2
\$\begingroup\$

I have connected the outputs of two TSOP4838 IR receiver modules using diodes to prevent reverse current. The idea is to mount them on a moving model vehicle at 180 degree orientation to one another so that the field of reception is closer to 360 degrees. If this is a poor idea and there is a better solution I would be interested to know.

The schematic is shown.

schematic

simulate this circuit – Schematic created using CircuitLab

The output at both points A, shown on the scope capture is very clean as expected:

Output at points A

However at point B the output signal is "dirtied" somewhat:

enter image description here

Can anyone explain to me why this is, and if there is a method of "cleaning-up" the output signal so that it closer matches the output at points A?

Note: I did not necessarily press the same button on the IR transmitter to generate the signals at points A and B - it's the "shape" of the output waveform that is of interest.

Datasheet: Vishay TSOP4838

\$\endgroup\$
7
  • \$\begingroup\$ That output should not even be possible if only one receiver pulls low and other keeps the output floating high. But it is possible if both outputs pull low. What is on node B or is it just floating with the scope probe? Which scope that is? \$\endgroup\$
    – Justme
    Commented Jan 15 at 15:29
  • \$\begingroup\$ Node B is floating with the scope probe. I am using PicoScope 2000. \$\endgroup\$
    – RickyBoy
    Commented Jan 15 at 15:30
  • \$\begingroup\$ What pulls node B low in the first place? \$\endgroup\$
    – winny
    Commented Jan 15 at 15:32
  • \$\begingroup\$ I am just pressing any button on a TV remote to generate a 38kHz input signal to the circuit in the schematic \$\endgroup\$
    – RickyBoy
    Commented Jan 15 at 15:33
  • \$\begingroup\$ How do you expect it to go low? \$\endgroup\$
    – winny
    Commented Jan 15 at 15:57

3 Answers 3

3
\$\begingroup\$

There are two problems with the circuit you presented:

  1. The slow fall of the signal is as described by other answers. The output is floating when both sensor outputs are low. You need a pull-down resistor. Something like 10 kohms is suggested. (As @RickyBoy noted in the comments the internal pull-up of 33k does not allow that so a higher value such as 100k is needed)

  2. The diodes are the wrong way round. As your traces show the no-signal value is a logic high, it goes low when there is a signal. If one sensor receives a signal and the other does not the logic high from the no-signal sensor will win and you will get just a sustained logic high at the output.

You need to reverse the diodes and put a pull-up resistor at the output for it to work as you require. Again 10 kohms is a suitable value.

As the internal circuit of the sensor is a transistor pulling down with a passive pull-up the same effect can also be achieved without using diodes by just connecting together the outputs of the two sensors. In that arrangement an external pull-up resistor would also not be needed as the internal ones should suffice.

Block diagram from datasheet Image from Vishay Semiconductor

TSOP4838 datasheet

\$\endgroup\$
4
  • \$\begingroup\$ This works well but not with 10k ohms pull-down resistor. TSOP4838 has 33 kohms internal pull-up, so I find using 100k ohms external pull-down (but leave 10 kohms pull-up at output) is a working solution. \$\endgroup\$
    – RickyBoy
    Commented Jan 15 at 22:14
  • \$\begingroup\$ @RickyBoy - Interesting. I looked at the datasheet but missed that. \$\endgroup\$ Commented Jan 15 at 22:23
  • \$\begingroup\$ Super answer, thanks \$\endgroup\$
    – RickyBoy
    Commented Jan 15 at 22:34
  • \$\begingroup\$ The datasheet you have is for slightly different component. I've put a link to the datasheet for the precise device I am using in the question. That said, the datasheet I was using was old and said 33 kohms, the up-to-date one says 30 kohms as in yours. Not that the 3 kohms difference changes much, but no harm in accuracy. Will try your revised circuit and report. \$\endgroup\$
    – RickyBoy
    Commented Jan 16 at 9:26
3
\$\begingroup\$

"Node B is floating with the scope probe. I am using PicoScope 2000."

That is expected with your setup, you said yourself the node is floating.

When IR receivers see a burst, they drop their output pins low. There is nothing else than the weak pull-down of the scope to discharge any capacitance from voltage so it will drop slowly.

Depending on which probe you have and how you use it, it will have 1M to 10M of pull-down resistance, and few to few tens of picofarads of capacitance, so you see the discharging RC waveform.

\$\endgroup\$
1
\$\begingroup\$

Following the answer of @Kevin White, I simulated the circuit as shown:

schematic

simulate this circuit – Schematic created using CircuitLab

The dotted rectangles simulate the TSOP4838 IR sensors. The pin numbers as they would appear on the packages are shown. Points A and B are used to control the outputs at pin 1 of each sensor. If A or B is connected to Vs then this simulates the presence of an IR burst of 38kHz and the corresponding output(s) are pulled low. If A or B is connected to 0V then it simulates the absence of IR and the pullup resistors within the sensors pulls the corresponding output(s) to logic high. The multiplexed voltage at OUT behaves as required.

The 600μs on/600μs off test pattern signal output:

enter image description here

In the present application OUT is connected to the input pin of a PIC microcontroller. The particular MCU has the option of a weak pull-up internally. However, this is not strictly necessary because as @Kevin White says, the internal pull-ups of the TSOP4838 devices are sufficient.

\$\endgroup\$
1
  • \$\begingroup\$ This site has built-in CircuitLab editor and simulator, so it's best to use that. Then the circuits can be copy-pasted, re-simulated by others, etc. \$\endgroup\$ Commented Jan 21 at 22:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.