#498

[@RexxSchneider]

At 5:00 "We don't know [the available output voltage swing] ... The datasheet wasn't characterised at ±12V." However, the typical performance graphs on the following pages of the datasheet give us rather more information. In particular, Figure 7 shows typical Maximum Peak Output Voltage vs Supply Voltage into a 10K load. At ±15V supply it's ±13.5V (as we already knew), and at ±12V supply it's around ±10.7V. From the graph, you can see that it is a linear relationship and at ±6V supply, it's a fraction under ±5V. However, to estimate worst-case values, you'd want to note that the worst-case at ±15V supply was twice the typical loss, and use that to estimate the worst-case at ±12V supply as ±9.4V and at ±6± supply as ±4V. That's a lot of output swing lost at low supply values (because it has a bipolar class A-B output stage) and that's why it's unusual to see TL072 series operated at much below the recommended ±15V. Also, the inputs are never going to reach the negative rail because the TL072 series has p-channel JFETs as its inputs, and current mirrors as their drain loads, which need some voltage, so as soon as you get within about half-a-volt or so of the negative rail, the JFETs' gate-source junction starts to become forward biased, and that's not a desirable mode to be operating JFETs in. As you've shown, these are really quite unsuitable for single-supply use with a negative ground. Perversely, they would be happy with a ground at the positive rail, because Vcc is within their common-mode range, but nobody ever designs circuits with positive ground any more. At 8:30 if you put in a 20K and a 10K, you're going to get a gain of x3 (non-inverting), so I assume you fixed that before going to the demo with the 'scope.

[@cnvogel]

How large is the output swing for different supply voltages? The datasheet at least has a figure (Fig 7)... (and the plots are usually of course "typical" values) I think people should use single-supply circuits way more often, it's just much more complicated these days to even *find* modern chips that survive 30V (old stuff ran at +/- 18V, so even 36V) total supply. With complementary output stages that go full rail-to-rail, and sticking to inverting configurations (that keep inputs often savely at or around Vref/virtual ground, it's not much more complicated than old-school design. And whatever you want to feed into (ADCs) nowadays often only accept 0.."few V" anyway. Dual supply is nice for audio-stuff, though ;-). You want your +18dBu output. Maybe you want to do a video about fully differential design?

[@IMSAIGuy]

You bring up a good point about mixed mode circuits. part analog part digital. Yes you need to be careful on both sides. The world has moved to single supply for a lot of things. When I was designing big boy circuits decades ago I used an A/D that had bipolar inputs (+/-5V input). Today I'd probably keep the analog part +/-12 and have to made sure just before the A/D to move the analog range to single ended. creating and maintaining virtual grounds in systems has its own problems