I have a hack, but unfortunately my wife and I have "temporarily misplaced" our digital camera, so I can't post pics yet. I figured I would ask some advice, so when I open it back up to take pics I can make some adjustments if y'all have any advice.
I have housed a home built mic-preamp in an Apple brand external SCSI drive enclosure (from 1996). The mid-90s beige decor nicely matches my audio-centric Color Classic Taco (which is functional again). The front power button and LED power indicator are functional. The plug-in jacks are mounted on the rear panel. The circuit is small, and I hope to build an amplifier to power small speakers in the remaining space.
The mic-pre is based on this circuit.
This is the first circuit I've ever built! Mine's built on perf-board, not PCB. The main differences in my circuit include: the 2.2 uF capacitors are not polarized (ratty shack didn't have any). I'm using a TL082 op-amp IC rather than the 5532 (everything was from whatever ratty shack had in stock). It's hard wired with a 1k resistor controlling the gain, rather than having a switch. I tried a couple different resistor values and I guess for my purposes and my mic I guess it needs that much gain. I am running it off of 9v, I originally wanted to use a 12v wall adapter, but I got a lot of hum that I don't get with batteries. Using a 9v battery was easier than stringing together 8 1.5v cells. Oh, and the LED I'm using is not in that original circuit. I don't know if that would have any effect.
The pre-amp is pretty noisy, a lot of hiss. I'd like advice on how to get that down. Is it that it's not on PCB? It is my amateurish soldering? Is it the lack of polarization on the DC filtering capacitors? Is it just to be expected from an op-amp circuit with no negative power supply? I know it's hard to diagnose things over the web, but if I make modifications or re-do it, I'd like to know what changes are likely to be most effective.
Thanks!
You should place a .1 microfarad cap on each power input of the op amp to ground as close as physically possible to the op amp. Actually it's kinda shocking the original design didn't include that. The author probably assumed it was so obvious he didn't need to include it.
also I'd place a 10 microfarad cap between the power ground and the signal ground. These grounds must be seperate as the signal you are outputting has a DC offset of 6 volts (bad).
Also for the love of god, don't cut the feedback line on pin 2.
I was wondering about that, as a book that I checked out of the library mentioned doing that.
I'm not sure what you mean by seperate grounds. The signal grounds are of course seperate, but they both go to the same ground rail that the whole circuit is on. Do you just mean that each signal ground has its own capacitor *before* the ground connection? And would that cap be polarized?
Oh, and is that the signal input or output?
Thanks!
It does not matter at all if caps are pollarized or not...
Oh btw, how does one work out what value caps to use in a circuit such as this?
0.1 mic caps are used to filter out AC noise from the line. Thats just a common size to use to do this.
Erm - But I thought caps filtered DC!! <--There again I suppose you mean AC noise can be allowed to conduct to ground...
Perhaps you mean an inductor...
They should not go directly to the earth ground. Normally, the signal grounds all go to the center tap on the transformer of the AC supply but since you are using a DC supply, I'd put all your signal grounds and power grounds back to your DC supply, and then put a cap between the DC ground and the Earth ground on your chassis.
Also, make sure all your ground points have their own wire going to the DC ground. Never share the same wire for 2 grounds on an analog circuit like this.
All EMI filter caps should be NON polarized. You are going to get noise on both sides of zero.
Signal input.
Caps allow high frequency AC noise to short to ground while appearing open to DC
Ideally you should do the math and figure out what the voltage of your noise is and how much current has to be passed in order to smooth it out and then calculate the cap value based on that. The closer to the correct value you use, the better the noise suppression will be. However, .1ufd seems to work pretty well 99% of the time.