Computer

USB vs. PS/2 Keyboards & Mice

Tube Circuits were First

Microprocesor

A microprocessor is a computer on a chip.  It has:

• data bus
• address bus
• control lines

The Intel 4004 was the first chip that had general purpose capability.  It was made for use in a hand held 4 function calculator and/or for use as a traffic light controller.  It used 4 bit words, enough for numbers but not enough for 8 bit ASCII characters.  The 8008 then 8080 chips were 8 bit data word chips.
The microprocessor works like a Turing machine (Wiki) in that it stores data and instructions in the same memory (Wiki: Computer architecture).

Microcontroller

The microcontroller is a much more modern part.  Instead of having the classic address and data buss I/O pins it has register pins that can be digital I/Os, analog inputs/outputs, serial buss signals, or other things that relate directly to hardware interfacing.  These are typically used in embedded control applications not for general purpose computing.  The PIC line from Microchip is a typical example (16F88).

The PIC family of microcontrollers uses the Harvard architecture (Wiki) where there are separate memories for data and instructions.  This not only prevents reading data as instructions or storing data in locations that are for instructions but also makes for faster operation.

Military

SWTP Kit

This was my first personal computer.  It came as a kit and there was no software available when I got it.  It came with the Motorola 6800 pocket hex instruction card.

The 6800 has a Von Newman architecture (Wiki) both in terms of software and hardware.   That's to say the RAM contains both instructions and data and the hardware uses memory mapping for the hardware interfacing.  This is different than the early Intel microprocessor chips that had separate electrical terminals for external hardware.  It's also different from the Harvard architecture (Wiki) used with the PIC micro controllers (PC6, LED1PC2, PICpkg, PIC16F88, PClock, Precicion Clock #4).

The 7 SWTP mother board slots each have the same 50 buss lines.  The mother board includes address decoding for the 8 I/O slots.   This means that the I/O boards do not need address decoding since they are already memory mapped.  Note the early Intel based home computers, like the Altair 8800 (Wiki) used the "S-100" buss that had twice the number of connections and no address decoding so all the cards had 100 connections making them much more expensive.

YouTube:  The 8-bit home computer from 1975 you've never heard of, 30:12, March 26, 2022 -
YouTube: deramp5113: SWTP 6800 - Teletype -

Fig 1 Shown with front and back trip strips just sitting in place. The expanded metal top/side cover is not shown.	Fig 2 The silver can cap on the right with white line cord wiring reasonates with a transformer winding.	Fig 3 MP-B Motherboard
Fig 4 50-Pin RAM Notice that the2102 static memory chips are soldered directly to the board.  Sockets were not reliable and so caused problems. I had a similar problem with plug-in hard drives. i.e. the socket caused a disk crash.	Fig 5 50-Pin RAM	Fig 6 50-Pin RAM This is the SWTP RAM board, the others are after market.
Fig 7 50-Pin MP-A CPU	Fig 8 50-Pin DMAF1 Dual Drive DS Floppy Disk I/O	Fig 9 MP-S Serial
Fig 10 MP-T Timer	Fig 11 MP-LA	Fig 12 Sound (resistor ladder & LM380)
Fig 13 MP-L Parallel Printer	Fig 14 Dual YE-Data 8" Floppy Drives	Fig 15 Audio Cassette Data Interface
Fig 16 Works with two cassette drives, one for writing & one for reading.

I started with the Motorola 6800 in a SWTP (South West Technical Products) kit.  The kit came with NO software and there was no software on the open market.
It did come with the folding Motorola wallet size machine language instruction list.  I learned how to program directly in hex machine code.  This was difficult because if you added or subtracted an instruction inside a loop you need to recompute the loop jump distance.
Motorola folding Pocket 6800 Instruction Set Card (Issue B):

1.	2.	3.
4.	5.	6.
	7.

• Special Operations
• Interrupts
• Hex & Decimal Conversion, Powers of 2, ASCII Characters
• Register Programming Model
• Accumulator and Memory Instructions
• imaIndex Register & Stack, Jump & Branch, Conditions Code Register Instructions
• Operation Code Numerical Listing -

Teletype Pocket Card with 8-level punched tape data format - you can read paper tape by eye  Front Side - Table Side (at 2X magnification)
Between bits 3 and 4 there is a row of small sprocket holes.  The special  USASCII (X3.4-1967) characters are:

• NULL - no holes punched the same as reading no tape (RUBOUT = all holes punched)
• SOH - Start of Heading
• ETX -  End Of Text
• EOT - End Of Transmission
• ENQ - ENQuiry
• ACK - ACKnoldege
• BELL - rings the bell
• HT - Horizontal Tab a long string
  
• BS - Back Space, can be used with rubout to fix a typing error.
• LF - Line Feed only rotates the platen one line (does NOT move carriage)
• VT - Vertical Tab
• FF - Form Feed
• CR - Carriage Return. This command takes the most time to complete. Therefore send CR followed by LF and maybe a couple of rubouts to be sure the carriage has returned.  People that didn't know would send LF CR and the first letter of the next sentence would be printed in the middle of the page.
• SO - Shift Out
• SI  -  Shift In
• DLE - Data Link Escape
  
• DC1 - Device Control 1
• DC2 - Device Control 2
• DC3 - Device Control 3
• DC4 - Device COntrol 4
• NAK - Not AcKnoledge - a handshake method where either NAK or ACK is always returned
• SYN - Synchronous Idle
  
• ETB - End of Transmission Block
  
• CAN - Cancel
• EM - End of Medium
  
• SUB - Substitute
• ESC - ESCape
• FS - File Seperator
  
• GS - Group Seperator
  
• RS - Record Seperator
  
• US - Unit Seperator
  

Motorola published the M6800 Microprocessor Applications Manual, an 8.5 x 11" book about 1.5" thick.  It contained sample code for many applications that may have been parts of a point of sale terminal.  This was a handy reference to learn many "tricks of the trade". My copy has seen a lot of use.

