The Cardboard Computer - IO is my name

My goal is a 4-bit CPU using recycled cardboard substrate and Diode Transistor Logic. This will develop into an educational platform for me.

Similar projects worth following
Attempt to build a four bit computer using recycled, re-purposed and just plain junk as the main construction material. Why? Well, the cardboard is free as my workplace puts it in the dumpster and I do not have the funds to buy better material. A very green build indeed. The first electronic parts to start this were given to me by a friend who passed away and what better tribute than to build a computer. Except for the ALU all gates are either NAND or Nor using DTL. The transistors used are NPN 2N2222 type or PNP 2N3906. Thanks to Yann Guidon and Jezz - Morning Star for additional parts. Very large construction is employed as to aid in my learning about logic circuits and also to make it easier to make changes as I find out what does not work. Each gate measures about 0.75 X 1.75 inches on cardboard cutouts. RISC will used and the clock speed will be very slow at about 3Hz to start. The main bus is 8-bits ( 4 for data and 4 for address ).

HACKADAY Wheels, Wings and Walkers 2017 contest entry.

Cardboard computer video update 1

Cardboard computer video update 2

Cardboard computer video update 3

IO - The Cardboard Computer update 4

  • 1 × 555 timer ic chip
  • 130 × 2n2222 npn transistor
  • 95 × 2n3906 pnp transistor
  • 230 × 1n914 Switching diode
  • 462 × Resistors Assorted common values

View all 8 components

  • July 16, 2017 - IO is getting crowded

    Dr. Cockroach4 days ago 1 comment

    The multiplexer is installed and space is filling up.

  • July 16, 2017 - MUX here we come

    Dr. Cockroach07/02/2017 at 07:36 8 comments

    The checkout of the MUX is completed and I am amazed that there are no wiring errors and it works like a champ. I can now start pushing 4 bit data around on the data buss :-)


    This is the finished 4:1 MUX. Now to do a complete checkout...

    Here is the current MUX circuit I am going with. This is for one bit but will expand as needed for four bits.

    Currently building up the needed logic gates for the 4:1 MUX and here is a sneak peak of one of the four input NAND gates.

    July 6, 2017 - Here is the layout of all 22 gates on their cardboard base. Two on left are the select gates and the four across the top are the outputs. Next is to hot glue them in place then start the wiring.

    July 8, 2017 - The power and ground wiring is now in place.

    July 10, 2017 - The first bit is wired and passed the selector test between four different inputs to the single output . The wiring looks a bit messy but that's the fun I have with this adventure :-)

  • June 22, 2017 - The move is done and IO is alive and well

    Dr. Cockroach06/23/2017 at 00:18 7 comments

    I have spent the past month making a move to a new home here on the North Carolina coast and am glad to report that IO has survived the trip. I am now getting the work table and IO set up and ready to continue melting solder :-)

  • May 22, 2017 - Construction on short hold

    Dr. Cockroach05/22/2017 at 23:53 1 comment

    Not to worry, the adventure with IO will go on, however My wife and I are going to move in the next 30 days and I will need to box up all my junk plus IO. This is sudden but needed. Staying in the same county but into a larger home with a real yard to mow :-) Also an out building to store lots of my stuff :-)

  • May 16, 2017 - Servo display panel

    Dr. Cockroach05/17/2017 at 00:14 2 comments

    This is a rough idea for a display panel using a servo to show the output of the ALU. Input is being made with IO's front panel data switches.

    This is Uglytech at its best :-)

    This is a cleaned up version. Not as ugly but still Uglytech.

    The ALU feeds its output to the 4 to 16 line decoder on the left. The output from the decoder goes to a resistor bank on the lower right and then to the astable multi-vibrator upper right that then sends its output to the servo.

    This is how my Winscope sees the servo signal. Not bad for a simple two transistor astable :-)

  • April 21, 2017 - Revised instruction card

    Dr. Cockroach04/21/2017 at 19:38 3 comments

    I have found out that after several cards were constructed that the cards were too stiff and warped out of shape enough to not give me a correct read every time. I have gone back to a simple design that gives a view of what is going on with the circuit and the magnet serves as both a graphic presentation of the instruction but also serves as a grip for placing and removing the card from the reader panel. I still need to wire the rest of the panel.

