Analog switch or SSR?

Don't be so quick to throw in the towel. Turns out there's a simple solution. I sketched it out and tried it before I went out on the town with customers this evening. I'll post it tomorrow. ...Jim Thompson

Like, use a switchmode regulated current source? Those are not uncommon for LED drive, and aren't much different from switchmode regulated voltage sources. I'm thinking of playing with Micrel MIC2287, it just about would work for your 300 mA source.

Grin.. well driving in today I was thinking I could try turning the current source on and off... rather than switching the current around. I'll try it both ways.

George H.

Energy, But we'll measure power over a known pulse time. So need to know I, V and time.

george h.

You can, but you have to use two of them back-to-back, like this:

Version 4 SHEET 1 1352 680 WIRE 896 176 736 176 WIRE 896 208 896 176 WIRE 736 224 736 176 WIRE 848 224 800 224 WIRE 736 336 736 304 WIRE 896 336 896 304 WIRE 800 416 800 224 WIRE 848 416 800 416 WIRE 992 432 896 432 WIRE 800 448 800 416 WIRE 896 528 896 512 WIRE 800 560 800 528 FLAG 736 336 0 FLAG 896 528 0 FLAG 800 560 0 FLAG 992 432 switch SYMBOL nmos 848 336 R0 SYMATTR InstName M1 SYMATTR Value 2N7002 SYMBOL nmos 848 304 M180 SYMATTR InstName M2 SYMATTR Value 2N7002 SYMBOL voltage 736 208 R0 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V2 SYMATTR Value PULSE(5 -5 5m 1u 1u 5m) SYMBOL res 880 416 R0 SYMATTR InstName R7 SYMATTR Value 1k SYMBOL voltage 800 432 R0 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V3 SYMATTR Value PULSE(-10 10 100u 1u 1u 1m 2m 20) TEXT 704 584 Left 2 !.tran 12m

Cheers

Phil Hobbs

Switch at OpAmp, avoid high current switching...

...Jim Thompson

Why not do that the simple way?

___ +VS >--+-|___|--+-------------. | | | | | |\| | | _ '-|-\ | | / \ | >-----||-+ '-(+ -)----|+/ ||-> \_/ |/| ||-+ | | .-------+ | | | V | - | | | | | .-. ||-+ | | ||-----||-+ '-' | | | | === === GND GND

That's the "diversion" method, first proposed by John Fields. Works quite well, BUT current is always flowing... and dissipating.

My method turns the FET off, but leaves the feedback loop in the active region so the OpAmp, doesn't rail, and stays happy ;-) ...Jim Thompson

Or turn the current source on a few microseconds before releasing the clamp.

An RC on the FET gate might well be good enough.

Cheers

Phil Hobbs

Yep. All kinds of solutions to the timing since George did say "1ms". ...Jim Thompson

___ +VS >--+-|___|-+-----------------------. | | | | | ___ |\ | | _ '-|___|--+--|-\ | | / \ | | >-----||-+ `-(+ -)----------|--|+/ ||-> \_/ .-. / ||-+ | | | | | | '-' ,-------+ | | | Q1||-+ | D1 V ||-----||-+ `-(+ -)----------|--|+/ ||-> \_/ .-. |/ ||-+ | | | | | | '-' | | -. ||-+ | | ||-------||-+ '-' | | | | === === GND GND

Grin.. well driving in today I was thinking I could try turning the current source on and off... rather than switching the current around. I'll try it both ways.

Yeah something like that. (It does get more complicated.)

George H.

Thanks Fred, Say, why the diode in the load line? Is it there to reduce the leakage current through the load when the FET is shunting the current around the load? (I had a similar idea.)

George H.

See upstream for nice versions of Fred's circits... (I'm not going to clean up the google groups mess.)

Hey That's interesting! I was thinking of switching the voltage reference on and off, and not the feed back path. Does your method work better? I guess I worry about not getting the current source to turn all the way of f.... Or if I force it off hard, the opamp drives itself to the opposite ra il. (And then has to recover when recover when I turn it on again.)

Thanks. George H.

3/4 of a package of 74HC4066 analog switches is "complicated"? ...Jim Thompson

When you don't have a layouter to carve it into silicon, yes...

Yup. Reduces that to sub uA level...

This _is_ just a test bench isn't it? No "compaction" requirements? ...Jim Thompson