The MFOS SCOOP is being sent at your request to keep you in the loop regarding MFOS. I'll try to get something out every month or two. MFOS' mission is to be the most informative synth-DIY site on the planet. We're here to help you keep imagining, keep inventing, and stay ingenious!

If you have any suggestions about what you would like to see in an issue of The MFOS SCOOP email me.

Please forward this to any of your friends who have an interest in synth-diy.

	Power Supply Kits
	ULTIMATE SUPER Deal Clincher
	News Letter Specials of the Month!
	Improve Your Op Amp Skills
	Check Out The New Combo-Packs!
	Coffee Shop Web Apps

New! Selected 2N5172 noise transistor. Noise level greater than 3V pp with non-inverting gain of 100.

AD633AN Analog Multiplier
Precision 0.1% 100K Res.
Precision 0.1% 200K Res.
SSM2210 Matched Pair NPN
PT094 Tempco
PN4391 N-Channel JFET
MPF102 N-Channel JFET
NJM13700D (LM13700)
2N5172 Noise Transistors

NEW! Metal film tempcos are on the way! Stay tuned for more excellent parts for your projects.

The kits contain all of the parts for your synth's power supply except for the wall wart and/or transformer. The LM317/LM337 kit can power an entire case of modules with it's adjustable voltage and 1.5A maximum capability. The wall wart kit can hold its own as well and can easily power an ULTIMATE and ULTIMATE EXPANDER or a Sound Lab Mini-Synth no problem.

Adjustable LM317/LM337 1.5A Supply Kit	Wall Wart Power Supply Kit (+/-12V)
The LM317/LM337 kit comes with (4) 15,000 uF @ 25V caps for the main supply filtering. I suggest a transformer with a center tapped secondary the output of which is between 24 and 26 VAC with at least 1A of output current capability.	The Wall Wart Kit comes with (6) 4700 uF @ 35V caps for the main supply filtering and LM7812/LM7912 regulators. I suggest that the wall wart not supply more than 9 to 12 VAC output with 500mA to 1A of output current capability.

Powering An Op Amp From One 9V Battery

This circuit makes use of what is known as a virtual ground. The virtual ground is a voltage level that is 1/2 the supply voltage (the supply voltage is 9V in this case). The junction of the two resistors and two capacitors becomes the actual virtual ground point in the circuit. When you measure between the virtual ground (using your meter's black lead) and the + side of the battery (using your meter's red lead) you see +4.5V. When you measure between the virtual ground (using your meter's black lead) and the - side of the battery (using your meter's red lead) you see -4.5V. This is just what we need to power an op amp (or several op amps). Notice that we use the virtual ground for the sleeve connections of the jacks as well. We isolate the input and output from the outside world capacitively. We use a 1uF on the input and a 10uF on the output. This circuit faithfully reproduces whatever signal you inject times 10. If you inject too hot of a signal the op amp will clip at about 3.5V to 4V. Just keep in mind that the virtual ground is where all circuitry gets its ground reference in the circuit.

Some Op Amp Insight

In essence when we use an op amp in the inverting configuration as shown in the illustration the op amp's main job is to try and keep the inverting input at the same voltage as the non-inverting input. So when we connect the non-inverting input to ground the op amp tries to keep the inverting input at the same voltage level. This phenomenon is the same reason that the inverting input configuration makes such a great signal mixer. The inverting input is referred to as a virtual ground (not the same as the discussion above) but it is a reference to the fact that when the non-inverting input is grounded and we use negative feedback the inverting input is held at ground by the action of the op amp's internal circuitry and the feedback resistor.

Consider the circuit in the illustration. If we apply +1V to the input and remembering that the op amp is trying to keep the inverting input at ground we know that there are 10 uA flowing though the 100K input resistor to the inverting input. What mechanism keeps the inverting input from rising to 1V? The feedback resistor and the op amp's internal circuitry are the culprits. We intuitively know that the output of the op amp will be -2V since we have a 200K feedback resistor and we know the gain formula for the inverting op amp configuration is (feedback resistor value)/(input resistor value). But lets look deeper. When the output of the op amp is -2V we are pulling 10uA in the other direction through the 200K feedback resistor and toward the output. This is what keeps the inverting input voltage from rising to 1V and what causes the output to go to -2V. The output via the feedback resistor pulls away the necessary amount of current to balance what arrives at the inverting input.

