Return to Roger's Home Page
1 May 2023
; "Noakes Data Communications")
Click images for larger view
1973-1987 Noakes Data Communications 3330 Stovall St, Irving, TX 75061 214-790-1050
Noakes Data Communications was a technology company owned by Gilbert F. Noakes Jr in Irving Texas during the 1970s-80s.
Noakes Data Communications, 3330 Stovall St, Irving, TX 75061
I'm writing this page to be the memory of Noakes Data Communications which was a small tech company in Irving Texas in the 1970s and 80s. I know very little about the higher-up things that went on there. If you would like to contribute information or just say hi, contact me at: roger AT arrick DOT com.
Incorporation Date: 13 June 1973
Dissolution Date: 20 January 1987
I got hired at Noakes around January 1978 (age 16) as a technician and left around January 1981 to take a position at Percom in Garland. My job was mainly repairing 8080-based industrial computer boards but I eventually did some hardware design and wrote a CP/M BIOS for the system.
Noakes sold and repaired terminals and modems, and built an industrial computer used for data collection and other things. At one point we built acoustic couplers also. Gil Noakes was the founder. At the peak, which I think was around 1979, Noakes had about 30 people.
;)
Some of the people I remember who worked there are:
Click on pictures for a larger view
I saved these few rare pictures from Noakes. All of these were taken at the 3330 Stovall Building in Irving around 1980. They were probably all taken on halloween (my birthday).
This picture is Jaynell Tarvin who answered the phones. Out the window you can see a cadillac. There was a red one and a blue one parked out front - one was Gil's and one was Ken Brown's.
;)
This is some of the crew. Mostly production. Jaynell is on the far right.
;)
This picture is Kathleen. I don't remember what she did, maybe Mary Ann's assistant.
;)
This picture is a girl who worked in engineering and I don't remember her name.
;)
This is Shelia who worked as a technician fixing boards with me. I don't remember her last name.
;)
Here are some media spots I've found by hunting. Click them for a larger view.
Computer World 1982 - Noakes aquires Envax
;)
;)
I have this acrylic panel in my old junk bin but I don't remember what it could be, I'm guessing "DA-COL" was Data Collector. I probably hand-lettered it with rub-offs.
;)
Noakes distributed, serviced and also manufactured modems. I believe the ones we manufactured were from a small company Gil bought because I don't remember any engineering to design those. This is the only picture I can find. It has the Noakes logo on it.
;)
At Noakes we use to service a lot of these Singer printers. They are a 5x7 dot matrix character printer if I remember right and took tractor-feed or roll paper. The speed was 30 characters per second and it had an RS232 port on the back. They were mounted to a heavy stand with wheels. One day in 1978 or 79 one of these heavy printers rolled out from under me and crushed my toes resulting in an ER visit - good times. If you've ever seen one of these or know of any data, please let me know.
;)
Noakes built a ton of cables, mostly RS232 with those old-style hoods. The pins were so expensive back then that a 25 pin D-Sub connector would only have pins that were needed, maybe 9 total. The pins were gold plated and crimped on a big machine.
;)
Noakes designed and built a proprotary 8080-based bus-based industrial computer called the Micro-80 Emulator. This is my primary interest because I spent a lot of time working on these and playing with them. Eventually I got CP/M running on them which was a lot of fun, and the beginning of my computer career.
One of the big customers for this computer was possibly American Teleprocessing Corporation in Houston.
The Noakes computer had an 8080 CPU board that plugged into a 72 pin bus along with RAM, ROM, and I/O boards. It was big and heavy, but state of the art for the mid-70s.
I don't have any pictures of a real system to show and know of no advertisements in computer magazines.
If you happen to run across some of this old hardware please contact me.
;)
There were 2 sizes of motherboards made, the bit 12-card board shown here, but also a smaller one that I think was 6 slots. Here's the bus pinout I reverse engineered in 2017 from the schematics. Card edge connectors with fingers on .156 centers. 3/4" between cards. Each card was 8-1/8" wide and 6.5" tall, standard .062 PCB double-sided. Most of the boards were made by Irving Circuit Board who was across the street.
;)
;)
There were 2 types of enclosures, small and large.
The large enclosure was a 19" cabinet 7U tall, with rack ears. The metal was gold alodine plated This one was in production first. I think there was only a reset switch and LED on the front panel. A black plastic vacuum-formed front panel. The enclosure housed a card cage with a 12-slot motherboard with vertical cards and a large linear power supply on the right side. The back had D-sub cutouts for connectors.
This same large enclosure was also used as a dual 8" floppy disk drive system with Shugart SA800 drives and a linear power supply and small regulator board. There was a solid-state relay inside that could turn off AC to the drives so they wouldn't spin. This relay was controlled with a pin on the RTC/parallel board.
The small enclosure was a frame with a 6? slot motherboard. There was a standard linear power supply in the back. The cards were horizontal and plugged through the front which was covered by a smokey plexiglass front that velcro'd on. You could see LEDs from the cards through it. Then a black metal C-shaped shell slid over the frame having curved edges and a handle on top. This enclosure was introduced about 1980 or so. I think it had the SDLC card and was marketed as a data collector or data concentrator.
The card cage was an aluminum frame with slots on the sides for air-flow between the boards. Grey plastic card guides fit into holes on each side. This whole frame was bolted to the motherboard. The large motherboard had 12 slots and the small one had about 6 or so. Each card slot was a 72 pin card-edge bus on .156 centers. Spacing between cards was 3/4".
Here is a picture of the 12-slot card cage in my home-brew cabinet:
;)
The power system was a standard linear type supply starting with AC entering a Corcom filter module having a PH163 oval 3-prong connector. The filter was routed to a fuse holder and a big toggle switch for power on/off. There was a single big transformer with 3 outputs driving two large bridge rectifiers. There was 1 giant filter cap for the +10V and 2 smaller filter caps for the +20/-20 rails. These 3 rails and ground were wired directly to the motherboard. Current capacity was around 25 amps for the 10V rail and 5 amps for +20 and -20. Later we put large bleed resistors across the filter caps. Boards had their own 3-terminal regulators to produce 5V for the chips and +12/-12 for the RS232 level translators.
Here are the pitiful remains of what I have left:
;)
This is the memory map used to boot CP/M on the Noakes computer. 2K of ROM sat at the top with System I/O handlers. 2K below that was Mon80 which was a simple monitor to modify memory and look at I/O ports but it could also read/write disk sectors. RAM started at 0000 where CP/M would load up to F000 maximum.
Memory Map for booting CP/M F800-FFFF 2K 2716 PROM with System I/O Handlers F000-F7FF 2K 2716 PROM with MON80 0000-EFFF RAM
I designed a circut to modify the CPU board so that upon reset a latch would hold the top 4 address lines high making F000 appear at location 0. The could would then do a jump to fix the program counter, then do an access to the on-board interrupt controller which would reset the latch and release the address lines. This way reset would jump to F000 where the CP/M boot loader was. CP/M loaded at 0. Here's a drawing of that circuit:
;)
The 8080 chip has only 256 bytes of I/O so they are used sparingly. Most boards could be configured for any address but usually they were configured like this: (addresses in hexidecimal)
Dual Serial Board with 2 8251 UART chips: (Console) 8251 #1 data register: 80H 8251 #1 control/status register: 81H (printer) 8251 #2 data register: 82H 8251 #2 control/status register: 83H (modem) 8251 #3 data register: 84H 8251 #3 control/status register: 85H (terminal) 8251 #4 data register: 86H 8251 #4 control/status register: 87H 8214 Interrupt controller on CPU board: ECH? EFH? Parallel/Timer board with 8255 and 8253: 8253 clock 0: 78H 8253 clock 1: 79H 8253 clock 2: 7AH 8253 control and status register: 7BH I don't know where the 8255 was, maybe 24H Floppy Disk controller and DMA board (my prototype board): 8257 probably consumed F0-F8H FDC 8257 DMA Ch 2 buffer address register: F4H FDC 8257 DMA Ch 2 byte count register: F5H FDC 8257 DMA mode/status register: F8H FDC 1771 control/status register: 20H FDC 1771 track register: 21H FDC 1771 sector register: 22H FDC 1771 data register: 23H FDC drive select register: 24H (some sort of latches) FDC motor control register: 25H FDC door lock register: 26H FDC son-alert register: 27H
There was only 1 type of CPU board, an 8080 running at 1.8Mhz, completely with interrupt controller and power regulators, nothing else. The part number was NDCS-80-CPU
I have one 2 of these boards. The one in my homebrew system works and was modified by me in 1979 to hold the reset addresses high so it would boot to upper ROM so it could load CP/M which needs RAM open at the bottom of the memory map. I also have a schematic for this board although I'm not sure if there were minor revisions after this.