    The Instruction set reading left to right , top bottom

    code - instruction

    1 - Subtract one from the current cell value

    2 - Add one to the current cell value

    3 - Move pointer back one position

    4 - Move pointer forward one position

    5 - Begin while loop

    6 - Return loop until current cell = zero

    7 - Input value to current cell

    8 - Output value from current cell

    All cell values are set to zero at the start of a program run.

  • April 16, 2017 - Instruction card updated

    Dr. Cockroach04/16/2017 at 08:31 0 comments

    I have updated the instruction card buy reducing the magnets down to two. One for the reader pad and the other for the card. I have also added foil to the reader pad contacts and now have very good reliable contact between card and pad. I am also enclosing the card using the #Origaime (Cardware-bis) method.

    Ceramic magnets are now going to be used as the pull is now enough to provide reliable contact.

  • April 12, 2017 - Brainwarp, The program entry panel

    Dr. Cockroach04/13/2017 at 00:19 7 comments

    This is a scale mock-up of the program entry panel known as Brainwarp. A very RISC instruction set based on Brain*uck is used with some notable changes. There are eight instructions and program cards will be used with each card setup for one instruction.

    Below is a working example of how the card works.

    The card with the desired instruction is placed onto the program pad.

    And in this case, the code for this instruction is 3

  • April 9, 2017 - Parts update and something new

    Dr. Cockroach04/10/2017 at 00:03 1 comment

    I have just received over 1000 1n4148 diodes and now work on IO can begin once again. I have been brain storming with the team and I have decided on a RISC instruction set for this project. However, I am not yet ready to reveal the instructions and the method of program data entry. But I will say that this is going to be a very novel way to program IO and to read the instructions. I do not want to tip my hand on this and spoil the surprise for everyone :-)

  • March 29, 2017 - A late welcome to Yann Guidon

    Dr. Cockroach03/29/2017 at 23:16 2 comments

    I am a bit late but I want to welcome the newest member of the Cardboard Computer team, Yann Guidon.

    Yann has been a helpful supporter of this project from the first days of my joining HaD. Yann has donated to the

    project 500 transistors to help further the build. Thank you Yann :-)

    @Yann Guidon / YGDES

View all 34 project logs

  • 1
    Step 1

    My basic nand gate

    I guess that this could be called a single 2 input nand gate IC . Just very large scale.

    Start with cardboard and brass fasteners

  • 2
    Step 2

    Layout the parts needed.

  • 3
    Step 3

    Solder the parts to the cardboard base

View all 5 instructions

Enjoy this project?



Dr. Cockroach wrote 04/21/2017 at 19:58 point

Want to welcome @Dylan C.

To the Cardboard Computer team :-)

  Are you sure? yes | no

David H Haffner Sr wrote 03/21/2017 at 11:29 point

Hey Mark, I see you entered your project for the 2017 prize! I'm really glad you did and the best of luck my friend :)

  Are you sure? yes | no

Dr. Cockroach wrote 03/23/2017 at 23:45 point

Hey thanks David, If nothing comes of the contest then at least I am having fun. It's interesting to see just how low tech I can be and still do this. Like in the T.V. ad,"  Even a caveman can do it " :-)

  Are you sure? yes | no

esot.eric wrote 02/21/2017 at 07:57 point

Interesting quote, there in the details regarding public-labs. And you already know I think this is a cool project.

  Are you sure? yes | no

Dr. Cockroach wrote 02/21/2017 at 08:12 point

Hey there :-) The quote just struck me the right way and I really hate to conform so David let me use it :-) I like working and thinking out of the box ( Ohhh I just had to say that ) :-D

  Are you sure? yes | no

David H Haffner Sr wrote 02/20/2017 at 23:56 point

Hey Mark, your "baby" is looking very cool, actually I think I'm going to find some old pics of the CP901 (similar to the CPU in the P-3 Orion back in my day) man this thing is really looking the best way!

