THS770006 Datasheet, PDF (16/39 Pages) National Semiconductor (TI) – Broadband, Fully-Differential, 14-/16-Bit ADC DRIVER AMPLIFIER

THS770006
SBOS520B – JULY 2010 – REVISED JANUARY 2012
www.ti.com
TEST DESCRIPTIONS
The following sections describe how the tests were performed, as well as the EVM circuit modifications that were
made (if any). Modifications made for test purposes include changing capacitors to resistors, resistors to
capacitors, the shorting/opening of components, etc., as noted. Unless otherwise noted, C1, C2, C9, and C13
are all changed to 0.1µF.
Frequency Response: 200mV_PP, 2V_PP, 3V_pp
This test is run with and without transformers in the signal path.
For tests with transformers, the standard EVM is used and only the gain magnitude is shown. A network analyzer
is connected to the input and output of the EVM with 50Ω coaxial cables and set to measure the forward transfer
function (s21). The input signal frequency is swept with the signal level set for the desired output amplitude. The
use of transformers gives better magnitude response that correlates best with detailed design simulation in terms
of peaking in the response due to better control of parasitic capacitance at the device output pins, but also
results in excess phase shift. So only magnitude is plotted.
For tests without transformers, the standard EVM is used, with the gain magnitude and phase shown. A network
analyzer is connected to the input of the EVM with 50Ω coaxial cable, the output is terminated with a 50Ω load,
and a high impedance differential probe is used for the measurement. The analyzer is set to measure the
forward transfer function (s21). The analyzer with a probe input is calibrated at the input pins of the device and
signal is measured at the output pin, thus effectively removing the transformers from the transfer function. The
input signal frequency is swept with signal level set for desired output amplitude. Not using transformers gives
better phase response that correlates best with detailed design simulations, but as a result of extra parasitic
capacitance at the device output pins gives significantly more peaking in the magnitude response. The –3dB
points of the magnitude response measured without transformers correlates better with measured slew rate, so
both magnitude and phase are plotted.
s-Parameters: s11, s22, and s12
The standard EVM is used with both R15 and R17 = 24.9Ω, and R16 = open, to test the input return loss, output
return loss, and reverse isolation. A network analyzer is connected to the input and output of the EVM with 50Ω
coaxial cables and set to measure the appropriate transfer function: s11, s22, or s12. Note the transformers are
included in the signal chain in order to retrieve proper measurements with single-ended test equipment. The
impact is minimal from 10MHz to 200MHz, but further analysis is required to fully de-embed the respective
effects.
Frequency Response with Capacitive Load
The standard EVM is used with R15 and R17 = R_O, R16 = C_LOAD, C9 and C13 = 953Ω, R21 = open, T2
removed, and jumpers placed across terminals 3 to 4 and 1 to 6. A network analyzer is connected to the input
and output of the EVM with 50Ω coaxial cables and set to measure the forward transfer function (s21). Different
values of load capacitance are placed on the output (at R16) and the output resistor values (R15 and R17)
changed until an optimally flat frequency response is achieved with maximum bandwidth.
Distortion
The standard EVM is used for measurement of single-tone harmonic distortion and two-tone intermodulation
distortion. For differential distortion measurements, the standard EVM is used with no modification. For
single-ended input distortion measurements, the standard EVM is used with with T1 removed and jumpers
placed across terminals 3 to 4 and 1 to 6, and R5 and R6 = 100Ω. A signal generator is connected to the J1
input of the EVM with 50Ω coaxial cables, with filters inserted inline to reduce distortion from the generator. The
J3 output of the EVM is connected with 50Ω coaxial cables to a spectrum analyzer to measure the
fundamental(s) and distortion products.
Noise Figure
The standard EVM is used with T1 changed to a 1:2 impedance ratio transformer (Mini-Circuits ADT2), R15 and
R17 = 24.9Ω, and R5, R6, and R16 = open. A noise figure analyzer is connected to the input and output of the
EVM with 50Ω coaxial cables. The noise figure analyzer provides a 50Ω (noise) source so that the data are
adjusted to refer to a 100Ω source.
16
Submit Documentation Feedback
Product Folder Link(s): THS770006
Copyright © 2010–2012, Texas Instruments Incorporated