Instrumentation amplifier for high accuracy load cell

Thread Starter

varuntejavarma

Joined Apr 16, 2014
36
Hello,
I have to design an Instrumentation amplifier for a Load cell with a resolution of 0.1 gm and I am going to use a 24 bit ADC at the o/p end of IN amp. please tell me if its better to build an Instrumentation amplifier or buy one for this application and why.
If i can build then,
Is it better to divide the designing in two parts: difference amplifier( 2nd part) and buffer stage(1st part). But i do not know on which factors should i select the resistor values and the right opamp for each part and how to check the output of each part.
It would be helpfull if someone could give me a step by step process on where to start and what to know for designing an Instrumentation amplifier. I am trying the way of learning by doing. so, please help me
Regards
 

ericgibbs

Joined Jan 29, 2010
18,848
hi,
For 100mGm resolution and 24bit ADC I would not consider building a Instrumentation Amplifier, I would buy a suitable IA.

The reason is the high tolerances required in matching the resistor values for a self designed IA.

I would not use a Difference OPA configuration for this application.

E
 

crutschow

Joined Mar 14, 2008
34,430
It's generally very difficult and expensive to build an instrumentation amplifier as good as the better integrated circuit amps. There's no good reason to build your own.
 

joeyd999

Joined Jun 6, 2011
5,283
You don't need an IA. Chose a 24 bit a/d with a digital PGA, and run the load cell directly into it. Here is one I like:

TI ADS1242

There are many others that would be acceptable.
 

sanuzr

Joined Mar 29, 2012
7
I agree with eric and crutsshow. It is very difficult to design an IA of your own having such matched performance. You better use an IC like MAX4208. I am using it for similar king of application and getting a very good performance.
 

joeyd999

Joined Jun 6, 2011
5,283
I agree with eric and crutsshow...
And not me????!

Sigma Delta converters have really progressed in the last 10 years or so. They have made many front-end analog designs redundant and/or obsolete.

It's best these days to convert to digital as close to the transducer as possible. Analog drifts. Digital doesn't.
 

crutschow

Joined Mar 14, 2008
34,430
And not me????!

Sigma Delta converters have really progressed in the last 10 years or so. They have made many front-end analog designs redundant and/or obsolete.

It's best these days to convert to digital as close to the transducer as possible. Analog drifts. Digital doesn't.
I agree as long as the A/D has input signal conditioning including needed gain and differential input if required.
 

Thread Starter

varuntejavarma

Joined Apr 16, 2014
36
Thanks for the replies. I am using an HYCON 3102 A/D conv. It is a 24 bit ADC with a built-in PGA with 21 bit ENOB. Is it sufficient or should i also buy an IA IC.
 

crutschow

Joined Mar 14, 2008
34,430
Thanks for the replies. I am using an HYCON 3102 A/D conv. It is a 24 bit ADC with a built-in PGA with 21 bit ENOB. Is it sufficient or should i also buy an IA IC.
It probably is adequate, depending upon the voltage characteristics of the load cell signal. The PGA allows going down to a full-scale input as low as ±10mV. Figure 4 in the data sheet shows it being used with a bridge sensor.
 

Thread Starter

varuntejavarma

Joined Apr 16, 2014
36
I have an IN AMP ic from analog devices. I have tested it with a load cell of 1mv/v sensitivity with max load 500 gms. and with supply voltage of 5 v, gain of 1000 and 500gms(full load), it gives an output of 4.47 v ( max voltage).

Shouldnt i be getting 5v. is there something i should do to get 5 v or is it the maximum value.

http://datasheet.octopart.com/AMP04FP-Analog-Devices-datasheet-4256.pdf

I have just connected a res(100ohms) across 1 and 8 for gain and a capacitor b/w supply and ground. I have connected ref(pin 5) to ground. Is there any connections that i am missing
 
Top