Final Project - ECE 421L 

   

Authored by Do Le (led2@unlv.nevada.edu)

25th of November, 2020
   

Project Goal

Design, layout, and simulate a digital receiver circuit that accepts a 

high-speed digital input signal D and Di (a differential pair connected to your circuit from, for example, a twisted pair of wires such as in an 

Ethernet cable). D and Di are complements so, for example, if D is 5V then Di is 0V and output = 1. Another example, when D is 1V and Di is 2V

then output = 0. At high-speeds and long distances the voltages received aren't full digital logic levels (i.e., 5V and 0V), hence the need to design, 

and use, a high-speed digital recevier circuit. Ideally, when D > Di the receiver outputs a 1. When D < Di the receiver outputs a 0. Base your 

design on the topology seen in Fig. 18.23. Try to design for high-speed and low-power. Characterize your design (in sims) and the trade-offs. 

For example, show that you get higher-speed if you use more energy (burn more power). See if you can get, in this 500 nm process, 250 Mbits/s

(a bit width of 4 ns) with an input voltage difference of, for example, 250 mV (with D and Di swinging back and forth between 2.75V and 3V, 

for one of many examples, your circuit outputs the correspondingly correct values). Note that while Fig. 18.23 shows one inverter on the output 

you may find, for example, that two inverters work better (at the cost of power). Use a table to summarize your design's performance.

  

Part 1: Operation 

Building the Modules:

    The heart of the digital receiver are input buffers. To start, a self-biasing diff-amp using NMOS inputs is created.

    N-flavored Diff-Amp 

    The buffer circuit is then simulated using a symbol.

N-flavored Diff-Amp Simulation

    The V- input is held at a constant 2.875V and the V+ oscillates between 2.75V and 3V. The resulting waveform

    looks like this (and is taken from the Vout- terminal making it inverted).

N-flavor Diff-Amp Waveform0

    This waveform functions correctly but the output does not reach full logic levels. To solve this, a buffer stage

    is added to the output of the diff-amp using an inverter; the inverter also inverts the output so that it matches the

    input logically.

N-flavor Diff-Amp Simulation Buffer

    Now the output waveform reaches near-full logic levels.

N-flavor Diff-Amp Waveform1

    Some issues remain. The output does not reach a full 5V logic level, and transitions for low voltage outputs are not as clean.

    To fix this, an input buffer using PMOS gates as inputs is made.

P-flavor Diff-Amp

    Simulation shows superior operations to the NMOS based schematic at low voltages, but suffers from simuliar problems when transitioning

    at high voltages.

P-flavor Diff-Amp Schematic

P-flavor Diff-Amp Waveform

  

    A really useful input-buffer combines both types of input buffers to create a more decisive waveform.

    In the simulation below, solid waveforms are produced by combining the NMOS input and PMOS input 

    self-biased diff-amps.

Diff-Amp

Diff-Amp Waveform0

    With the current changes, combining the NMOS and PMOS diff-amps and using a buffer on the output, this is the input buffer so far.

Receiver0

  

    The current design works, however the waveform is choppy for high frequency. Lowering the input waveform period to 8ns yields the following

    results. Attempting to drive a load would result in a worse waveform.

Diff-Amp Waveform1

    When trying to drive a 1pF load, the following occurs on the output:

Diff-Amp Waveform2

    This can be resolved by buffering the output using more inverters at the expense of power dissipation from the extra inverters.

Diff-Amp Buffers

    The result is a more clean output waveform that reach full logic levels and can drive a 1pF load for other logic gates

    while supporting a bit width of 4ns.

Diff-Amp Waveform3

    The circuit can also operate with a bit width of 2ns with reduced performance.

Diff-Amp Waveform4

Limitations:

    There are some limitations to the configuration. The first being that the input differential signal cannot operate

    at any range. The optimal range for the differential signal is to bounce about 2.5V as shown in the simulation below.

    At low (relative to Vdd/2) DC operation, the logical high dominates the outputwaveform, taking longer to decrease. While at high

    voltage (relative to Vdd/2) the logical low dominates. This may be due to mis-biasing of the transistors.

    (Voffset is the voltage that the differential signal swings around, assuming a 250mV difference).

Diff-Amp Waveform5

 

    VDD is also mostly stable at 5V. When the voltage drops, the logical high bit width becomes degenerated. If VDD is too low,

    especially lower than the input DC voltage, the output completely collapses as shown in the following graph (where Vsource is VDD voltage).

