|
enough to bias up a 6550. Of course they could be wired up in parallel, which might work, albeit at great expense, if the dissipation and voltage limits (1W and 100 volts) for the devices were not exceeded. If a more complex circuit is permissible, then more options become available. The next level of complexity would be the commonly available 3-pin adjustable voltage regulators. An LM317 would make an excellent current source. It contains a fairly precise, fairly low noise voltage reference which is used to define a fixed voltage differential between its output and its adjustment pin. If a resistor is placed in between these two pins, the current limit will equal the reference voltage divided by the resistor value: Current = 1.25v / R .
For example, a resistor value of 125 ohms will yield 10 mA's of current, 0.01 = 1.25 / 125; a 25 ohm resistor, 50 mA's of current, 0.05 = 1.25 / 25. The disadvantage to this device is that it might require a heat sink if the dissipation is too great, and that it has a fairly low voltage differential limit of 40 volts. This last limit is eased up (50 volts) in the high-voltage version, the LM317HV. For even greater voltage headroom, the Texas Instruments TI783 works, as its voltage limit is 125 VDC. For an even greater voltage limit, a cascoding of any of these regulators with a high voltage MOSFET or tube would do the trick. Alternatively, a zener or zener string could be placed in series with the regulator,
|
|