Analog interface

Mar 23, 2005 8 Replies

Optos tend to vary pretty badly in their current transfer ratio from one part to the next, some of them by 4:1 or so, and are linear in current, not in voltage. Getting the system accurate will be a pain, and will most likely require you to supply power to the isolated side to be able to have some active circuitry over there to help.

That being the case, as long as you're going to have to supply power to the isolated side, you might as well just move the A/D over there as well. Digital signals are much more forgiving of trying to push them across an isolation boundary; a poorly linearized 1 or 0 is still a 1 or 0.

You could go with a pair of high-impedance, high-voltage dividers to get the signals to within, say, 5 volts of ground, then use a diffamp to take out the common-mode stuff. You could make the dividers out of a string of 1% or 0.1% resistors, or buy some nice kilovolt-rated precision voltage dividers from Caddock or those other guys.

I do this in my electric power meters. I have a die-cast box with the dividers inside, with some zeners and stuff for added safety. Electricians can run 480 vac wires into one side of the box, and the electronics guys wire up to the low-voltage side. There's a metal divider in the middle, sort of a labor/cultural divide.

John

Hello all,



Someone has asked me to design a system that can take multiple "oscilloscope" type isolated differential inputs ranging form 0 to 500volts AC or DC. In reality they want to be able to hook up to a 3 phase motor or a 0 to 5 volt signal line, or a current clamp that provides 0-1 volt output. Frankly I'm a little stumped. The AD converter, multiplexer and automatic gain selection issues are easy enough to handle, it's the isolation I'm having trouble with. I can't use a transformer because of DC. I want to use an opto-isolator, like the Farichild H11F3, but I have the roughly 1 volt drop on the input LED that gives me a problem with the current clamp or anything in the 0 to about 1.4 volt range. I have considered using the TI (Burr Brown) INA148, but I don't think it really provides the isolation I need if I have 3 of them connected differentially across a 3 phase motor.



Anyone got any ideas?



Thanks,



Steve


I read in sci.electronics.design that steve wrote (in ) about 'Analog interface', on Tue, 22 Mar 2005:

Chopper followed by a transformer. Details left to the student. (;-)

Regards, John Woodgate, OOO - Own Opinions Only. There are two sides to every question, except 'What is a Moebius strip?' http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

There is Analog Devices AD7400 and 7401

Robert

Hello John,

Way to go, especially if the frequency range has to exceed audio. Even if not I'd do it xfmr/chopper because of cost.

Regards, Joerg

formatting link

"Joerg" a écrit dans le message de news:EZ_%d.14522$ snipped-for-privacy@newssvr14.news.prodigy.com...

It may not be totally necessary to have galvanic isolation. If only high input impedance is possible, then one way is a pair of inverting amplifiers with common mode feedback, followed by a difference amplifier.

Much simpler than the chopper and its floating input stage power supply and, and, and...

Thanks, Fred.

Have a look at the LEM range of Hall Effect current sensors. The 'instantaneous' type go from DC to 100KHz, small signal. Normal current sensors for each of the phase currents, and a low-current sensor (with series resistors) for the voltage pickup.

Tony Williams.

Google for IL300. You'll have to include an isolated supply to the hot side.

Paul Burke

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required