From uchinews!vixen.cso.uiuc.edu!uwm.edu!cs.utexas.edu!uunet!pipex!demon!chrism.demon.co.uk!chris Thu Sep 16 10:27:46 CDT 1993 Article: 48553 of alt.folklore.computers Newsgroups: alt.folklore.computers From: chris@chrism.demon.co.uk (Chris Marriott) Path: uchinews!vixen.cso.uiuc.edu!uwm.edu!cs.utexas.edu!uunet!pipex!demon!chrism.demon.co.uk!chris Subject: Re: Old vs. new calculators Distribution: world References: <009724E3.6B22038E@pomona.claremont.edu> Organization: None Reply-To: chris@chrism.demon.co.uk X-Newsreader: Simple NEWS 1.90 (ka9q DIS 1.21) Lines: 36 Date: Sat, 11 Sep 1993 20:37:55 +0000 Message-ID: <747779875snz@chrism.demon.co.uk> Sender: usenet@demon.co.uk Status: R In article <009724E3.6B22038E@pomona.claremont.edu> dpeschel@pomona.claremont.edu writes: >I've noticed some threads here about old calculators... also, while looking for >the _Scientific American_ issue on information, I cam across some ads for >scientific calculators. > >Since I've never used a really old calculator, that brings up the question: > >Do older models have any neat features that have disappeared from newer models? >(e.g., quality parts, legible screens, ample decimal places, programming) > >If people reply by mail, I will post a summary. > >just making everyone glad they read news, >-- Derek dpeschel@pomona.claremont.edu > The WORST calculator I've ever used was the first one I bought - a "Sinclair Cambridge Programmable", circa 1983 (?). It had what must have been the world's worst algorithms for trig functions, correct to about 1 (yes ONE) significant figure. Sine 30 degrees, for example, was given as 0.48! The battery (one of those 9-volt rectangular ones) lasted about an hour. Unless you had the delicate touch of a brain surgeon, pressing a key resulted in multiple entries from the keys. You had to lurk in dark corners to read the LED display. I do NOT have font memories of that calculator! Chris -- -------------------------------------------------------------------------- | Chris Marriott | chris@chrism.demon.co.uk | | Warrington, UK | 100113.1140@compuserve.com | | Save the whales. Collect the whole set. | CompuServe: 100113,1140 | -------------------------------------------------------------------------- From uchinews!linac!uwm.edu!spool.mu.edu!uunet!munnari.oz.au!metro!basser.cs.su.oz.au!swift!jeremy Mon Jan 4 10:01:48 CST 1993 Article: 32861 of alt.folklore.computers Newsgroups: alt.folklore.computers Path: uchinews!linac!uwm.edu!spool.mu.edu!uunet!munnari.oz.au!metro!basser.cs.su.oz.au!swift!jeremy From: jeremy@sw.oz.au (Jeremy Fitzhardinge) Subject: Re: VCS processor (was Re: Game machine processors) Organization: Softway Pty Ltd Date: 4 Jan 93 11:14:42 GMT Message-ID: References: <1993Jan2.180042.1@kean.ucs.mun.ca> <8505@charon.cwi.nl> <1i63h4INN7tu@uniwa.uwa.edu.au> Sender: news@softway.sw.oz.au (Usenet) Lines: 66 Status: R In <1i63h4INN7tu@uniwa.uwa.edu.au> john@gu.uwa.edu.au (John West) writes: >Any other evil hardware hacks in production machines? Things that really >should never have been done? Well, since you mentioned them, the Sinclair machines are a bountiful source of such goo. The Sinclair Spectrum has a chip known as the ULA (Uncommitted Logic Array - some programmable device that never seemed to have its own real name) which does pretty much everything the Z80 doesn't, like video. When the ULA is pumping out video, and the CPU wants to use the 16k block with video memory in it, the ULA stops it. How? Not by interrupting it, not by adding wait states - it just turns off the CPU's clock (derived from the video clock in a fashion that seems to insure maximum possible amounts of digital hash in the video output). This was OK until I tried to put a 8MHz Z80H into the machine. I could never work out how to make sure my 8MHz clock could be turned off by the ULA, since the necessary signals never came out pins. In retrospect, any number of things could have been done, but it didn't seem worth the effort. Also, Sinclair machines always had production line "red wire" to patch hardware faults. The revision 1 (A?) Spectrum boards had a 74ls chip on its own small PCB flying on wire to correct a timing problem. The revision 2 (b?) boards had a transistor across the CPU to fix some other probem with chip selection. The real mess was the "Interface One" adapter. Every one I ever saw had a meter or so of patch wire, and a few transistors and resistors scattered around amongst the cut tracks. And while we're here, how about some "software where hardware should have been" tales? OK, Sure! The IF-1 has an interface for the MicroDrive (100 feet of sliced videotape in a continious loop inside a box half the size of a matchbox), the network (very slow token-ringish thing), and a serial port (could be made to work with RS232, but not really standard). All three bits of hardware were driven with pure CPU bit bashing. For the microdrive this meant that if you hit NMI or reset while it was writing, then it would leave the head energised and keep the tape rotating, but without modulation. Data? Where? The problem was worse for the serial and network ports however. There was no interrupts used in the machine apart from the 50Hz frame interrupt, so you had to poll the serial port or network continiously if you wanted to use them. Off course, full-duplex was not supported by the ROM code, and very hard to do anyway. Also, there was a Sinclair tradition of flashing the screeb border colour whenever hardware was being polled or used. That meant that tape reads would flash red/cyan for headers and blue/yellow for data (but at >50Hz, so there were stripes - sctually at the freq of the sound on the tape), and the Interface One ports would flash black/white when they are used. It made assembling off MicroDrive very painful at 3am. Note the transition from present to past tense. Sweet memories turn bitter on comtemplation. J >John West >-- >For the humour impaired: Insert a :-) after every third word