UCD3138064A: ERROR of ADC

Hi Rana,

There are two types of ADCs in the UCD3138 devices.
One is the 12-bit multi-channel general purpose ADC that is used for mainly monitoring.
Second the low-res but fast error ADC that is mainly used to close the compensation loop.

Are you reffering to the general purpose ADC or to the EADC(Error ADC)?
The real EADC resolution is always 6 bits.
But depending on the AFE setting these 6 bits are shifted inside the 9 bits of the error result output.

Please refer to section 3.1 "Error ADC and FrontEnd Gain" in the Technical Reference Manual:

www.ti.com/.../sniu028a.pdf

Regards,

Yitzhak Bolurian

Hello,

Thank you for your Replay.

What I would like to know how to calculate 10-bit result from the (EAPx – EANx) voltage difference, as well as how will all the errors affect this measurement. I attempted something based on the Front End block

diagram, but I’d like to get the correct formula from your side.

Best Regards,
Rana

You don't need to calculate it.  You can look at the ABS_VALUE bitfield in the EADCVALUE register.

Also in the EADCVALUE there are two bits - EADC_SAT_HIGH and EADC_SAT_LOW.  If either of them are set, the ABS_VALUE will not be accurate.

The EADC is very fast, but its dynamic range is much less than full range.  So if the EADCDAC isn't set relatively close to the actual voltage input, the EADC will be saturated. 

Hi Rana,

I have tried to send you the formula that I found through the forum and it did not work.

Please send me your email, so I can send you the information that I have.