I’m using the MSP430AFE2xx Microcontroller 24-Pin Socket Target Board (MSP-TS430PW24) to test the noise performance of the MSP430AFE253 SD24A ADC. I’m not getting the noise performance I would expect from the ADC. I’m getting roughly 10 to 11 noise-free bits (typically 35 – 55 counts of noise) on the upper 16 bit data (total of 65535 counts). Here are the settings I’m using:
Differential input (SD24UNI=0)
Continuous sampling mode (also tried single conversion mode)
Gain of 1 (tried other gains)
Oversampling ratio (OSR) of 1024
Internal analog reference (resulting in an allowable +/- 0.6V max.swing)
Microprocessor is 3V powered by the USB/JTAG (also tried powering externally with a 3.3V battery, removing the JTAG connection, and adding a 470nF bypass cap, but this resulted in little or no improvement)
Pos/neg channel leads internally shorted using SD24INCHx=0x111 and tied to analog ground (also tried externally shorting all input leads)
While there are no performance specs listed for the SD24A, or at least none that I can find, I would expect much better performance from a 24-bit ADC part. I must therefore assume I’m not setting some parameter correctly or not interpreting the output correctly. Any help would be appreciated.
I did manage to find one source for SD24 noise performance -- in a TI energy tech. presentation for the MSP430F67xx, which includes the SD24 “B” ADC. In that presentation, the SD24B was reported to have a SINAD that approaches 90dB (~15 bits) at an OSR of 1024.
