What's wrong with this picture??

[ni-chrome]

Hello all! Could someone please help me with this simple project. I'm at my witts ends. It's supposed to detect machine movement failure in 10 seconds. The first part works fine (got from Radio Shack project book "event failure"), the second part I got from the seller of the piezo vibration sensor. One problem is the op-amp won't stop it's output when the piezo switch stops moving. Second, Q2 transistor doesn't switch even tho Q1 is getting .8vcd at base.
Anyone make sense of this? Any help appreciated.

 

[RadioRon]

Here are my observations on this circuit:

1) the sensor manufacturer, in their data sheet, shows the use of a variable threshold envelope detector after the op amp. Your circuit does not have this feature. I think you should add this in. It would be easy to add by placing a potentiometer between the opamp output (junction with R2) and the following diode.

2) if this sensor amp is recommended by the sensor maker, then pay no attention to this comment, but I'm uncomfortable with the lack of anything at the input to establish a DC level, to drain off charge and insure the input is at 0 V DC bias. It might be harmless to place a 10Mohm resistor from op amp pin 3 to ground to allay my fears.

3) The value of R4 is quite large and may be too large to provide sufficient base current for Q2. This may be why Q2 isn't turning on. You can fix this easily by raising the value of R6 which may be too low.

The way I see this working is...when the pan shakes, you get AC out of the op amp, which is turned into DC at C1 which turns on Q1. Q2 should then be held off since its base voltage is below turn-on threshold. The 555 monostable needs a falling voltage for the trigger input so it hasn't triggered yet. Q3 remains off also.

I presume this circuit is supposed to sound the alert when the machine stops moving, or in other words, when you get no output from the sensor. I would think that the sensor will continue to give some low level output even after you think the machine isn't vibrating any more, simply because it is sensitive to very minute vibrations. That's why you should have a potentiometer before the diode detector, so you can adjust for this.

Anyway, so the 555 output should remain high until the vibration decreases, right? Should this happen, then Q1 turns off, and Q2 turns on. When Q2 turns on, we get the edge to trigger the 555 and it begins trying to charge C3. What puzzles me is that you have Q3 there to short out C3 and Q3 will turn on when Q2 turns on, so the monostable output pulse is held on until 10 seconds after the motion is re-established. Is that the way its supposed to work?

 

[ni-chrome]

RadioRon, your observations on the function is right on. The pan moves scrap out of an unattended stamping machine. If the pan stops moving ,damage could result; hence the relay will stop machine.
This is my first foray into this hobby and chose this project to learn and make something useful. I sent the timing circuit to the vibration sensor manufacture, this is what he sent back.
In response to point 1): I will put a 1M pot just before the diode 2) yes, he wanted this particular op-amp, I'll add the 10M from pin 3 to ground. 3) I'll play with more resistance at R6. The rest of what you said is Greek to me. I'll try to figure out .
I'm not married to this way of accomplishing what I want. I thought a 555 timer would be good to keep from getting a false (no) signal otherwise I would just try to come out of the opamp to trip machine.
If you have an easier schematic for me to try on my breadboard, I would appreciate the input.
Thank you....

 

[RadioRon]

1 Mohm for the potentiometer is way too high a resistance. I think you should consider a 1K or 10K pot instead. The reason for this is that the diode detector needs to take some charging current through the pot and 1M is such a high resistance that it may not get enough. You could experiment with different values of pot if you have some in your junkbox, but don't go above 100K at most I think.

Wire the pot with one side to the opamp output, the other side to ground and the wiper as the output to your diode.

R6 can afford to go up to 100K ohms I think.

Continue on with this circuit as it seems OK. Its not terribly complicated and you won't find an awful lot simpler.

 

[ni-chrome]

I don't have a 1K or 10k pot , could I just use a carbon resistor inline before diode?

 

[RadioRon]

No, one resistor alone will not do the job. You need to create a voltage divider which requires two resistors. You can use two resistors to ground and attach the diode to the junction between them and then swap different resistors in and out, but really this is going to be a major pain in the a** so you really need a potentiometer there to make it easier to adjust.

 

[ni-chrome]

Thank you! .. looks like another trip to Radio Shack... tomorrow...
While I have someone who knows electronic on the line.... I'm doing this project for work as a side thing, I'm a tool designer by trade.
My home project is a solar sun tracker and piggy-backed my work purchase with home project purchases. This tracker calls for IRF7313 which is super small and won't "plug" into bread board....
How do you adapt those small surface mount components to fit a prototype board to test a circuit?

 

[RadioRon]

Thank you! .. looks like another trip to Radio Shack... tomorrow...
While I have someone who knows electronic on the line.... I'm doing this project for work as a side thing, I'm a tool designer by trade.
My home project is a solar sun tracker and piggy-backed my work purchase with home project purchases. This tracker calls for IRF7313 which is super small and won't "plug" into bread board....
How do you adapt those small surface mount components to fit a prototype board to test a circuit?

I normally don't use a plug board for such things. Instead, I prefer to grab a scrap of single sided blank pcb material (blank meaning all copper on one side with no patterns etched in) and use a knife or dremel tool to cut a quick layout. Its pretty easy once you get the hang of it. You just sketch out big rectangles of copper with tiny cuts between them in the right places, then solder components across these cuts. A bit of planning with pencil and paper beforehand and its surprising how much you can simplify a circuit for the minimum number of cuts.

Others use pre-made adapter boards which convert from surface mount to wires, but that's too inconvenient for me.