http://www.zxdesign.info Blog dedicated to the design and engineering of a ZX Spectrum cone, interesting software and techniques. en-gb Sun, 18 May 2008 17:24:23 UT Sun, 18 May 2008 17:20:13 UT This site and content free; Permission is granted to copy, distribute and/or modify the website under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation; with no Invariant Sections, no Front-Cover Texts, and one Back-Cover Text: "The contents of this blog, detailing the reverse engineering of the ZX Spectrum and the design of the Harlequin was produced and written by Chris Smith chris@zxdesign.info." A copy of the license is included in the section entitled "GNU Free Documentation License". Some website content, where specified, is covered by the GNU General Public Licence and as such you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. 60 http://www.zxdesign.info/images/specmbrhs.png http://www.zxdesign.info The ZX Spectrum ULA has been successfully replaced with a CPLD version! This announcment means that there is now no component of a Spectrum that cannot be repaired. http://www.zxdesign.info/cpld/ulaReplacementWorks.shtml http://www.nfluid.com/cpld/ulaReplacementWorks.shtml Sun, 18 May 2008 17:20:13 UT Hardware The first phase of the ULA replacement has been completed and hooked up to a ZX Spectrum. http://www.zxdesign.info/cpld/ulaReplacement1.shtml http://www.nfluid.com/cpld/ulaReplacement1.shtml Sun, 30 Mar 2008 21:04:11 UT Hardware The Harlequin DRAM interface has been implemented and tested, and shown to produce exactly the same snow effect as a genuine Spectrum. Pictures included. http://www.zxdesign.info/harlequinSnow.shtml http://www.nfluid.com/specDesign/harlequinSnow.shtml Sun, 16 Mar 2008 10:37:20 UT Hardware A full design of a dynamic memory interface for the Harlequin has been completed, and a full discussion of the Spectrum video addressing given. http://www.zxdesign.info/harlequinDRAM.shtml http://www.nfluid.com/specDesign/harlequinDRAM.shtml Mon, 10 Mar 2008 20:04:22 UT Hardware After a short break from Harlequin design, the prototype gets it's first dynamic memory board. http://www.zxdesign.info/dynamicRam2.shtml http://www.nfluid.com/specDesign/dynamicRam2.shtml Mon, 03 Mar 2008 21:02:12 UT Hardware Back in 1986/87 I did a some work for Mikro-Gen Software, and have had an unfinished game sitting in my attic since then. As this appears to be of interest to some, I have uploaded to the blog a small background of myself and how the game came about. Downloads included. http://www.zxdesign.info/software/skyway.shtml http://www.zxdesign.info/software/skyway.shtml Sun, 09 Dec 2007 18:47:21 UT Software The Harlequin Design Blog has moved to its own website! Old links via www.nfluid.com/specDesign/* should continue to work, but you should update your links/bookmarks to www.zxdesign.info/* http://www.zxdesign.info http://www.zxdesign.info Fri, 30 Nov 2007 11:46:25 UT Hardware Both the horizontal and vertical sync circuits have been greatly simplified, and a further optional Vertical Sync cheat presented which reduces the component count at the expense of breaking the specification. http://www.zxdesign.info/videoSimplification1.shtml http://www.zxdesign.info/videoSimplification1.shtml Fri, 23 Nov 2007 20:34:09 UT Hardware A description of the ULA dynamic RAM generator is presented, along with a description of the Snow Effects cause. http://www.zxdesign.info/dynamicRam.shtml http://www.nfluid.com/specDesign/dynamicRam.shtml Fri, 02 Nov 2007 21:32:09 UT Hardware Reducing the Harlequin logic into a CPLD or FPGA would mean a cost effective clone could be built. This also makes a ULA replacement possible. http://www.zxdesign.info/cplddesign.shtml http://www.nfluid.com/specDesign/cplddesign.shtml Thu, 01 Nov 2007 19:53:47 UT Hardware Several areas that are sensitive to propagation delays are examined in depth, plus a proof of where in a cycle the floating bus read takes place. http://www.zxdesign.info/clockAlignment.shtml http://www.nfluid.com/specDesign/clockAlignment.shtml Sun, 07 Oct 2007 10:57:17 UT Hardware The memory contention timing implemented by the Harlequin was thought to match that of the ZX Spectrum, This was not the case, and has revealed a few secrets of the ZX Spectrum. http://www.zxdesign.info/memContRevision.shtml http://www.nfluid.com/specDesign/memContRevision.shtml Tue, 25 Sep 2007 21:42:17 UT Hardware Several changes have been made to the timing of the memory and IO wait signals to ensure that the CPU clock is always glitch free. http://www.zxdesign.info/improveStabilityWait.shtml http://www.nfluid.com/specDesign/improveStabilityWait.shtml Sun, 16 Sep 2007 17:54:02 UT Hardware The spectrum FLASH mode is the only display component missing from the Harlequin. This has now been added and the latest video schematic released. http://www.zxdesign.info/flash.shtml http://www.nfluid.com/specDesign/flash.shtml Sun, 9 Sep 2007 09:19:07 UT Hardware Today saw the addition of ZX Spectrum 48K sound support which also allows the cassette loading to be monitored. http://www.zxdesign.info/soundbeeper.shtml http://www.nfluid.com/specDesign/soundbeeper.shtml Mon, 27 Aug 2007 14:42:35 UT Hardware Even though the Harlequin seemed to be 100% compatible with the ZX Spectrum, it failed to run MegaShock Demo 3 successfully. The cause of this has been investigated over the last few weeks, and the problem solved. Now the Harlequin really is 100% compatible! http://www.zxdesign.info/softwareTest.shtml http://www.nfluid.com/specDesign/softwareTest.shtml Sun, 26 Aug 2007 11:28:43 UT Hardware With the new contention model in place, the Harlequin demonstrates that it is 100% timing compatible with the ZX Spectrum! http://www.zxdesign.info/testSuccess.shtml http://www.nfluid.com/specDesign/testSuccess.shtml Sat, 28 Jul 2007 11:52:23 UT Hardware The new ULA compatible contention model has been tested and a lot of information has been captured.. http://www.zxdesign.info/contentionTest2.shtml http://www.nfluid.com/specDesign/contentionTest2.shtml Sat, 28 Jul 2007 11:03:41 UT Hardware The ULA contention and wait timings have been implemented by a dedicated wait generator. http://www.zxdesign.info/ulaContention.shtml http://www.nfluid.com/specDesign/ulaContention.shtml Mon, 16 Jul 2007 20:46:21 UT Hardware Further speculative details of the ULA timing during IO operations have been released, based on observed behavior and Z80 data sheets. http://www.zxdesign.info/portContention3.shtml http://www.nfluid.com/specDesign/portContention3.shtml Mon, 09 Jul 2007 21:00:47 UT Hardware Some speculative details of the ULA IO port operation have been added to the Harlequin blog. http://www.zxdesign.info/portContention2.shtml http://www.nfluid.com/specDesign/portContention2.shtml Sun, 08 Jul 2007 18:23:12 UT Hardware A detailed look at the ZX Spectrum's port contention is underway in an attempt to discover what the ULA is actually doing. http://www.zxdesign.info/portContention1.shtml http://www.nfluid.com/specDesign/portContention1.shtml Thu, 05 Jul 2007 18:21:12 UT Hardware The Harlequin port contention has been found to be incorrect, causeing Sidewise to perform badly. http://www.zxdesign.info/sidewizeTest.shtml http://www.nfluid.com/specDesign/sidewizeTest.shtml Sat, 30 Jun 2007 14:34:29 UT Hardware The interrupt acknowledgement