Author Topic: How does digital to analog conversion in a DC standard work? (Knick)  (Read 2099 times)

0 Members and 1 Guest are viewing this topic.

Offline e61_philTopic starter

  • Frequent Contributor
  • **
  • Posts: 962
  • Country: de
Hello,

at work we have a 5.5 digit Knick DC calibrator. I wonder how the digital to analog conversion works. I can't find any DAC ICs. Is there a method like the dual slope method (A->D) in DMMs used in such devices which build a discrete digital to analog converter with great resolution?

Philipp

Edit: Sorry upside-down
« Last Edit: November 04, 2013, 07:42:11 pm by e61_phil »
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16356
  • Country: za
Re: How does digital to analog conversion in a DC standard work? (Knick)
« Reply #1 on: November 05, 2013, 04:48:29 am »
Probably a lot of R-2R BCD resistor networks on ceramic substrates that have been laser trimmed, along with a summing network to sum the discrete digits. You will likely see a lot of ceramic blocks with them inside under a shielding cover.
 

Offline e61_philTopic starter

  • Frequent Contributor
  • **
  • Posts: 962
  • Country: de
Re: How does digital to analog conversion in a DC standard work? (Knick)
« Reply #2 on: November 05, 2013, 06:42:24 pm »
No, there aren't any ceramic resistor networks as far as I can see.

I can't believe a resistor network will work with 20bit resolution (5.5 digit = 1e6 counts ~ 20bit). The MSB accuracy have to be better than 1/(2^20)?
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16356
  • Country: za
Re: How does digital to analog conversion in a DC standard work? (Knick)
« Reply #3 on: November 05, 2013, 07:07:20 pm »
I used to use some BCD ceramic networks that were made specially for the use. Came in a square ceramic block with 20 odd leads on the sides to give a voltage from 0 to Vref ( which was a 5 VAC 400Hz sine wave from the main AC reference generator) that was used to convert the input from a 5 digit BCD switch which had as the MSD a polarity switch. It was pretty stable, the LSD never drifted more than a count over the full operating range ( or from a bath in freezer followed by a cooking from a hot air gun) during testing. There was one trimming resistor for the unit, used to adjust the offset to zero. Typically I used a 1M resistor there to ground, as you needed a resistor in the schematic, but had a choice of connecting the chosen value to ref or GND.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf