64 analog inputs to arduino. Mayhew ADC shield.

Other Hardware Sensors
1

Hi everyone,

I am currently starting a project that involves analog to digital conversion. The main concerns are:

  1. From 8 -minimum- to 64 -maximum- channels, both unipolar (64 max.) and differential (32 max.). The analog inputs range -10/0/+10 Volts.

  2. Resolution: 12 bits; 14 bits are desirable.

  3. Maximum conversion time 1 ms. It refers to the time the whole 64 channels being converted into the ADC themselves; the subsequent reading into the microcontroller is not so critical.

  4. The planned microcontroller to use is a DUE. Their AD inputs are discarded for they are not differential and are 0 - 3.3 Volts ranged. Therefore, the use of external ADC(s) is foreseen.

  5. DUE (or finally selected uC) will connect to a Raspy. Due will make some arithmetics on the converted (analog to digital) values that will be (the calculated values) subsequently sent to the Raspy (that will hold the human interface).

  6. If possible, the project is to be developed with commercial components.

Has anybody used the Mayhew converter?:

Thanks in advance to all the ones that are going to help (and to many others).

Best regards.

2
  1. Maximum conversion time 1 ms.

That sounds like you need a system with an ADC per channel that are either sampling continuously
or you need a trick that can trigger all ADC's very fast. NB a simple loop over 64 devices could
easily take 100 uSec.

just thinking...

3

Could be of use?

Analog Devices Catalog -> Filter For ADCs

4

1milliSecond (mS)? 1000 samples/second?
I think you'll need chips like these

for 12 bits, +/- 10V input, differential input
You'll also need +/-12V power supply for the analog part.
These are parallel output devices, so plan a lot of pins, or a couple shift registers to store the output and shift them in to the uC.

5

Hi,

Thanks Rob, Johnny010 and CrossRoads.

Excuse me for failing in inserting the link to the Mayhew ADC; I'll try again (two ways):

"http://mayhewlabs.com/products/extended-adc-shield2"

(I'm afraid I keep on failing inserting the direct link :confused: )

As you can see it uses the Linear Technology LTC185x chip and announces "Up to 100 ksps sampling rate". In addition it works at (only) 5 Volts and connects directly to the +-10 sensors and ¡is arduino R3 compatible!.

In spanish we would say that it is "una perita en dulce" ("a sweet small peer", approximately).

What do you think?

Regards

6

Looks good. You are going to mount a bunch of them side by side, connect the control lines in parallel, and have the Uno talk to one at a time with a unique control line for each?
"RD (Pin 27): Read Input. This active low signal enables the digital output pin SDO."

7

you need to think how to power all of them too.

8

Hi,
Thanks again.
I am developing the project with a colleague (that is the one that manages it). I have posed him my doubts about the mechanical "stability" of the assembly. I've got a couple of ideas on the concern, but I find it a bit clumsy (i like this word: "clumsy". It appeared in my first english course: "Harry: this inflatable umbrella is clumsy and ridiculous").
Well, back in the subject: it looks that you can direct each card independently to read them. The trigger would be common; this is not a problem as this is just what we need: to read the 64 sensors in a ms (the project is aimed to measure a mechanical part and accept/discard it, apart from storing the readouts, some calculations and so on. A similar system -based in a PC and a suitable ADC card and . . . . believe me or not . . . . ¡in FreeDOS!- is currently working in many factories. The plan is to substitute it).
Regards

P.S.: the power supply concern is still under study.