Modulated IR Detector

[gunrunnerjohn]

I'm looking at building a modulated IR detector circuit. I'm figuring on around 40khz modulation. Obviously, driving the IR emitter at 40khz is pretty easy. I'm looking for an easy way to detect the 40khz signal using the PIC processor. Obviously, the point of the exercise is noise immunity to avoid false triggers of the IR receiver from other light sources. I'm using a PIC 16F1825, but other similar PIC processors are also possible. I may want to step up to a 20 pin part to enable multiple detection channels as well as the other I/O planned. I'm thinking along the lines of being able to handle four IR sensing send/receive pairs with fairly low processing overhead.

I have also considered external filters, but in order to keep the parts count down, I'd like to do it on the chip.

 

[dr pepper]

You can get cheap ready made ir detector ic's with everything built in and logic output.
Do a search for tssop modules.

 

[gunrunnerjohn]

I suspect I need more than "tssop modules" as that doesn't return anything useful. Do you have an example link?

 

[dr pepper]

https://www.vishay.com/docs/81926/tsop4038.pdf

I thought these were available in 40 kc but it looks like 38kc is nearest.

 

[gunrunnerjohn]

OK, I got a little smarter about searching.

I found the Vishay TSSP58P38 that looks interesting at Mouser.

https://www.mouser.com/ProductDetai...ha2pyFaduhxHTPehEigO7%2b5MW2HnWg9QrRJMEWPtOY=

 

[gunrunnerjohn]

FWIW, 38k is fine, I'll be generating the drive for the sensor anyway. I just want a presence detector, so I'll have a continuous 38K driving the IR LED(s), and the detectors will just let me know when they see the signal. This is bouncing it off a model train to detect passing trains.

 

[alec_t]

I'll have a continuous 38K driving the IR LED(s)

Don't those TSSPs need bursts of 38kHz with gaps between?

 

[audioguru]

Yes, if the IR receiver receives continuous 38kHz IR then its AGC reduces its gain a lot because it does not want to be triggered by continuous 38kHz IR from a compact fluorescent light bulb.
Then it also might ignore your continuous 38kHz IR.

The datasheet linked by Mouser is very brief. Most datasheets of IR receivers from Vishay explain bursts of pulses and minimum gap times between them for full sensitivity.

 

[gunrunnerjohn]

Pulses are no problem, I have a processor generating them. I'll just send short bursts with a brief interval between them. The fact that they will only react to the 38khz is the key for this application.

I'll have to find a better datasheet for the part. I'll get a few in and tinker with them to see how they react.

Thanks for the idea, better than trying to filter it with the processor!

 

[languer]

There are a couple of examples on the following links that show a similar implementation:

• https://www.electro-tech-online.com/blog-entries/ir-beam-break-test-bed.156/
• https://www.electro-tech-online.com/articles/basic-ir-communications.361/

You could also look at a reflectance sensor (e.g. QRE1113, QRD1114 - https://www.sparkfun.com/products/9454) if you can locate it between the train tracks and the train.

 

[gunrunnerjohn]

Thanks for the ideas. I like the second article, looks just like what I'm doing.

The QRE1113 is probably not going to do it, those need the reflective surface pretty close. That's the chip that is used in a number of model train speed control systems with a striped flywheel, the spacing is pretty critical to insure triggering. Also, it doesn't have any frequency discrimination, so the processing would have to done in the uP, I like the idea of the one-chip solution for the decoding.

 

[dr pepper]

You can get pics with a data signal modulator built in that will generate the signal for the led.

Last year I did some work on a scalextric layout, part of which was a car detector, infra red wasnt my choice, I used one of those ultra sonic modules, transmitting continuous pings and waiting for a response, it works that well I could dectect ad discriminate between cars on the track next to the sensor and the far side of the track.
If you were using this idea as a break beam then it'd probably be ok, however reflecting the infra red of different model trains may give varying results.

 

[gunrunnerjohn]

It's only necessary that it sense some of them, truthfully. There is a timeout before the status changes, so I just need to pick one up every few seconds to be effective. I have a cheap eBay reflective sensor on a couple of signals, and it's pretty reliable, so I can't imagine a higher reliability solution being that bad. I'll have to look at the PIC choices, since this is a new project, I can change horses.

The reason for the reflective operation is I don't want to have to have two separate pieces. There sill just be on mile post at the side with the two sensors in it, that's what I use now.

 

[dr pepper]

Sounds like you sussed it.

 

[gunrunnerjohn]

Thanks everyone, I did indeed get a lot out of the thread. I hadn't though about the canned option, that takes a lot of the work out of the task.