Electronic – Using two voltage regulators (780x) in series to prevent heat sink

Almost every TO220 packaged IC that you can obtain will have the most negative circuit point on the centre terminal and tab. "Insulated tab" versions are often available in cases where an exposed negative tab is liable to be a major disadvantage in many applications.

The reason for the negative tab is that this point is at the potential of the IC substrate or die - it is the lowest potential layer in the silicon construction for power purposes. Small signal voltages lower than this may occur BUT they are dealt with specially to ensure that the substrate is not reverse biased . Connecting this ground power level directly to the tab makes most sense thermally. Insulating it from the tab has a thermal penalty and makes cooling more difficult.

In the 7805 positive regulator the most negative point = ground = tab

In the 7809 negative regulator the most negative point = -Vin = tab as it is a negative regulator.

Exception that demonstrates the rule:

An apparent exception to this rule (and an actual exception when the IC is in use) is the LM317 positive regulator which is a "floating design - the actual regulator is all above ground in normal use. In this case the most negative point when the LM317 is used as a voltage regulator is the "adjust" terminal BUT the tab is the output - which is 1.25 V more positive than the adj terminal when in use. A look at a block diagram f the IC shows that even this one obeys the "tab = substrate = most negative point" "rule". The digram below shows that tab = substrate = output BUT that the adjust terminal is then pulled to 1.25V below this level during operation. A minor node inside the IC is pulled to one Vbe below IC "ground" with the -1.25V appearing at the end of an internal resistor.

ground = tab

Two things to look at (but I don't promise either one is the solution):

• Most simulators like every node in the circuit to have a dc path to ground. In your circuit, the primary side of the transformer is fully isolated from ground. You can solve this by arbitrarily connecting one of the two sides of the ac source to ground, or by connecting one of them to ground through a high-value resistor.

• 430 mF is a very large capacitor. This value could result in an ill-conditioned problem for the simulator. When the capacitor is charged up (assuming you're doing a transient simulation and the capacitor starts out at 0 V) there will be very large currents, which could lead the simulator to giving up on the grounds that the results are unreasonable. Try reducing this value or adding a small series resistance (like 1 ohm or so, maybe less) to C1.