0-600 vac to binary

Leds are current driven devices, i.e. they should give *some* output even at very low current (microamps). So, how about using an optoisolator in a linear mode, then amplifying the resulting mains frequency ripple?

Is op-amp allowed?

The complexity is all on the low-voltage side. Perhaps you can multiplex the receiver photodiodes/transistors into a single more complicated circuit? You don't need any RC time constants if you can throw some software at it. You can use some of the spare I/O's you have gained to drive the mux!

Use a better optoisolator. You can get ones spec'd with 0.7 CTR at

Hi You mean use a voltage divider then send the result through a 8 to 1 analog multiplexer? And as you say use software to sample? That's a thought. I will need to draw some current though to be able to tell induced voltage from a true on condition. I am using the equipment in noisy enviroments. Monitoring contactor coils for on and high limit switches for open and so on. In HVAC. equipment.

I was talking about using the mux on the low voltage side. (I am assuming that you need a real actual safety-approved optoisolator to do the isolation. So cannot just divide down the 600V resistively say and measure directly).

So you mux the photocurrent into an opamp virtual ground and measure the output in software.

Or use a lower current opto-isolator like Spehro suggested. (But that's cheating!)

. I

ent

rs

Using a themoresistor bonded to a 1W resistor, you should be able to detect the heat change. You would need the op-amp for amplifications and micro for delta-sigma.

Well I was hoping for something I had not thought of. I appreciate the thoughts though. Measuring the heat rise is good and indirect. Using a lower current opto is good but 8 x 1.5 watts in a plastic box is not great. Thanks Tony

I like linnix's thermistor idea, and Spehro's opto is tops too. But if even 1W dissipation per channel is too much, why not go pulsed? Rectify the a.c., drive a relaxation oscillator, which pulses an opto. That could be done with milliwatts.

Cheers, James Arthur

Surely the thermistor idea is too slow? Tony was worried about "RC time constants" so I would think the thermal one for a 1W resistor would be much worse!

it. I

rrent

ts

tors

ut

rts

plex

to

he

I am cheap, so I like cheap solution. Themistors are cheaper than opto-isolator. You must even be able to go with milliwatts resistors. You would need to try out actual values.

Yep, slow. I was just skimming...missed that.

There are lots of ways...like a hi-z divider driving a

Cheers, James Arthur

uit. I

urrent

ia

tts

itors

put

ch

an

iplex

to

o 1

do

the

's

If the apps is a data logger tracking HVAC, time should not be that critical. However, the OP might need a separate dedicated micro with

How about:

signal >---[R]--[C]--->[bridge rectifier]--. | OPTO --- |/ \\ / ~~> | --- |>. | ===

Dissipation = low, caps = big.

Cheers, James Arthur

How about voltage divide the signals and digitize them on the hot side of the isolation barrior. Then send the SPI interface signals through the digital opticouples to the micro. Use an isloated powersupply for to power the A2Ds (8 channel). done

How about a resistor feeding a small transformer? You only need the inductance to be several uH if the secondary side is fed into a low impedance circuit. A HC4051 could be used to multiplex the secondary onto an op-amp circuit.