The Reconstructions of Konrad Zuse's Z3 Computer

Raul Rojas

The Reconstructions of Konrad Zuse's Z3 Computer

T0 review · 0 major / 6 minor · reviewed 2026-08-02 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper describes four functional reconstructions of Konrad Zuse's Z3, the 1941 relay computer destroyed in 1943, and argues that together they preserve the machine's operational heritage.

desk verdict A careful, useful survey of the four Z3 reconstructions; not a research contribution, but legitimate documentation with new details on Traber's machine. read the letter →

arxiv 2607.13089 v1 pith:3CWH4BQU submitted 2026-07-13 cs.GL cs.ARphysics.hist-ph

Raul Rojas This is my paper · ORCID

classification cs.GLcs.ARphysics.hist-ph

keywords KonradZuseZ3computerreconstructionrelayhistoryofcomputingpatentZ391Z3rDeutschesMuseum

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that the functional reconstructions of Konrad Zuse's Z3—built in 1962, 2001, 2010, and 2024—matter because the original 1941 machine was destroyed in an air raid, leaving only patents, diagrams, photographs, and memories. A sympathetic reader should see the paper as establishing that, despite their differing purposes and levels of fidelity, these machines collectively preserve the Z3's operational heritage. The paper shows how each reconstruction uses the surviving patent to recreate the Z3's binary floating-point relay logic, with trade-offs between authenticity, visibility, and reliability. If the paper is right, the Z3 remains experientially available to future generations even though the original is gone.

What carries the argument

The central blueprint is Konrad Zuse's patent application Z391 (filed in 1941 as Z 26476), which specifies the combination of a binary floating-point arithmetic unit, memory, and automatic program control. Each reconstruction implements this logical core in relays—mostly modern industrial relays—and diverges only in control implementation (electronic simulation, time-delay relays, or electromechanical stepping switches) and physical layout. This patent plus surviving circuit diagrams is the shared source that lets later builders recreate how the Z3 computed.

What would settle it

Finding a wartime photograph or drawing of the Z3's internal wiring that contradicts patent Z391, or a documented 1941 test computation that none of the reconstructions can reproduce, would show the inherited blueprint does not match the original machine.

Watch

Extended reading notes

Core claim

The paper documents four functional reconstructions of the Z3 built between 1962 and 2024: the 1962 replica produced by Zuse's own company, the 2001 Berlin hybrid with relay arithmetic and electronic control, Horst Zuse's Z3r, and Christof Traber's fully electromechanical machine. Each reproduces the Z3's logical and arithmetic operation from the surviving patent application and circuit diagrams, but they differ in purpose, fidelity, and control technology. The paper's central claim is that together these machines preserve the operational heritage of the 1941 computer, since the original was destroyed in 1943.

Load-bearing premise

The reconstructions inherit the 1941 machine's design only through the patent application Z391, surviving diagrams, and Zuse's recollection; if those documents describe an idealized or different machine, each replica preserves a plausible Z3 rather than the actual one.

Editorial extensions

If this is right

A working Z3 can be experienced by the public; the Swiss reconstruction is regularly demonstrated on the last Saturday of each month.
The Z3r's roughly sixty pushbuttons allow individual bits to be controlled, enabling step-by-step teaching of relay arithmetic.
The 1962 replica is the best documented, with construction records archived, so its circuitry can be studied precisely.
The four machines are not copies of one another; each embodies a distinct strategy for preserving a lost machine.
The 2001 hybrid shows that modern electronic control can faithfully reproduce the Z3's data path while using relay arithmetic.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The variety of reconstructions implies that 'the Z3' now exists as a family of interpretations anchored to the patent, not as a single physical object; historians should treat each replica as a hypothesis about the original.
A systematic comparison of the four machines' timing and arithmetic behavior against Zuse's early notes could test whether the patent describes the actual 1941 machine or an idealized design.
The didactic aims of later reconstructions may influence public perception, foregrounding programmability and floating-point arithmetic as the Z3's defining features.
The patent-based approach could be pushed further: a faithful reproduction of the original's dual-coil relays and switching drum might resolve remaining questions about its speed and control.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

First, the thing to know: this is a documentation piece, not a research paper, and it does what it sets out to do. It gives a clear chronological account of the four known Z3 reconstructions — the 1962 Zuse KG replica in Munich, Rojas's own 2001 Berlin machine, Horst Zuse's Z3r, and Christof Traber's newly finished Swiss machine — and it is honest about the differences among them: different goals, different fidelity, different trade-offs between historical accuracy and didactic clarity. The most useful new material is the description of Traber's machine, which has not had much coverage elsewhere: the 31-word memory, the replacement of hard-to-find dual-coil relays with a standard relay plus two diodes, and the prototypical 4-bit tape reader. I have not seen those details before in this compact form.