  Are you sure? yes | no

Dr. Cockroach wrote 02/21/2017 at 00:47 point

Hey thanks David. There is a lot more to go before even the first instruction flows through but it is coming along and I am learning a lot :-)

  Are you sure? yes | no

Morning.Star wrote 02/17/2017 at 11:02 point

Hi Mark!

Thank you for the invitation of contributor to your awesome project... Its a cardboard revolution! :-)

  Are you sure? yes | no

Dr. Cockroach wrote 02/17/2017 at 22:23 point


  Are you sure? yes | no

Dr. Cockroach wrote 02/21/2017 at 00:49 point

Hey Jez, That's a revolution on two fronts ;-)

  Are you sure? yes | no

J. Peterson wrote 01/18/2017 at 22:12 point

This is incredible work. My one caution is the choice of cardboard as a circuit substrate. If anything shorts or overheats, it’ll go off like a bomb. Corrugated cardboard is particularly bad, because the spacing within it ensures there’s plenty of oxygen to fuel the fire.

Please think about this. If nothing else, keep an extinguisher handy, and don’t leave it running unattended.

  Are you sure? yes | no

Dr. Cockroach wrote 01/18/2017 at 22:19 point

I am making sure that all stages are running cool and very low current, but perhaps placing milliamp fuses in various places would be wise :-) I had not thought about the fire hazzard till now :-O

  Are you sure? yes | no

K.C. Lee wrote 01/18/2017 at 22:47 point

Don't know how reliable this is.  With the good old Borax...

>This technique will only work for natural materials such as wood, cotton, jute, and paper or the like.

Read the comment part.

BTW stand alone vertical arrangement of paper burns the fastest, but there aren't much fuel.

  Are you sure? yes | no

Dr. Cockroach wrote 01/18/2017 at 23:18 point

The Borax is something to look into, will have to test first as to how conductive the Borax would be as well as how the solution would affect the glues in the cardboard. Thanks for the link :-)

  Are you sure? yes | no

Ted Yapo wrote 01/18/2017 at 14:45 point

Congrats on the blog writeup!

  Are you sure? yes | no

Dr. Cockroach wrote 01/18/2017 at 15:11 point

Thanks Ted, I still have a long way to go and lots and lots to learn :-) I am currently redoing the PC from ideas and nudges from you all. ( getting rid of the pulser boards...)

  Are you sure? yes | no

Ted Yapo wrote 01/18/2017 at 16:08 point

Nice.  I saw you mention you don't currently have a scope - that's probably where one would help most.  How fast do you think your clock will be?  If you can test it at less than 20 kHz or so, you can use a sound card as a low-frequency scope - a quick search should turn up some software solutions.  Some soundcards (like those used for some software-defined-radio projects) go even higher than 20 kHz.  I used a soundcard-based spectrum analyzer to test some audio noise generators, and it was surprisingly good.  I haven't played with a soundcard scope for a long time, but I'd be willing to bet there's something decent out there.

  Are you sure? yes | no

Dr. Cockroach wrote 01/18/2017 at 19:23 point

Hi Ted, I do have some sound card scope software here that works well, however right now my clock is around 3hz and that is below what the scope can see. Not a major problem yet and I will speed things up once various sections start talking to each other :-) And you knew about the blog before I did :-)

  Are you sure? yes | no

Dr. Cockroach wrote 02/21/2017 at 01:38 point

I am going to keep one pulser  for the sequencer circuit as that might be the only stage that will require edge triggering.

  Are you sure? yes | no

Yann Guidon / YGDES wrote 01/08/2017 at 08:46 point

Wait... Why would you even need a 555 ? :-D

  Are you sure? yes | no

Dr. Cockroach wrote 01/08/2017 at 09:04 point

Well, I was having my doubts about the transistor osc signal being clean enough ( Was still in the dark about level triggering ) and used the 555. I will change back to transistor later, just not a priority at the moment :-)

  Are you sure? yes | no

Yann Guidon / YGDES wrote 01/08/2017 at 09:35 point

Another cool circuit to try...
You know, you could use a basic classic astable oscillator with 2 transistors to generate a square-ish signal :-)

