Great explanation. Thank you for producing these. Love to see the differences and enhancements. Geeking out on this series. I'm starting a similar project to really learn how all of this works in practice and have what may sound like a silly idea in regards to the control lines and would love your take or pros/cons. I'd like to make my build using PCBs and be as modular as possible in order to change things, upgrade parts, etc. I'm going down a similar path or model or the control lines (I need more than 16), but instead of having the 238, thinking of adding a +4 bit dip switch on each board to "set an address" and have each board receive that address instead of the single signal. Using an adder or comparator to decode. I realize this adds more ICs for the decode part on each board (say a few $$ is not an issue), but gives me more flexibility in configuring. Do you have any initial thoughts on this model? Yeah/Nay...
Thanks! I am working through the microcode for my 8-bit instruction set that I am working on, and admittedly this Google Sheet approach doesn't scale well. It's great when inspecting a single instruction, but unwieldy when making 256 instructions. I will need to think of a better way 🤔
Glad you found it useful! I have at least one more video planned (though not a lot of time to make it), and I do intend to do a few more projects on the SAP-1 as there a few more things I want to learn.
Interesting conversation on the '238 HCT is in stock, but the lead time on HC is 35 weeks and LS is now unobtainium on this one! Active low seems to be much more common. The LEDs do add some complexity.
You could make the LEDs work with an active low signal by placing the LED's cathode on the signal line and then connecting the anode to Vcc, and placing a current limiting resistor somewhere in there. In fact, from an electrical design perspective, that is actually better because most ICs, and notably LS, can sink more current than then can source. Powering an LED takes a large fraction of the current the LS can source. Later in this project I got in trouble because of the active high LED sourcing that I am doing (start pulling too much current from an active high signal line). However, my goal here isn't designing the best solution, but more to work through the logic and structure of various CPU components. I find active high line to be conceptually simpler, hence my preference for them.
Very nice explanation. Around minute 29:30 you mentioned that you already upgraded your flags registers and there would be another video. I looked here and github and your posts on Bens Reddit but didn’t see anything. If it exists could you point me in the right direction?
Note that I am pulling off the inverted CLR signal from the switch feed too. True, I could have just used one inverter to do that, but since these inverters come in packs of 6, I just used a double inverter to feed the OR gate, generate the /CLR signal, and maybe clean up the signal a little.
I have been thinking about making one of these after watching Ben Eater. I wanted more in the ALU , so add, subtract, NOR XOR OR AND, and reset the step counter controlled by 3 control lines and logic gates. Registers, A,B,C,D,X,Y. C and D 8 bit input or output to external control. design the hard ware, design the software , more software needs more hardware and more hardware needs more software. 23 instructions so far, 24 control lines 8 bit bus Thinking about an EEPROM to save programs too, I dont DO arduino,. maybe program it through pc com port to the C register. If i can get the hardware and software to work on paper, then I will get started. I will not get chips made in prc as they are generaly not worth the cheapness. I was used to working with 4000 series chips with their wide voltage range, but they are getting rare. Have not done machine code programming for 35 years, (commadore 64 ) I was concidering putting the control registers in the middle, and rotating the chips to give the input/output lines to the bus facing the bus. I like your presentation and your computer build, there are some RATS NEST builds out there that would be a nightmare to trouble shoot.
I certainly will be working on adding ROM to my build to save programs to. It will be part of my RAM upgrade project which admittedly I haven't started yet. I too didn't want to mix an AVR chip into this build, as it is antithetical to the "learning" goals IMO. I have had some success with chips from China, but only from "reputable" vendors. There is certainly a lot of junk there too. None the less, for this project, I mostly order from Mouser because I am not ordering at volume that makes the savings worth it. While I certainly did not want a rats nest of wires, it also wasn't my goal to make a picture perfect build. I just didn't want to get lost in all the wires. This is why wiring all of the LEDs up front comes in handy. I now think about the control lines by their LED position (1 through 36), which is easy to track, rather than their EEPROM or 74'238 pin, which is harder to track.
Love your speaking style and great video! I wanted to point out that the '138 is actually a demultiplexer ruclips.net/video/pwLErAYZvzI/видео.html Also, I'd love to hear the reasons why you are using HC instead of LS. I think I know, but am curious just how low of a power we can get something like this to run using CMOS. Battery/solar/off grid? Maybe even a Peltier thermoelectric setup. So many ways to go with this.
Thanks. Actually, I was using a 74 HCT238, not a 74HC238. The reason for using the HCT is simply I couldn't get a '238 in LS, and HCT is compatible with LS. I'm not really concerned with power usage (right now), most focussed on learning what works.
