At any given frequency, the difference between Aol and Acl is the amount of global feedback. For comparison, the dotted lines show the AC characteristics of a typical conventional power amp operating with the same closed-loop voltage gain. 4, the heavy, solid lines are for the composite amplifier. The Bode plot of the amplifier is shown in Fig. Aol for the OPA-121 is typically 120dB, and for A2, itвЂ™s 78dB. All thatвЂ™s missing is the open-loop gain (Aol) specs for A1 and A2. You now have almost enough information to make a Bode-plot of the composite amplifier. You can determine the ampвЂ™s bandwidth by using the value of Cf and R2 in the equation, Fbw = 1/2ПЂCf R2 = 107kHz. The size of Cf is found using the following equation: I refer you to the previously mentioned application bulletin for more details.) (Both the mechanism of stabilization, as well as the reason for the minimum gain, require an explanation that is well beyond the scope of this article.

3, capacitor Cf stabilizes the composite, but for the stabilization method to work, the closed voltage loop gain of the composite must be ten or greater. Unfortunately, I was not able to utilize all that speed because the slew rate of A2 itself is about 12V/Ојs, and this slew rate sets the power bandwidth of the amp. The voltage gain of A2 at 20kHz is about 130, so the effective slew rate of A1 is 260V/Ојs. The composite solves the apparent slew rate shortage of this op-amp because it is multiplied by the gain of A2, as mentioned previously. The intended output power is 50W into 8О© up to 20kHz, which requires a minimum slew rate of 3.55V/Ојs (2). Since this project is a power amp, you must also consider slew-rate needs. All thatвЂ™s needed now to complete the composite is a compensated IC op-amp having an FT of 2MHz. Indeed, most of the design effort in this project was directed toward meeting that requirement.ДЄs things worked out, 2MHz was the highest unity-gain frequency I could achieve in A2 while still meeting the phase-margin requirements. Making the discrete amp unity-gain stable at some specific frequency with the required phase margin of 90В° turned out to be a tough job. Remember, though, that this is to be a 50W power amp, so A2 in this project is a discrete unit to handle the power. These requirements are easy to meet if you use a dual, compensated IC op-amp. At FT, each must have a phase margin of 90В°. However, in this type of composite, it is essential that A1 and A2 have the same unity-gain frequency, FT. I have learned from building this amp that the open-loop gains of A1 and A2 need not be the same. 3, A1 and A2 are identical, compensated op-amps, which makes it easy to predict the composite rolloff needed for AC analysis. Fortunately, there is a way to stabilize this composite. Two amplifiers in a loop virtually guarantees oscillation, since the individual poles combine to produce a two-pole rolloff. With composites, you should include the rolloff characteristics of both A1 and A2 in the AC analysis, and also have some means of providing phase compensation for the loop. Obviously, the composite must be stable for you to benefit from its use. Another advantage in my amp is the slew-rate boost: in the composite, the slew rate of A1 is multiplied by the gain of A2. My composite amplifier has a DC output-offset voltage of в‘mV, stays there regardless of the load placed on the output transistors, and uses no вЂњzero adjustmentвЂќ pot. Thermal feedback in a power amplifier driving heavy loads can cause errors due to self heating, but the composite eliminates this problem. This is an especially welcome feature in a power amplifier. In fact, because A1 does not drive the load directly, its DC accuracy is better than that of A1 alone. Since A2 is in the feedback of A1, the composite retains the DC characteristics of that amp. For example, you can combine the best qualities of both amps used in the compound, and DC performance of the composite is excellent. Figure 3: Composite amp circuit used for this project.

Burr brown op amp offset measurement