21 Aug 2004 - Shortly after getting my SWTP 6800 kit and building it I started learning Motorola assembly language from the pocket card and the 81/2 x 11 x 2" book that included a lot of source code and schematics.  Note there are only 2 buttons on the front of the SWTP 6800 On-Off and Reset.  When either of them is activated the 6800 either goes to a cold or warm start address in the operating system ROM and boots itself.  The 6800 has an instruction that is a relative 8 bit  jump and it was this instruction that allowed Microware to create their Real Time operating system ROM chip to replace the stock Motorola ROM used in the SWTP 6800.  Motorola, AMI and other vendors were big on the idea that you could write code that could be:

• Relocated anywhere in the address space.  This could NOT be done with the intel chips.  The idea of
  
• Recursive execution of code was also needed, that is to say that the variables associated with some routine would not be stored at some hard memory address, but would be stored using a stack.  This way the same routine can be called time after time without corrupting that data.  The calls would be from different routines.
  
• Re-entrant - means that the code can call itself.  For example a factorial routine.

The big advantage of Re-locatable Recursive code is that the routine does not need to be re-written when it's used in a different location allowing a library of routines to be built and called as needed.  When the 6800 uP came out it was common to write self modifying code (it is a Von Newman architecture where code and data share the same address space).  The PIC uC chips use a Harvard architecture where code and data spaces are separated thereby forbidding self modifying code.

This RT68 operating system evolved into OS-9 which was a very sophisticated op system with features that even today (OS-9000) are great.  It depends on the ability to have Relocatable Recursive code.

I remember when the 6809 came out in the Color Computer and I went to a local Silicon Valley meeting to see what some hackers had found.  One of the most interesting things about the 6809 as implemented in the CoCo was that although the operating system/Basic language was stored in ROM, it was moved into RAM for execution.  This means that other operating systems, I used OS-9 on my COlor Compugter, could be loaded and run.  The demo was to display on the TV screen the machine code in a dynamic way, page after page.  A Motorola app note for the 6809 contained the schematic for the CoCo.

Motorola had a program where if you contributed some 6800 assembly code to a pool of code you would be allowed to get code for the pool for free.  This pool of code was analyzed by Motorola and the 6809 had a more orthogonal instruction set as a result.  The 68000 also had a long relative jump making it easier to write relocatable code.  The 6809 probably had the long jump.

The Altair kit had a front panel with a lot of switches and lights that were used to manually define a loader program so that the computer could be started, very time consuming.  The SWTP front panel has just two buttons: Power on/off and Reset.  It contains a ROM chip with the bootstrap program, a simple debugger, and a way to enter programs in Motorola "S9" format.

The Altair kit has a mother board that has a 100 pin buss (with a hard to get connector) and no chips.  Each plug-in board needs to decode the full buss making the boards cost more than needed.  The SWTP has two buss structures, one "SS-50" 50-pin for RAM, DMA (Direct Memory Access), etc. (that used easy to get Molex type connectors) and one narrow I/O buss where the motherboard has already decoded some addresses to allow a simple I/O card to be used.  Many years later the IBM PC followed the SWTP lead using a separate I/O buss.

Operating System

I replaced the Motorola ROM with the first product from Microware called RTOS (Real Time Operating System).  The key to this was that the 6800 had instructions to allow jumping to an address RELATIVE to the current program counter.  The Intel chips only had instructions that pointed to fixed addresses.  The availability of a relative address instruction allows code to be written in modules that could be relocated in the address space.  Relocatable code is much easier to use than code that must be loaded in a fixed location. Reentrant, Recursive, and some other R words characterize much of the code written for Motorola chips at that time.

Later there was a Flex Disk Operating System.  I think it supported a flavor of Basic so you could write programs in a much easier way than using hex or assembly.
FLEX User Group -

ASR33 TTY (Automatic Send Receive)

This is a Teletype (brand name) machine that is 8 level (meaning that it worked with 8 bit data).  There were a lot of surplus 5 level machines available that would not work with 8 bit microcomputers.  The neat thing about the ASR33 was that it could do four things:

• Print on paper (this was before CRT displays existed)
• Keyboard for data entry
• Punch paper tape for data storage (this was before magnetic tape & disk storage)
• Read paper tape (some programs may take 30 minutes to load, so you leave the computer turned on all the time)

The interface was 20 mA current loop with ASCII data format.  There was an RS-232 I/O card for those who could afford an acoustic coupled modem.  (At that time AT&T had all of the phone system and it was a crime to connect directly to the phone lines.)

It could use pin drive or friction feed paper.  Friction feed allows the use of lower cost paper in rolls with no sprocket holes.  The nicer paper that has a perforation every 11 inches needs to have sprocket holes so that you can do a "Form Feed" to get to the top of a page.  You could get multipart paper for the ASR33 since it was an impact printer.  The printing head is a cylinder that moves up and down and rotates to get to the character needed.

