--- title: "Two Factories, One Computer: The Making of the Timex/Sinclair 1500" id: "69599" type: "post" slug: "two-factories-one-computer-the-making-of-the-timex-sinclair-1500" published_at: "2026-05-24T11:33:01+00:00" modified_at: "2026-05-24T11:39:40+00:00" url: "http://localhost/timex-sinclair/two-factories-one-computer-the-making-of-the-timex-sinclair-1500/" markdown_url: "http://localhost/timex-sinclair/two-factories-one-computer-the-making-of-the-timex-sinclair-1500.md" excerpt: "The Problem That Needed Solving By late 1982, Timex Computer Corporation had a gap problem. The TS1000 had launched in July of that year and sold in extraordinary numbers, over a million units by some accounts, but it was showing..." taxonomy_category: - "Timex/Sinclair" taxonomy_post_tag: - "Timex Computer Corporation" - "Timex Portugal" - "TS 1500" taxonomy_model: - "Timex/Sinclair 1500" taxonomy_indiv: - "Danny Ross" - "Louis M. Galie" - "Tom Belpasso" - "Will Collum" --- [/wp-content/uploads/2026/05/TS1500-Panphlet-inside2.jpg](/wp-content/uploads/2026/05/TS1500-Panphlet-inside2.jpg) ## The Problem That Needed Solving By late 1982, [Timex Computer Corporation](/company/timex-computer-corporation/) had a gap problem. The [TS1000](/model/ts-1000/) had launched in July of that year and sold in extraordinary numbers, over a million units by some accounts, but it was showing its limitations fast. American consumers, unlike their British counterparts, had little patience for the machine’s aggravations: the membrane keyboard that barely responded to touch, the RAM pack that could cause the machine to crash if you bumped the computer, the cassette interface that demanded careful calibration every time you wanted to load a program. In the UK, hobbyists were accustomed to fussing. In America, people simply would not put up with it. The color computer (what would eventually become the [TS2068](/model/ts-2068/) ) was in development under [Director of Engineering Lou Galie](/indiv/lou-galie/) , who had come over from Burroughs Corporation in 1982 specifically to lead that project. But it was going to take time. There was a gap in the product line, and the team wanted something to fill it. What they needed was a machine that was recognizably a TS1000 but fixed: better keyboard, built-in RAM, improved cassette loading. Through a companion accessory, the 1510 cartridge player, it might even gain a foothold in the cartridge-based gaming market that Atari was dominating. That machine would be the TS1500. And the story of how it got made involves two engineering teams on opposite sides of the Atlantic, a stack of telexes, several competing circuit board designs, and a prolonged argument about who, exactly, was going to determine what this computer looked like. ## Timex Portugal and What Made It Unusual Most people in the American Timex/Sinclair community had no idea the factory in Portugal existed until after Timex Computer Corporation shut down in February 1984. When that happened, enterprising members of the user community tracked down the Portuguese operation and discovered that deals could be made: floppy disk systems, spare parts, even some of the more exotic hardware that had never made it stateside. What they found was a remarkable facility. [Timex Portugal](/company/timex-portugal/) wasn’t just an assembly line. Unlike the Dundee operation, which was primarily manufacturing, Timex Portugal had an R&D team, a group of engineers who worked on real hardware problems and came up with their own solutions. [João Diogo Ramos](/indiv/joao-diogo-ramos/) , whose [Load ZX Spectrum Museum](https://loadzx.com/en/) has accumulated a significant portion of the documentary record from this period, spent years cultivating relationships with the men who ran that team. The museum’s visitor guide describes the origin: [Fred Olsen](https://en.wikipedia.org/wiki/Fredrik_Olsen) , owner of Timex, challenged António Gomes, Chief Executive of the Timex Portugal factory at [Quinta dos Medronheiros, Lazarim, Caparica](https://www.google.com/maps/place/R.