Continue to Site

Welcome to

Welcome to our site! is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

ad676 resolution and noise

Not open for further replies.


Full Member level 5
Sep 18, 2005
Reaction score
Trophy points
Activity points
Hello, I have two questions concerning ad676(here is datasheet as a matter of fact I've been trying hard to get full resolution(16 bit) of this device but not much success.
1- in the datasheet it has been noted that AGND and DGND should be connected at one point preferably at the ad676. I read somewhere else that's because if a voltage difference between AGND and DGND exists it can damage the IC. but if AGND and DGND be connected at ad676 then the ground return path of both analog and digital are exactly commom and all the noises in digital section will be present in analog section. right? why datasheet has recommended such scheme? then what is the use of separate AGND and DGND if we connect them at ad676?

2-I have made a circuit that uses ad676. I have made all the recommendations of datasheet: a) I have used two different grounds connected at ad676. analog ground for reference ad587, analog sections and digital ground for microcontroller and digital ICs. b)I have decoupled power exactly as datasheet has recommended.
c)I have used double sided board but with extensive ground zones(known in prrotel as polygons. both sides of the board have big Analog ground zones near ad676. d)signals are brought with shielded wires. but I still have lots of noise. when I connect input to analog ground and analog ground sense I get about 6 LSB (+/-3 in 65535 range) fluctuation.(Sampling rate is about 1KS/sec).
I founded(may be I'm wrong) that analog ground sense is extremely sensetive factor when I connect ad676 input to an analog voltage in another board(via
sheilded wire) and also connect AGND sense to the ground of that board which of course has the same AGND with my ad676 board noise level considerably increases. environment noise is not a strong factor although all the boards are in a metal box that connected to AGND. I founded that if I turn on a fluorcent lamp in vicinity of boards noise level dose not change considerably!
I will appreciate any hint or help or answer to any of my questions. thanks alot.

I think that your problem is with the noise isolation in the grounds. I think that the AD people were suggesting that you join the grounds together on the board because inside the chip, both the grounds are actually connected to the same substrate.

Definately, the analog ground is sensitive to all the stray inductances of all the wires and the probes. Hence, I would suggest that you use a different ground for analog and digital. In case, you cannot do so, you have to actually isolate the digital and analog grounds on the board and connect them seperately. Also you can provide a jumper on the board in case you have to connect the DGND and AGND

thanks vamsi, I also noticed that I had a noise component on signal which after removing that situation got somehow better but I still have about 6 LSB noise.
I have used different grounds for analog or digital but the main point that I don't understand is if they are to be connected at ad676 what is the use of their separation, what difference it can makes(in theory) if we have two different grounds that are actually connected at ad676? I thought the logic behind having separate grounds is to have different return paths for digital and analog current of the IC but if the are connected at ad676 then what's the use of them? thanks, I think if I understand this I can deal with noise better.

Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to