RHD5950 Datasheet, PDF (2/8 Pages) Aeroflex Circuit Technology – RadHard-by-Design RHD5950 16-Channel Multiplexed 14-Bit Analog-to-Digital Converter

Gain compression will occur near either power supply extremes but can be avoided if the references are more
than 200mV away from the respective supply terminals. The input span can be less than 4 volts at the expense of
ultimate resolution
The analog channel’s input impedance is primary capacitance (20pF typical). The input voltage charges a
track-and-hold hold capacitor through transmission gates. The input bandwidth is determined by the slew rate of
the hold amplifier and is adequate to allow input sampling in three clock periods (3uS nominal). The ultimate
bandwidth is determined by the aperture uncertainty associated with the closing of the sample gate
(approximately 5nS). The converter bandwidth is then determined by the sampling Nyquist frequency rather than
the input signal; change rate (dv/dt) and the LSB weight in volts as would be the case if there were no
sample-and-hold.
Start-Convert (STCNV_H), Busy (BUSY_L) and End-Of-Convert (EOC_H) status and control lines are
provided. The converter will operate in either continuous or single conversion modes. To operate in continuous
mode, STCNV_H should be tied to BUSY_L. The digital output register changes at the end of a conversion and
is latched when EOC_H is asserted High. The output data will remain latched until the next conversion is
complete and will be updated when EOC_H is asserted High. The output circuitry operates from a voltage
independent of the remainder of the chip such that I/O is compatible with digital systems from, less than 3.3 volts,
to 5 volts.
The converter divides the reference voltage into 16 segments with a linear weighted resistor network. The voltage
on any segment is passed to a linear 10-bit DAC for interpolation. The sampled input voltage is compared to the
output of the two stage DAC for a 14-bit successive approximation conversion.
All inputs are protected to both power supply rails by semiconductor diodes. Inputs should be constrained to VCC
+0.4 and GND-0.4 to avoid forward biasing protection paths.
The devices will not latch with SEU events to above 100 MeV-cm²/mg. Displacement damage environments to
neutron fluence equivalents in the mid 10¹⁴ neutrons per cm² range are readily tolerated. There is no sensitivity to
low-dose rate (ELDRS) effects. SEU effects are application dependent.
Notes:
- The STCNV_H is a dynamic input (positive edge triggered) and should not be tied to a static voltage.
- The input signals should be low pass filtered to reduce high frequency noise
- If Sleep mode is enabled (EN_H=0), when waking up (EN_H=1), the unit has to complete an entire
conversion cycle so the digital logic is in the proper state.
Ex. If using a 1MHz clock; after EN_H=1 and 20us after STCNV_H is applied.
SCD5950 Rev G 1/14/16
2
Aeroflex Plainview