+Q.ta+dos+Medronheiros,+Portugal/data=!4m2!3m1!1s0xd193552abfb27cb:0x8fa5a61e542d7737?sa=X&ved=1t:242&ictx=111) , to convert the company’s mechanical watch assembly operation into one focused on electronics. (The same period also saw Timex Scotland invest in the [Nimslo 3D camera](https://cameralegend.com/2019/09/27/cult-camera-collection-a-look-back-at-the-nimslo-nishika-3d-cameras/) , a measure of how aggressively Olsen was pushing the company to diversify.) Gomes recruited a consultant, Professor Álvaro Oliveira, and together they built the operation that would put Portugal on the Spectrum’s map. The museum obtained an extraordinary archive: hundreds of kilograms of paper, telexes, memos, schematics, bills of materials, and internal correspondence that had spent decades in Álvaro Oliveira’s personal collection when his family donated it following his death. Oliveira was a central figure throughout. He ran a company called Alpha Micro, was never formally a Timex employee but served as a consultant for the entire run of the computer operation, and had his hands in nearly everything: the Portuguese software library, the manual translations (his wife handled those), and the technical coordination between Portugal and Middlebury. He also, as the documents reveal, had access to internal American correspondence that by any normal logic should not have existed in Portugal at all: telexes between engineers in Middlebury and NCR in Ohio, memos discussing chip costs and schedule pressures, documents with no Portuguese names on them. He accumulated these as his relationship with the Timex USA organization evolved, becoming an informal link between the two operations even after American engineers moved on from the computer business. ## The American Vision: A New Chip TS1500 SCLD Prototypes The Middlebury approach to the 1500 was technically ambitious. Documents from January 1983 show engineers considering whether to integrate the gate array, ROM, and TTL buffer into a single custom chip, collapsing several components into one. This idea was discussed with [NCR](https://en.wikipedia.org/wiki/NCR_Voyix) , which had a facility in Miamisburg, Ohio and was the manufacturer of the [SCLD](https://en.wikipedia.org/wiki/Complex_programmable_logic_device) (Semi-Custom Logic Device) that would become the heart of the production 1500. The chip design went through multiple revisions. The Álvaro Oliveira archive includes the full logic translation documents for the SCLD, sheets labeled “TS 1500 Final Logic Translation” with a change log running from late 1982 through early 1983, with new revisions appearing nearly every week in January alone. By the time development was well underway, the chip was on revision H. When the third pass of silicon from NCR came back unacceptable, the team was forced to a fourth pass, causing at least a two-week slip in the overall schedule. This was a real chip design effort. The 68-pin NCR chip replaced the Ferranti ULA that had been used in the ZX81 and TS1000, requiring less power, improving display quality, and reducing spurious RFI emissions. It also fixed the “ghost image” problem caused by the original ROM not being properly decoded to the lowest 8K of address space. The engineers working on this included Lou Galie (Director of Engineering), Simon Zetto, [Tom Belpasso](/indiv/tom-belpasso/) , [Victor Schiff](/indiv/victor-schiff/) , and Dan Yoder. Lou Galie also served as the primary liaison with Portugal: he was the name at the top of the most important telexes, the person António Gomes negotiated with, the human connection between the two operations. [Daniel Ross](/indiv/ross-dan/) , as Vice President of Timex Computer Corporation, was involved at the business level rather than the engineering level. At this stage, Middlebury’s concept for the physical product was conservative: take the TS1000 case and keyboard, paint it silver instead of black, and fit a slightly enlarged circuit board inside it to accommodate the eight RAM chips. The machine would look like a TS1000 variant. That was the plan. One other piece of American-side work shaped the 1500’s character: a new ROM. Programmer [Will Collum](/indiv/will-collum/) modified the original TS1000/ZX81 ROM to add support for the [TS1510 cartridge player](/product/ts-1510-cartridge-player/) and corrected some pre-existing bugs in the Sinclair code. He also, unfortunately, introduced three new ones. LOAD failures in certain conditions jumped execution to the middle of an instruction at address 03E5h, corrupting memory or crashing the machine outright. The cartridge check routine assumed no RAM existed at address 2000h and could crash if that location happened to hold a 1. And a sufficiently long program, pushing toward 16K, would drive the display file to address 32767, causing the machine to report out of memory even with a RAM pack installed and usable space remaining. None of these were caught before the machine shipped. ## The Portuguese Vision: Use What You Have By March 1983, with the 1500’s launch approaching and a substantial number of decisions still unmade, Portugal came in with a counterproposal. A telex dated March 11 from António Gomes to the American team lays it out plainly: “Based on local development, Portugal