100 Words per minute is 500 characters per minute or 8.333 characters per second.  In 30 minutes you could read 15,000 characters and the tape would about about 125 feet long.  If wound on a 1" core the paper tape spool would be about 5" in diameter.  I had a drafting 115VAC powered electric eraser and an attachment with a flat face and some pins that would wind up a 100 foot tape in under a minute.  Some care was needed to not tangle the tape and avoid a paper cut.

The one illustrated at http://www.nadcomm.com/33asr.htm has a rotary dial and built in modem that was not present on mine.

TTYPUNCH: punch a perforated ITA2 TTY tape -

BROOKE CLARKE HTTP://WWW.PRC68.COM
 B    R  O   O   K  E        C   L  A   R   K  E        H   T   T   P   ://                      W  W  W  .               P   R  C   6   8   .               C   O   M

 There's something wrong with the TTY Punch program.  When a character that's not in it's alphabet something strange happens.  The input was 34 characters and the output is 40 characters.

Also see:
HP 2748B High Speed Rack Mount Paper Tape Reader -
Tape-ARD® Paper Tape Key Punch Model 173A
Telegraph - Teletype
KD-100 Key Tape Disintegrator

ADM-3A Glass TTY Kit (Wiki)

The PCB is on a diagonal which is at the hing line between the top (white) and bottom (Blue).  In the kit version that PCB was filled with ICs.  There was a later version with the same size PCB, but it had maybe a couple of ICs and the rest was just blank space.

Sirus Cybernetics VGA-1 50-Pin Printed Circuit Board

This was part of a kit to add a VGA color display instead of using an I/O board to drive a TTY machine (real or glass).
The PCB is dated 1980 and is raw, i.e. no parts have been installed.  Maybe it required an expensive chip.

Ferro Resonant Power Transformer

I replaced the stock power transformer with a ferro resonant model.  This involved also adding an oil filled cap that connects to an extra winding on the transformer.  The idea is that the resonant circuit rejects narrow spikes and dropouts and also supplies some energy for short brown outs.

Intel 2102 Static RAM Memory

The large type card for the SWTP computer could be purchased that were all covered with 2102 memory chips.  They made a point of NOT using sockets because the reliability of the socket was lower than the chip.
2 pins for power, 4 control pins (DO, DI, /CD and R/W) and 10 address pins for a total of 16 pins.  It has a capacity of 1 bit x 1024 = 1k bits.

AC-30 Audio Cassette Tape Data Storage

The next innovation was a box made by SWTP that allowed storing data on audio cassettes by means of tones, very similar to the 1200/2400 Hz modems that were starting to appear.
Wiki: Kansas City standard.  This required the SWTP AC-30 interface box.  Two common portable cassette tape recorders and three cables going to each recorder: Mike, Earphone & Motor control.  You need two recorders so copy from one to the other.  While it can be done with one recorder it takes much more time.

8" Floppy disk DMAF1

This was a big improvement on the cassette tape system.  The 8" data transfer rate (500 k bits/sec: IBM 8-Inch) was too high for the microcontroller to bit-bang (Wiki), so a wide SWTP card with DMA (Wiki: Direct Memory Access) was used to move data directly to/from RAM to the floppy disk.  Motorola has a chip (6844?) just for doing the DMA operations that could also be programmed to move data in the RAM space.

The early 8" disks were single sided and later there was a double sided 8" disk and SWTP packaged two of these side by side.  You need a way to copy from disk to disk and using two drives is much easier than buffering the data and manually installing and removing disks.  The same applies today to DVD and Blueray disks.

YouTube: CouriousMarc: All you never wanted to know about 8 inch floppy drives, 37:40 -

AFAICR the data rate was 500 k bits per second or 2 microseconds between bits or 16 microseconds between bytes after a serial to parallel chip converted the data. 

My next door neighbor was Reed Anderson (Wiki) founder and president of Information Terminals Corp (formerly Anderson Jacobson Inc (Patents 1971 - 1984) (Wiki)), maker of telephone modems and floppy disks (later renamed Verbatim).  He told me about the 8" double sided YE-Data floppy drives.   I bought his 8" disks and talked to him about what drives were available.  One day he brought Victor Borge (Wiki) to my house so that he could see a home application of the floppy disks.  Later Victor did commercials for Verbatim disks.  Reed Anderson was a very class act and gentleman.

YouTube: CuriousMarc: More about 8" floppies: recovering data from odd diskette formats I didn't even know existed!, 15:59 -

5 1/4" Floppy Disk

Replaced the 8" disk with a pair of 5-1/4" disks because they were smaller and ran at half the speed allowing for bit-banging, thus greatly lowering the cost.  These had a slower data rate so the microcontroller could handle the data without needing the complexity of DMA.  At 250 k bits per second or 32 microseconds between bytes there's enough time to bang bytes with a 6800 chip.

3.5" Floppy Disk SSSD

Replaced the 5-1/4" disks because of even smaller size and more rugged because of the hard plastic shell.  Replaced the use of tape for write protection on the 8" and 5-1/4" disks and instead used a sliding plastic door that could be read by an optical sensor.

These started as Single Sided Single Density and had the same interface as the 5 1/4" Floppy disks.  They progressed through Double Sided and finally Double Density.  (I paid $50 for a box of 10 DSDD disks from HP when they were introduced.)

