Maker Pro
Maker Pro

Zero offset via PIC

O

ozzy

Jan 1, 1970
0
Hello!

I would like to address a problem I'm mulling over at the moment. I
have a micro-controller based circuit with a front end consisting of an
Instrumentation amplifier. I thought of applying zero volts to the
inputs (via pins from PIC) as a way of finding the sum of offset
voltages - via A/D converter. My worry is that this could affect the
high CMRR (120dB) needed at the input - can this be quantified?

Another alternative is to use Comparators of the PIC, but I'm already
using one of them. This circuit is battery operated (9V split supply)
and would therefore have to perform (lowest point) at 3V - ruling out a
CMOS switch.

These are just a few thoughts whizzing around me at the moment. Any
help on which way to go would be helpful.


Regards,


Ozzy.
 
D

Dave

Jan 1, 1970
0
ozzy said:
Hello!

I would like to address a problem I'm mulling over at the moment. I
have a micro-controller based circuit with a front end consisting of
an Instrumentation amplifier. I thought of applying zero volts to the
inputs (via pins from PIC) as a way of finding the sum of offset
voltages - via A/D converter. My worry is that this could affect the
high CMRR (120dB) needed at the input - can this be quantified?

Another alternative is to use Comparators of the PIC, but I'm already
using one of them. This circuit is battery operated (9V split supply)
and would therefore have to perform (lowest point) at 3V - ruling out
a CMOS switch.

These are just a few thoughts whizzing around me at the moment. Any
help on which way to go would be helpful.


Regards,


Ozzy.

Use a relay instead to short the inputs. Much easier to keep that high CMRR,
as a relay is a nearly perfect switch. They come pretty small these days as
well.

Dave
 
O

ozzy

Jan 1, 1970
0
Dave said:
Use a relay instead to short the inputs. Much easier to keep that high CMRR,
as a relay is a nearly perfect switch. They come pretty small these days as
well.

Dave


Dave,

Thanks for the suggestion. My circuit will operate off a 9V battery
(+/- 4.5V rail to rail), so i'm really looking at something that could
switch with at least 3V - I'm sure a relay will need more than this.
That's why I was thinking back to the Comparator, as it will provide a
low output impedance - as if the connection was directly to ground.

The CMRR won't be important here (contradicting my previous statement!)
as no signal is passing through - strictly 0V DC needed. By the way my
idea was not to have this continually nulled - mimicking a chopper amp.
But to have this done only on Power ON.
A time out after 10-15 minutes will ensure that the drift of Vos with
time and temp will not be significant.


regards,



Osman
 
J

John Larkin

Jan 1, 1970
0
Hello!

I would like to address a problem I'm mulling over at the moment. I
have a micro-controller based circuit with a front end consisting of an
Instrumentation amplifier. I thought of applying zero volts to the
inputs (via pins from PIC) as a way of finding the sum of offset
voltages - via A/D converter. My worry is that this could affect the
high CMRR (120dB) needed at the input - can this be quantified?

Another alternative is to use Comparators of the PIC, but I'm already
using one of them. This circuit is battery operated (9V split supply)
and would therefore have to perform (lowest point) at 3V - ruling out a
CMOS switch.

These are just a few thoughts whizzing around me at the moment. Any
help on which way to go would be helpful.


Regards,


Ozzy.

There are lots of cmos analog mux chips that will work at very low
supply voltages.

John
 
J

Joerg

Jan 1, 1970
0
John said:
There are lots of cmos analog mux chips that will work at very low
supply voltages.

For something like a simple shorting of the input a 2N7000 or BSS123
should be fine. If it absolutely has to get to lower Z at 3V then IRF
offers some nice FETs. In case capacitance is an issue there are RF
dual-gate MOSFETs.
 
O

ozzy

Jan 1, 1970
0
Joerg said:
For something like a simple shorting of the input a 2N7000 or BSS123
should be fine. If it absolutely has to get to lower Z at 3V then IRF
offers some nice FETs. In case capacitance is an issue there are RF
dual-gate MOSFETs.


Joerg,

I was'nt aware that MOSFETS could switch so well at 2.7V - RDS(on) of
30mOhms!
Is this partly due to the increase in surface area provided by the HEX
structure (similar in a way to U-MOS by ZETEX, although i'm not sure if
they ever took off - huge problems in gate oxide integrity delayed
project by 2 years).


regards,


Ozzy
 
J

John Larkin

Jan 1, 1970
0
Joerg,

I was'nt aware that MOSFETS could switch so well at 2.7V - RDS(on) of
30mOhms!
Is this partly due to the increase in surface area provided by the HEX
structure (similar in a way to U-MOS by ZETEX, although i'm not sure if
they ever took off - huge problems in gate oxide integrity delayed
project by 2 years).


regards,


Ozzy

Don't short the signal, just disconnect it and short the inputs of the
signal-conditioning front-end. That doesn't require a very low on
resistance. So you need a spdt or maybe a dpdt analog switch.

John
 
J

Joerg

Jan 1, 1970
0
It's market driven. Stuff has to work with newer 3.3V or lower voltage
node logic. Then there are all the gadgets that run off 4V LiIon and
thus need a low Rdson switcher FET. Not that mankind really needs all
these gadgets but...

AFAIR the lowest Vgs FET I have seen has a guaranteed Rsdon limit at
1.8V Vgs. Wasn't even terribly expensive.

Don't short the signal, just disconnect it and short the inputs of the
signal-conditioning front-end. That doesn't require a very low on
resistance. So you need a spdt or maybe a dpdt analog switch.

I usually do short it but with a resistor up front :)

If Ozzy decides on DPDT switches and uses the typical CD405x or 74HC405x
then he should take extra care of all the noise paths. It can't hurt to
RC filter the VCC, plus some filtering on the control lines maybe. Don't
know about leakage of these devices.

Of course, the best of all switches is the SD5400 but these have become
rare and expensive. Must have fallen from grace.
 
J

John Larkin

Jan 1, 1970
0
It's market driven. Stuff has to work with newer 3.3V or lower voltage
node logic. Then there are all the gadgets that run off 4V LiIon and
thus need a low Rdson switcher FET. Not that mankind really needs all
these gadgets but...

AFAIR the lowest Vgs FET I have seen has a guaranteed Rsdon limit at
1.8V Vgs. Wasn't even terribly expensive.



I usually do short it but with a resistor up front :)

If Ozzy decides on DPDT switches and uses the typical CD405x or 74HC405x
then he should take extra care of all the noise paths. It can't hurt to
RC filter the VCC, plus some filtering on the control lines maybe. Don't
know about leakage of these devices.

Of course, the best of all switches is the SD5400 but these have become
rare and expensive. Must have fallen from grace.


These sorts of things are interesting:

http://www.fairchildsemi.com/pf/NC/NC7WB66.html

Lots of people make gadgets like these... TI, Fairchild, Onsemi. Some
are sold as logic bus switches, but they still make nice analog
switches, if you're not too afraid of charge injection.

John
 
J

Joerg

Jan 1, 1970
0
John said:
These sorts of things are interesting:

http://www.fairchildsemi.com/pf/NC/NC7WB66.html

Lots of people make gadgets like these... TI, Fairchild, Onsemi. Some
are sold as logic bus switches, but they still make nice analog
switches, if you're not too afraid of charge injection.

These are indeed great. It always feels good for an analog guy to hitch
a bargain ride on a digital chip. Almost like the advent of the
unbuffered CMOS inverters a few decades ago. Initially I had shunned bus
switches because of their high prices but that has come down nicely.
 
Top