AMC3330-Q1: How to run Monte Carlo/Worst Case analysis with IC tolerance?

Sky Hsieh48

Prodigy 130 points

Part Number: AMC3330-Q1
Other Parts Discussed in Thread: TIDA-010232, AMC3330, TINA-TI, PSPICE-FOR-TI

Hi TI,

I would like to know design tolerance of TIDA-010232 reference design.

I only know how to run a worst case with tolerance of resistive bridge.

How can I get worst case result with tolerance of resistive bridge and AMC3330?

Our customer ask us to to do WCCA (worst case) report by simulation to make sure the reference design can meet their requirment.

over 1 year ago

0 Alexander Smith over 1 year ago

TI__Guru 50641 points

Hi Sky,

Are you looking to understand the worst-case error for the AMC3330?

If so, please see this excel calculator: https://www.ti.com/lit/zip/sbar013

0 Sky Hsieh48 over 1 year ago in reply to Alexander Smith

Prodigy 130 points

Hi Alexander,

1. How is "Typical Error RSS" calculated?

2. What does "2-point gain" mean?

3. I would like to know does SPICE model of AMC3330 use typical values for simulation?

4. If I want to change resistors with 0.5 or 0.1% tolerance, how to get Maximum Error by calculation or simulation?

0 kai klaas69 over 1 year ago in reply to Sky Hsieh48

Guru 140200 points

Hi Sky,

Sky Hsieh48 said:
4. If I want to change resistors with 0.5 or 0.1% tolerance, how to get Maximum Error by calculation or simulation?

In the free TINA-TI you can use the "Control Object" method:

sky_tolerances.TSC

The "Monte Carlo" method can only be used in the paid version of TINA.

Kai

0 Alexander Smith over 1 year ago in reply to kai klaas69

TI__Guru 50641 points

Hi Sky,

We specify both typical and maximum error values in our datasheets and "RSS" refers to "root sum of squares". This video has additional details: https://training.ti.com/ti-precision-labs-adcs-statistics-behind-error-analysis

This post explains error sources and calibration methods: https://e2e.ti.com/support/amplifiers-group/amplifiers/f/amplifiers-forum/937333/amc1302-q1-what-do-tceg-nonlinearity-and-tceg-nonlinearity-drift-mean/3463286#3463286

The model is specified to operate near typical values.

0 Sky Hsieh48 over 1 year ago in reply to kai klaas69

Prodigy 130 points

Hi Kai,

I would like to run worst case simulation with tolerance of resistors and AMC3330 by PSPICE-for-TI.

I think the result above only include tolerance of resistors, there is no tolerance when I use IOP1 (ideal OPA), is that correct?

Can I add tolerance of IOP1 (or AMC3330) in my simulation?

If possible, could you provide the simulation file of PSPICE-for-TI?

0 Sky Hsieh48 over 1 year ago in reply to Alexander Smith

Prodigy 130 points

Hi Alexander,

The video helps me a lot. Thank you.

According to the video, typical value is between positive or negative sigma.

What is the typical value (positive or negative) in SPICE model? Why?

0 kai klaas69 over 1 year ago in reply to Sky Hsieh48

Guru 140200 points

Hi Sky,

Sky Hsieh48 said:
I think the result above only include tolerance of resistors, there is no tolerance when I use IOP1 (ideal OPA), is that correct?

Yes.

Sky Hsieh48 said:
Can I add tolerance of IOP1 (or AMC3330) in my simulation?

I don't think so.

Kai

0 Alexander Smith over 1 year ago in reply to kai klaas69

TI__Guru 50641 points

Hi Sky,

Our simulation models are designed to behave consistently within datasheet parameters, not with random statistical variables according to datasheet max/mins.

I think calculation is the correct way to perform WCCA.

Amplifiers

Amplifiers forum

AMC3330-Q1: How to run Monte Carlo/Worst Case analysis with IC tolerance?