Bipolar PSU solution '?'

[petemoore]

  A regulators requires a load to work.
  10k is a common test load for regulators, data sheet [yuea I know] might tell more about exactly how much load is required, but I'm lazy and leaning on the other members here, hopefuly putting this up'll let me or someone else view to remember it.
  Generally a circuit is requiring a regulated output before the regulated output circuit is built...
  Would a reading of the circuits non-powered resistance say, right at the battery clip, give a 'right on' or ballpark load value to calculate if it will sufficiently load the regulator circuit?
 

[Paul Perry (Frostwave)]

Yes, because the apparent resistance or a random assemblage of semiconductors & resistors will almost always be lower with a higher voltage. Since the meter tests with less volts than the circuit will see, you will be OK. It's a very very rare problem, though.

[axg20202]

Not necessarily. In my recent thread on bipolar power, my +'ve regulators were putting out the correct voltage with no load at all, while the negative reg was not and was well over the target voltage. A 10k load was also not enough load to pull the negative regulator into...er...regulation. I don't know what the threshold is, because I ended up putting a much larger load (a low voltage filament bulb) on there to test that a significant load would do the business. It did, so I investigated no further.

[the_random_hero]

Not necessarily. In my recent thread on bipolar power, my +'ve regulators were putting out the correct voltage with no load at all, while the negative reg was not and was well over the target voltage. A 10k load was also not enough load to pull the negative regulator into...er...regulation. I don't know what the threshold is, because I ended up putting a much larger load (a low voltage filament bulb) on there to test that a significant load would do the business. It did, so I investigated no further.

Did you try a resistor before the caps to lower the input voltage feeding the regulators?

[George Giblet]

Not sure what you exact problem is but in general that is not true.  Whether a regulator needs a load or not depends on the specifics of the circuit/device.

For example,

A zener regulator will produce a regulated voltage without a load.

If you look at a zener regulator which has an additional transistor for increased output current, such as that used on some chorus units, then those regulators do need a load.    They don't actually *need* a load but what happens is the the voltage can rise a quite a bit under light loads.

If you look at say the  LM317 datasheet it specifies a 240ohm output resistor.  The resistor is necessary because it forms part of the feedback network.  The reason that it is such a low value is because some current flow out of the adj pin, that current varies with conditions.  The low resistor causes enough current to swamp the adj pin current which improves regulation etc.  If you look at say an 7805 regulator with the resistors wiredlike an LM317 to increase the output voltage the regulation/ripple isn't as good because the adj pin (actually gnd pin) current is much higher on these devices.