are submitting for appraisal a version of TS based on current technology which we believe will meet the requirements of marketing, finance and FCC.” This was not the American design, and it was not a silver TS1000. TS1016S Prototype The stated objectives were three: counter Timex’s main competitors, utilize excess inventory held in Portugal as a result of the [TS1016S](/product/ts-1016s/) cancellation, and utilize the excess labor that same cancellation had created. Some context is required here. The original TS1016 (styled exactly like the ZX81 16K RAM pack) had launched successfully in the US and sold well, because 2K was simply too little RAM for most purposes. There was a second TS1016, a sleek and lower-profile redesign (labeled internally as the TS-1032), that was the one that never made it past prototype stage. It was the cancellation of this second version that left Portugal holding components and workers with capacity to spare. The Portuguese team was making a practical argument: do not develop something new when you have parts and people sitting idle. Their proposed machine used existing TS1000-era components, including the eight-chip DRAM layout the American design was trying to rationalize down to two chips. The RAM cost differential was significant. Eight chips ran about eight dollars in component cost, and the American design’s two-chip approach represented a real saving at volume. Portugal pushed back on how to calculate this, arguing that they were going to sell the existing memory inventory at a profit rather than absorb it as a write-off, which changed the math considerably. [/wp-content/uploads/2026/05/IMG_2487.jpeg](/wp-content/uploads/2026/05/IMG_2487.jpeg) TS1500 Circuit Board The Portuguese design also featured an improved serial interface running at 4,000 baud, an attempt to address the cassette reliability problem that had plagued the TS1000. This was particularly acute in Portugal itself, where pirated software on dubious tape stock made loading miserable. Games were copied in stores using whatever blank cassettes happened to be on hand. Friends would use 60- or 90-minute music tapes and mark the counter to find the right spot. The quality was bad enough that collectors today report that original UK tapes load reliably while many Portuguese copies don’t. The proposal also included automatic gain control for the fast loader, which Portugal argued would eliminate loading problems not just on the 1500 but retroactively improve the experience for TS1000 owners already in the field. ## The Case That Changed Everything [/wp-content/uploads/2026/05/IMG_2484-scaled.jpeg](/wp-content/uploads/2026/05/IMG_2484-scaled.jpeg) TS1500 The most consequential element of the Portuguese proposal was the case. Rather than the silver-painted TS1000 enclosure Middlebury had been planning around, Portugal proposed adopting the reworked ZX Spectrum case: nearly the same mold Sinclair was using for the Spectrum, fitted with the Spectrum’s membrane keyboard and rubber key overlay, but with legends for the TS1500 instead of Spectrum key labels. This was a radically different product proposal, visually and ergonomically. The Spectrum case was larger and more refined. The keyboard had better tactile response than the original TS1000 membrane. Portugal had an in-house designer named Eric Suzano, described by Ramos as their equivalent of [Rick Dickinson](https://en.wikipedia.org/wiki/Rick_Dickinson) , the designer responsible for the distinctive look of Sinclair’s UK machines. Suzano worked in what they called the “two room,” a prototyping and design space within the factory. He later gave Ramos a silver-painted TS1000 case bearing a prototype TS1500 label, physical evidence of design work done before the final product was determined, but also a reminder of what the 1500 could have looked like had Middlebury’s original concept prevailed. The Álvaro Oliveira archive contains handwritten schematics for a 1500-circuit-board sized to fit inside a standard TS1000 enclosure. This smaller form factor was explored and eventually set aside. There were also keyboard samples arriving from Taiwan, including a dark gray variant. Case paint was being tested for RF interference characteristics: nickel acrylic paint versus aluminum metallization. These were not abstract discussions. They were testing actual sample units. ## The RAM Remapping Problem One specific technical question generated considerable back-and-forth between the teams. What happened to the 1500’s built-in 16K when a user plugged in an external RAM pack? The external pack’s memory position in the address space was fixed. Existing software expected to