Author Topic: Analog front-end for a multi range ADC, help needed!  (Read 1420 times)

0 Members and 1 Guest are viewing this topic.

Offline IU8BBMTopic starter

  • Newbie
  • Posts: 6
  • Country: it
  • HW Designer in Naples (ITA)
    • Personal webserver
Analog front-end for a multi range ADC, help needed!
« on: December 16, 2024, 03:21:47 pm »
Hello folks,
I've been asked for a design of an analog acquisition module which have to replace what actually it's being done with a DSO: live capturing 4 signals from near field E\H probes and outputting the corresponding RMS value in real time (no triggering is needed: every 100ms output the rms value of the previous 100ms).

The customer asked for these minimum specifications, in square parenthesis I'm also adding the optimal requirements (can be waived):
- 4 differential channels
- 12 bits [16bits]
- 1 MSa/s [10 MSa/s]
- Range (every channel is indipendent) ±0.1/0.5/1/2V [.../5V]

Please let's just ignore for the moment the purpose and how the values have to be manipulated for the RMS to be calculated which is under my expertise.  Let's just discuss about the analog frontend and the ADC(s) that have to be connected to the MCU (a STM32H7). At the moment I'm struggling with the multi range capability.

The customer shared with us an interesting solution, the NewAE ChipWhisperer Husky which, as open hardware project, share its schematics which report a VGA+ADC combination syncronized with a MCU+FPGA via a PLL. Unfortunately it does'nt match the minimum requirements.
I was planning on 4x(VGA+ADC) connected straight to the MCU via parallel output which would be the easiest solution but my search engine knowledge also gave no clue on app notes\example projects to take as free inspiration.

Have you had any experience doing this? Could you suggest a similar solution to follow?
Help me EEVbloggers, you're my only hope.
Luca
 

Offline jwet

  • Super Contributor
  • ***
  • Posts: 1168
  • Country: us
Re: Analog front-end for a multi range ADC, help needed!
« Reply #1 on: December 16, 2024, 03:59:03 pm »
I think if I were given this task, I take a good look at the Digilent Analog discovery boards.  The first generation was based on ADI parts but the latest generation is TI.  They don't have much for a front end but might be a good start.

Maxim and ADI make multiranging front end ADC's.  They're mainly for PLC and industrial automative with flexibile inputs.
 
The following users thanked this post: voltsandjolts

Online David Hess

  • Super Contributor
  • ***
  • Posts: 19228
  • Country: us
  • DavidH
Re: Analog front-end for a multi range ADC, help needed!
« Reply #2 on: December 16, 2024, 05:57:17 pm »
- 4 differential channels
- 12 bits [16bits]
- 1 MSa/s [10 MSa/s]
- Range (every channel is indipendent) ±0.1/0.5/1/2V [.../5V]

You say that it is replacing a DSO.  Does it require 1 megohm differential inputs?  What kind of bandwidth is required?

What kind of input voltage range is required?  This is especially important for differential inputs because it has to account for the input common mode voltage which gets rejected.
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 17265
  • Country: de
Re: Analog front-end for a multi range ADC, help needed!
« Reply #3 on: December 16, 2024, 06:08:35 pm »
A sampling ADC would want a AA filter and in most cases some oversampling to reduce the demand on the filter. The bandwidth may be more relevant then the actual sampling rate. With a target of 1 MSPS for a smapling ADC one may be in a range where a SD ADC is also possible with a slightly lower sampling rate and still getting comparable bandwidth.
For just the RMS result one could even use subsampling in a limited range, so the filter needs may be different from a classic AA filter.
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 19228
  • Country: us
  • DavidH
Re: Analog front-end for a multi range ADC, help needed!
« Reply #4 on: December 16, 2024, 07:25:34 pm »
A sampling ADC would want a AA filter and in most cases some oversampling to reduce the demand on the filter. The bandwidth may be more relevant then the actual sampling rate. With a target of 1 MSPS for a smapling ADC one may be in a range where a SD ADC is also possible with a slightly lower sampling rate and still getting comparable bandwidth.
For just the RMS result one could even use subsampling in a limited range, so the filter needs may be different from a classic AA filter.

A filter is only required to tailor the frequency range over which the RMS measurement is made.  Aliasing is irrelevant in this application.  Bandwidth will depend on the analog signal chain, and the sampling aperture time which depends on the ADC.  Many suitable ADCs support bandwidths which are many times their Nyquist frequency.
 

Offline xvr

  • Frequent Contributor
  • **
  • Posts: 926
  • Country: ie
    • LinkedIn
Re: Analog front-end for a multi range ADC, help needed!
« Reply #5 on: December 17, 2024, 04:52:26 pm »
Quote
I was planning on 4x(VGA+ADC) connected straight to the MCU via parallel output
This MCU has its own ADC (with high enough speed and resolution). All you need is external PGA (4 pcs). There are a lot of different PGA and VGA from Analog Devices (very high quality, high speed and high cost  :-// ), also TI makes a lot of different PGA.

https://www.digikey.ie/en/products/filter/instrumentation-op-amps-buffer-amps/687?s=N4IgjCBcpgbB0QGMoDMCGAbAzgUwDQgD2UA2iACwAcFATLAJwiEVgMDMADAKzMjsNOcTiAC6hAA4AXKCADKUgE4BLAHYBzEAF8dQA
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf

 

-->