high speed full wave rectifier

WHAT saturation? Are you tying to run it as single-supply? Don't ;-)

...Jim Thompson

Brain engaged, mouth (OK, fingers) now in neutral - as opposed to the other way round.

doh.

Thanks Jim

Nope ;-)

And it takes substantial OpAmp slew-rate at 20MHz to get good performance.

I've only used the design up to around 1MHZ.

...Jim Thompson

Several people are now making RF detector chips, with linear or log outputs, and bandwidths into the GHz.

John

interestingly enough, the datasheet for a schottky BAT54 has a 5ns Trr ?!

a BAT81 is pretty good though.

an LM7171 does a reasonably good job. an LM6152 is terrible. and yeah, its the slew rate all right.

interestingly enough, when I go for a 200kHz bipolar sawtooth, an LM6152 is almost as good single-supply as with bipolar supplies, and is not much worse than the LM7171. But I'll stick in a bipolar supply anyway, it'll only cost me $0.20 or so, and this is a proof-of-concept thing

the application is a crazy smps topology that puts terrible demands on a CT (50Hz - 250kHz BW) - so bad I havent made one work well yet. HF or LF performance, but not both.

And the current is bipolar, so a DCCT isnt really feasible - although a current-output (eg LEM LAH-xx) DCCT can in theory drive a schottky bridge, in practice the phase reversals cause evil R-R glitches - probably because when all diodes are off the impedance gets very high. no amount of snubbing or (realistic) loading really helps. a shame really, cos it was a nice (OK, kinda nasty) idea....

when we go to an ASIC it'll get a lot easier (it becomes the chip designers problem ;). unless we use the same french idiots who've made a hash of the other ones. I've suggested we use a guy in Arizona, it'll make arguments about politics that little bit more interesting :)

Cheers Terry

Those ratings are kinda weird.. but I suppose it makes practical sense to report whatever looks like it. Reverse recovery could be modelled as a big capacitance at Vf ~= 0V, after all...

The thing that doesn't make sense to me is the schottkies that say 100ps and some specific test (what was that called..) that doesn't exist except perhaps in the annals of the IEEE. What good is a number when you aren't told WTF it is!?

Tim

-- Deep Fryer: A very philosophical monk. Website @

"unless we use the same french idiots who've made a hash of the other ones"

Suck butt Nigger.

DNA

Tektronix has a patent on a technique using feedforward compensation specifically for sawtooth rectification at high frequency...

Terry Given wrote: > I've been playing with full wave rectifiers. one of the nicest designs > is the EDN idea on Jims website. but for 20MHz operation (200kHz > sawtooths) the opamp is a bit problematic - its response is all about > how the opamp comes out of saturation, even with the extra diode. >

See bottom of page 19 of the Analog Devices AD8036 datasheet. It does what you want within the restriction of 5 volt power supplies. Or do you want Jim's circuit at 20MHz?

GG

Hi Glenn,

thanks, that was one solution I came up with when I went a-googling. turns out I just needed to pick better diodes, and I can do it with a variety of opamps. I'd have given my left nut for that clamp amp in 1992 though, when I designed a (not too flash) output stage for a pulse generator I built.....

Cheers Terry