find RAM pack memory at a specific address, and you could not change where the external hardware appeared. What you could change was where the internal RAM appeared. Portugal worked out a solution using a switch and a NAND gate: when a 16K RAM pack was attached, the internal 16K would remap to appear after the RAM pack’s memory, not before it. The total usable RAM would be 32K, with the external pack occupying the lower page and the internal RAM occupying the upper page. A December 1982 memo from the American side shows this same problem being examined, with a proposed “remapping scheme” that the author indicated he would attempt to implement on a breadboard without impacting schedule. A handwritten note on the memo reads: “forget third party problems.” The practical reasoning was that Timex’s own 16K pack wouldn’t conflict with internal RAM, and third-party packs with higher capacity were not Timex’s concern. The production 1500 handles this remapping automatically, detecting whether a pack is plugged in and adjusting without requiring a manual switch. That said, [not all RAM packs worked](/article/compatibility-of-machines/) . Timex changed a signal on the edge connector, and even with an extender card, a number of third-party packs were incompatible. The rearranged power and cassette connectors also physically prevented longer units, like [Memotech’s](/company/memotech/) , from connecting at all. Users who wanted to expand beyond 16K needed to know which packs were compatible, and [a few POKEs were required](/article/using-ram-packs/) to make the system recognize additional RAM in any case. ## The Knowledge Transfer The Portuguese engineers’ ability to do independent hardware work at all was not accidental. Early in the Timex Portugal computer operation, Álvaro Oliveira and others had traveled to a Timex lab in Cupertino, California (where much of the computer engineering work was being done, separate from the watch headquarters in Middlebury) and spent several weeks absorbing the technical knowledge they needed to run an R&D operation back home. Some engineers also traveled to Dundee to learn manufacturing processes and equipment. That transfer was real enough that by 1983, Portugal had built its own test equipment for the assembly line. One piece João discusses is a jig tester built specifically for the TS1000: a device the board would plug into for automated short circuit and open circuit testing. The initial defect rate coming off the Portuguese assembly line was around 27%, by one engineer’s recollection, and the test equipment was part of how they drove that number down. The communications infrastructure tying all of this together was the telex. Fast by 1983 standards, slow by any other measure. Action items generated follow-up telexes asking whether the action items had been acted upon. Requests for test results produced responses saying samples were in transit and results would be available in two to three weeks. A document asking Portugal to define test points for the 1500 appears to have been drafted by António Gomes’s secretary and routed through Álvaro Oliveira rather than Gomes himself, a sign of how communication pathways had developed and how central Oliveira was to them. ## Three Weeks to Launch: The FCC Problem On April 29, 1983, eighteen days before the TS1500 was announced on May 17, the American team tested a prototype from Portugal. It failed FCC Class B certification at several points. A machine that can’t pass FCC emissions testing cannot legally be sold in the United States. The Jack Isaacs organization, a third-party testing lab, was brought in to analyze the failure modes. Álvaro Oliveira was included in that analysis. Telexes went back and forth about what was wrong with the RF shielding and the modulator. Portugal was running its own RFI testing in parallel, shipping sample units with different case metallization approaches: some with nickel acrylic paint, some with aluminum metallization. Around the same time, a memo about prototype changes to the modulator was documenting component-level fixes: capacitor additions, resistor value changes, component substitutions. These units were still being called “prototypes.” Two months before the planned retail launch of a product in development since at least December 1982, the machine wasn’t clearing its primary regulatory hurdle. It cleared eventually. The FCC filing lists the applicant as Timex Computer Corporation and the manufacturer as Portugal, with emissions testing conducted by Atlantic Research Corporation. But the timeline was tight enough that the decision about which version of the 1500 would actually go into