Skip to content

Stanford-BIS/BenTasktic

Folders and files

NameName
Last commit message
Last commit date

Latest commit

ย 

History

12 Commits
ย 
ย 

Repository files navigation

BenTasktic - Tasks for Ben

Braindrop

Tasks for October run

Master Bias Generator (PTAT)

  • Need startup circuit?
  • Work out Synapse EXC/DC/INH matching
    • Three different DACs, not guaranteed to be matched
    • But, the three DACs can be set independently
    • May need to scale D by E/D and I by E/I and scale the two back at the output
    • In this way, D and I's matching can be improved by using more DAC bits
    • This way, single bits correct more for D and even more for I.
    • Need to do the math.
  • Test and validate schematic
    • Finalize I_unit (inside the core FB circuit) [1pA]
    • Finalize I_out: 1pA (along with V_CAS for each) as input to DAC
    • Finalize resistance
      • 4.x MOhms & 50mV drop
    • PTAT performance across corners
      • Error w.r.t. PTAT < +/- 0.5% over 20-30C
    • Finalize transistor sizes (for mismatch)
      • *[Note] Gain stages are 2.5n x 1/10x1/250 = 1pA
      • [Note] A single stage mirror needs about 25 Tx on each side to get 95% within +/- 10%
      • [Note] Have models in Mathematica that need to be optimized
    • PSRR
      • Small-signal analysis matches AC for mirrors.
      • With the voltage buffer, the bias generator starts peaking around 10KHz and has 0dB PSRR beyond that.
  • Layout

DAC

  • Schematic
    • Check linearity in SPICE
    • Custom additions to specific biases
      • Size global diode transistors (e.g., 4X if synapse's n-m = 4, add/subtract I_W from I_DC from two biases)
  • Layout
    • Analog core
    • Trim metal
    • Short and drive duplicate wires
    • Provide power landing pads [in M3]
    • Export LEF (and merge in Voltage Buffer)

Voltage Buffer

  • Test and validate schematic
    • Size transistors
    • Test for NMOS and PMOS biases
    • Test across corners
    • Merge with DAC
  • Layout

Neuron

  • Schematic
    • Update 1 Synapse + 2x2 Soma tile
  • Layout
    • Export LEF
  • Plan global bias and power lines

Synapse

  • See if I can use MOM (1-5) cap, block all the metal up to 8 and get the same capacitance as the MIM (9-10).
    • Can use MOM for Synapse PE: 22.8fF with ~ 3.4๐œ‡mx4๐œ‡m area
  • Schematic
    • Finalize I_LK range (minimum I_LK = 100fA => maximum = 100pA)
    • Finalize I_PE range [for t_PE >= 10๐œ‡s or 500us period, use 1000pA to 20pA]
    • Finalize I_DC, +/- I_W range [70pA +- 65pA]
    • Finalize I_CM range (if synapse gain != 1)
    • Finalize capacitance (for tau <= 100ms and t_PE >= 20๐œ‡s)
      • C_tau
      • C_PE: MOM cap M3-M5, 22.8fF, ~ 3.4๐œ‡mx4๐œ‡m area
    • Finalize initial transistor sizes (for mismatch), adjust if space available
  • Layout

Soma

  • Schematic
    • Finalize I_REF range
    • Finalize I_OFFSET range (100fA to 100pA) [if I_OFFSET not in Synapse]
    • Finalize I_R and I_G range
    • Finalize gain & offset bits (if any) [gain = {1, 4}, offset = {-3, -2, -1, 0, 1, 2, 3}]
    • Finalize initial transistor sizes (for mismatch), adjust if space available
  • Layout

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published