is suspected of being the cause of the intermittent floating bus errors. http://www.zxdesign.info/intAck.shtml http://www.nfluid.com/specDesign/intAck.shtml Sat, 30 Jun 2007 13:52:17 UT Hardware The phase of the CPU clock is critical to correct floating bus behaviour, and was found to be unpredictable at power up. http://www.zxdesign.info/floatbusUnstable.shtml http://www.nfluid.com/specDesign/floatbusUnstable.shtml Sat, 30 Jun 2007 10:27:42 UT Hardware A glitch was discovered in the horizontal clock reset signal. This was causing erratic interrupts. http://www.zxdesign.info/hrstGlitch.shtml http://www.nfluid.com/specDesign/hrstGlitch.shtml Sat, 30 Jun 2007 07:40:19 UT Hardware The Harlequin is showing some strange interrupt timing which is under detailed analysis. http://www.zxdesign.info/interruptTest.shtml http://www.nfluid.com/specDesign/interruptTest.shtml Fri, 29 Jun 2007 19:17:23 UT Hardware Having completed the cassette interface design and prototype, a selection of games have been successfully loaded by the Harlequin. However there appears to be a timing issue with the interrupts, and this is under investigaion. Full details and screenshots are available on the Blog. http://www.zxdesign.info/loadingTest.shtml http://www.nfluid.com/specDesign/loadingTest.shtml Thu, 28 Jun 2007 18:56:14 UT Hardware Full design and analysis of the Harlequin cassette interface has been completed. http://www.zxdesign.info/cassette.shtml http://www.nfluid.com/specDesign/cassette.shtml Sun, 24 Jun 2007 14:27:32 UT Hardware Today saw a successful test of the Harlequin keyboard interface and interrupts, allowing the first programs to be run! A cassette interface is underway and will appear soon. http://www.zxdesign.info/keyboardtest.shtml http://www.nfluid.com/specDesign/keyboardtest.shtml Fri, 22 Jun 2007 21:30:02 UT Hardware A description of the Spectrum interfaces and a keyboard interface design has been completed. http://www.zxdesign.info/interfaces.shtml http://www.nfluid.com/specDesign/interfaces.shtml Mon, 18 Jun 2007 20:49:12 UT Hardware The latest schematics - those implementing the revised video byte fetch (one per T-state), have been released. http://www.zxdesign.info/waitTooLate.shtml http://www.nfluid.com/specDesign/waitTooLate.shtml-1.7 Thu, 14 Jun 2007 12:45:09 UT Hardware The CPU interface page has been updated to contain more detail. Previously it skipped over many details such as the important Z80 control signals and how they are produced, clock, CPU wait generation etc. http://www.zxdesign.info/cpu1.shtml http://www.nfluid.com/specDesign/cpu1.shtml-update Thu, 14 Jun 2007 07:12:01 UT Hardware The latest state diagrams failed to display with IE 6 or 7. This has been fixed. Firefox, Konqueror etc were okay. http://www.zxdesign.info/waitTooLate.shtml http://www.nfluid.com/specDesign/waitTooLate-IEFix.shtml Tue, 12 Jun 2007 11:16:21 UT Hardware The previous contended memory instability has been solved, and the Harlequin now boots successfully. http://www.zxdesign.info/waitTooLate.shtml http://www.nfluid.com/specDesign/waitTooLate.shtml Mon, 28 May 2007 14:21:43 UT Hardware Booting the Z80 from a ZX Spectrum ROM worked to a degree, but highlighted some potential timing problems. http://www.zxdesign.info/bootingFromRom.shtml http://www.nfluid.com/specDesign/bootingFromRom.shtml Tue, 22 May 2007 19:53:11 UT Hardware Connecting the Z80 and ROM to the video circuit requires special treatment of the buses and control signals. http://www.zxdesign.info/cpu1.shtml http://www.nfluid.com/specDesign/cpu1.shtml Fri, 18 May 2007 21:03:12 UT Hardware The alignment issues associated with the timing of the vertical