Don Lancaster TV Typewriter Kit

This is a way to use a standard NTSC TV monitor as a display device.  The problem is that with a TV bandwidth you can not display a line of 80 characters.  I have the book but didn't make it.  Prior to this the main interface was a Teletype ASR 33.

Lear Siegler ADM-3A Dumb Terminal (Glass TTY) Kit

This is just a keyboard and CRT to replace the TTY keyboard and paper printer (but not the paper tape punch & reader).  24 Rows by 80 Columns.  Punched cards have 80 columns so the display works for input and output of data that may have been formatted for cards.  It has no graphics capability, only alphanumeric text. Came as a kit that uses what seems like a hundred ICs.  Later Glass TTY terminals used only a handful of custom IC chips and cost much less.

You could do graphics the same way they did for the Teletype machines, i.e. by printing characters in a graphics pattern. A period has a small amount of ink in a character space while a letter M has a lot of ink.  It was a tedious process to make up a still image in this way but was popular around Christmas time.

TI 810 Dot Matrix Printer

This floor standing dot matrix text only printer takes fan folded 132 column paper and has pin drive.  A later version had graphics capability.  I seem to remember that is cost something like $2,000 back when a dollar was worth something.

Epson FX80 Dot Matrix Printer

This pin drive printer takes 8.5" wide paper and has graphics capability.  Mine has a special add on chip that increases it's capability.

Wood Cased Version

I installed the boards into a custom wood case rather than keep using the factory SWTP case.  Note I seemed to always have the cover off the SWTP case in order to get access to do hardware hacks.

Home Brew Computer Club (Wiki)

The meetings were held at the auditorium at the Stanford Linear Accelerator Center (SLAC) on Page Mill Rd. in Palo Alto, a few minutes drive from where I lived in Los Altos Hills.

The auditorium had each row of seats about 1 foot higher than the row in front so that everyone could easily see the black boards, speaker, etc.  I think it's built on the side of a hill.

The format would be:

• General Announcements (Lee Felsenstein Wiki)
  
• Guest speaker or speakers
• audience members would pose questions or offer things
• Random Access where we would interact.

At one of the meetings a couple of kids with greasy hair had a card table at the auditorium door showing off their computer on a board that had a built in video display, built in basic as well as a tape interface, they called it the Apple.  Many of the Silicon Valley microcomputer types were attending these meetings.

It was here that Tom Pitmam (Wiki, Itty Bitty Computers) found me.  I had a SWTP 6800 computer and a Model 33 Teletype.  At the meeting I got two paper tapes, one for a 6800 editor program and the other for a 6800 assembler.  I had already coded a hex a disassembler that would read Motorola 6800 machine code and print out a pseudo assembly listing.  This allowed Tom and myself to figure out how to use the undocumented paper tapes.  This allowed Tom to write Tiny Basic (Wiki) based on an article in the Computer Journal about a basic language with a metaphor about layers of a onion.  Because I was interested in using my SWTP 6800 computer to interface hardware one of the things I asked for were instructions for PEEK and POKE to allow looking at or modifying any memory location.  This was a capability of the HP RMB language.

At this time I was going to the Home Brew Computer Club meeting at the Stanford Linear Accelerator Center (SLAC) auditorium on Page Mill Road in Palo Alto.  Inside they would have a guest lecturer, and time for questions, wants, etc. then a Random Access time when you could go to someone with a question, want, etc. and talk one (or many) to one.

Computer called the Apple

There would be an exhibit at the entrance door.  One time it was a couple of kids showing their computer, on a card table, that was all on one board, included a basic language in ROM and had a TV interface, they were calling it the "Apple".  I think it used the 6502 chip that was very similar to the 6800 but with some enhancements (color display and sound).  This was a raw board, no cabinet.

Motorola Assembler and Editor

At another one of the Home Brew Computer Club meetings someone had the paper tapes for the Motorola Editor and Assembler (that's 2 separate programs).  And I got the tapes, no documentation or anything else.  By this time I had written a source code generator program.  It would read a machine language paper tape and print out the source code and machine language mnemonics.  This allowed me to reverse engineer the editor and assembler to find out how they were handling the I/O.  It turned out they were written by a professional and had a header where the I/O ports were defined.  It was a simple matter to reassign the ports to match the SWTP TTY I/O port.  This was a great tool, now you could write code without worrying about recomputing loop jump addresses.

Later a generic assembly language became popular called "C".  "C" is a low level language it's just a little bit removed form assembly language where the language is very closely tied to the registers of the target chip.

I think most (all?) of the computers that use the 68xxx chips use the same method so it's straight forward to have one box emulate another box.
A very high level language is LabVIEW by National Instruments.

San Francisco Computer Fair (Wiki)

There was an annual show that was for small home type computers in SF.  Jim Warren (Wiki) put it on and used roller skates to cover the large area of the show floor.  There were papers given and exhibits.  Those two kids with their Apple computer had a booth.

Tiny Basic

Tom Pittman wanted to develop a version of Tiny Basic as written up in Doctor Dobbs Journal.  He used my computer by working into the wee hours for a number of days.  When he was done I got a copy.  I also asked him to include hooks (PEEK & POKE) to allow the language to bang bits so that it could be used to control hardware, not just be used as software.