The author is a participant in one of the reconstructions, and he says so; the tone is not defensive. The comparative framing — if fidelity is the criterion, the 1962 machine is only logical, the 2001 machine is a hybrid, Z3r goes back to full electromechanical operation at near-original scale, and Traber's tries hard to stay electromechanical but with modern parts — is useful for curators and historians. The photographs are a bonus.

The soft spots are proportionate to the genre. This is a survey built on already published facts plus personal knowledge, so there is no new archival research. The central premise — that the patent application Z391 and surviving diagrams let us reconstruct 'the' Z3 — is acknowledged as a limit in passing but never examined. If those documents were idealized or incomplete, the reconstructions preserve a plausible Z3 rather than the 1941 artifact. The paper could have spent a paragraph on that epistemic question; it doesn't. Also, several technical figures (2,500 relays in Z3r, 31 words in Traber's memory) are stated without citations. The reference list is thin: four items, two of which are by the author or edited by him. None of this is fatal, but it narrows the paper's usefulness as a scholarly baseline.

Bottom line: this is a solid, well-written survey for people who need a compact overview of the Z3's reconstructions, and the author's insider status adds primary detail. It is not a major scientific contribution, but it is a legitimate historical documentation article. I would send it to a suitable venue like the IEEE Annals of the History of Computing rather than desk reject it. A good referee should check the technical details against the sources and push the author to address the fidelity-recovery caveat more explicitly.

Referee Report

0 major / 6 minor

Summary. The paper surveys four functional reconstructions of Konrad Zuse's 1941 Z3 computer: a 1962 replica by Zuse KG, now at the Deutsches Museum; a 2001 hybrid relay/digital reconstruction directed by Rojas, now at the Konrad Zuse Museum in Hünfeld; Horst Zuse's Z3r, presented in 2010; and Christoph Traber's electromechanical reconstruction (2021–2024) at the Enter Technikwelt Solothurn. The central claim is that, because the original machine was destroyed in 1943, these reconstructions collectively preserve the Z3's operational and cultural heritage, each with different goals and differing degrees of fidelity. The paper presents chronological, detailed descriptions of each machine, including technical choices, locations, and photographs.

Significance. If the descriptions are accepted, the paper offers a useful consolidated account of the known Z3 reconstructions, of particular value to historians of computing and museum curators. Its strength is that it does not overstate fidelity: it explicitly notes that the 1962 replica used newer relays and only one memory cabinet, the 2001 machine relies on a PC and digital control, the Z3r replaces stepping switches with time-delay relays, and Traber's prototype tape reader is limited to four bits. The reader's stress-test concern—that the reconstructions may preserve a plausible rather than the actual 1941 Z3—does not land as a fatal flaw, because the paper's claim is about preserving functional operation and heritage, and the paper openly discloses the evidentiary limits of patents, photographs, and memory. No internal inconsistency or unsupported load-bearing assertion was found.

minor comments (6)

[Introduction / Berlin Reconstruction] The patent application is identified as 'Z 26476' in the introduction and as 'Z391' in the Berlin Reconstruction section and reference [3]. If these are the same document, please reconcile the numbering; if they differ (e.g., internal Zuse number vs. patent-office file), say so explicitly.
[Summary] The claim that the Z3 is 'the world's first functional, freely programmable, fully automatic binary computer' is a strong historical statement and is currently unsupported by a specific citation. Reference [1] is cited for the Z3's status as a pivotal artifact, but a direct citation or qualification would strengthen the paper.
[Berlin Reconstruction of 2001] The author's own role as the leader of the 2001 reconstruction is described in the third person. For editorial transparency, add a brief competing-interest or author-contribution statement noting that the author directed that project and edited the volume cited as reference [3].
[Figures] Figure captions could be more complete: for Fig. 1, the source is only 'Wikimedia Commons' without a date or photographer; for Figs. 2–4, the photographer is named but not the date or location. Adding this information would help readers assess the materials.
[Throughout] Some specific technical quantities (e.g., '2,500 modern industrial relays' for Z3r, '31 words' for Traber's memory, 'approximately sixty pushbuttons') are stated without a supporting reference or derivation. If these figures come from the authors' direct inspection or from cited technical reports, a short note indicating the basis would be useful.
[Title and formatting] The title and some body text contain OCR-like artifacts (e.g., 'Reconstruc-ons', 'Mathema2cs', 'sta2s2cs'). Please proofread the final text and ensure that mathematical symbols and hyphens are properly typeset.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: descriptive historical survey grounded in externally documented artifacts.

full rationale

This paper is a descriptive historical survey rather than a derivation chain. The central claim—that four functional Z3 reconstructions exist and collectively preserve the machine's operational heritage—is supported by specific artifacts, photographs, named builders, institutions, dates, and detailed technical descriptions. The paper does not fit parameters, derive equations, or invoke uniqueness theorems. Its explicit reliance on the patent application Z391, surviving circuit diagrams, photographs, and Zuse's memory is disclosed as a limitation, not used as an input that is later relabeled as a prediction. The author's self-citations ([1], [3]) provide historical context and primary-source transcriptions, but they are not load-bearing: the existence and characteristics of the reconstructions stand on the independently documented machines themselves and on the distinct implementations described. No circular step, no self-definitional result, and no fitted-input-called-prediction pattern is present. The paper is self-contained as a review and therefore receives score 0.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

