Automatic soap dispenser

[tonytnnelectrotech]

Hi,
I made the attached circuit to make an automatic soap dispenser ... it works well except that sometimes when the timed out ends and you take your hand from under the IR sensor the timed out is reactivated ...

some advice? thank you.

 

[Diver300]

You need a timer for the minimum off time, so that only if the PIR has been deactivated for say 2 seconds can the right hand LM555 be triggered.

I would use a processor.

 

[AnalogKid]

Is this a reflective system, where the presence of a hand reflects light from IE1 into PH1? If so, then the problem is a bounce in the PH1 output as the light level decreases relatively slowly when the hand moves away. First thing - delete D1. D1 is there to shorten the recovery time, but you need a longer dead band between triggers. Without D1, you have a dead band of approx 1 ms. Still way too short. Next, increase C5 to 1 uF and see if that filters out the false triggers out of PH1.

Thanks for the schematic, and extra thanks for the reference designators.

ak

 

[tonytnnelectrotech]

Is this a reflective system, where the presence of a hand reflects light from IE1 into PH1

yes I think the problem is this ...

what do you think to put too a 1µF electrolytic capacitor between pin3 of PH1 and gnd?

 

[Nigel Goodwin]

It's also a very poor design, as it transmits a continuous beam of 38KHz, the TSOP is designed for modulated pulses, and doesn't work properly or reliably with a continuous beam. You should also feed power to the TSOP via a resistor, with a decoupling electrolytic.

But as in post #2, use a processor - you could replace the entire circuit with an 8 pin PIC (12F1840 for example), an IR LED, a TSOP, a couple of capacitors, and a couple of resistors - and another resistor and LED if you want the LED on the output. With such acircuit it would be easy to cure all your issues, and add much more functionality and reliability.

 

[AnalogKid]

According to the datasheet, the TSOP output is open-collector. So ...

1. You can eliminate R2.

2. Adding a cap from pin 3 to GND is a better solution than increasing C5.

ak

 

[AnalogKid]

And, since no one has thought to ask - do you have any of the hardware, software, or programming experience it would take to convert this project to one based on a uC?

If yes *to all three*, then why did you decide to go with an all-discrete solution?

ak

 

[Nigel Goodwin]

According to the datasheet, the TSOP output is open-collector. So ...

1. You can eliminate R2.

You need R2 BECAUSE it's open-collector.

 

[AnalogKid]

You need R2 BECAUSE it's open-collector.

Don't think so. The data sheet shows an internal 30 K resistor to Vcc. It's on the internet, so it must be true.

ak

 

[JimB]

The data sheet shows an internal 30 K resistor to Vcc.

In which case, it is not open collector.

JimB

 

[Nigel Goodwin]

Don't think so. The data sheet shows an internal 30 K resistor to Vcc. It's on the internet, so it must be true.

ak

So why did you say the exact opposite? - I even quoted, and highlighted, your reply.

 

[tonytnnelectrotech]

And, since no one has thought to ask - do you have any of the hardware, software, or programming experience it would take to convert this project to one based on a uC?

If yes *to all three*, then why did you decide to go with an all-discrete solution?

ak

I have been programming atmel microprocessors for about 15 years ...

it seemed to me simply excessive to use such a solution ... but now I realized that I had not considered all these facets ...!

Since I have already made many PCBs, I have to solve the problem better with the circuit that I have ...

Thanks to your valuable advice, it seems that the problem is solved! I eliminated D1 and added a 1µF electrolytic capacitor between pin3 of PH1 and GND.

I confirm that R2 can be deleted ...

However tomorrow I will do more tests ....


If you have any other suggestions to improve this circuit, welcome ...

Thanks again.

 

[Nigel Goodwin]

I have been programming atmel microprocessors for about 15 years ...

it seemed to me simply excessive to use such a solution ... but now I realized that I had not considered all these facets ...!

This is the 21st Century, an 8 pin microcontroller can replace most applications for 555's far better and far cheaper. Nothing 'excessive' at all, as it greatly simplifies the project, and makes it work far better as well.