Radio Shack Color Computer (Wiki)

CoCo Computer	Multi Port Adapter

This was based on the Motorola 6809E Chip.  The Microware OS9 operating system was available for this computer.  It used a color TV as an output device and can be used to make titles for video tape.  Motorola has an app note on the 6809 chip with the design of the COCO.  This was a low cost computer with a lot of capability.

The way the CoCo was made the operating system is stored in ROM, but at power on it is transferred into RAM.  This means that you can write code to replace the operating system.  Since the standard TV display is just a part of memory you can change the start address of the display map and that allows you to look at any RAM address dynamically.

There was a company in Sunnyvale that was buying the COCO from Radio Shack, removing the case and embedding it into a commercial product that was involved in routing telephone calls (like to a doctor's office after hours).

When the 6800 chip came out Motorola had a firmware program where if you donated some code  you could access their 6800 library of other donated code.  They analyzed that code and based on what they found improved the orthogonality of the 6809 code as well as that for the 68000 series.

Multiport Adapter

By plugging this into the game slot, you could have 4 game slots available.  There were a number of after market products that fit the game slots, like a Glass TTY type display that would drive a wide bandwidth monitor to allow 80 characters per line which is more that you can get with a TV display.

One of the cards in the display adapter is to drive a CRT monitor so that I can display 80 characters in one line (way more than the TV display allows).

What's in a Name?

Displays

Lamp

This is a binary counter made using Fairchild RTL ICs.  I  lived in Mountain View, California at this time and Fairchild started out on Ellis St.  A local phone call would get a BIG three ring binder full of RTL data sheets.  Some surplus stores sold seconds that did not have labels but instead a color stripe and these were lower in cost than the official parts.  One button increments the count and the other is the reset.



This is a prototype Johnson (Twisted Ring) Counter (Wiki).  The lamp indicates a divide by ten, i.e. is the carry out.  A push button mounted in the drilled steel practice panel increments the count.
A Johnson counter may be the fastest counter you can make since the toggle rate for each IC is much lower than the input unlike a straight binary counter where the input stage is toggled at the input rate.

Neon

A group of 10 NE-2 Neon bulbs were arranged in a row with a photographic film overlay showing the Arabic digits 0 through 9 in a straight line.  This AC-4 used 4 dual triode vacuum tubes.

Nixie

A Nixie tube (Wiki)  has the Arabic digits formed by shaping wires and they are all stacked front to back inside a Neon filled glass tube.  You only see the digit that's on. 

I made a Johnson ring counter using 5 Fairchild 923 RTL Flip-Flops (Semi Museum) on one Printed Circuit Board (on the right).   A decoder PCB has RTL gates on the back.  These only need to have two inputs since decoding a 5 stage Johnson counter is much easier than a binary counter.  Each gate output drives high voltage transistor that turns on a Nixie element.  The artwork was done using black tape on plastic drafting sheets at 4 X life size.  A blueprint shop then makes a high resolution negative that's life sized.  A PCB etching house makes the single sided board.

The metal shop at Aertech had some machines that engineers were allowed to use (mess up) but most of the machines were strictly off limits and rightfully so.  The manual punch press and sheet metal brake were fair game.  So it's possible I made the sheet metal part or that a machinist did it in the late 1960s, or whenever RTL was introduced by Fairchild.  The sheet metal would hold a two digit counter.

Nixie Tube Patent

2138197 Electric discharge lamp, Pressler Hermann, Richter Hans, Radio Patents Corp, 1938-11-29, - but not different symbols, this is more like a neon sign.
2142106 Signaling System and Glow Lamps Therefore, Hans P. Boswau, Jan 3, 1939, 315/184 ; 235/23; 313/245; 313/250; 313/257; 313/582; 315/167; 315/170; 315/299; 315/324; 315/DIG.1; 377/113 - looks like Nixie tube, but  has a problem described below
2618760 Glow tube anode construction, Robert D Hancock, Charles R Williams, Glenn E Hagen, Chris S Andersen, Northrop Grumman Corp, 1952-11-18 - the only citation is to 2142106. "It is, therefore, an object of the present invention to provide a means whereby the distance between the cathode and anode of a multiple cathode ionic glow tube or the the read-out type is maintained nearly uniform, over` the entire length of the cathode, and for all of the cathodes."

LED

One of the first solid state display devices was the Light Emitting Diode.  A relative had a Bulova Accutron (Wiki) watch with an LED display, but the problem was that it took a lot of power and used up batteries quickly.  Later TI sold an LED wristwatch and when the battery went dead it was cheaper to buy another watch than have someone replace the battery which required special tools and skill.

LCD

The Liquid Crystal Display consumes a very small amount of power and there is a type that only consumes power when changes are made (see Kent below). 

Kent Displays - Zero Power LCD display

Memory Chips & Programmers

    GQ-4X

MCUmall GQ-4X programmer supports  EPROM, EEPRO/Flash, Serial EPROM, MCU, SRAM, GAL/CPLD devices.  Purchased to read EEPROMs in old  Trimble GPS receiver and other equipment that use memory chips.

