Project Logs

Collapse

• Finally - A CPU

  Tim • 12/11/2021 at 13:45 • 0 comments

  Read more »

• Post Layout Simulation, Testing Actual Designs

  Tim • 12/11/2021 at 12:13 • 0 comments

  Again, it turns out that introducing post-layout simulation was a very good thing. I found a bug in the cell library that caused the counter to fail after placemenent. Everything works now!
  

  Read more »

• Microcell implementation

  Tim • 12/11/2021 at 09:03 • 0 comments

  Now that we managed to define the logic gate implementation and synthesis into a netlist, it is time to figure out how to implement the gates on a PCB. The placementtool PCBPlacer.py in PCBFlow does this by breaking down logic gates into microcells that represent very basic circuit unit. The rationale is to create a minimum set of microcells to simplify the desing. A secondary contrain is that each microcell should be roughly square shaped so that they can be arranged in a grid.

  Currently supported gate types:

  • NOT
  • NAND2
  • NAND3
  • NAND4
  • D-Flipflop

  Read more »

• Analog Simulation of Synthesized Design

  Tim • 12/07/2021 at 21:24 • 0 comments

  It's nice to simulate individual gates, but better the check whether they work together as well. I created a spice library describing the individual gates the I simultated in LTSpice before. The description of the NAND gate and Latch can be seen below:
  

  .SUBCKT ne_NAND2 A B Y
  XU1 N001 N002 N001 VCC Y NC_01 VCC 0 TLC55X
  C1 N002 0 1n
  C2 VCC 0 100n
  R1 VCC N001 {RL}
  D1 N001 A RTL_DIODE
  D2 N001 B RTL_DIODE
  .ENDS ne_NAND2
  
  .SUBCKT ne_LATCH E D QN
  XU1 N001 N002 N001 VCC QN NC_01 VCC 0 TLC55X
  C1 N002 0 1n
  C2 VCC 0 100n
  R1 N001 0 {RL}
  Q1 D P001 N001 0 RTL_NPN
  R2 VCC N001 {RL}
  R3 E P001 {RL}
  .ENDS DLATCH

  Simulating the synthesized design turned out to be quite an ordeal. The TLC555 spice model makes extensive use of parametrised switches. This are described differently in different versions of spice, specifically LTSpice and NGSpice. I found adoption of the TLC555 spice model to both NGSpice and LTspice. Somehow I did not manage to get the full counter design to simulate properly with NGspice due to problems with the operating point analyises.
  

  Read more »

• Plugging into PCBFlow - Synthesis

  Tim • 12/05/2021 at 18:29 • 0 comments

  Good thing I already have a design-flow to transform digital designs in exotic logic styles into PCBs: PCBFlow.

  Now that all basic gates of our NE555 logic technology have been defined, I can start implementing the necessary technology description files to implement it into the flow.

  The first step in the flow is the synthesis, which requires a description of the basic gates as a liberty file.  The description of NOT and NAND2 gate is found below. The area designates the number of NE555 required to implement this gate.
  

  Read more »

• NE555 based latches

  Tim • 12/05/2021 at 18:00 • 0 comments

  Latches and D-Flipflops have cause the biggest headache in every logic familiy I have implemented so far. This is no different with NE555 based logic. A straightfoward way would, of course, be to form a polarity hold latch out of 3 NAND2 gates. This would require 3 NE555, however, and seems to be a bit of a waste.
  
  The NE555 has an integrated RS flipflop, meaning that a single NE555 should be sufficient for a latch. What is a bit tricky is to gate the data with the enable signal, so that the RS flip flop is either set for input=1 and EN=1 or cleared for input=0 and EN=0.
  

  In the end I found a hacky way, using an NPN transistor as a pass gate. Since the NE555 logic is slows anyways, it does not matter that the pass gate is quite slow as well. The circuit is shown below.

  Read more »

• Spice Simulation of Combinatoric Gates

  Tim • 12/05/2021 at 13:27 • 0 comments

  Testing more detailed gates will be easier in spice. First step is to replicate the inverter. Testcircuit shown below.

  Read more »

• Some Experiments: Building an Inverter and NAND gate

  Tim • 12/05/2021 at 11:23 • 0 comments

  It appears my NE555 are from a few kits I ordered on Aliexpress a while ago. They have some kind of distorted TI logo on them, so I am pretty sure they are not genuine devices, but they should work nevertheless. Time to pull out the breadboard. The resistor capacitor network you can see on the right side is to set the level of my function generator correctly to 0/5V.
  

  Read more »

• First Assessment

  Tim • 12/05/2021 at 10:07 • 0 comments

  Ok, I may have played around with the NE555 years ago and also owned some kind of circuit book for it.  But I always only copied circuits from somewhere else. So, time to study its details a bit more. It's not my first obscure logic style, so we ought to find something, right?

  The block diagram of the N555 from the TI datasheet is shown above. What do we get on a functional level? Two comparators with fixed thresholds, an RS flipflop with reset input and an output driver. The threshold input should be good to act as switch for an analog summing input network and the RS flipflop could become part of a latch. This does not look too bad. CONT and DISCH are probably not too useful for logic circuits, but we will see.

  The functionality is described as this in the datasheet:

  I don't get why the RESET voltage threshold is not listed. Luckily it is defined later in the datasheet:

  A minimum of 0.3V appears quite low. This means that no diodes or anything causing a large voltage drop may be between the reset input and ground. Something to keep in mind for later.

  Ok, all in all this does not look too bad. We can probably construct both registers and combinatoric logic out of the NE555 and some additional components. In the mean time I also found some NE555 lying around. So let's get to some experiments.
  

View all 9 project logs