Post by thread:Proximity detector

I've been asked to look at developing a simple proximity detector. I kinda have in mind using a PIC to pulse a series resonant LC once per minute or so. The LC would radiate a damped 49 MHz sine wave based on its self resonance. The PIC would just drive this LC with a step that changes state once per minute or so. I need to detect this damped sine wave up to 10 feet away. I'm thinking of another LC (probably using circuit board traces for L) driving a single chip RF amp set up as a Schmitt trigger so it only detects stuff over a certain level. This would drive a diode detector which would then probably drive an analog in on a PIC. Software would check the timing of the damped RF pulses to identify the particular unit transmitting. Anyone have experience in this area? I am somewhat concerned about the whole thing being jammed by a cordless phone but don't want to build a superhet receiver to try to get more selectivity. Perhaps a ceramic or SAW filter after the LC? Even then, with shared spectrum I'm likely to find interference. Comments? Thanks! Harold

I go through old Everyday with Electronics mags every now and then. I think I saw a article in one of the 1996 issues that went on about location-sensing with inductive loops. I'm sorry that I can't be more specific about which issue it was, but it sounded much like what you want. good luck! Harold Hallikainen wrote: {Quote hidden}> > I've been asked to look at developing a simple proximity > detector. I kinda have in mind using a PIC to pulse a series resonant LC > once per minute or so. The LC would radiate a damped 49 MHz sine wave > based on its self resonance. The PIC would just drive this LC with a > step that changes state once per minute or so. > I need to detect this damped sine wave up to 10 feet away. I'm > thinking of another LC (probably using circuit board traces for L) > driving a single chip RF amp set up as a Schmitt trigger so it only > detects stuff over a certain level. This would drive a diode detector > which would then probably drive an analog in on a PIC. Software would > check the timing of the damped RF pulses to identify the particular unit > transmitting. > Anyone have experience in this area? I am somewhat concerned > about the whole thing being jammed by a cordless phone but don't want to > build a superhet receiver to try to get more selectivity. Perhaps a > ceramic or SAW filter after the LC? Even then, with shared spectrum I'm > likely to find interference. > Comments? > > Thanks! > > Harold -- eric van es spam_OUTvanesTakeThisOuTilink.nis.za cape town, south-africa http://www.nis.za/~vanes/

At 15:44 02/07/97 EDT, Harold Hallikainen wrote: > I've been asked to look at developing a simple proximity >detector. I kinda have in mind using a PIC to pulse a series resonant LC >once per minute or so. The LC would radiate a damped 49 MHz sine wave >based on its self resonance. The PIC would just drive this LC with a >step that changes state once per minute or so. >[...] > Anyone have experience in this area? I am somewhat concerned >about the whole thing being jammed by a cordless phone but don't want to >build a superhet receiver to try to get more selectivity. Perhaps a >ceramic or SAW filter after the LC? Even then, with shared spectrum I'm >likely to find interference. > Comments? Why don't you measure the inductivity instead of the damping of an LC? Could be done e.g. with an oscillator measuring its frequency, or a bridge circuit. ------------------------------------- Gerhard Fiedler <.....gerhardKILLspam@spam@pobox.com> S‹o Paulo - Brazil

On Thu, 3 Jul 1997 19:50:03 +0200 Eric van Es <vanesKILLspamilink.nis.za> writes: >I go through old Everyday with Electronics mags every now and then. I >think I saw a article in one of the 1996 issues that went on about >location-sensing with inductive loops. I'm sorry that I can't be more >specific about which issue it was, but it sounded much like what you >want. > >good luck! > Thanks for the responses... Guess I didn't make the requirements clear. What we're trying to do is make a very small transmitter and a very small receiver. The receiver needs to let us know if the transmitter is "out of range" (about 10 feet away). Distance itself is pretty noncritical. since this is a very short distance for RF, I was thinking I could just do a damped sine wave generator by stepping a PIC output from 0 to 5 volts or back. This output would probably drive a series LC tuned to some RF frequency (as allowed by part 15 of the FCC rules for unlicensed devices). The idea is that at the resonant frequency there would be a high circulating current in the LC, causing radiation from the L as the current decays due to losses. We're kinda pulling an RF harmonic off the step waveform. That's kinda what I have in mind for the transmitter. For the receiver, another LC driving something. Perhaps a current input type RF amplifier. The L would again serve as the receive antenna. The RF amp would drive some sort of detector which would drive the receiving PIC where some coding would identify the unique transmitter. this is a little like keyless entry stuff. Could also use "transponder" technology, but it seems that a separate transmitter and a separate receiver is more cost effective here since they would always be sold as pairs. I think transponders are typically used where there a bunch of real cheap transponders and a few relatively expensive units for polling them (like theft prevention tag systems). So... anyone have some ideas in this area? Cost is very critical, but if we are only transmitting a few feet, it doesn't seem that we should have to go to triple conversion superhet receiver design! Harold

On Sat, 5 Jul 1997 23:02:38 EDT Harold M Hallikainen <.....haroldhallikainenKILLspam.....JUNO.COM> writes: [...] > Thanks for the responses... Guess I didn't make the >requirements >clear. What we're trying to do is make a very small transmitter and a >very small receiver. The receiver needs to let us know if the >transmitter is "out of range" (about 10 feet away). Distance itself >is pretty noncritical. since this is a very short distance for RF, I >was thinking I could just do a damped sine wave generator by stepping >a PIC output from 0 to 5 volts or back. This output would probably >drive a series LC tuned to some RF frequency (as allowed by part 15 of >the FCC rules for unlicensed devices). The idea is that at the >resonant frequency there would be a high circulating current in the >LC, causing radiation from the L as the current decays due to losses. >We're kinda pulling an RF harmonic off the step waveform. A single damped sine wave is going to be impossible to differentiate from impulsive noise, which is common in the 49 MHz band. At the very least, send a series of several pulses with known timing. Also at 49 MHz the typical PCB mounted coil's physical size will be too small to act as much of an antenna. Some sort of wire antenna will likely be needed. You're also going to have to fight with a lot of strong FM signals from cordless phones, baby monitors, fire trucks and police cars, etc. A higher band would have less interference. At 312 MHz, about all that could interfere are garage door openers and the Air Force. {Quote hidden}>That's kinda what I have in mind for the transmitter. For the >receiver, another LC driving something. Perhaps a current input type >RF amplifier. The L would again serve as the receive antenna. The RF >amp would drive some sort of detector which would drive the receiving >PIC where some coding would identify the unique transmitter. this is >a little like keyless entry stuff. Could also use "transponder" >technology, but it seems that a separate transmitter and a separate >receiver is more cost effective here since they would always be sold >as pairs. I think transponders are typically used where there a bunch >of real cheap transponders and a few relatively expensive units for >polling them (like theft prevention tag systems). So... anyone Theft prevention tags are just nonlinear antennas that resonate with the transmitter's strong signal and generate harmonics which are detected by the receiver. The tags themselves are completely passive. Some of them have a coil and a diode and others have a piece of mu-metal biased with a magnet. Transponder tags which send a code such as for toll payment or implementing the New World Order use an antenna connected to with a switch driven from open to short by the code at about a Kbaud. The polling system has a transmitter and receiver tuned to the same frequency. The transmitter sends a continuous signal and the receiver looks for AM modulation in the returned signal resulting from the variation in the reflection from the tag. Both of these systems make the polling unit expensive because (a) transmitter power on the order of watts is required and (b) precise RF components are needed to keep the transmitted signal from overwhelming the receiver. >have some ideas in this area? Cost is very >critical, but if we are only transmitting a few feet, it doesn't seem >that we should have to go to triple conversion superhet receiver >design! Go with AM in the 312 MHz garage-door opener band. The transmitter is just a single transistor oscillator with a printed coil operating as both the inductor and the antenna. The receiver is a regenerative design with a single transistor doing all the RF, typically followed by an op-amp for baseband gain. Since the receiver is an oscillator, it is conceivable that the same circuit could be used for both transmit and recieve. However, because the receiver is an oscillator, it will likely interfere with other nearby receivers (garage door opener receivers often use an RF amp to improve sensitivity a bit and also reduce the regen oscillator leakage back out the antenna.

Mike Keitz wrote: > Go with AM in the 312 MHz garage-door opener band. The transmitter is > just a single transistor oscillator with a printed coil operating as both > the inductor and the antenna. The receiver is a regenerative design with > a single transistor doing all the RF, typically followed by an op-amp for > baseband gain. You mean a Super-regenerative (Superregen) receiver. These things play havoc with my UHF CB (467 MHz) when you pass or park near a car with one of these burglar alarms! Very suspicious! > However, because the receiver is an oscillator, it will likely interfere > with other nearby receivers (garage door opener receivers often use an RF > amp to improve sensitivity a bit and also reduce the regen oscillator > leakage back out the antenna. The best ones use an RF Monolithics (or equiv.) SAW filter between RF stage and superregen, as well as a SAW resonator locking the transmitter frequency. Cheers, Paul B.

At 11:02 PM 7/5/97 EDT, you wrote: {Quote hidden}>On Thu, 3 Jul 1997 19:50:03 +0200 Eric van Es <EraseMEvanesspam_OUTTakeThisOuTilink.nis.za> >writes: >>I go through old Everyday with Electronics mags every now and then. I >>think I saw a article in one of the 1996 issues that went on about >>location-sensing with inductive loops. I'm sorry that I can't be more >>specific about which issue it was, but it sounded much like what you >>want. >> >>good luck! >> > > > Thanks for the responses... Guess I didn't make the requirements >clear. What we're trying to do is make a very small transmitter and a >very small receiver. The receiver needs to let us know if the >transmitter is "out of range" (about 10 feet away). Distance itself is >pretty noncritical. since this is a very short distance for RF, I was >thinking I could just do a damped sine wave generator by stepping a PIC >output from 0 to 5 volts or back. This output would probably drive a >series LC tuned to some RF frequency (as allowed by part 15 of the FCC >rules for unlicensed devices). The idea is that at the resonant >frequency there would be a high circulating current in the LC, causing >radiation from the L as the current decays due to losses. We're kinda >pulling an RF harmonic off the step waveform. > That's kinda what I have in mind for the transmitter. For the >receiver, another LC driving something. Perhaps a current input type RF >amplifier. The L would again serve as the receive antenna. The RF amp >would drive some sort of detector which would drive the receiving PIC >where some coding would identify the unique transmitter. this is a >little like keyless entry stuff. Could also use "transponder" >technology, but it seems that a separate transmitter and a separate >receiver is more cost effective here since they would always be sold as >pairs. I think transponders are typically used where there a bunch of >real cheap transponders and a few relatively expensive units for polling >them (like theft prevention tag systems). > So... anyone have some ideas in this area? Cost is very >critical, but if we are only transmitting a few feet, it doesn't seem >that we should have to go to triple conversion superhet receiver design! > >Harold > It sounds a lot like an electronic (invisible, RF) pet fence to me. You might want to buy one at any pet supply shop and reverse engineer it. Such a fence is a transmitter emitting a low frequency (~150KHz) and when the dog gets within a few feet of a wire buried in the ground around the perimeter of the property, the dog gets a shock from a receiver in his collar. Please forgive me if I tell you something you already know, but many people on the piclist are foreigners and may not be familiar with such devices. :-) I have designed a proximity detector (capacitance sensing) for a client, but it only has a range of a few inches (all the application needs). BTW I also have an electronic security device that is a keyring receiver that beeps if my briefcase gets more than a few feet away. You can buy a similar device at a baby supply store to alert you if your child gets away from you while you are looking away. They can't be expensive. I hope this gives you some ideas.