Post by thread:snappy transistors

Hi all, ive just spent the weekend trying to build an infrared comm link between the pic and the serial port. The reciever end is a radioshack photodiode that switches the base of a 2n2222 transistor which runs through a diode to the rs232 Rx line. The problem is that the computer cant decode the characters. it gets the right number of gibberishes, in the right order, though. i think that the resistor is too sloppy to give clean voltage transitiions. i looked in catalogs and cant find a shottky device with less than 14 pins, and i only need to clean up one signal, and hopefully cram everything into a d-9 case. Does anyone have any sound advice on how to produce cleaner levels for the port using just a few parts? or is sending ir ascii a hopeless approach? thanks, alice

Dear Alice, 1) What is the comm speed are you using? 2) Which polarity of the RS232 voltage the 2n2222 is switching for bit '1'? up or down?, it must be down, from zero to -12Volts. 3) I understand you are not using any modulation signal, just pure bit 1 = IR tx led solid irradiation, right? 4) What is the distance from TX ro RX led? I recommend start this tests in a closed very short distance TX-RX (less than 2 inches), inside a tube or pipe, to avoid external interferences. Note that IR can travel through several materials, even black or solid colors for IR is just as clear as pure glass... :) Once you have this "close test" working fine, go for taking them apart and them you will deal with bigger distances... I bet you would end up using 30-50kHz modulation pulses for distances bigger than 1 meter. Remember that Radio-Shack or Digikey has ready to use IR modules that nicelly remove the modulation and delivery pure digital ones and zeros. You just need to modulate the TX led and data with 30-50kHz. Wagner Alice Campbell wrote: {Quote hidden}> > Hi all, > ive just spent the weekend trying to build an infrared comm link > between the pic and the serial port. The reciever end is a > radioshack photodiode that switches the base of a 2n2222 > transistor which runs through a diode to the rs232 Rx line. The > problem is that the computer cant decode the characters. it gets the > right number of gibberishes, in the right order, though. i think > that the resistor is too sloppy to give clean voltage transitiions. > i looked in catalogs and cant find a shottky device with less than 14 > pins, and i only need to clean up one signal, and hopefully cram > everything into a d-9 case. Does anyone have any sound advice on how > to produce cleaner levels for the port using just a few parts? or is > sending ir ascii a hopeless approach? > thanks, > alice

Hi Alice, Try the MN1381. You can get it at Digi-key and elsewhere. it's meant as a voltage detector reset device, but it works fine as a general-purpose schmitt trigger. They come in many voltages, but the important thing to pay attention to is the maximum operating voltage, which I recall being 7 volts. The CMOS output version is what you want. It comes in a TO92 transistor package. Cheers, Bob http://www.bobblick.com/

Alice, There is a Basic Stamp application note ( App Note #11 Infrared Communication ) that implements the very thing Wagner recommended. It uses a 555 timer and a few pots to generate the 40 khz signal. The testers claim the circuit is reliable at speeds of 1200 baud over distances over 10 feet indoors. Using a higher performance LED could increase the range. Parallax makes these app notes available to the general public on their website ( http://www.parallaxinc.com ). In fact, in general, the complete TX/RX circuit is applicable, just use PIC pins instead of BS-I pins <g> ! HTH, Eric {Original Message removed}

> or is sending ir ascii a hopeless approach? > thanks, > alice Sending IR ascii is not hopeless, but your simple circuit will be extremely sensitive to background IR. Try is in full dark and it might work. IR remotes use a modulated approach (information often on-aff modulated on a 30 - 60 kHz carrier). Simple canned receivers, e.g. siemens SFH506 and sharp exist, but those are not cheap but much cheaper than the equivalent build from separate parts. Look on the web for RC5. Wouter.

You can implement a Schmitt trigger with 2 transistors. Email me privately if you wish to pursue this. Also. in similar pulse type applications you can often get magical (*) improvements in pulse shapes with a simple 2 pole active filter using an emitter follower, 2 R's and 2 C's. Presumably, "The art of electronics" covers both of the above. regards Russell McMahon From: Alice Campbell <spam_OUT1502amcTakeThisOuTLO.SCSENG.COM> ive just spent the weekend trying to build an infrared comm link between the pic and the serial port. The reciever end is a radioshack photodiode that switches the base of a 2n2222 transistor which runs through a diode to the rs232 Rx line. The problem is that the computer cant decode the characters. it gets the right number of gibberishes, in the right order, though. i think that the resistor is too sloppy to give clean voltage transitiions. i looked in catalogs and cant find a shottky device with less than 14 pins, and i only need to clean up one signal, and hopefully cram everything into a d-9 case. Does anyone have any sound advice on how to produce cleaner levels for the port using just a few parts? or is sending ir ascii a hopeless approach? thanks, alice