Attenuators and amp loads

[Enzo]

I just read a couple threads discussing the above topic.

Common attenuators we use are basically resistors. Oh sometimes we hang caps or inductors on them to emulate speakers, but mainly they are resistors.

I know in the rarified world of engineering, they have sophisticated things like programmable electronic loads for powr supply testing. Especially when supplies might crank out 50 amps or something. These guys don;t just hang a resistor on a supply and get out the volt meter.

I know little about it, so i am hoping the engineering brain trust around here can educate me.

Seems to me that electronic loads comprise something like power MOSFETs or even bipolars wired to shunt current across the output of a powr source.. the MOSFET can act like a resistor by conducting as much as you tell it. The MOSFET would be on a heat sink and cooled, but it ain;t a resistor.

I mean, if I wired an MJ15024 across some power source and biased it on, it would load that source, right? And I could control how much.

Have i at least got that much straight?

SO would this same sort of approach work for a dynamic source like an amplifier output? Or might there be interactions I am not forseeing?

[diagrammatiks]

you'd need to be able to match the impedance for a tube amp.

i think.

[Steve Conner]

If you just bias a transistor on, the resistance you get is non-linear. It varies a lot with current and temperature. You would need some sort of feedback to control it.

The logical conclusion would be a complete solid-state power amp with current-mode feedback, and another feedback loop wrapped around that to vary the current demand signal in proportion to the input voltage. Voila, a really big and expensive resistor.

By adding filters to the outer feedback loop, you could emulate any reactive load you wanted.

Using a Class-D amp could be interesting because the power absorbed in the "resistor" would be returned to the DC rails instead of being dissipated.

[kg]

just use a resistor.

[Enzo]

Oh, I do. It isn't a project, I just wondered about the technology.

[Tom Phillips]

The electronic loads that I have read about are for DC systems only. They are used in engineering development and for manufacturing test. In addition to testing DC power supplies they are useful to test battery systems, power components and solar cells. A significant feature is that they are programmable so the user can simulate things such as turn on surges and pulsed loads with various slew rates. You could even program a custom load curve. Some units have built in DVMs that report results via a computer data interface. The cost for programmable electronic loads starts at ~$500 and quickly climbs into the multi $k range. For information on the technology behind electronics loads I think you will find a trail at DC Electronic Loads | Agilent . I see links on that page to manuals and application notes.

Regards,
Tom