Byte Shop (Wiki)

Calculators

HP came up with a way to calculate hyperbolic functions digitally to many decimal places of precision.  This was the first time that could be done and made books of tables obsolete.
3766370 Elementary floating point cordic function processor and shifter, J Walther, HP Inc, 1973-10-16, - "Three arithmetic units including three shifters are operated in parallel and controlled by a microprogram stored in a read-only memory to provide an improved elementary function floating-point processor. The microprogram includes a set of routines for calculating 20 elementary functions including arithmetic, exponential, hyperbolic, logarithmic, square root, and trigonometric functions. Each shifter is capable of reading a fixed plural number of consecutive bits, beginning with any bit position, from an associated data storage register."
Cordic Functions (Wiki).

3839630 Programmable calculator employing algebraic language, E Olander, R James, I Larson, R Watson, F Yockey, W Covington, J Walden, F Wenninger, H Russell, HP Inc, 1974-10-01, - HP 9810A
3859635 Programmable calculator, Robert E Watson, Jack M Walden, Charles W Near, HP Inc, 1975-01-07, - HP 9810A

HP 9100 (Wiki)

This was the first calculator that worked with things like logs, trig functions, unlike the prior calculators that only did add or subtract (and some could do multiply and divide or square root based on adding and/or subtracting).  The next generation was the HP 35 (Wiki) calculator.  It had the same computing power but would fit into a shirt pocket.

At Aertech in Mountain View, California we hand made tunnel diodes.  After the key electrical parameters were measured on each diode the cutoff frequency and another calculated parameters were needed.  The 9100 could do the calculations using a program that was loaded into the calculator.  Since it used core memory the program was always there.  We made a plywood top for a typewriter cart that had countersunk holes for the feet of the  9100 and there was a strap holding the 9100 to the cart.  This way we could wheel it to the test station where it was needed in a safe way.  This calculator was made before ICs were available.  It was about this time that a start up company called Fairchild, located in the same city, was offering RTL logic in round plastic packages that ran on 3.3 Volts.

The display was a CRT which showed  4 lines of the stack (as far as I can remember).  This made using reverse polish notation much easier for a beginner.

About this time the sports car magazines were carrying ads for the Curta Calculator (Wiki) that was a mechanical device that looked a lot like a pepper mill.  This was used for Time & Distance Rallies.

3176144 Selective signaling system, Robert O Gunderson, Edmund F Klein, NCR, 1965-03-30, - related to program control and core memory so may be related to the 9100?

HP 35 (Wiki)

This was the first hand held calculator that would do scientific calculations and as an engineer working on microwave products, some of which were bound for the Vietnam conflict, I ordered an HP 35 with a DX-A2 priority, which meant HP had to ship my calculator prior to any shipped for civilian use.  This calculator does not have any programming ability, you can only use the canned functions.  The LED numeric display was a battery eater.
3855461 Calculator with key code association and display features, R Stockwell, HP Inc, 1974-12-17 -
3863060 General purpose calculator with capability for performing interdisciplinary business calculations, France Rode, Jr William L Crowley, Alexander D R Walker, David S Cochran, HP Inc,1975-01-28 - HP 35
3892958 Inverse/complementary function prefix key, Chung C Tung, HP Inc, 1975-07-01 -

HP 41 (Wiki)

The 41 uses an LCD alphanumeric display and can be programmed or can use plug in ROM modules.  Someone found a bug in the HP 41 that would allow getting into the native machine code.  This allowed writing code in machine language and the creation of the "Zen Rom" that has a huge number of new routines.  Printers, mag card readers and bar-code readers can also be plugged into the module slots.  The alpha aspect of the display make this a very powerful pocket computer.

HP 48GX (Wiki)

This is a later version HP scientific calculator that also has an RS-232 port that I interfaced to a Motorola GPS receiver.  Has a graphic LCD display.

This is my everyday calculator, but early June 1023 the display shows '(Greek symbol for Pi)' instead of the number 3.1415......
Pressing Purple Shift Clear causes CLEAR to be displayed.  Pushing <- (left arrow) seven times erases it.
So far have not been able to get back to normal operation.  Let me know what's going on.

HP 80 Series Calculators

These were being sold during the time when the first models of the IBM PC were being sold.  The "C" in PC means computer and there were a number of government contractors that had strict rules about who could buy "computers".  By calling the 80 series machines "calculators" the purchasing process was greatly simplified.  These calculators have the ability to control instruments using the HP-IB interface allowing test automation.  Had a CRT display but I don't remember if it did graphics.

TI Speak & Spell (Wiki)

This was a very high technology item when first released in 1978.  Made use of Linear Predictive Coded Speech chips (Wiki).

Versions

Ver 1. Chliclet keyboard, Vacuum Florescent display
Ver 2. Membrane keyboard, Vacuum Florescent display
Ver 3. 2019 Membrane keyboard, no ext power jack, no headphone jack,  LCD display with narrow viewing angle, very small internal PCB.

Photos

Fig 1 Version 1, Chiclet (Wiki) keyboard (Wiki).	Fig 2 Box.	Fig 3 Fun with Words! book.
Fig 4	Fig 5	Fig 6   The daughter board is probably the power supply  (Reddit: PSU) and includes a VFD output.
Fig 7	Fig 8	Fig 9 Power Supply

Power Supply

First powered up using external power supply set to 6V and current limited to 1 Amp.  Grunge sound and bad ju-ju smoke smell.  This is very similar to the experience of Reddit user TheMightyMadman (Speak & Spell PSU).  But my main board and the power supply daughter board (1033305-3, date code 8013) are different than the Reddit main and daughter board (1033305-5, date code 8304).
The 8 connections to the daughter board shown are:
1. Speaker 1
2. Speaker 2
3. -3.5 Volts
4. -12 Volts
5. V battery (+6 V)
6. Ground
7. Start Up
8. Power Down

Keyboard

4 Rows x 10 Keys = 40 keys total.  Implememted as a matrix of 4 x 5 (20) keys.  4 row wires + 5 column wires = 9 wires per group of 20 keys.
The white plastic for the group is marked (Fig 8) "Texas Instruments, Incorporated, Pat. Pend."
The white plastic also has 5 test points on a diagonal.

