Okay- now you're talking. That will give you an AC gain of -100/(0.1+Rp) where Rp is parallel resistance of (-) input divider in K Ohms. Trying hanging a big decoupling cap off the (-) divider and see what happens:_)
Has nothing to do with being any genius - more to do with people looking for a free lunch.
47K is your series input R to the '393 (+) and 100K is the '393 feedback resistor, 47+100=147. When the '393 pulls low, the voltage at the feedback end of the cap is -Vc where Vc is its steady state net charge due to the Vcc->2K->100K->47K->741 output loop.
I attached the schematic and cross posted to A.B.S.E. That should end any confusion about the circuit. I put it into switchercad today and played around simulating it. Since Linear doesn't give you a 741 or a 393 I subbed them out with Linear parts that I just picked pretty much at random. After simulating, I find that depending upon the amplitude and frequency of the sine input signal I get differing results. Some frequencies and amplitudes rapidly follow the sine signal and then finally the comparator snaps low and stays there for a while. Other frequencies and amplitudes will give me the exact result I want by snapping low at the beginning and staying that way. I surmise that given the nature of clock sounds, this is probably not going to work out very well in the real world. At least not without allot of fiddling around. A software solution is not really desired, so I will likely tinker with a non-retriggerable monostable multivibrator.
If you feel that I'm trying to weasel some free engineering out of you, then don't put yourself out.
Look at the schematic and tell me if this is still right.
Right- that's what you finally got across. Your signal amplifier gain is only x20 because R1||R2 is in series with your speaker source at signal frequencies. The ac-gain is therefore -R3/(R4+R1||R2)=-20, and the lower passband edge is ~13Hz because C1 and C2 are in series for an equivalent
2.35uF into 5K equivalent input resistance. To get closer to a x1000 gain, take C2, remove it from speaker lead and connect it from junction R1+R2 to GND, connect that speaker lead to the junction of R1+R2. You might want to increase C1 to 10U. This should blast your comparator input with a 4Vpp right off. If the delay inequality is still not good enough, then go with a window comparator.
Actually, you can take the speaker lead from which you removed C2 and connect it directly to GND. Then make C1 a 10U with (+) terminal connected to R4. I would still leave the 4.7U C2 from junction of R1+R2 to GND.
Aha, I think your problem might be that the positive feedback in the comparator circuit is being overdiden, its so much easier to see whats going on when you have something to look at, I saw this straight away, this is why your diode made a diference ... it stoped it from overiding the feedback in one direction.
I often use a dual op amp with one half functioning as a comparator (I have loads of lm358s), the risetime of course is pretty slow compared to comparator but works ok otherwise, and saves a package. you can also get op amp+comparator combinations, although they are not so easy to get.
The circuit of the lm393 given in the datasheet is very simple circuit to folow and is clear why what fred said is so, ... If one input is held below
+vc -2 then that input's transistor will always be biased on as long as the other input is higher, even if it reaches +vc.
btw as i mentioned in a reply to the circuit you posted your comparator is acting like a schmit triger as much as a monostable, full negative op from the 741 will cuase it to change state. a simple diode in series from the 741 op will block the negative excursion, it then needs a pulldown resistor so the 0.33 cap can discharge.
Thanks for the input. I'm having trouble understanding why a cap at the junction of the resistor divider (R1/R2) to ground would have an effect. I don't generally think in terms of anything "happening" there (with no audio signal being applied, just the divider going to the + input of the op-amp). What's up with that? IOW, why would the op-amp want to see a low impedance voltage there? BTW, the caps are tantalum, does that matter?
yes basicaly, when the comparator op goes high the other end of the 0.33uf capacitor goes from aprox 0.5vc to 1.5vc, if the op of the 741 falls to 0v the center of the two resistors wil be about 0.5vc, as soon as the capacitor has discharged only a small amount it will become lower than the - input and cuase the output to go negative again.
You really need a one-shot with lockout for the '393 circuit- and that's a dual comparator with lots of components. I would go with a 555 instead. Below is your same circuit idea with the 555 as the lockout one-shot. The 10K of CONT pulls the 555 threshold down to Vcc/2 and also uses this to bias the '741 output- assumes bipolar 555, adjust as necessary for CMOS. The 680K pulls the 741 output bias point down a few hundred mV's. The gain is the same, and since the speaker is across the differential input, no blocking capacitor should be needed. If you have too much sensitivity, replace the 100 ohm input with a 1K pot and adjust. The gain as shown is x1000. The 555 output is normally high, and when the 741 output goes a few hundred mV high, the 555 THRESH is crossed which triggers the 555 output low for 120K x 1U~ 120ms etc, all the while no longer responding to the 741 drive on THRESH. View in a fixed-width font such as Courier.
Scratch that last one- you do need a cap in series with the speaker to block the DC current due to 741 offset voltage, which could make for a hefty output offset with that large gain:
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.