--- title: "Notes on SCLD Failures and potential solutions" type: "article" slug: "notes-on-scld-failures-and-potential-solutions" url: "http://localhost/article/notes-on-scld-failures-and-potential-solutions/" markdown_url: "http://localhost/article/notes-on-scld-failures-and-potential-solutions.md" published_at: "2020-10-27T17:14:39+00:00" modified_at: "2026-11-08T09:54:08+00:00" featured_image: url: "http://localhost/wp-content/uploads/2022/04/20230809-041332.jpg" excerpt: "Victor M.S. AcuñaBuenos Aires, October 16, 1993 UPDATE MAGAZINEP.O. BOX 1095Peru, IndianaU.S.A. Dear Frank: I am sending my subscription fee ( in cash because I must pay $ 10. plus, for a check), and a few tech. notes about my circuits in repairing TS-2068's. Some words about my curriculum: I am a Superior Electronics Technician…" post_tag: - name: "Full Text" slug: "fulltext" taxonomy: "post_tag" url: "http://localhost/tag/fulltext/" - name: "Hardware project" slug: "hardware-project" taxonomy: "post_tag" url: "http://localhost/tag/hardware-project/" - name: "SCLD" slug: "scld" taxonomy: "post_tag" url: "http://localhost/tag/scld/" - name: "TS 2068" slug: "ts2068" taxonomy: "post_tag" url: "http://localhost/tag/ts2068/" model: - name: "Timex/Sinclair 2068" slug: "ts-2068" taxonomy: "model" url: "http://localhost/model/ts-2068/" indiv: - name: "Victor Acuna" slug: "acuna-victor" taxonomy: "indiv" url: "http://localhost/indiv/acuna-victor/" publication: "Update Magazine" publication_r: id: 10283 title: "Update Magazine" type: "periodical" url: "http://localhost/periodical/update-magazine/" authors: "Victor Acuna" issues_articles: - id: 26490 title: "Update April 1995" type: "issue" url: "http://localhost/issue/ts-2068-up-date-april-1995/" pages: "5-7" pubdate: "April 1995" archive_link: false volumeissue: "vn" --- # Notes on SCLD Failures and potential solutions [Victor M.S. Acuña](http://localhost/indiv/acuna-victor/) Buenos Aires, October 16, 1993 UPDATE MAGAZINE P.O. BOX 1095 Peru, Indiana U.S.A. Dear Frank: I am sending my subscription fee ( in cash because I must pay $ 10. plus, for a check), and a few tech. notes about my circuits in repairing TS-2068’s. Some words about my curriculum: I am a Superior Electronics Technician ( a tertiary level grade), and actually I am a student at CIE (Cleveland Institute of Electronics), in Electronics Engineering course. I work since 1986 in the computers field, and really I am an expert in repairing TSs. The topics I send in this letter are over repairings of SCLD IC. Evidently the troubles generator in the TSs is the SCLD IC. In the service shop I encountered that about 60% of TS’s fails are over it, and like it change cost about $ 30.- ( I have until many of them, and I did a special iron to remove them), when the price of the machine began to fall, an option was design circuits specifically to repair parts of it. Now a working TS 2068 cost around $ 50. And the TS repair market, practically don’t exist. The total number of machines imported to my country was around 5,000. Principally, the SCLD problems are: 1) No Interrupts ( every ~17ms) from SCLD to CPU, and then “no keyboard” fail. 2) One or more rows of keyboard not work. SCLD not read or it’s data lines are “open”. 3) No cassette. Not read data from the cassette player. 4) The refresh circuit build into the SCLD for the high memory bank ( 32,768-65,535) A7R don’t work. The machine initialize, but the programs crash. THIS IS THE MOST “DESTRUCTIVE” FAIL, and commonly only reparable by changing the SCLD. Yes it’s my great discovery !!! Thanks, thanks,……. 1) The SCLD provide, via INT signal (Pin 16 Z80) the signal necessary to scan the keyboard. This signal are generated in every field of the screen ( 1/60th sec.). But if it aren’t, the CPU never read the keyboard. More explain about it are in the Technical Manual. Originaly, I tested taking the signal from vertical synchronism, but it don’t work well. Then I designed a circuit from a 555 timer, and it worked O.K. The circuit is: [![555 timer interrupt circuit and output signal](http://localhost/wp-content/uploads/2020/10/update-22123-1-1024x491.png)](http://localhost/wp-content/uploads/2020/10/update-22123-1.png)555 timer interrupt