Chips

Computer on a Chip TMC0271-N2L (a TMS1000 (Wiki), Data_Manaul.pdf) variant); data code 8022;

Speech Synthesier TMC0281

TMC035INL (2ea) ROMs.

Converting speech into an electrical signal and converting an electrical back to speech is very important in the field of voice cryptography, machines.

Patent numbers on Speak & Spell (but related to calculators)

	3921142 Electronic calculator chip having test input and output, John D Bryant, Glenn A Hartsell, TI, 1975-11-18, -
	3934233 Read-only-memory for electronic calculator, Roger J. Fisher, Gerald D. Rogers, TI, 1976-01-20, -
	3955181 Self-refreshing random access memory cell, Joseph H. Raymond, Jr., TI, 1976-05-04, -
	3987416 Electronic calculator with display and keyboard scanning signal generator in data memory, Jerry L. Vandierendonck, Roger J. Fisher, Glenn A. Hartsell, TI, 1976-10-19, -
	3988604 Electronic calculator or digital processor chip having multiple function arithmetic unit output, Joseph H. Raymond, Jr., Individual, 1976-10-26, -
	3989939 Electronic calculator or digital processor chip with combined functions for constant, keyboard and control bit, Joseph H. Raymond, Jr., Individual, 1976-11-02, -
	3991305 Electronic calculator or digital processor chip with multiple code combinations of display and keyboard scan outputs, Edward R. Caudel, Joseph H. Raymond, Jr., individuals, 1976-11-09, -
	4005293 Pushbutton keyboard switch assembly having individual concave-convex contacts integrally attached to conductor strips, Henry J. Boulanger, TI, 1977-01-25, - Phosphor bronze dome
	4064554 Microcomputer with code conversion, Graham S. Tubbs, TI, 1977-12-20, - to save memory space
	4073006 Digital processor system with simultaneous branch and OP code function, Graham S. Tubbs, TI, 1978-02-07, -
	4074355 Digital microprocessor system with shared decode, Graham S. Tubbs, TI, 1978-02-14, -

Speak & Spell and Related Patents

	4188626 Method for scanning a keyboard and for actuating a display device via common conductors, Gene A. Frantz, George L. Brantingham, TI, 1980-02-12, -
	4209844 Lattice filter for waveform or speech synthesis circuits using digital logic, George L. Brantingham, Richard H. Wiggins, Jr., TI, 1980-06-24, -
	4295181 Module for an integrated circuit system, Richard R. Chang, Gene A. Frantz, William R. Hawkins, TI, 1981-10-13, - ROM module
	4310831 Pulse width modulated, push/pull digital to analog converter, Alva E. Henderson, Gene A. Frantz, TI, 1982-01-12. - 30 pages or figures -
	4357489 Low voltage speech synthesis system with pulse width digital-to-analog converter, Alva E. Henderson, Gene A. Frantz, TI, 1982-11-02, - Speak&Spell - 30 pages or figures -
	4398059 Speech producing system, Kun-Shan Lin, Kathleen M. Goudie, Gene A. Frantz, TI, 1983-08-09, - uses 127 Allophones in a Linear Predictive Coding scheme.
	4380371 Liquid crystal display device, Gene A. Frantz, TI, 1983-04-19, - not Vacuum Fluorescent Display?
	4403965 Electronic teaching apparatus, William R. Hawkins, TI, 1983-09-13, -  Touch and Tell
	4507750 Electronic apparatus from a host language, Gene A. Frantz, Kun-Shan Lin, TI, 1985-03-26, - Speech to Text
	4516260 Electronic learning aid or game having synthesized speech, Paul S. Breedlove, James H. Moore, TI, 1985-05-07, -
	4624012 Method and apparatus for converting voice characteristics of synthesized speech, Kun-Shan Lin, Alva E. Henderson, Gene A. Frantz, TI, 1986-11-18, - adds age and sex to speaker's voice.
	4631748 Electronic handheld translator having miniature electronic speech synthesis chip, Paul S. Breedlove, James H. Moore, George L. Brantingham, Richard H. Wiggins, Jr., TI, 1986-12-23, - Touch and Tell
	4685135 Text-to-speech synthesis system, Kun-Shan Lin, Kathleen M. Goudie, Gene A. Frantz, TI, 1987-08-04, - 39 pages of rules for LPC
	4731847 Electronic apparatus for simulating singing of song, Gilbert A. Lybrook, Kun-Shan Lin, Gene A. Frantz, TI, 1988-03-15, - one version box has 12 note keyboard - adds pitch to speach
	4749353 Talking electronic learning aid for improvement of spelling with operator-controlled word list, Paul S. Breedlove, TI, 1988-06-07, -
	4970659 Learning aid or game having miniature electronic speech synthesis chip, Paul S. Breedlove, James H. Moore, George L. Brantingham, Richard H. Wiggins, Jr., TI, App: 1978-04-28, Pub: 1990-11-13, -
	7006995 Secure distribution of digital data, Roy I. Edenson, Peter F. van KesselGregory J. Hewlett, Paul S. Breedlove, William B. Werner, Keith H. Elliott, TI, 2006-02-28, - TIRIS Registration and Identification System for optical disks (copyright protection).

