So it is not critical that I use 2.5 volts for the +Input of the LM741? I was under the impression that it must be half the supply voltage to allow the op amp to be stable. Example: 5v = 2.5 or 10v = 5v.
If I can run a lower voltage to pin 3, I would get a wider range of variation? If so, that’s good news. Right now using 2.5v as my vref, my debug window shows a definite change in value depending on the noise, but the changes are very small. For example, in my debug window it shows a value between 0.99 (no sound) to 1.99 (loud coke can popping), or something along those lines, I can’t remember exactly what the numbers were.
When I ran 1.6 volts to pin 3 in one of my pervious tests, the LED would respond to sound. I have not however, tried this lower voltage reference yet with the new ADC part.
No, the op-amp will work as long as the +Input is not too close to either Gnd or Vcc.
Yes, you would have the capability for a wider variation. To actually get a wider variation means that you need more gain in the op-amp. If you are using 1K and 200K for the two Rs that meet at the -Input, then you want a higher value for the 200K guy (recall that I’m using around 1 Meg now in my circuit, which gives a gain of about 1000X, because 1000000/1000=1000).
If you have 2.5V on the +Input, then the output will be capable of going from 2.5 to almost 5. But then you’re ‘wasting’ the space between 0 and 2.5.
If we assume these are the true values, note that your no-sound value is about 1, rather than being close to 0. But if the +Input was at (say) 0.5V, then your no-sound would become (say) 0.2, and your max would still be 1.99. In short: changing the +Input voltage simply changes your no-sound value in the A/D.
At the moment, I’d say your biggest problem is the 200K R3. If you can crank that up by a factor of 3 to 5, then you’ll see that your mic is much more sensitive to sound. And when that happens, it’ll be easier to hit the upper limit (1.99). And then the fix for that will be to put a lower voltage on pin 3.
It all makes perfect sense to me now Pete! I will buy a larger resistor for the gain and put a lower voltage to the positive input. These two changes should give me exactly the results I’m looking for.
How’s your project coming along? When you are finished, I can create gerber files for your circuit so you can have a professional board made with green solder mask and silk screen. If you can do this already, thats ok also. As a return favor, I want to offer my assistance to help you develop the final circuit board. To give you an idea of cost, My board will only cost $12.95 for one board from spark fun electronics.
If you decide you want me to create gerber files for your design, all I will need is your schematic drawing using ExpressSCH.
Thanks for your offer of PCB setup help. It’ll probably be awhile before I am ready for that - I will keep the circuit in prototype form until I have tested it on the real bot platform.
My platform (a Lynx 4-leg 3dof) is maybe arriving tomorrow. The only plug-and-play HW I’ll have is an SSC-32, so I’ve got a bunch of work to do to setup power, a ‘brain’ (a PIC of some sort), collision sensors, etc. For power I’ll use a cable from an AC supply initially, while I’m working on the SW for the walking gait and such. For the PIC brain, I may be able to ‘borrow’ a board that I have here at work, which includes an 18F8720. The board has other useful HW, including I2C stuff and an RS232 port. And, I have access to a Hi-Tech C compiler (which costs $950 - glad I didn’t have to buy it!), and I own the code that runs on this board, so it should be a good starting point.
I will also recommend you to use LM324 in your application. Its an embedded operational amplifier and has 4 channels embedded in a single chip. You can quite easily operate 4 amplifiers from a single chip. If you are working on it then you should have a look at Introduction to LM324 and download its Proteus simulation, which will help you in understanding How it works.