  Are you sure? yes | no

Yann Guidon / YGDES wrote 01/08/2017 at 09:40 point

however the astable oscillator is more delicate/fragile/picky than a relaxation oscillator or an inverter-based, hysteretic oscillator... Just add an amplifying stage (or two) to create proper edges :-)

  Are you sure? yes | no

Dr. Cockroach wrote 01/08/2017 at 10:05 point

The astable was what I started with but again, I was/am new to this and was not sure at the time :-) Well, time to get breakfast ready for the residents at work. Yumm

  Are you sure? yes | no

Morning.Star wrote 01/02/2017 at 13:33 point

When an artist's work grows too big to move, he calls it an Installation...


  Are you sure? yes | no

Morning.Star wrote 01/01/2017 at 13:18 point

Awesome project! Seeing this took me back to my early teens reading about Babbage and Turing in the village library, playing with electronics kits and discovering the school had a BBC Micro in a cupboard in the Maths Room that nobody knew how to use.

I'd like to draw your attention to Nyles Steiner's ( experiments with Memristors. These are simply-constructed doped crystal junctions that exhibit programmable resistance. The device then is a passive 2-pin solid state memory cell that could be used to provide EEPROM type memory for your processor, perhaps to store commands in?

In operation, a single gate is programmed with a high forward voltage (9v) which makes the junction thereafter conductive to a low forward voltage (3v) and highly resistive to a low reverse voltage. A high reverse voltage makes the junction thereafter conductive to a low reverse voltage and highly resistive to a low forward voltage. The device is behaving like a Tunnel Diode that can be made to change polarity by over-driving it. The gate can then be tested with a smaller voltage to determine polarity and thus state... Which is stored physically in the crystal lattice, requiring power only to read or change its state.

I have successfully followed Nyles' experiment and built a 6 bit 'flash' memory cell using an Atmel, a handful of 3904s, LEDS and some scrap copper and aluminium, but it appears you have more ambition and patience than I ;)

Hacky New Year...

  Are you sure? yes | no

Dr. Cockroach wrote 01/02/2017 at 01:42 point

Thank you for the awesome :-) I have looked at the idea of memristors a little bit but for this project I will stay with what parts I have on hand. Back in the late 70's A group of friends and I built a Imsai 8080 and had a ball learning to catch the rabbit on the led display. No keyboard or video then for us :-) I would at least like to repeat that program on this machine.

Hacky New Year to you as well..

  Are you sure? yes | no

Morning.Star wrote 01/02/2017 at 09:04 point

I just have to admire the look of Cray super, with a hint of Practical Electronics that has about it besides what it actually does. Literally old school... You sir, are an artist.

It was just those brass studs begging for a little sulphur and a tinfoil hat that appealed to my artistic nature as well as my sense of humour. There is a delicious irony in constructing things from ingredients, its a sort of alchemy... I have a thing for copper myself, it is beautiful, structural and its oxides sensitive to heat, light and pressure as well as having electrical memory. I've built many experimental devices and sensors using copper, household chemicals, heat and water just to learn from. 

One must however respect cardboard. ;-)

Hmmm. Around then I remember being ushered into Careers, a room with a large table littered with cards; farmer, postman, lawyer... In the corner was a box with a keyboard and a screen, a CBM Pet. It wasnt even switched on.

"Oh, thats a computer," the Careers Officer said. "Pay no mind to that boy, it will never catch on." I'd love to go back and show him the waterproof battery powered touchscreen computer I wear on my wrist today, a direct descendent of the BBC that inspired me to wait all these years for it to become small enough.

I wish you luck with your endeavour.

  Are you sure? yes | no

Dr. Cockroach wrote 01/02/2017 at 11:44 point

Good morning, Well, I guess that every thing we do or construct is alchemy in one form or another. Just keep that cup of coffee away from my cardboard ;-D  I do wonder if my studs will work that way, might have to remove any protective coating first. That melts away when I solder. I sure would like to get my hands on a Pet. My first comp of my own was a trs-80 and a few zx-81's. As the years move on, I find I no longer can keep up with the internal workings these days. I am 60 and want to go back to more basic times so this project might scratch that itch :-)

I might have to change the shape as I can see that I will run out of room. Might have to put the panels on a wall but I do like the cray type shape. It will be heavy to move around.