References

Retroist: A History of the Speak & Spell;
Circuit Bending Wiki: Speak & Spell -
TI Voice Synthesis System manual (S%26S_IC_datasheet.pdf)
Reddit: Speak & Spell PSU

Federal Screw Works Votrax (Wiki)

This was a phoneme-based speech synthesizer IC used in arcade games, amateur radio stuff & much more. SC-01 was the first one.
Herb Caen in the San Francisco Chronicle (Wiki) would mention the sign for Bay City Screw & Bolt (Wiki).

	3836717 Speech synthesizer responsive to a digital command input, R Gagnon, Scitronix Corp, 1974-09-17, -  - 32 phonemes
	3908085 Voice synthesizer, Richard T Gagnon, Individual, 1975-09-23, - 64 phonemes
	4128737 Voice synthesizer, Mark V. Dorais, Federal Screw Works, 1978-12-05, - - 128 phonemes
	4130730 Voice synthesizer, Carl L. Ostrowski, Federal Screw Works, 1978-12-19, - 64 phonemes RE30991
	4264783 Digital speech synthesizer having an analog delay line vocal tract, Richard T Gagnon, Federal Screw Works, 1981-04-28, - 64 phonemes
	4301328 Voice synthesizer, Mark V. Dorais, Federal Screw Works, 1981-11-17,  - 128 phonemes
	4433210 Integrated circuit phoneme-based speech synthesizer, Carl L. Ostrowski, Bertram White, Federal Screw Works, 1984-02-21, - 64 phonemes
	4470150 Voice synthesizer with automatic pitch and speech rate modulation, Carl L. Ostrowski, Federal Screw Works, 1984-09-04, - 64 phonemes
	4532495 Speech digitization system, Richard T. Gagnon, VOTRAX, 1985-07-30, - speech encoding system - AD7520 12 bit D/A converter,
	4829573 Speech synthesizer, Richard T. Gagnon, Duane W. Houck, VOTRAX, - 64 phonemes

HP & Agilent Rocky Mountain Basic Workstations

These came long before the PC and were typically used to control instruments using HP-IB, although some Computer Aided Drafting (Wiki: CAD) packages were written for them as well as other scientific uses.  Based on the Motorola 68000 series uP.  HP used their RMB workstations with their own CAD software to design HP products.

Magnetic Storage

When working in the 1970 we could buy direct from the government surplus equipment.  One of the things offered was a drum memory.  Never did get it, but it's a very early example of magnetic data storage.
2862199 Magnetic drum storage system, John E Scott, Sperry Rand, 1958-11-25, -

Later while at HP/Agilent I worked on the Hard Disk Spin Stand used to test heads and media.  I used air bearings.  But there was a design flaw, i.e. you can not use 50 Ohm coax cables to drive a write head.  They are OK for the read head.

Curta Calculator (Wiki)

YouTube: Adam Savage's Curta Calculator Gets CT Scanned!, 19:56 - first part is ad for scanning company.
@11:09 Visualization
The rest is mostly an ad.

Patents

		DE747073C/en Calculating machine with a single staggered roller surrounded by setting wheels, Herzstark Curt, App: 1938-08-19, W.W>II, Pub: 1944-09-06, -
		DE747074C/en Calculating machine with only one staggered roller, Herzstark Curt, App: 1939-04-13, W.W.II, Pub: 1944-09-06, -
		CH266406A/en Adding machine, Herzstark Curt, 1950-01-31, -
		2533372 Zeroizing Mechanism, C. Herzstark, Contina, 1950-12-12, - RE23553 Zeroizing Mechanism, C. Herzstark, Contina, 1950-09-30, - Same as DE873000C/en
		2588835 Independent Actuator Tens-transfer Mechanism, C. Herzstark, Contina, 1952-03-11 - DE860420C/en Ten switching device, Herzstark Curt, Contina Ag, 1952-12-22, -  uses Fig instead of Abb so there's a US version of this patent.  Found it.
		DE872999C/en Small format calculating machine, Herzstark Curt, Contina Ag, 1953-04-09, -Type 2
		DE873000C/en Counter clearer for calculating machines, especially miniature calculating machines, Herzstark Curt, Contina Ag, 1953-04-09, - Type 2

YouTube

1950 Curta Calculator, 4:48 - Model 1,  some example computations
How the CURTA Works, 13:10 -

Web Pages

Curtamania.com - Type and Age of your Curta -

Patents

Delay Lines (Wiki) date to the 1920s (Wiki).  Mercury delay lines were developed for RADAR in W.W. II (Wiki)
2629827 Memory system, Jr John Presper Eckert, John W Mauchly, Eckert Mauchly Comp Corp, Oct 31, 1947 - digital data in Mercury

Magnetic Drum Memory dates to 1932 (Wiki).

2611813 Magnetic data storage system (disk drive), Jr Edwin S Eichert, Thomas K Sharpless, Technitrol Engineering Company, May 26, 1948, 360/49, 192/84.1, 360/72.1, 360/98.1 - first disk drive patent?

Links