Sigma Delta vs ADC chip?

rwgast_logicdesign

So I was just listening to first spin, and they were kind of talking about Sigma Delta not to detailed though. It made me wonder if there's any advantage to sigma delta ADC over something like an mcp320x.

Im wondering becuase Im designing this board with an mcp3208 for robotics, I was planning on putting 100 ohm resistors on all the user i/o prop pins for current protection, I know this will kill any chance of a prop sigma delta, so im curious if it even matter when you have an actual ADC chip

Mark_T

Well it reduces the number of chips...

Phil Pilgrim (PhiPi)

rwgast_logicdesing wrote:

I was planning on putting 100 ohm resistors on all the user i/o prop pins for current protection, I know this will kill any chance of a prop sigma delta, ...

No, it doesn't kill it at all. The sigma-delta converter has ~100K resistor looking into it, so the 100 ohms won't make much of a difference.

If you need absolute DC accuracy greater than eight bits, a dedicated ADC is probably the way to go. But for AC apps, like audio, and low-precision DC apps, the sigma-delta method will save money and board space.

Here's a sigma-delta app note that explains the technique in more detail:

http://www.parallaxsemiconductor.com/an008

-Phil

Tymkrs

Hey rwgast We're totally aware it wasn't even a surficial glance at sigma delta - we'll be going more into it in the next few episodes. I've done a bit of research on it as well, and the first few posts will come on Friday (I believe) at tymkrs.tumblr.com if you're interested. It's a lot more complicated than I thought!!

Addie

Mark_T

Phil Pilgrim (PhiPi) wrote: »

No, it doesn't kill it at all. The sigma-delta converter has ~100K resistor looking into it, so the 100 ohms won't make much of a difference.

If you need absolute DC accuracy greater than eight bits, a dedicated ADC is probably the way to go. But for AC apps, like audio, and low-precision DC apps, the sigma-delta method will save money and board space.

Here's a sigma-delta app note that explains the technique in more detail:

http://www.parallaxsemiconductor.com/an008

-Phil

The sigma-delta components need to be mounted right on the pins ideally - wouldn't bringing the sense pin out to a connector mean it would pick up more noise and compromise sigma-delta performance (whether or not theres 100 ohm resistor)? Haven't used it myself but I've seen lots of caveats about tightness of layout needed for sigma/delta circuit.

Phil Pilgrim (PhiPi)

Mark is right about keeping a tight layout. The feedback resistor and filter caps should be surface-mounted and placed right at the Prop's pins. I would also exclude the feedback pin from any additional connections, e.g. to an external header.

-Phil

jmg

rwgast_logicdesign wrote: »

Im wondering becuase Im designing this board with an mcp3208 for robotics, I was planning on putting 100 ohm resistors on all the user i/o prop pins for current protection, I know this will kill any chance of a prop sigma delta, so im curious if it even matter when you have an actual ADC chip

If you have a 12b ADC already, then a Prop sigma delta will not add much.

If you need to go above 12 bits, a device I like the look of, is the NAU7802 - in stock at Digikey, in DIP16 and SO16, and under $2 @ 10+
This is a dual 24b i2c ADC claims
* 23 bits effective precision analog-to-digital converter
* Simultaneous 50Hz and 60Hz rejection (reaching -90dB)
* RMS Noise:
50nV in 10 SPS data output rate and PGA gain = 128
150nV in 80 SPS data output rate and PGA gain = 128
* Programmable PGA gains from 1 to 128

cavelamb

jmg wrote: »

If you have a 12b ADC already, then a Prop sigma delta will not add much.

+1