Take care

  Are you sure? yes | no

David H Haffner Sr wrote 12/30/2016 at 11:58 point

Hey Mark, I think I understand a part of what you are doing now, you are utilizing a large part of your present inventory of components. I should have read further in the details section, you did mention the '2222's, I am trying to think of something simpler that can be used that is just as effective. 

  Are you sure? yes | no

Yann Guidon / YGDES wrote 12/30/2016 at 03:38 point

If your wife complains, tell her you got 1.6k views already. Amazing !

  Are you sure? yes | no

David H Haffner Sr wrote 12/30/2016 at 00:34 point

Well Mark, you are certainly thinking outside the box...and I luv it!

  Are you sure? yes | no

Dr. Cockroach wrote 12/30/2016 at 00:41 point

I hate say to say this but instead of thinking out of the box.... I'm using it :-D

  Are you sure? yes | no

David H Haffner Sr wrote 12/30/2016 at 00:50 point

Ha ha, yes, you are right, a hint...using a 2n2222a (NPN)  can be used for small signal switching, if that helps.

  Are you sure? yes | no

Dr. Cockroach wrote 12/30/2016 at 01:52 point

Currently I have used up all of my 2222's . That is why I switched to the 3906 pnp's I still have. I still have about 180 of them left. I still have around 130 1N914 diodes.

  Are you sure? yes | no

Yann Guidon / YGDES wrote 12/30/2016 at 03:34 point

If you need a few hundreds of BC550 and 1N4148, just tell me where to send them ;-)

  Are you sure? yes | no

Dr. Cockroach wrote 12/30/2016 at 09:50 point

You got Yann :-)

  Are you sure? yes | no

David H Haffner Sr wrote 12/29/2016 at 21:56 point

This is very clever! 

  Are you sure? yes | no

Dr. Cockroach wrote 12/29/2016 at 23:03 point

Thanks, I am thrifty by nature and am having fun and learning at the same time on the cheap.

  Are you sure? yes | no

David H Haffner Sr wrote 12/29/2016 at 23:24 point

Man. this is a lot more than "thrifty," I almost think it boarders on ZEN-like! My present project is the first 1 of mine, that has the sophistication that it has, though most of the parts I design myself to save money.

I was thinking about those bimetallic blinkers for Xmas tree lights as make shift switches, but I think their response time would be way too slow to be even effective for this project. I don't know, I'm always thinking of the weirdest things!

  Are you sure? yes | no

Dr. Cockroach wrote 12/29/2016 at 23:48 point

Weird is good ( sometimes ). I have made a lot of ham radio projects that were no more than 3 or 4 transistors so this computer is way over the top for me. My wife asks me how are we going to move this thing if we relocate in a few years. I have no idea. One of many good things about HaD is thinking out of the box. By building ourselves, we can flex our thinking as much as we want to and share with other weird folks :-)

  Are you sure? yes | no

Yann Guidon / YGDES wrote 12/30/2016 at 03:39 point

You nailed it, Mark :-)

  Are you sure? yes | no

Dr. Cockroach wrote 12/29/2016 at 19:27 point

I went shopping today at a Habitat restore ( thrift store ) and bought a string of holiday white LED's for 1 dollar. It's a rats nest of wires but seems to be over 200 LED's so not a bad deal. At a initial look each LED runs at 3 to 4 volts at 3.75 ma. I self destructed one at 5 volts. Now I have plenty for this cardboard project for my output display panel.

  Are you sure? yes | no

Ted Yapo wrote 12/29/2016 at 20:26 point

Great idea - they have leads already attached, too :-)

I have often thought about cheap incandescent holiday bulbs for homemade PROMs inside a diode array - apply 12V to program (burn out the zero bits), and normal logic levels to read.  Just replace the burned bulbs to erase.  Incandescent strings  are becoming tougher to find, though...and you would have to avoid those burn-proof ones with the shorting links around the inner terminals.