No mathematical derivation or empirical fit is involved. The review rests on historical-domain assumptions about the reliability of the surviving documentary record and the public availability of the machines.

assumptions (3)

domain assumption The surviving patent applications (Z26476/Z391), circuit diagrams, photographs, and Zuse's memory are sufficient to reconstruct the functional architecture of the original 1941 Z3.
Invoked in the Deutsches Museum and Berlin sections; if false, the reconstructions' fidelity to the lost original is undermined.
domain assumption The original Z3 was completely destroyed in 1943 and none of its components survived.
Stated in the Deutsches Museum section and in the summary; underlies the paper's claim that reconstructions are the primary surviving access to the machine.
domain assumption The four described machines exist and are publicly accessible as stated.
The paper gives museums and dates; independent verification would require visiting the collections.

how reviews work

0 comments

Cite this review

Pith. "Pith review of The Reconstructions of Konrad Zuse's Z3 Computer." pith.science (2026). https://pith.science/paper/3CWH4BQU

@misc{pith2026260713089,
  author       = {Pith},
  title        = {Pith review of: The Reconstructions of Konrad Zuse's Z3 Computer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3CWH4BQU}},
  note         = {Machine review of arXiv:2607.13089}
}

read the original abstract

This paper reviews the various reconstructions of Konrad Zuse's Z3 computer, built in 1941. The first is the reconstruction made by Zuse himself in the 1960s, which is now housed at the Deutsches Museum in Munich. Second is the 2001 reconstruction, made in Berlin, and which has a didactic purpose. Third is the Z3r, a Z3 reconstruction built by Horst Zuse in honor of his father's 100th birthday. Fourth is the reconstruction by Christoph Traber in Switzerland. All of these machines strive to preserve the cultural heritage that the Z3 represents.

Figures

Figures reproduced from arXiv: 2607.13089 by the authors.

**Figure 2.** The 2001 reconstruc5on of the Z3. Modern relays were interconnected by conduc5ve tracks on circuit [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Horst Zuse’s Z3r at an exhibi5on at the Heinz Nixdorf MuseumsForum. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Christof Traber’s reconstruc5on in Switzerland. The three racks contain, from le= to right, the memory, [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

3 extracted references

[1]

The Reconstruc-ons of Konrad Zuse’s Z3 Computer Raúl Rojas Dept. of Mathema2cs & Sta2s2cs University of Nevada Reno What mo2vates engineers to reconstruct historical computers? Vintage Computer Fes2vals are held regularly in Europe and the United States, where early computers are presented to new genera2ons of programmers and enthusiasts. This interest ca...

1941
[1943]

calcula2ng device

This meant the loss not only of the machine itself but also of the most important material evidence of its opera2on. Only a small number of people had seen it working. On June 16, 1941, Zuse ﬁled a patent applica2on with the German Patent Oﬃce for a “calcula2ng device” under applica2on number Z 26476 [2]. The applica2on described the fundamental combina2o...

1941
[2011]

3: Horst Zuse’s Z3r at an exhibi5on at the Heinz Nixdorf MuseumsForum

Fig. 3: Horst Zuse’s Z3r at an exhibi5on at the Heinz Nixdorf MuseumsForum. Photograph by H. Zuse. Horst Zuse used approximately 2,500 modern industrial relays manufactured by Finder. These are mounted on easily accessible rails, can be replaced individually, and are equipped with LEDs indica2ng their switching states. The historical stepping switches wer...

2021

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.

[1] [1]

The Reconstruc-ons of Konrad Zuse’s Z3 Computer Raúl Rojas Dept. of Mathema2cs & Sta2s2cs University of Nevada Reno What mo2vates engineers to reconstruct historical computers? Vintage Computer Fes2vals are held regularly in Europe and the United States, where early computers are presented to new genera2ons of programmers and enthusiasts. This interest ca...

1941

[2] [1943]

calcula2ng device

This meant the loss not only of the machine itself but also of the most important material evidence of its opera2on. Only a small number of people had seen it working. On June 16, 1941, Zuse ﬁled a patent applica2on with the German Patent Oﬃce for a “calcula2ng device” under applica2on number Z 26476 [2]. The applica2on described the fundamental combina2o...

1941

[3] [2011]

3: Horst Zuse’s Z3r at an exhibi5on at the Heinz Nixdorf MuseumsForum

Fig. 3: Horst Zuse’s Z3r at an exhibi5on at the Heinz Nixdorf MuseumsForum. Photograph by H. Zuse. Horst Zuse used approximately 2,500 modern industrial relays manufactured by Finder. These are mounted on easily accessible rails, can be replaced individually, and are equipped with LEDs indica2ng their switching states. The historical stepping switches wer...

2021