counter reset have been re-analysed and the previous fix found to be incorrect. http://www.zxdesign.info/alignment2.shtml http://www.nfluid.com/specDesign/alignment2.shtml Wed, 09 May 2007 18:22:26 UT Hardware Harlequins video byte fetch while quite efficient, does not match that of the ZX Spectrum. This has been redesigned and should give accurate floating bus behaviour. http://www.zxdesign.info/videoFetchRedesign.shtml http://www.nfluid.com/specDesign/videoFetchRedesign.shtml Sat, 05 May 2007 22:13:26 UT Hardware The video memory interface has been prototyped and some real ZX Spectrum loading screens have been displayed from EPROM. http://www.zxdesign.info/firstScreenPicture.shtml http://www.nfluid.com/specDesign/firstScreenPicture.shtml Tue, 01 May 2007 20:05:11 UT Hardware The video output circuit and video memory interface has been designed and is being put together. http://www.zxdesign.info/memoryToScreen.shtml http://www.nfluid.com/specDesign/memoryToScreen.shtml Mon, 30 Apr 2007 23:02:31 UT Hardware The noise on the Vertical Clock line has been tracked down and was also the cause of a faint vertical line on the screen. The stability of the vertical clock has also been improved at power on. http://www.zxdesign.info/timingProblems.shtml http://www.nfluid.com/specDesign/timingProblems.shtml Tue, 24 Apr 2007 21:10:03 UT Hardware The component values of the emittor-follower driver need to be selected for true ZX Spectrum colour output. To help a spreadsheet has been produced. http://www.zxdesign.info/ddrivedesign.shtml http://www.nfluid.com/specDesign/ddrivedesign.shtml Wed, 18 Apr 2007 20:13:43 UT Hardware The previous testing hack of connecting the red RGB line to the TV through a resistor worked, but doesn't give very much voltage control, current control or protection. Several solutions are available and one particularly suits our needs. http://www.zxdesign.info/displaydrive.shtml http://www.nfluid.com/specDesign/displaydrive.shtml Tue, 17 Apr 2007 18:21:07 UT Hardware The ZX Spectrum has three distinct RGB signals: Border, Ink and Paper. These must be routed to the display at the appropriate time.... http://www.zxdesign.info/colourselect.shtml http://www.nfluid.com/specDesign/colourselect.shtml Fri, 13 Apr 2007 19:13:23 UT Hardware As the Harlequin is now generating proper TV pictures, the time has come to look again at the ZX Spectrum display and start designing the video memory interface. http://www.zxdesign.info/memdisplay.shtml http://www.nfluid.com/specDesign/memdisplay.shtml Thu, 12 Apr 2007 14:21:00 UT Hardware A small adjustment to the horizontal sync timing has centred the main display rectangle into a far more pleasing display position. http://www.zxdesign.info/displaycenter.shtml http://www.nfluid.com/specDesign/displaycenter.shtml Sun, 08 Apr 2007 20:20:00 UT Hardware After a few days of design work the Harlequin Video Sync generator has been completed which, after a bit of debugging, has generated a clean static TV picture! http://www.zxdesign.info/vertcontrol.shtml http://www.nfluid.com/specDesign/vertcontrol.shtml V2 Thu, 05 Apr 2007 08:45:00 UT Hardware Today saw the completion of the horizontal video prototype, and some oscilloscope pictures have been uploaded to show the important signal timings. http://www.zxdesign.info/horizPics1.shtml http://www.nfluid.com/specDesign/horizPics1.shtml Fri, 23 Mar 2007 12:49:11 GMT Hardware A schematic implementing the horizontal video control signals so far discussed has been produced and made available for download. http://www.zxdesign.info/horizcircuit.shtml http://www.nfluid.com/specDesign/horizcircuit.shtml Mon, 19 Mar 2007 14:19:27 GMT Hardware