Author Topic: Looking for a very accurate voltage controlled PWM duty cycle chip or circuit  (Read 1539 times)

0 Members and 1 Guest are viewing this topic.

Offline AQUAMANTopic starter

  • Regular Contributor
  • *
  • Posts: 101
  • Country: gb
I have data from an NTC and RTD that I need to pass over an isolation.

What I am thinking of doing a normal voltage divider for both and then turning this voltage into a duty cycle or frequency signal to pass through a digital optocoupler. Probably looking at about 10kHz

In designing this circuit money is my least concern, but footprint (both height and area) and accuracy are my main requirements. I need it to be very accurate and and I need the components height less than 2.5mm. So for resistors, caps, NTC and RTD I'm using 0603 or 1005 SMD.

I have been looking at LTC6992 (https://www.analog.com/media/en/technical-documentation/data-sheets/LTC6992-1-6992-2-6992-3-6992-4.pdf)
I like this chip, it's easy to use, small footprint, but I worry it will not be accurate enough. It says plus minus 4.5% accuracy in the datasheet. It is also only 0.1-0.9V control voltage so I think it makes my readings from NTC and RTD more susceptible to noise because I have to lower the sensing current to get it to this level

Are there any other ICs like this, or would I be better off trying to find op amps in small footprints - or would to do it all on a small micro controller like ATTINY85 by better?
« Last Edit: January 31, 2020, 09:59:50 pm by AQUAMAN »
 

Offline jmelson

  • Super Contributor
  • ***
  • Posts: 2852
  • Country: us
Would a voltage to frequency converter be better?  There can be some distortion of pulse widths when passed through optical isolators.

Jon
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
 
The following users thanked this post: I wanted a rude username

Offline iMo

  • Super Contributor
  • ***
  • Posts: 5572
  • Country: va
« Last Edit: January 31, 2020, 10:15:57 pm by imo »
Readers discretion is advised..
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Besides mentioned isolated ADC solutions other way would be to use generic (i2c or SPI) SD ADC with bus isolator. Also there is "programmable SD ADC", msp430f2013. Any (discrete) voltage to PWM or frequency converter will be much less accurate than ADC isolation approach.
 

Offline AQUAMANTopic starter

  • Regular Contributor
  • *
  • Posts: 101
  • Country: gb
https://www.analog.com/en/parametricsearch/11101
http://www.ti.com/isolation/isolated-adcs/products.html?pqs=paqs&familyid=2021#
Hi
Thanks I'll have a look, but most of these are low input range of less than 1V

I also have to send the reading over 1 metre since to the control unit
 

Offline AQUAMANTopic starter

  • Regular Contributor
  • *
  • Posts: 101
  • Country: gb
I've been looking into these also
I've been simulating this relaxation oscillator in this PDF http://ww1.microchip.com/downloads/en/appnotes/00895a.pdf
 

Offline pwlps

  • Frequent Contributor
  • **
  • Posts: 372
  • Country: fr
Thanks I'll have a look, but most of these are low input range of less than 1V

I also have to send the reading over 1 metre since to the control unit

You also have purely analog isolation buffers like the ISO224 (+/-12V input) or for higher precision the ISO124 (+/-10V).   This might save some space since no adc/transceiver is needed on the isolated side.   This will be less precise than the isolated adc solution though.
 

Offline Mechatrommer

  • Super Contributor
  • ***
  • Posts: 11714
  • Country: my
  • reassessing directives...
but footprint (both height and area) and accuracy are my main requirements..
or would to do it all on a small micro controller like ATTINY85 by better?
what accuracy are we talking about? how much delay you can accept? attiny10 is much cheaper than LTC6992, can take 5V directly, 16bit PWM and 8bit ADC, 1 bit error is 0.4% error, averaging can give better and stable output but more delay, you choice... pro: you can do many more than your V->PWM convertor, incl V->FREQ... con: you need to program it.
Nature: Evolution and the Illusion of Randomness (Stephen L. Talbott): Its now indisputable that... organisms “expertise” contextualizes its genome, and its nonsense to say that these powers are under the control of the genome being contextualized - Barbara McClintock
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf