1. Home
  2. Electronics Design
  3. need op amp pet detector circuit check

need op amp pet detector circuit check

I see several problems with that schematic. One is that the first stage has an inductor directly across its inputs. The DC feedback around that stage will only suffice to get the output active if the input offset is small enough that, when multiplied by 3.3M/Rcoil, the result is somewhat less than 1.5V. That stage should be AC coupled like the others. The second problem is that the threshold of detection is poorly determined. Whether those two independent but nominally identical voltage dividers end up producing the same level is a matter of chance. The third is that the circuit's output offset will be excessively sensitive to the op-amp bias current. The LM324 is not the best choice given that sensitivity. A fourth problem, related to the second, is that it will detect pulses in one direction a lot better than in the other direction. But there is nothing in the setup as described that justifies any particular polarity of the signal produce by nearby pet motion. A fifth issue is not really a problem, but a possible improvement. The use of a relay coil for a pickup is not really clever and probably gives less sensitivity than could be had with a larger pickup coil made with a few wraps of wire surrounding a larger area.

The problem with polarity can be mitigated with some experimentation and arranging that the magnet placed on the pet is in a (more or less) known best orientation. The result of those experiments may also influence the coil orientation relative to where the pet is expected to be when this gizmo is activated.

--
--Larry Brasfield
email: donotspam_larry_brasfield@hotmail.com
Above views may belong only to me.
Reply to
Larry Brasfield
Loading thread data ...

A few years ago, I posted a message asking for ideas for a like the ones used by the automatic door manufacturer's. I got a great reply describing one that the poster had disassembled, but no circuit diagram. I have been able to get back to working on the project, and, since I am not a degreed engineer, I would like to verify that I interpreted his description correctly. Unfortunately, the email address I have for him is no longer valid. If some kind soul could read the description below, and click the link to check my schematic, I would be eternally grateful.

Description: Here is the explanation of the pet proximity detector where the pet wears a magnet embedded in a plastic medallion on its collar. I tested the sensing range to be from 12 to 18 inches. I thought the use of the relay as a source for an inexpensive coil was clever. Obviously, the contacts are not used. The front end amplifier consists of three sections of a LM324 op amp. Each section pretty much the same. All + inputs are connected to the bias generated from two resistors that divide the 3V power supply in half. The sense coil, having a 0.1 uF across it, connects from the bias supply to the

- in of the first amp. The op amps are coupled through 10uF caps--they each have 3.3Mohm feedback resistors with 0.1uF across them. The fourth amp is used as the level detector set to detect approx. 1.1 volt signal. Note: the output DC level is at bias supply level. You can do what ever you want from there.

Schematic:

formatting link
Thanks! Tom

Reply to
Tom Kupp

The + and - inputs are drawn the wrong way around on the three amplifer stages.

The final comparator stage is a bit suspect. At the very least, the voltage divider R6/R7 should not be the same as R1/R2; the default output polarity will be unpredictable and there could be false triggering. From the wording, it sounds like centre of R6/R7 should be 1.1V above or below the bias set by R1/R2 i..e 0.4 or 2.6V.

The low frequency gain of that circuit must be quite high. 12 to 18 inches ey? I'm curious enough to reach for a breadboard ...

Reply to
Andrew Holme

Thanks again. Sounds like it's time to tinker!

Reply to
Tom Kupp

Thanks for the feedback. Think I'll grab that breadboard, too!

Reply to
Tom Kupp

Tom Kupp wrote:

Faraday's Law states that V=-N*dPHI/dt where N is # turns and PHI is the magnetic flux cutting the coil area. The amplifier is forcing V=I*Rcoil by drawing current I through C2, no current passes through C1, so that Vout of the first amplifier becomes Vout=Vbatt/2-5K*V/Rcoil-Integral(V/Rcoil)/C2=Vbatt/2-V/Rcoil*5K+(N/Rcoil)*integral(dPHI/dt)/C2 and this is also dVout=(N/(Rcoil*C2))*integral(dPHI/dt). After the coil area is saturated with PHI,max, the induction voltage drops to zero and the charge is trapped on C2 to bleed off with a time constant of R3*C2=0.33 seconds or until the magnet moves far enough past to make dPHI/dt transition in the opposite direction, removing the charge on C2 and building a residual of the opposite polarity. The basic response also shows that stray fields due to RF or utility frequencies average to zero output- so that the detector time constant should be long in comparison to the lowest stray field frequency. So what you circuit sees is a small step output of first amp of amplitude~(N/(Rcoil*C2))*PHI,max*v/d roughly where v is component of magnet field velocity perpendicular to axis of coil, and d is dimension of coil. The succeeding two stages amplify this unipolar step output x100, but they also amplify the high frequency field response by the same amount, so you have done nothing to improve the situation in terms of signal to stray field noise. If you re-arrange those amps into low pass configuration with cut-off of 6Hz then the strays will be attenuated by a factor of 40dB while the step response due to pet magnet induction is amplified. If you replace C5/6 with R=330K, leave C3/4=0.1uF, then you will improve this signal to stray induction by

80dB- a factor of 10,000. Don't forget to add offset current cancellation resistors of 330K to the (+) inputs( which you have shown reversed). C1 does nothing BTW. You will need to ac-couple the coil as someone also mentioned- no chance for reasonable offset cancellation with the circuit shown.
Reply to
Fred Bloggs

Vout=Vbatt/2-5K*V/Rcoil-Integral(V/Rcoil)/C2=Vbatt/2-V/Rcoil*5K+(N/Rcoil)*integral(dPHI/dt)/C2

Reply to
Tom Kupp

Would these be in series with the (+) inputs?

Reply to
Tom Kupp

yes

Reply to
Fred Bloggs

Reply to
Tom Kupp

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.