;)
;)
This card had an 8255 parallel I/O chip (bottom left) which could be wired to drive a centronics parallel printer via a 26 pin connector on the top. It also had an 8253 timer chip (bottom right) which could be used for a real time clock or heartbeat interrupts. I don't have one of these boards but I'm pretty sure this is a fuzzy picture of it:
;)
This was an early board and used 2 8251 UART chips along with a 14411 Baud Rate generator chip. RS232 level shifting was done with 1448/1449 chips with a 16-pin DIP header that could configure the 26-pin header connector at the top for each channel. I lucked out and retained the schematic for this one, although I'm pretty sure it went through some minor revisions with the newer one losing the slide switches at the top.
;)
;)
The Quad Serial card used 4 8251 UART chips. It had 4 26-pin header connectors on the top with 2 on the front and 2 on the back. The chip in the upper left is the 14411 baud rate generator with a big crystal. This board came later, maybe around 1979. All I have on it is this fuzzy picture:
;)
The 4K RAM board used 2102 chips which are 1K by 1 bit. It takes 8 chips to make 1K, so 32 chips to make 4K. The regulator on this board got real hot. It was a 3A regulator if I remember right. That was before the 78H05 5A regulators came out. The 16-pin DIP header at the bottom is how we configured the board for various addresses. I remember a spell of bad chips and having to pull a lot of them.
;)
The 16K RAM board used 2114 chips which were 1K x 4 bit. It took 2 chips to make 1K bytes. There are 2 3-pin jumpers to configure the addresses. I have one of these boards running in my system.
;)
Something makes me think there was an attempt at a dynamic RAM board that didn't work and never went into production.
I don't have a picture of this board. If I remember right it had 4 sockets for 2708 EPROMS (1K chips)
Plain-Jane 16K EPROM board with 8 sockets to accept 2716 EPROMS (2K each). Giant.
;)
Plain-Jane 32K EPROM board with 8 sockets to accept 2732 EPROMS (4K each). As you can see it's almost an exact duplicate of the 16K board requiring just a few minor mods to take advantage of the new larger EPROMS. Giant.
;)
There was a DMA board using an intel 8257 DMA controller chip. I don't have any picture or documentation on it.
There was a floppy disk controller board with a WD 1771 controller on it. I don't have any picture or documentation on it.
This is a wire-wrap prototype of a floppy disk controller using a WD 1771 chip and an intel 8257 DMA controller chip all on one card. I believe it was the only one ever made and that these 2 functions were split into separate FDC and DMA boards in production. I think the problem was basically they wouldn't fit on a 2-sided board. This is the board I used to get CP/M V1.4 going on my machine.
;)
There was a prototype board we had made and you can see it was used for my FDC/DMA board above.
Later in the product line, maybe 1979 or 1980 there was an SDLC communications board. I don't have one or schematics or any documentation for it. This MIGHT be a partial fuzzy picture of it, but the intel 8273 chip is 40 pin and I don't see one on this board, so not sure:
;)
That's all I've collected about Noakes so far. If you run across anything please let me know so I can continue this preservation effort.
;){kind=link}
;){kind=link}
;){kind=link}