resistors to simulate the current load. A variac can then simulate maximum and minimum line voltages and the maximum and minimum DC voltages can be easily measured.
   Second, establish the maximum input voltage minus the output voltage, which determines the required voltage ratings of the pass device. For devices that are not tolerant of transients (essentially any solid-state device) it would be prudent to have the device withstand the entire maximum input voltage, plus a margin of safety. For vacuum tubes, a safety margin for maximum plate-cathode voltage isn't needed, and if the tube is run well below its maximum power dissipation, exceeding the maximum voltage ratings is permissible. For triode-connected pentodes or tetrodes, ideally the maximum screen voltage determines the maximum voltage rating. However, in practice, the screen voltage, in this case, can go as high as the maximum plate voltage, up to about 400 volts or so.
  Third, calculate the maximum pass device power dissipation: W = (max. input voltage - output voltage) * max. load current). If the bulk supply ahead of the regulator has poor regulation (i.e. its output voltage drops a lot under load), then you can use the maximum input voltage at maximum current, instead of at minimum current used for the voltage rating calculation. This corresponds to point E in the curves above. Since power dissipation results in heat and heat accelerates failures in both tubes and semiconductors, choose a device with a maximum power rating that is less than the maximum expected dissipation. The more the device is "de-rated," the more reliable it will be.
   Fourth, pick a device that will pass current at the maximum load current but minimum DC input voltage (point G, above). For semiconductors, this should not be a problem if the other maximum ratings have been observed. For tubes, the easiest way to determine this is from the characteristic curves. For triodes, determine if point G is in the region of negative grid voltage. For pentodes, be aware that the curves will vary with screen grid voltage. Also calculate the screen grid dissipation at point G (using the screen grid current curves) to make sure it is not exceeded. Screen dissipation is usually not an issue with triode-connected pentodes, since the screen current gets high only when the screen voltage is appreciably less than the plate voltage.
   Device selection is often an iterative process with several trial devices evaluated. Once a device or set of devices has been chosen, then other factors such a device speed, ease of driving, availability, cost, etc. come into play.
   A summary of the key operating points is shown in figure 6. In addition to the maximum and minimum voltages, the effects of poor input supply regulation is shown: a tilting of the operating points to the left at higher currents.
   One thing to be careful about when comparing characteristic curves: the slope of the triode curves depends

low-mu, and thus the overall gain of the regulator is limited. A very large grid swing will be needed which can be difficult to generate.
   The reason for the rp/mu trade-off is that the simplest way to decrease the rp is to widen the pitch of the grid wires. This lets more electrons flow to the plate, but the amount of control, and hence the mu, is reduced. Another way around the low-mu problem is to increase the cathode area. This is inherently what happens when triodes are paralleled. Another technique is to decrease the spacing between the plate and grid, although this can lead to limitations in maximum voltage ratings.
   An example of applying all these rp-reduction factors is the 6AS7G - the first tube explicitly designed to be a series-pass regulator, which was introduced by RCA in 1947. It has a large cathode area, very close plate spacing and a mu of only 2. The rp is only 280 ohms (at 125 mA) but the limited maximum plate voltage of 250 V and very low gain keep it from being the perfect regulator tube. Bigger brothers of the 6AS7G include the American 6336 and Russian 6C33C.

Choosing Pass Tubes
Before choosing a pass device, several circuit parameters need to be known:

  1. Maximum output (load) current
  2. Output voltage
  3. Minimum input voltage to the regulator
  4. Maximum input voltage to the regulator.

It is assumed in this discussion that the output voltage is fixed. If it is variable, then the design needs to be evaluated at the minimum and maximum output voltages.
  First, estimate what will be the minimum voltage drop across the regulator needed to maintain regulation. This will depend on subsequent design choices, but for now make an educated guess. The voltage drop will be significant with triodes and very small for solid-state devices. Then determine the minimum DC input voltage:
      V = output voltage + min. voltage drop.
Then determine the allowable variation in input line voltage. This is often taken to be +/- 10%, but if you live in a developing country with poor power, allow more variation. Design a bulk (unregulated) power supply that will supply this minimum DC input voltage at the minimum line voltage and maximum load current. If you are limited in your selection of power transformers, pick one that will supply at least the minimum input voltage needed. With the bulk supply designed, now calculate the maximum DC input voltage, based on the maximum power line input and minimum load current. This corresponds to point C in the curves above.
   Implementation note: once a candidate power transformer is chosen, it is worthwhile to solder or clip-lead together the bulk power supply circuit on the test bench and use power

pg. 6

< PREVIOUS

NEXT >