Author Topic: the Keysight U1272 ADC  (Read 798 times)

0 Members and 1 Guest are viewing this topic.

Offline niner_007

  • Regular Contributor
  • *
  • Posts: 197
  • Country: us
the Keysight U1272 ADC
« on: June 11, 2018, 08:22:49 am »
The Keysight U1272 multimeter is using an 24-bit ADC, the TI ADS1242, it seems like with this ADC, given its performance, they could've implemented a 5.5 digits meter (assuming probably better references, amplifiers etc than the U1272 is currently using). Why did they use such a high resolution ADC? They could've used an 18-bit ADC instead.
« Last Edit: June 11, 2018, 08:29:00 am by niner_007 »
 

Online Mr. Scram

  • Super Contributor
  • ***
  • Posts: 9664
  • Country: 00
  • Display aficionado
Re: the Keysight U1272 ADC
« Reply #1 on: June 11, 2018, 08:42:54 am »
The Keysight U1272 multimeter is using an 24-bit ADC, the TI ADS1242, it seems like with this ADC, given its performance, they could've implemented a 5.5 digits meter (assuming probably better references, amplifiers etc than the U1272 is currently using). Why did they use such a high resolution ADC? They could've used an 18-bit ADC instead.
Devices like the 34401A also make measurements are higher resolutions than needed. I think it's done to ultimately eleminate noise.
 

Offline wraper

  • Supporter
  • ****
  • Posts: 12199
  • Country: lv
Re: the Keysight U1272 ADC
« Reply #2 on: June 11, 2018, 10:39:48 am »
They could've used an 18-bit ADC instead.
Quote
21 BITS EFFECTIVE RESOLUTION
(PGA = 1), 19 BITS (PGA = 128)
It's not like this is too expensive component as well. It's not necessarily they would save money by using lower spec ADC. 18 bit ADC with effective 18 bit resolution likely won't be any cheaper. And some excessive resolution is useful for calibration. You could also skimp somewhere else in the circuit by not using full input voltage range of ADC because you have excessive resolution.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 11922
  • Country: us
  • DavidH
Re: the Keysight U1272 ADC
« Reply #3 on: June 11, 2018, 11:07:59 am »
Any extra digits would not be significant.  Even if a better reference and divider were used, the variable input resistance (1) would limit accuracy.  0.05% basic DC accuracy is 1 part in 2000 so even with "only* 30,000 counts, that is 10 times greater resolution than the accuracy will support which is typical for 4.5 digit meters and many of the other functions are 10 times worse than that.

(1) The specifications say "10 MΩ (nominal)" which is marketing speak for "not 10 MΩ and it varies all over the place".
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf