Compile a source follower amplifier, and show your resulting high-level file that created your experimental setup. Characterize the source follower for at least two bias values. Make a single plot for the transfer function with these bias values. Curve fit these curves to find the gain. Does the response change as a function of the bias values? Make an equivalent plot using the Xcos / Scilab simulation (level=2), put this graph on the same graph as your other plots, and compare the results.
From your data and analysis of the source follower, explain how you can find kappa as a function of source voltage. It turns out this configuration is a good measurement of how kappa changes for a fixed bias current. Discuss how this change in "kappa" would be related to changes in depletion capacitance.
Measure the gain of a high-gain transistor amplifier configured in a common-source configuration. Measure a rough transfer curve (say 25-50mV steps) between the amplifier input and the amplifier output. Notice where the data makes a sharp transition, and relate this value to measured parameters and bias voltages. Then make a detailed measurement in the gain region say at 1mV steps, etc. Make a single plot.
Next perform a dense sweep of the input voltage and measure the resulting output voltage around the high-gain region for both amplifiers. You need an input resolution around 1mV or smaller. Because of the high gain, you will want a voltage divider helps get cleaner data. The limited resolution of the DAC and ADC makes this step rather important. Dividing the input by a factor of 10 or 20 should be sufficient. There is a block in the library, which I would recommend you use. You will want a transfer curve of the divider block to line up your data.
Make a single plot of this data and curve fit to find the gain in the high-gain region. Measure the gain of each amplifier and compare it to typical values of sigma (Early voltage), kappa, and UT. You should compare to your numbers measured in your previous experiments. What current level do you estimate this circuit was biased at?
Program a large bias current near the edge of subthreshold operation (i.e. 100nA-300nA). First question is, how do we program and what are we programming? When you pull up your OTA block, there is a parameter asking for the bias current, which is set internally. These parameters program a Floating-Gate (FG) device on-chip, similar to a EEPROM or Flash memory cell, that acts as a current source for the transconductance amplifier. We will cover quite alot more about FG devices throughout the class. What you need to do is do a voltage sweep on the inputs, keeping one input fixed, and measure the output response.
We will examine the time--domain response of the OTA amplifier connected as a unity-gain device, which we call a follower--integrator circuit. Compile and program one of the internal OTAs as a follower circuit with a capacitive load. Since the switches already have capacitance, you might already think your previous circuit compilation should be sufficient, but this time with the - terminal connected to the output. Buffer the output of the IC (therefore only looking at the dynamics from the on-chip behavior). Program the device around 1nA of bias current.
Third, for one of your two biases, make a sweep of the output voltage versus input voltage (transfer function). What is the follower gain? Does anything unexpected happen as the input voltage approaches GND? Make a plot (linear scale) of Vout vs. Vin, with a curve fit over the linear region.
Apply a small amplitude step input---less than 40mV peak to peak when measured at the input. Make sure that you can see the dynamics of each circuit from your measurements. Display both input and output waveforms on a single plot. Repeat for a second bias current level.
Apply a larger amplitude step input---greater than 500mV peak to peak when measured at the input. Display both input and output waveforms on a single plot. Repeat at the a second bias current level (or close to that current level).
Determine the time constant t of the integrator. You should get this by curve fitting the data, and one should see a clear straight line fit. (as mentioned in the on-line lecture materials). Is there any difference between the rise and fall times? Curve fit response to linear RC response to determine the t of the integrator in both cases.