Some super sleuthing by Mark Florkowski @mrf found a NASA final report that includes some product cut sheets from QSI and the C-44 specifications. The bus pinout signal names and descriptions are spectacular. ntrs.nasa.gov/api/citations/19920015551/downloads/19920015551.pdf There are some key updates/corrections to things I said in the video: The name isn’t familiar, but QSI is probably the manufacturer in Utah I mentioned. According to these product sheets, QSI made a lot of nice support cards and racks up to 10 slots. Voltage range is 8 to 18 and I had pins 1&2 backwards. Fingers (L to R) are GND then Vin on left (1,2) and are GND Vreg on the right (21,22). There is another (obsolete) battery input on pin 12; The converter on the CPU board must supply 5v@75mA and when in low power mode 3v@10mA; There are two standby modes- a WAIT (which is just a halt) whereby the processor stops but system clock keeps going and bus power kept @ 5v. Then a STOP mode where the clock is stopped and bus power reduced to 3V. Not sure why the “wait” mode deserves mention since it is just a processor halt. To get out of either one the CPU board needs to either receive: a reset, an interrupt, or there is a WAKE-UP signal input from the bus Thanks again to Mark for finding these C-44 specifications!!
I love how old sockets and connectors came in blue, there's something very satisfying about that colour, and these days very retro feeling. I wish more stuff came in blue today.
This reminds me of proprietary bus systems in motion control and a filling computer of all things.. We also had GESPAC SBCs that also were used in a particle collider in the CERN but I‘m not sure about their availability. If they still are available then they may have classified use too.
I note the use of the 74HC4040 counter. It's possible that this is used as a very large clock divider to generate a wake-up interrupt of some kind. The length of the counter can allow a much slower interrupt rate than sometimes possible with regular peripheral timers, thus helping with very low power resting states with infrequent wake-ups. If the software knows the clock source, wake-ups can still be counted to enable basic time-keeping. I have seen such techniques used in other 8 bit computer hardware from the 1980s. If the application allowed, this method was often cheaper than using dedicated real-time-clock chips.
I think that is an idea definitely worth investigating. The counter in the 8155 is 14-bit, the 74xx4040 is 12-bit and they could be used in series. maybe an intermediate output of the 74xx4040 is used as a frequency divider to slow the CPU down to 1MHz to reduce power while still operating. There is a lot to explore on this board.
I wonder if C-44 could be a predecessor to the CAN bus? Being tuned for battery operation seems it would be a good fit for automotive, where CAN bus now dominates.
CAN BUS and it’s [originally] industrial counterpart RS485 are serial communications protocols using a differential signal on a twisted pair. (Think multi-drop, noise tolerant RS232). C-44 is closer to S-100 or any other full address/data/signal backplane bus.
Looks like it has the same form factor as a VIC-20 cartridge, also 44 pins. Pure coincidence ofcourse. Inserting this card in a vic 20 is not only useless but probably also very destructive: the power and GND pins on the VIC20 bus are switched around compared to this card.
This is a fantastic find ! I need to study and make a list of the things I got wrong, but the bus name pinout table and description page itself is worth its weight in gold. I have barely skimmed the document but nearly every paragraph seems to answer a question I had. Thank you immensely (and the NASA report writer that compiled the report and included original catalog pages)
Some super sleuthing by Mark Florkowski @mrf found a NASA final report that includes some product cut sheets from QSI and the C-44 specifications. The bus pinout signal names and descriptions are spectacular.
ntrs.nasa.gov/api/citations/19920015551/downloads/19920015551.pdf
There are some key updates/corrections to things I said in the video:
The name isn’t familiar, but QSI is probably the manufacturer in Utah I mentioned. According to these product sheets, QSI made a lot of nice support cards and racks up to 10 slots.
Voltage range is 8 to 18 and I had pins 1&2 backwards. Fingers (L to R) are GND then Vin on left (1,2) and are GND Vreg on the right (21,22). There is another (obsolete) battery input on pin 12;
The converter on the CPU board must supply 5v@75mA and when in low power mode 3v@10mA;
There are two standby modes- a WAIT (which is just a halt) whereby the processor stops but system clock keeps going and bus power kept @ 5v. Then a STOP mode where the clock is stopped and bus power reduced to 3V. Not sure why the “wait” mode deserves mention since it is just a processor halt. To get out of either one the CPU board needs to either receive: a reset, an interrupt, or there is a WAKE-UP signal input from the bus
Thanks again to Mark for finding these C-44 specifications!!
I love how old sockets and connectors came in blue, there's something very satisfying about that colour, and these days very retro feeling. I wish more stuff came in blue today.
Yes, I like retro green and blue. Anything but beige.
This reminds me of proprietary bus systems in motion control and a filling computer of all things..
We also had GESPAC SBCs that also were used in a particle collider in the CERN but I‘m not sure about their availability.
If they still are available then they may have classified use too.
I note the use of the 74HC4040 counter. It's possible that this is used as a very large clock divider to generate a wake-up interrupt of some kind. The length of the counter can allow a much slower interrupt rate than sometimes possible with regular peripheral timers, thus helping with very low power resting states with infrequent wake-ups. If the software knows the clock source, wake-ups can still be counted to enable basic time-keeping. I have seen such techniques used in other 8 bit computer hardware from the 1980s. If the application allowed, this method was often cheaper than using dedicated real-time-clock chips.
I think that is an idea definitely worth investigating. The counter in the 8155 is 14-bit, the 74xx4040 is 12-bit and they could be used in series. maybe an intermediate output of the 74xx4040 is used as a frequency divider to slow the CPU down to 1MHz to reduce power while still operating.
There is a lot to explore on this board.
I wonder if C-44 could be a predecessor to the CAN bus? Being tuned for battery operation seems it would be a good fit for automotive, where CAN bus now dominates.
CAN BUS and it’s [originally] industrial counterpart RS485 are serial communications protocols using a differential signal on a twisted pair. (Think multi-drop, noise tolerant RS232). C-44 is closer to S-100 or any other full address/data/signal backplane bus.
Looks like it has the same form factor as a VIC-20 cartridge, also 44 pins. Pure coincidence ofcourse. Inserting this card in a vic 20 is not only useless but probably also very destructive: the power and GND pins on the VIC20 bus are switched around compared to this card.
Yea, those 4-1/2” boards in a 44-contact 0.156” dual edge connectors go waaaaay back
ntrs.nasa.gov/api/citations/19920015551/downloads/19920015551.pdf - Pages 167-172 contain documentation on the C-44 bus itself
This is a fantastic find ! I need to study and make a list of the things I got wrong, but the bus name pinout table and description page itself is worth its weight in gold. I have barely skimmed the document but nearly every paragraph seems to answer a question I had. Thank you immensely (and the NASA report writer that compiled the report and included original catalog pages)