  Are you sure? yes | no

Dr. Cockroach wrote 12/29/2016 at 23:01 point

Not only the LED's but all of the wire will be used on the build as well. I went to Wally world also hoping to find the same but only the incandescent types were still on sale., the led strings were already gone for the season. Now back to soldering the dmux gates :-)

  Are you sure? yes | no

Dr. Cockroach wrote 12/23/2016 at 08:53 point

Well.... The ALU and decoder boards are using thumb tacks on wood bases. The wood I had to buy, the cardboard is free, lighter and easier to cut to size. As far as leakage, just have to wait and see. I could apply a non-conductive coating first but then I would have to spend money and I am a cheap builder :-)

  Are you sure? yes | no

Shaos wrote 12/25/2016 at 06:23 point

So it was thumb tacks initially? :)
I chose the same approach, but with "Cork Tiles":

  Are you sure? yes | no

Yann Guidon / YGDES wrote 12/25/2016 at 07:01 point

Apparently a new trend has been created ! :-D

  Are you sure? yes | no

Dr. Cockroach wrote 12/25/2016 at 09:17 point

Yup. I had constructed some radio projects with thumb tacks years ago. I can't take credit for the idea as I found a project online back 15 years that used them.

  Are you sure? yes | no

Dr. Cockroach wrote 12/25/2016 at 09:19 point

The thumb tack revolution has started. Yesssss !

  Are you sure? yes | no

Stuart Longland wrote 12/25/2016 at 23:13 point

Here I was thinking, "Well, I'll use thumb tacks for now but will replace it with something better later"…

Still haven't gotten around to the "something better" yet… and that antenna still goes.

Having seen the comments here, I don't feel so guilty.

  Are you sure? yes | no

Ted Yapo wrote 12/23/2016 at 01:40 point

At first I thought they were thumbtacks, and couldn't figure out how they stayed put :-)

Those fasteners are a great idea!

If you end up with leakage issues in humid environments, you could also substitute corrugated plastic board:

Although I'm not sure how well it stands up to soldering...

  Are you sure? yes | no

tomwsmf wrote 12/21/2016 at 15:21 point

Amazing project. Very hands on and very informative. This makes for a great deep look into how our tech shizzlewhizzle works under the epoxy blobs and TOS. Please keep adding notes and sketches here on Hackaday. 

  Are you sure? yes | no

Dr. Cockroach wrote 12/22/2016 at 12:30 point

Thanks, there are quite a few out there that have done this but for me it is both a learning aid and perhaps a strange looking piece of art.

  Are you sure? yes | no

Just4Fun wrote 12/21/2016 at 14:15 point

OMG!!!! :)

  Are you sure? yes | no

Yann Guidon / YGDES wrote 12/21/2016 at 03:15 point

Please, upload some schematics :-)

What do you use to mount the parts ? I've never seen this method before :-)

  Are you sure? yes | no

Dr. Cockroach wrote 12/21/2016 at 09:21 point

Good morning Yann, As soon as I figure out Hackaday I will get some sketches of the circuits I am using. The metal paper fasteners on cardboard is used for a few reasons. 1 I'm poor, no extra money to buy better material. 2 it spreads out the parts so I can better learn what is going on visually and 3 it looks kinda cool and almost a art form in itself. The gates are attached to the main cardboard frame using hot melt glue.

  Are you sure? yes | no

Yann Guidon / YGDES wrote 12/21/2016 at 13:43 point

You could write a log that explains your unusual but smart method :-)
For example, a step-by-step How-To with several pictures ?
Who knows who it might inspire in turn ?
Regards :-)

  Are you sure? yes | no

Yann Guidon / YGDES wrote 12/21/2016 at 13:44 point

Oooooops too late :-D

  Are you sure? yes | no

Dr. Cockroach wrote 12/21/2016 at 13:53 point

Yup, started posting some instructions. Now it's time to get dressed and go to work. I'm a chef at a assisted living facility and there are some days I feel like I need to live there :-)

  Are you sure? yes | no

coletonn wrote 12/21/2016 at 14:14 point

Yeah that would be nice. See how the thing ticks. 

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates