Wednesday, April 22, 2009

Regulated Power Supplies, Part 1 by Lee Bond, N7KC

April 22, 2009 Educational Radio Net, PSRG 48th Session

This is the first of a two part series on the subject of regulated power supplies. This initial part will introduce the idea of ‘regulation’ and contrast non regulated and regulated performance then include the linear or series pass regulator. The second part will deal with the switching supply and contrast the linear approach and the switcher approach to regulation.

To begin lets talk about power supplies which are not regulated. The easy example is the battery and a load. All batteries suffer from some finite internal impedance which affects the terminal voltage as a function of load. Some types of batteries exhibit extremely low internal impedance such as the lead acid type where this impedance might be in the tens of milliohms region. Other types, such as the carbon zinc cell, will have significantly higher internal impedance. The performance of this battery/load system is dictated by the relationship of the internal impedance and the external load.

For example, you cannot crank an automobile engine with a AA cell array making 12 volts but the engine will spin handily if a 12 volt lead acid battery is used. In any battery the chemistry producing the terminal voltage and the supply of electrons is in series with the internal impedance. The outside load is then in series with the battery innards and a complete loop, or circuit, is established. Since the internal impedance and the external load impedance are in series the total circuit current is governed by the combination of the two. Therefore, the smaller the internal impedance the less the external load affects the terminal voltage. The automobile engine starter is a low impedance load but much larger in numerical value than the internal impedance of the typical lead acid battery. As a result, the battery chemistry can provide the large cranking current without losses within the battery subtracting from the terminal voltage. The internal impedance of the AA cell is much larger than the lead acid impedance so a heavy load causes the internal impedance to drop a significant portion of the terminal voltage and the battery appears to be ‘dead’. Clearly, matching the type of battery and type of load is very important for good performance.

Earlier sessions of the Educational Radio Net have talked about voltage sources and current sources so lets review the definitions since these terms are commonly used to describe power supply performance. A box described as a voltage source will stubbornly maintain it’s set output voltage under all load conditions and would have zero internal impedance. A box described as a current source will stubbornly maintain it’s set output current regardless of load condition.

Modeling a battery is equivalent to providing a voltage source, i.e. the chemical process to enable work to be done, and including some internal impedance in series with the chemistry. The lower the internal impedance the more closely the battery looks like a voltage source at it’s terminals. The lead acid battery is an excellent facsimile of a voltage source under most load conditions. The AA battery is an excellent facsimile of a voltage source provided that the external load is much larger than the internal impedance. Ohm’s Law reigns supreme here when calculating losses associated with internal impedance.

Lets look at the lead acid battery with various loads to see how the math is done. Assume the internal impedance of the lead acid cell is 100 milliohms or 0.100 ohms and assume that the terminal voltage is 12 volts DC. Measuring the terminal voltage without any load using a voltmeter with normally high input impedance shows 12 volts. Adding a load of 1000 ohms shows a new terminal voltage of 11.998 volts. The difference from 12 volts, 0.002 volt, divided by the original 12 volts shows the regulation to be 0.0167%. That is, the change in terminal voltage when loaded compared to terminal voltage unloaded is 0.0167%. The same example using a load of 100 ohms yields a regulation of 0.0999%. Reducing the load to 1 ohm yields a regulation of 9.1%. These three examples show that the voltage regulation offered by a battery is good provided the load current in conjunction with the internal impedance of the battery does not disturb the terminal voltage significantly. To be a perfect voltage source the lead acid battery would need to have zero internal impedance. No such battery device is perfect.

At this point we know the definition of a voltage source and we know that unregulated power supplies offer a terminal voltage that is dependent on load and whatever internal impedance is present and, additionally, offer poor regulation unless very lightly loaded. Is there some way to compensate for internal impedance and maintain terminal voltage under all load conditions? The answer, of course, is yes and an example will illustrate a method and what is required.

Imagine a transformer with suitable plug to connect with a wall outlet. Following the transformer there is a rectifier assembly followed by a filter followed by a rheostat followed by a voltmeter followed by some sort of load. The transformer could be step up or step down depending on whether you want high voltage DC to run a vacuum tube device or low voltage DC to run your 12 volt DC radio.

Assume that you have the responsibility to ‘regulate’ the output voltage of this power supply manually at 12 volts DC. Assume that the load, such as a transceiver, is initially off and drawing no current. You have your hand on the rheostat knob and your eyes on the meter when someone plugs the cord into the AC outlet. All of a sudden the output voltage starts to increase, according to the meter which you are watching, and starts to shoot past 12 volts so you twist that knob as fast as you can to reduce the output to 12 volts. Ok, things are fine and you can watch the meter and tweak the rheostat as necessary to maintain the output at 12 volts with the very low drain on the system.

Now imagine someone turning on the transceiver load. All of a sudden the supply voltage starts to drop from 12 volts when there is a current demand and you frantically twist the knob to make the correction. In the process you notice that the voltage went too high when you over corrected, then too low when you under corrected, but you were able to eventually hit the 12 volt mark pretty well. In the meantime the transceiver went through hell because it really wanted good, clean, noise free voltage at 12 volts to operate correctly. Finally, someone is operating this transceiver in CW mode and the power supply must supply full current followed by minimum current then full current etc. You are twisting the rheostat control back and forth in vain trying to correct the power supply terminal voltage to the nominal 12 volts DC.

There is a better way and we will describe the necessary elements now. First, the available voltage must be greater than the required terminal voltage. Secondly, the power supply circuitry must be able to provide a maximum current that, at the least, equals and preferably exceeds the requirements of the load. Thirdly, there must be some reference voltage within the power supply for comparison purposes. Finally, there must be some means of comparing the power supply output voltage to the internal reference and adjusting some circuit element to maintain constant a output voltage. In the earlier example your eyes were comparing the metered output voltage with what you knew to be the desired output voltage and then your brain said to turn the rheostat in the correct direction to compensate for any error.

The first regulation method that we want to talk about is the series pass also known as the linear power supply. This methodology constantly samples the output voltage and compares this sample to some internal standard reference voltage. Any comparison difference is called an error voltage and this error voltage is used to adjust some series pass element in such a manner that the error is reduced to zero. The pass element is generally a power transistor or combination thereof located between the rectifier/filter and the power supply output terminal. The higher than output voltage reservoir preceding the pass element enables adding or subtracting from the nominal output voltage to maintain close regulation of the supply.

The down side of series pass regulated power supplies is poor efficiency. The pass element handles the entire supply current and gets hot so it is dissipating energy in addition to the energy used by the load. The closer the high voltage reservoir is to the output voltage the higher the power supply efficiency because losses in the pass element are smaller. Unfortunately the control circuitry requires a finite input/output differential to work correctly so there is a lower limit to how close these two voltages may be. Additionally, linear supplies tend to be transformer operated so large currents require large transformers and they tend to be very heavy in the higher current capacity models.

The up side of series pass regulated power supplies is very low noise generation and excellent regulation of the output voltage. Most of these supplies also offer fold back current sensing circuitry which protects the power supply and/or load from excessive and destructive currents.

The process of reducing an error voltage to zero requires what is called negative feedback circuitry. Additionally, carefully designed loop filters are required to minimize under and over shooting the desired output voltage in response to any dynamic changes. Modern supplies take full advantage of operational amplifiers and advanced filtering to achieve excellent performance.

This concludes the set up discussion for regulated power supplies part one. Are there any questions or comments with regard to tonight's discussion?

This is N7KC for the Wednesday night Educational Radio Net

1 comment:

G Kelly Robertson said...

Great Site Guys! Found it doing a Google search for APRS info.

Keep up the good work