Bat Detector Load Issue

[Raven Luni]

Greetings,

I've built a more compact version of my bat detector but seem to have run into some problems. Something seems to be knocking the rail splitter out of whack and so far I havent found any accidental shorts, cuts etc. on the board. I tested the splitter early on and it was working fine - more or less 50/50 split. In the finished circuit I'm getting the following readings (these are all rounded to the nearest volt):

Battery voltage: 18V (probably nearer 17 after a good bit of use)
Splitter total output : 16V
Splitter +ve / gnd : 6V
Splitter -ve / gnd : -10V

Since I've connected the logic circuitry to the positive side of the splitter, my guess is that its putting too heavy load on it, but I dont see why its not compensating. See the schematic and let me know what you think:

 

[Raven Luni]

Ah - found a problem in the splitter: 1 track cut missed and 1 connecting wire missed (strip board), result: op amp out is connected to V+ (should be cut) and not connected to the diode junction.

 

[KrisBlueNZ]

I think you can simplify your design quite a bit if you want to.

I don't think you need that complicated high-current rail splitter. The only components that really need a centre rail are the three op-amp inputs (the two amplifiers, and the comparator). This could be easily handled by a voltage divider or my pet preference the TLE2426.

You would need to return the electret microphone to 0V and reverse the polarity of the coupling cap. It could easily be a much smaller value too, provided the first stage input impedance is increased.

Your inverting amplifier stages seem to be drawn with the feedback resistor connected to the wrong end of the input resistor. Also, 47 ohms is a pretty heavy load to put on an electret microphone, and a combined voltage gain of 10k is probably a bit over the top. You might want to offset the comparator threshold slightly so it doesn't respond to noise below a certain threshold (such as the internal noise generated by the microphone).

I think a single 9V supply would be fine. You don't gain that much using an 18V supply. Perhaps as much as 10 dB higher voltage at the amplifier output before clipping. Big deal

I would also change the speaker driver to something that doesn't pull current continuously in either state. I suggest a simple unbiased complementary buffer (1x NPN, collector to positive, 1x PNP, collector to 0V, bases commoned and driven from the signal, emitters commoned to provide the output) feeding through a small capacitor, say 10 uF, to the speaker to ground.

Just some suggestions to consider

 

[Raven Luni]

Nice info - thanks. It shows how much I don't know about designing amplifiers

As for the feedback resistors, D'OH!! That was a stupid mistake

Good point about the comparator - I hadn't considered that.

The speaker output is a bit pathetic (taken from the first version which used earphones and a 64 ohm speaker). Your suggestion sounds better.

Also I found another mistake: the count down pin on the 193 should be held high not low.

 

[Raven Luni]

Luckily the resistor issue was just a mistake in the diagram - circuit itself is fine (I drew the diagram after in order to have something to illustrate in the post here). In theory all I should have to do is fix that counter pin and it should work. Then I can look at the comparator and add the high pass filter (got 1 spare amp left on the chip for that).