If we applied 2V to the input we would have 20uA flowing to the inverting input via the 100K resistor. How high would the output voltage have to go now to balance the current arriving at the inverting input? Well we need 20uA to flow away from the inverting input so the output would need to go to -20uA x 200K or -4 volts. By adding frequency selective components to the feedback network we can cause some signals to have more gain and some to have less gain (also known as filtering). Thus if you put a cap in parallel with the feedback resistor you are inviting higher frequency signals a ready path back to the inverting input and thus they op amp applies less gain to them and you end up with a simple low pass filter. With a better understanding of why the op amp applies gain to its output signal in the inverting configuration you can better understand how filtering networks do their thing. Although we are talking about applying simple DC levels to the input of the op amp remember that this is the same mechanism at work in AC op amp applications as well.

Well what if we don't have a feedback resistor? When we use an op amp without a feedback resistor it is generally because we desire the op amp to act as a comparator. Now lets say we apply +1V to the non-inverting input of the op amp and ground to the inverting input (we call this VREF) what happens? The op amp detects that we have an imbalance on the inputs and attempts to correct it. The non-inverting input voltage is higher than the inverting input voltage so the output of the op amp goes as high as it can (usually 1 to 1.5 volts from the supply voltage). If there was a feedback resistor the output would be trying to balance the inputs and pull the inverting input up to the level of the non-inverting input. But since there is none and no one bothered to tell the op amp it tries in vain to balance the input voltages by pushing the output as high as it can go. Now if the non-inverting input sees -1V the output will go as low as possible again in vain since there is no feedback resistor to help balance the inputs. Now you know why the comparator does what it does.

What's positive feedback? It's when you feed a portion of the output signal back to the non-inverting input. You see this in filters from time to time to help give the filter the desired response (filters are a whole other advanced topic). You also see it in comparators. In a comparator the effect of positive feedback is to add what is known as hysteresis to the response of the comparator. With no hysteresis a comparator can be a very finicky beast near it's threshold (VREF). Any noise on the supply can find it's way to the comparator's reference (VREF) so that it may chatter (oscillate high and low) when about to change states instead of smoothly jumping to either the high or low state. Let's say we are making a comparator and we want to apply the reference (VREF) to the inverting input and the signal that causes the comparator to go high or low to the non-inverting input (as shown in the illustration). This is a non-inverting comparator. When the input is above the reference (VREF) the output goes high and when it is below the reference the output goes low. If we now add a positive feedback resistor from the output of the op amp back to the non-inverting input we cause the comparator to have hysteresis. It is important to understand that in order for everything to work properly you need to insert a resistor between the signal you are detecting (the level that causes the comparator to change states) and the non-inverting input. Also the ratio of the input resistor to the positive feedback resistor is what sets the amount of hysteresis. The lower the value of the positive feedback resistor the greater the amount of hysteresis. If the feedback resistor if too low in value the comparator may never change states. In general a feedback resistor of 10 to 20 times the input resistor results in an acceptable level of hysteresis for general comparator applications. There are, of course, times when you want more or less than that.

Here is what happens with a positive feedback resistor. Lets say the non-inverting input is higher than the inverting input (VREF) and the op amp's output is high. Now the output is holding the non-inverting input high through the feedback resistor. The signal fed into the non-inverting input via the 100K resistor now has to go low enough to pull the non-inverting input below VREF despite the help that the non-inverting input is getting to stay high from the op amp's output via the positive feedback resistor. Once the signal goes low enough the output shoots low. In this case the output is helping to speed up the transition to the low state because it is tied to the non-inverting input remember. And just as you thought now the input has to go high enough to fight the feedback resistor which is holding the non-inverting input low. A trick to use to get the fastest response out of your op amp comparator is to add a small value cap (1pF to 20pF) in parallel with the positive feedback resistor. This causes the edge of the rising or falling output to get fed back to the non-inverting input almost immediately which speeds up the rise or fall time to just about as fast as the op amp will go.

I hope these tips help improve you understanding of op amps and comparators. Cheers, Ray