New Thread (From Mtg itad) & Isolated Rail Detection Discussion

@barnun posted:

...

Stan, the optoisolator place holder in the schematic is a prefab 2 channel board of which I have 4 or 5 on hand.  When designing my board I simply placed two 4-Pin female headers the proper distance apart so the prefab (male pins) drop in for easy replacement.  Looks like the chip on the board is D216.  Yes, my intent is to provide a 5v signal level output to an Arduino Mega 2560 Pro Mini.

...bricks, tied to a single ZW-C and properly phased.

...  These LTV-8x4 chips seem to be spot on for this application.  Any thoughts on the current limiting resistors?

Regarding the current limiting resistors, I was just going to comment that you choose them the same way you did for your existing optoisolators.  But if I understand it, your existing optoisolators are purchased modules that you did NOT "design" it yourself?

That said, just do a search for optoisolator or optocoupler design.  It's not just the input resistor(s) to play well with the 18V AC track voltage but also the output resistor (and smoothing capacitor if needed) to play well with the 5V Arduino connection.

I could just give you some ballpark values that would probably work with a typical AC optoisolator but better to confirm which exact optoisolator chip you plan to use.  Also, if you want to only use components you already have like the existing resistor and 330uF capacitor then this would be more challenging.

The value of the smoothing cap on the output side is dependent on the load on the opto output, I'm assuming 60hz as the input voltage.

@stan2004 posted:

.... your existing optoisolators are purchased modules that you did NOT "design" it yourself?

Stan, that is correct.

@gunrunnerjohn posted:

The value of the smoothing cap on the output side is dependent on the load on the opto output, I'm assuming 60hz as the input voltage.

GRJ, 60hz is correct.  On the output side the Arduino should not sink more than 20mA at 5VDC.

Gentlemen, as always thanks for your input.  Filing this away for the moment.  Since I have 5 of the original boards with the bridge rectifier, 330uF cap and the optoisolator design on the way I thought I would wait until they get here, assemble a prototype and do some testing.  Based on those findings I will look at designing a board with either the LTV-824 or the ACPL-214-500E chips and more appropriately sized capacitors and resistors.

Thanks again,

Barry

Something along the lines of 10uf caps will be sufficient I suspect.  My sound board uses a 1uf cap into a 10k load, and it's filtering 60hz from the opto just fine.  The output voltage from the opto when triggered with 60hz hardly moves at all on the 'scope.

@gunrunnerjohn posted:

Something along the lines of 10uf caps will be sufficient I suspect.  My sound board uses a 1uf cap into a 10k load, and it's filtering 60hz from the opto just fine.  The output voltage from the opto when triggered with 60hz hardly moves at all on the 'scope.

GRJ, excellent!  Now I have to dig into the stash to see if I have any 10uF caps.  With my luck I will most likely have some (but at the wrong voltage).

Most any value will work, and the voltage just has to be above 5V if you're feeding it to the Arduino.

If it wasn't dark I'd go out and buy a PowerBall ticket!  The 4th bag I opened had (12) 10uF 25V cap's in a 5x11 Radial leaded configuration.  Perfect.  I also performed an additional search (obviously using some magic words heretofore unused in my many, many searches) and came across this link which has a ton of information.  Thanks again,

Barry

Batter UP!!  Next flavor of an Isolated Rail Detector using the LTV-824 chip.  Tomorrow I'll port the schematic over to a PCB layout and post a photo.  Now it's time to go watch football (and drink a beer).

PCB layout from the schematic above.  Barry

Did some clean up work on the schematic (90% cosmetic, 10% functional).  Swapped the inputs for track 1 and track 2 (along with the labels) which straightened out the traces at H3.  Currently holding at Def Con 3 ... Confidence is high.

Looks fine to me, pretty simple flow through the optos.

In case any interested party runs across this in the future .... Received the pcb's & parts, got a prototype board together and tested.  The first iteration was using 1K resistors on the AC Line side (R1A & R2A) and failed the finger test (man, were they toasty).  Subsequently tried 1.5K, 2K, 2.7K and wound up using 3k.  The board seems to be working as designed and ran a 24hr. test with no problems.  Attached are the updated schematic and pcb picture, hopefully I can put this one to bed.

Barry

@barnun posted:

... The first iteration was using 1K resistors on the AC Line side (R1A & R2A) and failed the finger test (man, were they toasty).  Subsequently tried 1.5K, 2K, 2.7K and wound up using 3k.

Right.  With, say, 18V on the track, the opto diodes only require about 1V.  So the resistor is dropping about 17V.

Power (Watts) = Volts x Volts / Ohms = 17 x 17 / 1000 = 0.3 Watts.  Those look like 1/4W resistors?  But whether 1/4W or 1/2W they're still burning 0.3 Watts which can make even a 1/2 Watt resistor toasty!

At 3K, the resistor is only burning 0.1 Watts which is fine for a 1/4W or 1/2W resistor.

Stan,

You are absolutely correct (and observant), those are 1/4w.  Since I have to order some resistors as I move from prototype to production for the layout, in the better safe than sorry vein, I'll order 1/2w.  As always, thank you for your input.

Barry