Great explanation. Thank you for producing these. Love to see the differences and enhancements. Geeking out on this series. I'm starting a similar project to really learn how all of this works in practice and have what may sound like a silly idea in regards to the control lines and would love your take or pros/cons. I'd like to make my build using PCBs and be as modular as possible in order to change things, upgrade parts, etc. I'm going down a similar path or model or the control lines (I need more than 16), but instead of having the 238, thinking of adding a +4 bit dip switch on each board to "set an address" and have each board receive that address instead of the single signal. Using an adder or comparator to decode. I realize this adds more ICs for the decode part on each board (say a few $$ is not an issue), but gives me more flexibility in configuring. Do you have any initial thoughts on this model? Yeah/Nay...
Great work! The idea of using Google Sheets to show who controls what is very cool
Thanks! I am working through the microcode for my 8-bit instruction set that I am working on, and admittedly this Google Sheet approach doesn't scale well. It's great when inspecting a single instruction, but unwieldy when making 256 instructions. I will need to think of a better way 🤔
Hello mate found you from the Reddit post
I hope I will see more videos on the breadboard computer
Glad you found it useful! I have at least one more video planned (though not a lot of time to make it), and I do intend to do a few more projects on the SAP-1 as there a few more things I want to learn.
Interesting conversation on the '238 HCT is in stock, but the lead time on HC is 35 weeks and LS is now unobtainium on this one! Active low seems to be much more common. The LEDs do add some complexity.
You could make the LEDs work with an active low signal by placing the LED's cathode on the signal line and then connecting the anode to Vcc, and placing a current limiting resistor somewhere in there. In fact, from an electrical design perspective, that is actually better because most ICs, and notably LS, can sink more current than then can source. Powering an LED takes a large fraction of the current the LS can source. Later in this project I got in trouble because of the active high LED sourcing that I am doing (start pulling too much current from an active high signal line). However, my goal here isn't designing the best solution, but more to work through the logic and structure of various CPU components. I find active high line to be conceptually simpler, hence my preference for them.
Great work..,thanks to share
Very nice explanation. Around minute 29:30 you mentioned that you already upgraded your flags registers and there would be another video. I looked here and github and your posts on Bens Reddit but didn’t see anything. If it exists could you point me in the right direction?
It's in my video about increment registers here: ruclips.net/video/Ps2bChnnHDg/видео.html
22:10 do you really need second inverter? Why not tie it's outputs to the switch itself?
Note that I am pulling off the inverted CLR signal from the switch feed too. True, I could have just used one inverter to do that, but since these inverters come in packs of 6, I just used a double inverter to feed the OR gate, generate the /CLR signal, and maybe clean up the signal a little.
I have been thinking about making one of these after watching Ben Eater. I wanted more in the ALU ,
so add, subtract, NOR XOR OR AND, and reset the step counter controlled by 3 control lines and logic gates.
Registers, A,B,C,D,X,Y. C and D 8 bit input or output to external control.
design the hard ware, design the software , more software needs more hardware and more hardware needs more software.
23 instructions so far,
24 control lines
8 bit bus
Thinking about an EEPROM to save programs too, I dont DO arduino,. maybe program it through pc com port to the C register.
If i can get the hardware and software to work on paper, then I will get started.
I will not get chips made in prc as they are generaly not worth the cheapness.
I was used to working with 4000 series chips with their wide voltage range, but they are getting rare.
Have not done machine code programming for 35 years, (commadore 64 )
I was concidering putting the control registers in the middle, and rotating the chips to give the input/output lines to the bus facing the bus.
I like your presentation and your computer build, there are some RATS NEST builds out there that would be a nightmare to trouble shoot.
I certainly will be working on adding ROM to my build to save programs to. It will be part of my RAM upgrade project which admittedly I haven't started yet. I too didn't want to mix an AVR chip into this build, as it is antithetical to the "learning" goals IMO.
I have had some success with chips from China, but only from "reputable" vendors. There is certainly a lot of junk there too. None the less, for this project, I mostly order from Mouser because I am not ordering at volume that makes the savings worth it.
While I certainly did not want a rats nest of wires, it also wasn't my goal to make a picture perfect build. I just didn't want to get lost in all the wires. This is why wiring all of the LEDs up front comes in handy. I now think about the control lines by their LED position (1 through 36), which is easy to track, rather than their EEPROM or 74'238 pin, which is harder to track.
Love your speaking style and great video! I wanted to point out that the '138 is actually a demultiplexer ruclips.net/video/pwLErAYZvzI/видео.html Also, I'd love to hear the reasons why you are using HC instead of LS. I think I know, but am curious just how low of a power we can get something like this to run using CMOS. Battery/solar/off grid? Maybe even a Peltier thermoelectric setup. So many ways to go with this.
Thanks. Actually, I was using a 74 HCT238, not a 74HC238. The reason for using the HCT is simply I couldn't get a '238 in LS, and HCT is compatible with LS. I'm not really concerned with power usage (right now), most focussed on learning what works.