ClausB

I based the calculations on polynomials from a paper, supposedly good to one minute of arc. I don't know how well they extrapolate to today, so I compared them with CalSky.com for today at 0h UTC: Body err RA err Dec ----- ----- ----- Sun 0.7 m 1' Mercury 0.3 m 1' Venus 0.3 m 1' Moon 8.6 m 35' Mars 0.5 m 2' Jupiter 0.0 m 0' Saturn 0.0 m 0' Uranus 0.0 m 0' Neptune 0.1 m 1' Pluto 0.8 m 2' Well, the moon is off a lot (1 m of RA is 15' of arc, or a quarter degree, or half the moon's diameter). The sun and Pluto are off a bit, but the rest are not too bad, for 37-year old polynomials!

Ah the good old days! We used similar tricks to save RAM on our high school's Altair 8800As. They had 16K but the MS cassette BASIC took 10K of that for itself.

Here's an astronomy program I wrote in 1983: ASTRO.BAS No graphics, just text. It calculates Julian date and sidereal time, it locates planets (including Pluto), and it converts between celestial and altazimuth coordinates. (Yes, my degree is in astronomy and physics.) FAFMUL.ATR

No, there's only one signal there that's not on some other slot. It's called EXSEL (different from the /EXSEL signal on slots 0 and 3). It disables the motherboard's main address decoder which, in turn, disables all RAM, ROM, and I/O registers (except ANTIC). I presume it's there so the tester can insert its own RAM and ROM into the address space. Interesting that the 400's test connector doesn't have that signal.

Been reading up on this machine. Pretty nice for its time. Custom 4-bit CPU runs at 0.6 Mips. Up to 32KB RAM can be read 8 bits wide but written only 4 bits at a time. 42KB program ROM is 20 bits wide and implements full BASIC and I/O. Keyword atoms (tokens) save RAM space. Variables are 8-byte BCD floating point with 13 digits precision, typical for a calculator company back then. I would guess it's roughly comparable in performance to a TRS-80 Model 1 Level II. (Fun fact: the Z80 has only a 4-bit ALU inside.)

For what system?

Sad to see it go. Looks like you used the original 32K bank version, not the later 16K bank version I distributed after the article came out. BTW, the LS158 goes back in the motherboard socket, not the LS153.

The Uno cart problem might be due to the pull down resistors on the RAM deselect lines. On the 800 they are 680 Ohm where on the XLs they are 1K. IIRC replacing those fixes the problem.

I've been there! Made a customer visit to Shingle Springs. Flew into Reno, visited a friend there, drove around Lake Tahoe, took a hotel in Placerville. Stopped at a park where you could pan for gold but didn't have time for that so I just bought some gold flakes for the kids. Beautiful country there!

Yes and the link goes to my gallery of machines I used in the 70s or which had their origins in the 70s. The Atari 8-bits originated in the 70s but grew up in the 80s.

Here's the article. The thread has background info as well as the updated upgrade.

Sure, but almost no one saw it before 1980. And Star Raiders remained great well into the 80s.

Bitmap graphics modes show bits as pixels. GR. 4 and 6 two-color modes use one bit per pixel, so it's fun to use them as windows into the Atari's memory. Type: GR.6 B=PEEK(560)+256*PEEK(561)+4 POKE B,0:POKE B+1,0 Now you see the first 1.5K of RAM. See the timers in page 0 ticking away. See the stack in page 1 flickering. See system variables change as you type something. Plug in a joystick and watch bits change when you move it. Type: POKE B+1,208 Now you see the Atari's special hardware registers. GTIA, POKEY, PIA, ANTIC. See POKEY timers and random number generator. Type keys and watch bits change. Move the stick and see PIA change. Watch the fire button bit in GTIA and the light pen register in ANTIC. Run this program and watch all the memory scroll by. Line 30 skips blank RAM. Change it if you like. Lines 60-85 print the page number in hex. If you prefer decimal just print A or H. Change line 1 to GR.4:S=10 if you want bigger bits. Have fun.

The 2-bit Gray code is also called "quadrature" encoding, because the two bits are 90° out of phase. The two bits are often called A and B. One trick is to wire A to an edge-triggered interrupt. The interrupt routine would simply INC position if B is high or DEC position if B is low. Simple, but you get less position resolution that way.

What's the VGATE?

Assuming the program is correct, 37120 is above 32K, so it's in the third 16K board. To verify, swap the second and third boards and see whether the error address goes below 32K.

Yes, especially since the period is not an ASCII period, it's hex 11. Also there's no other BASIC code visible. 6800 derivatives were used widely in automotive controllers. Or it could be some industrial controller.

The bytes from addresses 06 through BB look like a LUT (look-up table) of some function which increases less as it grows, like a square root. I typed a cross-section of the table into a spreadsheet, squared the values, scaled down by 256, and took consecutive differences. It does indeed look like a square root approximation table!

Anotated ROM listing: https://archive.org/details/complete-timex-ts1000-sinclair-zx81-rom-disassembly All the clever tricks and RAM management explained.

Sinclair did an impressive job with the ZX 8K BASIC. That went a long way toward compensating for the limited hardware. That keyword entry feature and the various cursor modes made the membrane keyboard tolerable.

Looks like a TI keyboard there. I noticed the TI keyboard's matrix is very similar to the TS1000/ZX81's, making the substitution fairly easy. Was that a common mod in the day? Where's your TS1000 board, under the wood? Any more photos of that setup?

Great topic for Memorial Day!

Anything by Robert Miller? He's a friend of mine and long ago he gave me some utilities he wrote, like sunrise/set calculator and world map projections. I don't know how widely distributed they were.

My first computer was a TI SR-56 programmable calculator from 1976. I sold it in 1978 to buy a TI-59, which I still have. In 2010 I got another -56 off eBay, replaced the batteries, and still use it often.