>>> An analog scheme would involve a multiplier or similar function..
>>> conceptually as below, for example:-
>>
>> A simpler analog scheme that does not include a multiplier is to make a
>> switching power supply. =A0Put a known resistor on its output. =A0To var=
y the
>> power, you vary the supply output voltage. =A0Note that in this case the=
 power
>> will be proportional to the square of the voltage.
>
> That is a good idea. I am not so sure if this is how the
> "constant power mode" of the electronic load works.
>
> I am still interested in what Michael Watterson has in his mind when
> he mentioned constant power sink. Maybe he has some other
> nice ideas.

The several general concepts described are all the sort of thing that
Michael's comment brought to mind.

Using a buck regulator to drive a fixed resistive load at constant
voltage is simple and cheap and with not too much effort efficiency
can be high and reasonably constant. Constant efficiency with varying
Vin allows an accurate constant power load to be implemented. Absolute
efficiency is less important.

A reasonably good solution is a subset of the general buck regulator case.
It works by storing a fixed amount of energy in an inductor and then
discharging it N times per second. The charge rate varies with Vin but
the energy used is fixed.
An inductor is charged from Vin until Iinductor reaches some preset limit.
The inductor is now discharged into an arbitrary load.
if this operation is repeated N times per second power is N.I^2.L
As long as the charge / discharge cycle is sufficiently short to allow
a complete cycle in less than 1/Nth of a second at lowest used Vin,
power can be varied smoothly across a wide range.

At a given load, conversion efficiency should be reasonably constant
as the criterion for inductor charging are always the same (I =3D
I_limit).

This is different from the usual regularor drives load constant power
system as the power OUT is not measured or cared about. The energy
transfer efficincy is also irrelevant - only the ability to transfer
the same amount of energy into the inductor on every cycle counts.
[Avoiding residual energy in the inductor can be an issue that needs
due design care.]


 Russell

--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.