Taking a signal from cutting a notch out of an incoming DC voltage level

Thread Starter

rfagen

Joined Jan 19, 2013
6
I'm pondering a project where:
  • there are a group of eight photodiodes, each tuned to emit one volt when "illuminated"[1]
  • the output of these photodiodes is aggregated together to produce an output from one to eight volts

What I'd like to do is:
  • send an output to a latch to set it when the voltage is greater than 2.7 volts and less than 4.3 volts
  • send an output to reset that latch when the voltage is outside that range

I had a vague idea of having the combined voltage of the photodiodes get fed to a zener diode tuned to break down above 2.7 volts, and then have another zener diode that breaks down above 4.3 volts. Then use the voltage from the three points (before, middle, after) and run those through AND and NOT to produce a signal on either the "set" or "reset" for the latch.

Apologies in advance for any vagueness, incorrect terminology, or outright misunderstanding of why electricity doesn't work like water in a pipe :)

I guess it wouldn't hurt to mention that what I'm trying to build is a design for a RAICL[2] Block

Thanks in advance!

[1] where "illuminated" is seeing the expected amount of light, and the diode will be in an environment where it can only be exposed to a specific source
[2] Redundant Array of Inexpensive Conway's Life Blocks -- that is, blocks that can fit jigsaw style into an arbitrary number of tiled blocks where each has an LED in the center, and based on the eight photodiodes' input from their neighbors, will follow the rules of Conway's Life to either illuminate or die on a "next" cycle.
 

MrChips

Joined Oct 2, 2009
31,068
Look for a different approach, for example, digital instead of analog.
Treat each photodiode as one bit. Use digital circuitry to count the number of 1s.
Or you can sum the 1V analog signals in a summing amplifier.
 

Papabravo

Joined Feb 24, 2006
21,304
What does "cutting a notch out of an incoming DC voltage level" even mean? I guess translation software isn't all it might promise.
 

Thread Starter

rfagen

Joined Jan 19, 2013
6
Yea, I'm curious how the OP intends to aggregate the photodiodes to deliver the desired outputs considering how they operate.
It might only work this way in my imagination, but something like this (where PD is the photodiode and R is a resistor that causes 1v to come out of the PD at the desired level of illumination):
Code:
light -> 1v
---PD-----R-+
            |
light -> 1v |
---PD-----R-+
            |
dark  -> 0v |
---PD-----R-+
            |
            +---> 2v relative to ground?
 

Thread Starter

rfagen

Joined Jan 19, 2013
6
Look for a different approach, for example, digital instead of analog.
Treat each photodiode as one bit. Use digital circuitry to count the number of 1s.
Or you can sum the 1V analog signals in a summing amplifier.
I was wanting to do this analog because I didn't want to use a whole Arduino (or whatever) for each tile (imagining it would be too expensive). However, looking at all the ICs needed for a window comparator, maybe not? I guess I need to actually look up the prices for all these parts.

The "summing amplifier" also looks promising as the way to aggregate all the PD outputs -- https://www.electronics-tutorials.ws/opamp/opamp_4.html

Thanks!
 

ElectricSpidey

Joined Dec 2, 2017
2,881
That link in post #9 is not a good example of a window comparator as it's using two Op-Amps instead of open collector "comparators" that then require the diodes.

(no diodes needed with open collector comparators)

The inverter may or may not be needed depending on the logic you need following the WC. (also a small signal transistor can do the inversion)

Dual comparators can be gotten two in an 8-pin package. (LM393 for example)

Window comparator circuit (Rev. A) (ti.com)
 
Last edited:

sparky 1

Joined Nov 3, 2018
783
We sometimes look at a pulse design by its field shape, the time it takes to invert, its ability to do work which differs with application.
For example, an arduino project needs to scale the signal to transmit over a distance wireless might require good wave shape or field.

Years ago, some people were trying to see how fast they could spin a spherical magnet.
If you consider all the different pulsing methods, the field itself and the location where the magnet was placed are some of the factors.
The magnetic field is included in evaluating a notch in practical application. For many signals optical sense and hall effect can have different properties. The design of a signal with notch that reduces switching noise to get a nice, filtered DC has been approached by taking advantage of the sensor's isolation and pulse forming. I have seen some good arduino projects sit because the signal did not scale up as easy as the designer thought it would.

Personally, the digital hall effect isolated filtered and inverted pulse would be my first choice.
 
Last edited:
Top