Detecting IR sensor at higher speed

Question

I made a circuit which selects one of 4 sensors and determines if a valid data stream is present on that sensor and counts the number of times data is valid. I narrowed my problems to hardware.

In the original circuit, I did not include the LED and 100 ohm resistor in series (see thick wire in schematic) because I thought I didn't need it. Then I started adding it temporarily and everything seems to run more smoothly with it in.

So at this point with the resistor and LED in, I can detect light status on any sensor of my choice at a rate of 1 every 2 ms. Ideally, I want to detect at a rate of 1 every 125 us.

I believe my problem exists because of some delay caused by the math resulting from the internal resistor on the micro's RXD pin plus the 1k resistor plus a capacitor somewhere in any of the IC's?

The reason I added the 1k resistor is because if by fat chance I ground the RXD pin via software and output W is high and I didn't have any resistor then I'd have a short circuit which I do not want.

But what baffles me is that I get success with the diode and LED circuit even though output W would return a high or low since the part is a 151 not a 251 or does that chip output still have a resistor in series?

My other thought is to replace the AT89C2051 with an AT89LP4051 but that's asking for more money and another programmer.

Anyways, other than my thoughts, Can I configure this setup hardware wise to allow a signal to pass through from the photo transistor to the micro via a multiplexer at a rate of 1 bit per 125us? (like change resistor values and/or add/remove resistors)? etc.

Currently I have the entire circuit minus the LED+resistor in series mounted on a PCB with tracks at least 12 mils wide and power tracks at least 40 mils wide.

The supply voltage averages at 4.8V.

UPDATE

I have tried all of the following without success:

1. Replaced the series LED+resistor combo in original circuit with just a 1K resistor and used 220 ohm for series resistor

2. Removed the series LED+resistor combo and instead connected the GPIO pin to ground via 22K resistor and again used the 220 ohm for series resistor

Now I haven't gone as low as 75 ohms for series resistance that someone pointed in their diagram but I'm still a little fuzzy. Maybe there's an entry in a datasheet I need.

Now I'm curious, considering this micro has (based on literature I read) GPIO pins in open drain format only, could I get away with removing the resistors and LED and replacing the series resistor with only one reverse-based diode (so anode connects to GPIO and cathode connects to multiplexer output)? If so, can I get away with a standard 1n4148/1n914 diode?

Answer 1

It's not entirely clear what you're asking, but I think you're trying to receive short pulses from the W output of IC2 into the P3.0-RXD pin of IC1, and you're discovering that 2ms pulses are detected at IC1, but 125uS pulses aren't.

the 1K resistor is far too large, especially since it's having to work against the LED and its 100 ohm current limiting resistor.

The simplest thing to do to fix your circuit is to connect the LED and the 100 ohm resistor between the Y (pin 5) output of IC2 and ground (noting the polarity will be different). Change the 100 ohm resistor for something bigger (for a red led you probably want something more like 330 ohms).

With that modification you won't be loading the output excessively.

Next reduce the 1K resistor. The ratings for the 74HC151 give maximum source or sink current as 25mA, and the AT89C2051P lists a maximum source or sink current of 25mA as well. With a 5V supply, 25mA gives you a limit of 200 ohms, so swap the 1K resistor for a 330 ohm one.

Answer 2

If anything the LED pullup ought be 1k with no series R to expect good margin for Vol.

If you need an LED indicator with 100 ohms, use another inverter like the 'HC14 so as not to apply this much pullup current.

The opto slow rise time will delay the output due to the 50% hysteresis by 4~5us due to 8pF=Cout * 500k

^{simulate this circuit – Schematic created using CircuitLab}

This optical design needs to use a path loss analysis with more precise Rx gain than a phototransistor such as a PD with an OpAmp for gain.

There are also far better solutions using Sharp/Vishay IRDA Tx Rx IC's. Optical gain is done by using narrow-angle beamwidth.

This is what I mean up to 115.2 kbit/s (SIR) at 1m or 8 kb/s @ 12m range @ 0 deg https://www.vishay.com/docs/81965/tfdu4301.pdf

There are better ways if you define range and beamwidth.