R1; multiple compound PNP/NPN blocks (Figs. 3–4); optimized two- and three-transistor amplifier circuits at Figs. 5–6; explicit treatment of V_BE tracking, thermal matching of pairs, and source-impedance/CMRR tradeoffs.
• Williams '873 — Push-pull vacuum-tube stage, triodes 26/30, plate load resistors 36/44, per-tube cathode resistor pairs 38/40 and 46/48 with the output tapped at the cathode junctions; cross-neutralizing tunable capacitors 52/54 (plate of one → grid of the other) and A.C.-short capacitors 56/58 (load swing of one → cathode junction of the other); "corresponding components in each stage being matched in value." Notably Williams has no common cathode node and no constant-current tail — it is a matched push-pull/neutralized stage, not a constant-tail differential pair.

Not verified — flagged as a real limit on confidence:

• US 3,275,944 (Lavin) — I retrieved only bibliographic data (Bendix; filed 1963‑11‑26; issued 1966‑09‑27; classes 330/69, 330/71; Int. Cl. H03F 3/42, 3/45). The title ("High voltage d.c. coupled differential amplifier including series energized transistors") is itself a teaching of a stacked/series-energized transistor pair in a DC differential amplifier — i.e., the transistor analogue of Lucas's tube stack. I could not read its figure or its grid/base connections.
• US 3,124,762 (Reaves) — my search did not return usable content. I have only the patent's own characterization (see §3).

Critical caveat: whether the "voltage clamp means" limitation is met turns on where Lucas's V5/V6 grid electrodes are connected, which I could not read. If Lucas drives V5/V6 grids from the input pair's common cathode node (or from a node that follows it), Lucas alone approaches the claim. If Lucas drives them from a fixed bias, Lucas teaches the function (common-mode-insensitive plate potential) but not the bootstrap topology.

3. Admissions in the specification itself

These are usable as § 103 evidence independent of the cited references:

1. "The anodes of the tubes are connected to the cathodes of such tubes by nearly unity gain bootstrap circuits which aid in this cathode follower action." — the applicant describes the clamp as "bootstrap circuits," a named, pre-existing technique.
2. "The transistors 14 and 16 can be replaced by vacuum tubes to provide a pair of conventional cascode amplifiers on opposite sides of the differential amplifier." — an admission that the stacked second pair, per se, is conventional.
3. The applicant's own prior-art discussion concedes that Lucas and Reaves "transmitted the common mode signal through vacuum tubes or transistors before rejecting such common signal," i.e., an admission that these are prior art differing only in where the common-mode signal is absorbed.
4. "It will be obvious to those having ordinary skill in the art that various changes may be made…"

The net effect: the applicant concedes that (a) the cascode stack is old, (b) bootstrapping is old, and (c) the prior art differ in the locus of common-mode absorption. That narrows the novelty to the two-current-source, fixed-voltage coupling impedance sitting between the input pair's common node and the second pair's common control node.

4. Element mapping

Claim 1 (broadest — "signal translating devices")

Claim 2 — adds "source of substantially constant direct current" (Trilling device 26; Hilbiber D1/R1 bias) and the express cascode recitation (admitted conventional). Differences from claim 1 are minimal; claim 2 is at least as vulnerable as claim 1 on the same combination.

Claim 3 — tube input / transistor output hybrid, high resistance to a DC supply (not to a constant-current source). Trilling + Hilbiber + Lucas/Lavin supply every element; the "hybrid" character is a design choice over Lucas's all-tube stack (transistors replacing tubes for level shifting was routine by 1964, and admitted by the applicant).

Claim 4 — two constant-current sources bracketing a coupling impedance between the two common points. This is the claim whose added limitation is closest to pure design principle: fixing the voltage across an impedance requires only fixing the current through it (Ohm's law). It is the strongest KSR candidate.

Claims 5–6 — hybrid cascode recitation + load resistors at the collectors + coupling resistor. Load resistors at the output collectors are taught by Trilling ('066: 41/42, 46/47).

Claim 7 — RC coupling impedance in parallel + variable-resistance balancing bias. Trilling's potentiometer 43 with wiper arm 48 is a balancing network; paralleling a resistor with a capacitor for HF gain stability is a routine compensation technique (and the '685 spec explains it as exactly that: "to maintain the unity voltage gain of the bootstrap signal at high frequencies").

5. The combinations, and why the PHOSITA would make them

Combination A (lead): Lucas '155 + Lavin '944 + Hilbiber TP‑16, optionally + Trilling '066

Rationale. Lucas supplies the problem and the mechanism: stack a second device above the input pair so the input pair's plate potential is "virtually insensitive to the common mode." Lavin supplies the transistorized version of exactly that stack ("series energized transistors" in a DC coupled differential amplifier). Hilbiber supplies (i) the constant-current-bias doctrine for minimum drift and maximum common-mode tracking, and (ii) the motivation to make performance independent of device matching — the same design goal the '685 patent states.

Motivation to combine (KSR-compliant):

• Same field, same problem, overlapping time. All are 1958–1963 DC differential-amplifier disclosures aimed at drift and common-mode rejection. This is not analogous-art stretching; it is the same art.
• Direct substitution. A transistor stacked above an input device in place of a tube stacked above an input device was routine by 1964 — and the '685 specification itself concedes the interchangeability ("transistors 14 and 16 can be replaced by vacuum tubes"). KSR: substitution of one known element for another, predictable result.
• Two known techniques, combined. Lucas's "make the input device's plate potential immune to common mode" + Hilbiber's "hold the input device's current constant with a current source." The '685 clamp is what you get when you combine them with a follower: the second pair's control node follows the input common node, and the fixed-voltage impedance (constant current through a resistor) guarantees the tracking is exactly unity in DC terms.
• "Obvious to try" with a finite solution set. Given the desire to hold the input device's V_AK constant, the options are (a) fixed-bias the upper device (fails over large common-mode swings — the very problem stated in the patent), or (b) make the upper device's control node follow the input device's common node. There is one sensible way to do (b): couple the two nodes with an impedance and force a constant current through it. That is claim 4 read literally.

Combination B: Trilling '066 + Lavin '944 (+ Hilbiber)

Trilling is the closest architectural analogue: constant-current tail, common input-pair node, a second device pair stacked above, and a balance network. Its distinguishing feature is that it senses the common-mode voltage across resistor 27 and feeds it back to the current source. The obviousness argument is that Trilling's common-mode feedback and the '685 clamp are alternative, known solutions to the identical problem; KSR holds that "if a technique has been used to improve one device, and a person of ordinary skill would recognize that it would improve similar devices in the same way, using the technique is obvious." A PHOSITA seeking to avoid the device-matching sensitivity of common-mode feedback would be led to clamping, because clamping eliminates the matched-pair tracking requirement entirely — which is precisely the patent's stated insight.

Weakness of Combination B: Trilling also teaches away in a soft sense — it deliberately permits the common-mode voltage to develop across resistor 27 and then corrects for it. One can argue Trilling signals that this is the acceptable approach. That argument is available but weak under KSR, and the examiner's citation of Trilling against this very patent suggests the Office saw it only as general background.

Combination C: Williams '873 + Lavin '944 + Hilbiber

Williams supplies the matched push-pull philosophy and — importantly — the cross-neutralization capacitors that correspond to the '685 capacitors 68/70 (cross-neutralization of stray capacitance between inputs, the common node, and grid-anode capacitance). The claim to which this matters most is none of them directly (capacitors 68/70 are not claimed), but Williams is useful as evidence that cross-neutralization and matched-stage design were routine. Combination C is weaker than A on the clamp element.

Combination D: Any of the above + the acknowledged "bootstrap" technique

Because the specification calls the clamp a "bootstrap circuit" and describes the upper device as an emitter follower with "nearly unity voltage gain," a challenger need only cite one reference teaching bootstrapping to hold a device's terminal voltage constant (a technique with deep pre‑1960 roots in sweep-generator and constant-current-source design) to complete the motivation: make the anode follow the cathode, and the tube no longer sees the common-mode swing. I did not retrieve a specific bootstrap reference in this session, so I present this as the obvious gap-filler and flag it as unverified.

6. Claim-by-claim vulnerability ranking

Overall: the whole claim set rises or falls on one limitation — the fixed-voltage coupling impedance bridging the input pair's common node and the second pair's common control node. Every other element is in the record, and several are admitted by the applicant.