Diff-Amp Waveform7

 

Power Consideration:

    To examine the power dissipation, the power dissipated by the two diff-amps is examined, then the

    power dissipated by the buffers. To do such, the current through the bottom NMOS and top PMOS

    is examined (the current through the circuit) and multiplied by the VDD value.

    (P = I * V)

    This is done at multiple L values first. Here is the calculator data from CADENCE along with the formula used.

    This is done for a bit-width of 4ns.

Diff-Amp Power Versus L

            L                         P

1.2µm6.314mW
2.2µm4.588mW
3.2µm3.650mW
4.2µm3.047mW
5.2µm2.622mW
6.2µm2.306mW
7.2µm2.061mW
8.2µm1.865mW
9.2µm1.705mW
10.2µm1.571mW
11.2µm1.457mW
 
    The data shows that an increase in L increases power due to the higher channel resistance.
 
    The next sweep shows values for the power dissipation as a function of the input period (twice the bit width).
    T            P
5ns2.341mW
10ns2.325mW
15ns2.321mW
20ns2.319mW
25ns2.318mW
30ns2.317mW
35ns2.317mW
40ns2.316mW
45ns2.316mW
50ns2.316mW
 
    The data shows predictable results that an increase in frequency (lower T) increases the power consumption.
 
    The next three data tables show the power dissipation of the three inverters used to buffer the output
    (12µ/6µ, 24µ/12µ, 48µ/24µ inverters) as a function of the period.
 
    12µ/6µ Inverter
    T    P
5ns4.012mW
560ns2.783mW
1.115µs2.782mW
1.670µs2.782mW
2.225µs2.782mW
2.780µs2.782mW
3.335µs2.782mW
3.890µs2.782mW
4.445µs2.782mW
5µs2.782mW
 
    24µ/12µ Inverter
    T                 P
5ns2.206mW
560ns257.3µW
1.115µs242.7µW
1.670µs238.3µW
2.225µs236.4µW
2.780µs235.5µW
3.335µs234.9µW
3.890µs234.5µW
4.445µs234.3µW
5µs234.1µW
 
    48µ/24µ Inverter
    T                P
5ns1.4mW
560ns122.8µW
1.115µs82.77µW
1.670µs66.32µW
2.225µs61µW
2.780µs56.61µW
3.335µs53.6µW
3.890µs51.48µW
4.445µs49.86µW
5µs48.6µW
 
    The trend shows decreasing power consumption with each consecutive inverter
    due to the increasing width which lowers the resistance. Whether extra
    inverters to square the buffer is worth the extra hardware is dependent on
    the power dissipation up to that inverter and the speeds acquired from it.
 
    The last table shows power consumed by the circuit as a whole by examining
    the power given off by the source.
    T            P
5ns9.908mW
560ns4.314mW
1.115µs4.285mW
1.670µs4.275mW
2.225µs4.271mW
2.780µs4.268mW
3.335µs4.266mW
3.690µs4.265mW
4.445µs4.264mW
5µs4.263mW
 
 
Operating Temperature:

    The following simulation shows the result of operating the circuit at various different

    temperatures. There is little difference, however, as temperature increases, the delay

    of the output appears to increase.

    Results at 8ns clock.

Diff-Amp Wavform8

 

    

   

  

  

   

  

  

Part 2: Layout

    In the second part of the project, a layout is made. Three modules are laid out—the N-flavor, P-flavor diff-amps,

    and the final receiver module.

 

    Shown here is the layout of the N-flavor diff-amp. The diff-amp passes the DRC and LVS check.

N-flavor Diff-Amp Layout

N-flavor Diff-Amp LVS

    
    Next is the P-flavor diff-amp. DRC'd and LVS'd.
P-flavor Diff-Amp Layout
P-flavor Diff-Amp LVS
 
    With the components laid out, the full receiver layout can be made.
Receiver Layout
    A final LVS verifies that the layout matches the circuitry.
Receiver LVS
 
     
    The workings of the input-buffer allows for receiving of data transmitted through differential pair signals.
    This method of data transmission enables transfer protocols through long wire with potentially noisy environments
    where full logic level transmission is impractical.
     
    The design directory of the input-buffer can be found here: proj_dvl_f20.zip.

     

 

  

  

Return to led2 lab directory