circuit and output signal 2) If the keyboard fail only in some rows, and one are sure that the problem is in the data lines, ( via oscilloscope level test, for example) and not in the keys itself. Then the solution are build a circuit that overlap the SCLD keyboard circuit. To remember you, the keyboard use D0 to D4 to scan the KB0 to KB4 lines, in the port 254. This correspond to A0 = 0. Then the circuit is: [![74LS33 keyboard data line circuit with key column table](http://localhost/wp-content/uploads/2020/10/update-22123-2-1024x363.png)](http://localhost/wp-content/uploads/2020/10/update-22123-2.png)74LS33 keyboard data line circuit with key column table 3) The cassette input view from CPU is the same port that the keyboard, but in the D6 bit. All we need to do, is amplify the audio input, decode the 254 port, and accouple it to D6. Then a “dead” cassette TS come to work. The circuit is: [![Cassette input amplifier circuit (74LS367)](http://localhost/wp-content/uploads/2020/10/update-22123-3-1024x414.png)](http://localhost/wp-content/uploads/2020/10/update-22123-3.png)Cassette input amplifier circuit (74LS367) EXTRA :Like some TSs are more deaf than others, and in general all are “hard” to ear ( Sure I designed, sold, and made money with a signal improver, LG-02), seeing in the cassette input, Timex put only two diodes to ground -CR25/26- (I don’t know why, only two), for bias the input SCLD circuit, to protect the input electronics, and to conform the input signal. The solution I encountered, was to add one diode in series between them. Commonly I desolded one extreme of one diode and add one more between it and the hole in the board, – remember in the same direction -. Conclusion: with three diodes my TS become an excellent machine !!! [![CR25/CR26 diodes before and after adding a 1N4148](http://localhost/wp-content/uploads/2020/10/update-22123-4-1024x388.png)](http://localhost/wp-content/uploads/2020/10/update-22123-4.png)CR25/CR26 diodes before and after adding a 1N4148 4) The refresh problem was hard to find in the beginning, but with an oscilloscope and understanding of memory and Z80 devices, finally I win. The TMS 4416 was a new chip when it was used in the design of the TS2068. And been the Z80 the best 8 bits micro, than that it have inside, the refresh counter necessary to maintain the data in the dynamic memories, the combination was perfect. But when it was designed, only seven bits was used for this counter. It’s O.K. for 4116, 4164 family (it’s need only seven bits for refresh all cells), but in the “new” 4416 family eigth bit was needed. In the low memory block, the constantly reading to form the video, make the necesary refresh reading. But in the high memory block the SCLD make the eighth bit necessary for a complete refresh cycle (the A7R). When it fails, the machine can’t maintain it’s data integrity, and crash. The first thing to do is confirm that the problem is this, I made a little program that display this problem in the form of lost UDG data, simply making all bits set in UDG, the lost information are see like white dots in them. First input this program line: ``` 10 PRINT "ABCDEFGHIJKLMNOPQRSTU" ( all letters in UDG mode) ``` Then input this command line: ``` FOR A = UDG "A" TO UDG "U": POKE A,255: NEXT A (ENTER & RUN) ``` Now all letters must be black boxes, and in the upper left side of the screen you can see a coarse black line, 21 characters long. Every time you type RUN and ENTER you refresh the display area, tacking the UDG information from its original address ( high memory block). Running a program every minute, if you can see how the white spots come to appear, you machine have the refresh problem !!! Second build and install the next circuit: [![Address 7 generator for Z80 CPU (74LS157, 74LS73)](http://localhost/wp-content/uploads/2020/10/update-22123-5-1024x553.png)](http://localhost/wp-content/uploads/2020/10/update-22123-5.png)Address 7 generator for Z80 CPU (74LS157, 74LS73) The circuits given represent hours of high tech. research, they are copyrighted to my mind. I send they to you for a help in yours troubles, (and for my TS’s USA folks, of course). You should not use them commercially, without my acceptation.