7. Rebuttal arguments the patent owner would raise (and their weight)

1. Secondary considerations / unexpected results. The asserted CMRR of 100,000:1, measured as high as 500,000:1, DC–100 kHz, plus stability "for rejection ratios as high as 100,000 to 1." This is the applicant's best material. But it requires a nexus to the claimed combination and a showing that the result is unexpected relative to the closest art (Lucas + Hilbiber) — not merely good. The patent's own data are given with no comparison to a Hilbiber-equivalent circuit, so the nexus is asserted rather than demonstrated.
2. Teaching away. Trilling's common-mode-feedback approach can be argued to steer the PHOSITA away from clamping. Under KSR, "teaches away" requires more than a disclosure of an alternative path; Trilling's existence as a different solution is not enough.
3. No bodily-incorporation requirement. In re Keller, 642 F.2d 413 (CCPA 1981) — the references need not be physically combinable; only the claimed subject matter must be obvious from their teachings.
4. Non-analogous art. Not available — all references are the same art.
5. Means-plus-function scope. Claims 1–2 recite "means for connecting a high resistance" and "voltage clamp means." Under pre‑AIA 35 U.S.C. § 112 ¶ 6 the Office must construe these as the corresponding structure in the specification and equivalents (In re Donaldson, 16 F.3d 1189 (Fed. Cir. 1994)). For § 103 purposes that narrows the claim to the disclosed bootstrap structure (coupling resistor 32 / impedance across nodes 22 and 42) — which, if anything, helps the challenger by removing breadth arguments, but requires the prior-art reference to disclose structure performing the function rather than a bare function.

8. Bottom line

• A § 103 case is constructible and, on the merits, non-trivial: Lucas '155 (series device making the input pair's plate potential insensitive to common mode) + Lavin '944 (series-energized transistor stack in a DC differential amplifier) + Hilbiber TP‑16 (constant-current input biasing for CMRR/drift), with Trilling '066 as either a fourth reference or the primary architectural teaching and Williams '873 as evidence that matched-stage and cross-neutralized design was routine.
• The vulnerability is concentrated in claims 1, 2, 4, 6 and 7. Claim 4 is the most exposed because its added limitation is the least technical: reach a fixed voltage across an impedance by forcing a fixed current through it. Claims 1 and 2 are exposed because they claim the function broadly while the specification admits the cascode and the "bootstrap" are conventional.
• The hinge is a factual gap, not a legal one: no reference I verified expressly connects the input pair's common node to the second pair's common control node through a fixed-voltage impedance. Lucas '155 teaches the function by a series stack, but I could not confirm from the retrieved text that its upper tubes' control electrodes are driven from the input pair's common cathode node. If Lucas (or Lavin, whose specification I could not retrieve) discloses that connection, the prima facie case against claims 1, 2, 3, 4 and 6 becomes strong; if it does not, the case rests on the KSR argument that bootstrapping is the predictable alternative to Trilling's common-mode feedback.
• Confidence: high on the reference characterizations I retrieved (Lucas, Trilling, Hilbiber, Williams) and on the pre‑AIA § 102 date analysis; low-to-moderate on the completeness of the obviousness case, because the two references most likely to close the "clamp" gap — US 3,275,944 (Lavin) and US 3,124,762 (Reaves) — were not available to me in text. A defensible § 103 opinion requires their full specifications and figures, plus the Hilbiber Figs. 5–6 circuits (the reference actually cited by the examiner "specifically page 4, FIG. 5"). I flag rather than fabricate those disclosures.
• Evidentiary odds and ends to note: (i) the patent's own OCR renders the Hilbiber citation as "pp. l10," while the publication data give pp. 434–439 for the December 1961 IRE Transactions version and a February 1961 Fairchild TP‑16 application note — both are the same author/work, and the patent cites "FIG. 5," which I have located in TP‑16 but not yet confirmed at "page 4" of the note's pagination; (ii) three of the cited references ('873, '066, '944) issued after the '685 filing date, so any rejection must be predicated on § 102(e) using their US filing dates, not their issue dates; (iii) no litigation or Board record exists for this patent (per the earlier sections), so there is no claim-construction or estoppel history to work with or against.