ADUC7020

Hey, I don't want to hear about your problems... I'm trying to get an answer from Analog Devices about the AD7699! I think I have finally got them to admit that reference drift does *not* explain zero offset TC.

The interfaces and datasheets of mixed-signal products are typically horrible.

John

Oh-oh. Guess who else had looked at that datasheet.

Lucky me, I had already decided "perhaps not" If I have to use just an ARM with external converters, odds are I'd go with Linear's converters. I had been looking at the AD 8 channel DAC and a

This is one place where the folks at Silabs do a fairly good job. They document the analog parts as though they were external things. It is like they get the guy who does the analog section to write up something that goes to the technical writers.

Hmm. A customer of mine is planning to use that part to digitize very low frequency stuff--seismometer and tiltmeter data and so on...could you say a bit more about what the issues are?

CCD chips are the worst, IME.

Cheers

Phil Hobbs

I am seeing a huge amount of zero drift with temperature. Chris Augusta, the guy at ADI who wrote the datasheet, helped me find the problem. I now owe him beers.

In bipolar mode, you apply a differential input between a pair of pins, like IN0 and IN1, pins 16 and 17, for instance. Turns out that the IN1 pin must be within +-0.1 volts of Vref/2 or the zero TC goes to hell. It says that on the datasheet, but I missed it somehow. I'd got too used to pseudo-differential charge-balance ADCs actually having lots of common-mode rejection. My bad, I guess.

On my board, I'm using a 2.5 volt bandgap and a C-load follower opamp to generate the Vcm used by all 16 ADCs. I can change that to a 2.048 bandgap and get legal again.

Here's the board:

Now I can crank the accuracy specs up to less embarassing numbers.

John

Oh, if they plan to run this ADC at 500 KHz, the timing is right on the bleeding edge. Rob has working VHDL if anybody needs it.

John