Re: [CAD_CAM_EDM_DRO] AGND+DGND+EARTH=NOISE....true or false?
Posted by
Jon Elson
on 2002-09-12 22:43:00 UTC
Babar wrote:
over ordinary OP-amps. It is about 1000 times better than the plain
op-amp, and is supposed to be much better than the 3-op-amp
instrumentation amp that lots of devices use for this function.
The AD620 is not a traditional op-amp. You use one external resistor to
set the gain. What you want to do is connect the signal to the + input, and
the signal ground to the - input. The output will be a very true difference
between the two voltages, regardless of noise. If the common-mode
voltage exceeds the power supply rails to the AD620, you can make up
a voltage divider for both the + and - inputs. Use 1% resistors, or even
matched resistor sets for this, to preserve CMRR.
Do not connect the ground signal from the CNC control to the servo amp
ground, as this will cause current to flow in the ground wire, defeating the
purpose of Common-Mode rejection. You DO need to have some form
of common ground between the CNC control and the servo amp, but it should
be a separate wire (which will have a current-induced voltage across it.)
If you use the AD620 for a DC tach input, you need to provide a ground reference
to keep the voltage within the Common-Mode rejection range of the AD620. A
1K resistor to ground from either (or both) tach leads to servo amp ground
should be fine.
See the AD620 data sheet (available online from www.analogdevices.com)
for more info, details, sample circuits, etc.
Jon
> Jon How to use that AD620? (can I use any other OP-AmpIf you have common mode problems, the AD620 is a great improvement
> with high CMRR) shorting output and -ve input and
> giving input at +ve input of Op-Amp????
over ordinary OP-amps. It is about 1000 times better than the plain
op-amp, and is supposed to be much better than the 3-op-amp
instrumentation amp that lots of devices use for this function.
The AD620 is not a traditional op-amp. You use one external resistor to
set the gain. What you want to do is connect the signal to the + input, and
the signal ground to the - input. The output will be a very true difference
between the two voltages, regardless of noise. If the common-mode
voltage exceeds the power supply rails to the AD620, you can make up
a voltage divider for both the + and - inputs. Use 1% resistors, or even
matched resistor sets for this, to preserve CMRR.
Do not connect the ground signal from the CNC control to the servo amp
ground, as this will cause current to flow in the ground wire, defeating the
purpose of Common-Mode rejection. You DO need to have some form
of common ground between the CNC control and the servo amp, but it should
be a separate wire (which will have a current-induced voltage across it.)
If you use the AD620 for a DC tach input, you need to provide a ground reference
to keep the voltage within the Common-Mode rejection range of the AD620. A
1K resistor to ground from either (or both) tach leads to servo amp ground
should be fine.
See the AD620 data sheet (available online from www.analogdevices.com)
for more info, details, sample circuits, etc.
Jon
Discussion Thread
Babar
2002-09-09 21:29:46 UTC
AGND+DGND+EARTH=NOISE....true or false?
Jon Elson
2002-09-10 21:40:56 UTC
Re: [CAD_CAM_EDM_DRO] AGND+DGND+EARTH=NOISE....true or false?
Babar
2002-09-12 20:45:57 UTC
Re: [CAD_CAM_EDM_DRO] AGND+DGND+EARTH=NOISE....true or false?
Jon Elson
2002-09-12 22:43:00 UTC
Re: [CAD_CAM_EDM_DRO] AGND+DGND+EARTH=NOISE....true or false?