production, the American SCLD approach or the Portuguese current-technology approach, appears to have remained open past the point where a comfortable schedule would have had it resolved. Documents from late March show Portugal still asking: what is the status of the 1500? What approach has been chosen? The response from the American side was noncommittal. ## Who Won? The archive doesn’t contain a clear decision memo. The gap in the Álvaro Oliveira collection from May to October 1983 swallows whatever conversation settled the question. What exists on the other side of that gap is the product itself, and from the product you can read the outcome. The production TS1500 uses the SCLD chip, the American-designed custom logic device that went through four passes at NCR. That’s a clear win for Middlebury’s technical approach. But the machine is in the Spectrum-style case Portugal proposed, with the Spectrum membrane and rubber keyboard carrying 1500 legends. Every TS1500 ever sold came out of Portugal, without exception. The FCC certification was filed with Portugal as manufacturer. The keyboard membrane work, except for the special rubber overlay itself (sourced externally, as the specific printing process required for rubber keys was never established in Portugal), was developed and produced in Portugal. What the record actually shows is that both sides contributed something essential. Middlebury provided the custom silicon that made the 1500 technically viable at its price point. Portugal provided the physical design, the manufacturing process, the regulatory pathway, and the engineering work that turned a chip and a circuit board into a product that could sit on a retail shelf. Neither team could have shipped the machine without the other. The silver TS1000 never existed. Portugal’s Spectrum-style proposal won on the hardware side completely. Timex had planned to launch the 1500 in July. Manufacturing delays pushed the actual launch to September. The retail price was $79.99. [Timex simultaneously lowered the TS1000 price to $49.95.](/article/timex-developments/) ## Reception The press and existing enthusiasts were puzzled. Timex Sinclair User, [covering the 2068 preview](/article/the-new-arrivals/) , ran a sidebar titled “Who Needs the Timex Sinclair 1500?” The answers, which the editors said took weeks to figure out, were weak at best. [Softsync’s marketing director Ken Coach was blunt](/article/the-bargain-basement-blues/) : “Six months ago, it would have been great. Now, the moment has passed.” Prices for the TI 99/4A and VIC-20 had already dropped below $100. The 1500 was a computer with no color graphics launched into a market that had decided color graphics were the baseline. [Wayne Green](/indiv/wayne-green/) , publisher of several computing magazines, reported a “rising chorus of frustrated Timex users” telling friends not to waste their money on Timex computers generally. [Timex itself remained confident](/article/timex-introduces-the-1500/) . Daniel Ross, Vice President of Timex Computer Corporation, said at the announcement: “We believe the many advanced features of the T/S 1500 will generate enthusiastic consumer and retailer response, as they did our Timex Sinclair 1000.” The 1500 did find its audiences, just narrower ones. The [National Education Corporation](/blog/the-nec-technical-literacy-series/) built a Technical Literacy Series around it, bundling the 1500 with the TS2020 cassette recorder, eleven tutorial booklets, and a briefcase, selling the kit for $295. [Parker Electronics](/tag/parker-electronics/) adapted the 1500 as the user interface for a Carrier Corporation HVAC controller, which is perhaps the most unexpected deployment of a rubber-keyboard home computer on record. ## After the Launch: The Quality Fight By October 1983, returned units were arriving back at Middlebury with problems. A memo from that month documents the findings: four defective 1500s returned from a customer evaluation. Two had keyboard defects, one a loose membrane and one with the membrane not properly seated. The picture quality on the remaining units wasn’t acceptable. Victor Schiff checked the TV channel frequencies and found them out of specification. The memo’s author was asking whether the tuning problem was systemic, whether lots from Portugal were routinely shipping with channel tuning off. The language is careful and professional, but the subtext is clear. Someone in Connecticut was looking at the work coming out of Portugal and finding it wanting. Viewed from forty years out, this is ordinary organizational dynamics in a cross-continental manufacturing operation. In the moment, for engineers who had been working on this machine for over a year across a telex connection, it was surely less easy to take in stride. ## The Long Tail When Timex Computer Corporation closed in February 1984, it didn’t close everything. Timex Portugal kept going. And the documents show that even as the Portuguese operation was building TC 2068s and pursuing new markets in Latin America, work on the 1500 continued. A May 1985 memo discusses a cost-reduction version using two RAM chips instead of the original eight-chip layout, with fresh FCC certification underway. A July 1985 document confirms the approval came through, notes the ROM lead time of eight weeks, and asks whether Portugal can supply an additional 5,000 TS1000/RAM pack sets while moving forward on the 1500 revision. António Gomes himself traveled to Middlebury with the goal of bringing the TC 2068 to the American market. Middlebury was still legally in the picture as the parent operation. Nothing the Portuguese branch did commercially was automatic or unilateral. What this long tail says is that the people at Timex Portugal had never understood their role as simply executing decisions made elsewhere. They had engineered things, proposed things, pushed back on things, and when the parent organization lost interest, they kept going. António Gomes passed away in 2020. Álvaro Oliveira died before him. The archive their work left behind, now held at the Load ZX Spectrum Museum, is why we know any of this at all. ## Timex Connection Video Series In 2023, Joao Diogo Ramos and David Anderson sat down, over the internet, and recorded a series of videos exploring the Timex/Sinclair 1500 through documentation preserved by the [LOAD ZX Spectrum museum](https://loadzx.com/en/) . The seven part series is an eye-opening view into the development process and Timex’s internal politics. [TIMEX Connection: TS1500, Pt. 1 May 01, 2023 João Diogo Ramos (LOAD ZX Spectrum museum founder) and David Anderson (book author and founder of the timexsinclair.com website) bring you a new Youtube show, called TIMEX Connection, that delves into the world of TIMEX computers.](/video/timex-connection-ts1500-pt-1/) [TIMEX Connection: TS1500, Pt. 2 June 17, 2023 In this second episode of "Timex Connection," João Diogo Ramos (LOAD ZX Spectrum museum founder) and David Anderson (book author and founder of the timexsinclair.com website) examine both hardware and documentation related to the Timex/Sinclair 1500.](/video/timex-connection-ts1500-pt-2/) [TIMEX Connection: TS1500, Pt. 3 July 21, 2023 This episode examines Timex 1500 hardware and February 1983 technical documentation, revealing intensive development work three months before launch, including keyboard specs and RAM remapping discoveries.](/video/timex-connection-ts1500-pt-3/) [TIMEX Connection: TS1500, Pt. 4 August 07, 2023 This episode covers chronological TS-1500 development boards, March 1983 project management charts, and competing US/Portuguese designs racing toward May launch.](/video/timex-connection-ts1500-pt-4/) [TIMEX Connection: TS1500, Pt. 5 August 07, 2023 This episode examines April 1983 documents revealing Portuguese prototypes failed FCC certification just weeks before May launch, plus marketing strategy targeting extreme low-cost computer market.](/video/timex-connection-ts1500-pt-5/) [TIMEX Connection: TS1500, Pt. 6 October 27, 2023 This episode examines April-May 1983 final launch push, revealing US analysis comparing Portuguese versus Middlebury designs, with Portugal's version having cost advantages but severe technical problems.](/video/timex-connection-ts1500-pt-6/) [TIMEX Connection: TS1500, Pt. 7 November 03, 2023 This last episode completes a comprehensive documentation analysis, examining October 1983-July 1985 documents revealing continued cost reduction efforts and TS-1500 production extending surprisingly into 1985.](/video/timex-connection-ts1500-pt-7/) *Additional US-side material came through [Eric Johnson](/indiv/eric-johnson/) (whose uncle worked at Timex Middlebury), preserved by [Neil Cohen](/indiv/neil-cohen/) , and shared with David Anderson.* ### Tags [Timex/Sinclair](/category/timex-sinclair/) [TS 1500](/tag/ts1500/) [Timex Computer Corporation](/tag/timex-computer-corporation/) [Timex Portugal](/tag/timex-portugal/) ### People [Danny Ross](/indiv/ross-dan/) [Louis M. Galie](/indiv/lou-galie/) [Tom Belpasso](/indiv/tom-belpasso/) [Will Collum](/indiv/will-collum/) ### About The Author [David Anderson](/author/david/) I'm a big fan of the Timex/Sinclair computers: my first was the TS 1000. I love building things for my TS 2068 and sharing them with others on this site and andertone.com.