Thanks for the replies. It gives the 19.5 volts with no problems, just not the voltage with a 2 amp current reduction.
Power supply likely uses a flyback topology, using a single switching transistor. Each time the transistor switches on, current flows into the primary of the transformer, building up magnetic flux in the core. When the transistor switches off, the magnetic field collapses and energy is transferred into the secondary winding.
Load regulation is achieved by controlling the pulse width (thus varying the amount of energy transferred with each pulse), or for very low loads, the controller will usually go into discontinuous mode, where it skips pulses.
The transformer design is really the single most important aspect of the power supply. Everything from the turns ratio to winding inductance, parasitic capacitance, the ferrite core characteristics, etc.
By trying to operate it at a much higher than designed voltage, you're forcing it to operate in an area where the required energy simply can not be properly transferred through the transformer, at duty cycles much higher than would normally be required for a given amount of load, thus it is no longer running optimally. There's a limit to how much flux can be built up with each pulse before the transformer core saturates, and that will effectively limit how high of a duty cycle the power supply can switch at.
Every part of the design of an SMPS is a delicate balancing act, from switching frequency and duty cycle, to component choices, transformer design, etc. While there are some standardized transformers available for SMPS, most of the time they are designed and manufactured for a specific use.
Even if you were to theoretically modify the transformer windings to suit the new voltage, you would only be able to get about 4.1A out of it at 19.5V, because the amount of power (watts) the transformer is capable of has not changed.
Best way to see exactly what is going would be with an oscilloscope, but be careful, never hook up the scope to any part of the primary circuit unless you have either an isolated differential probe or the power supply connected to a suitable isolation transformer. Observing the switching waveform you would likely see that the duty cycle starts off rather high (as the power supply is attempting to maintain a higher output voltage), and as load is increased it would quickly reach a maximum and no longer increase, this is the point at which the output voltage starts to fall off as the load current is increased.
I hope this explanation makes some sort of sense.