Adafruit Industries, Essential service and business: NYC – Executive Order 202.6 - Read more. Accepting essential orders - here's how.

Out of ideas - Inaccuracies with HCPL-7520 IC
Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.

Out of ideas - Inaccuracies with HCPL-7520 IC

by robodude666 on Wed Dec 01, 2010 6:26 pm

Hey guys,

For the past several months I have been working on a power monitoring system, inspired by the Tweet-a-watt and PowerBox projects. I have another thread that I started regarding keeping things safe, and it eventually went into debugging some of my issues I've been having.

To quickly summarize, I'm using the HCPL-7520 optoisolator. This IC performs optical isolation, amplification and biasing of the signal. For current sensing, the voltage drop across a current-sensing resistor (0.015 Ohms) is fed to the input of the IC. The output is linear and is scaled to VREF. The input can range from -0.256mV to 0.256mV and will output between 0V to VREF, where 0.00V input is about VREF/2. The analog output is put into the Arduino's ADC, sampled and the actual current is calculated.

Everything is working more or less okay, and I'm pleased with the results. However, there's one thing that's bugging me:

When the input is brought to grown, aka 0v is the input to the IC, the output suggests there is a 2.22mV signal present at the input. Technically if the input is less than 2.22mV it appears as 2.22mV.

I have no idea why. I've tried decoupling capacitors on the power lines, RC filters, but nothing seems to work. Is there anything I can do to remove the "presence" of this false signal? While deal-able for the HCPL-7520 that's used for the AC voltage measurement, this is a significant bit of noise for my current sensing IC. A 2.22mV input equates to 150mA of current (using my .015Ohm current-sensing resistor), making it very hard to detect things that are consuming less than 20W.

Any ideas? I'm starting to think that my issues are being caused by the input limitations needed by this IC (+/- 256mV). The Tweet-a-watt project is pretty accurate, and the Solar Logger also does a very good job. Should I simply abandon these ICs and use a current transformer or use a regular op-amp and ignore the whole "making things safe via optical isolation" bit?

Posts: 118
Joined: Sun Mar 28, 2010 12:12 am

Re: Out of ideas - Inaccuracies with HCPL-7520 IC

by ImaginaryAxis on Wed Dec 01, 2010 7:30 pm

It is either an offset voltage or you are limited to how far the output stage can swing to ground.

Posts: 267
Joined: Sat Nov 13, 2010 3:16 pm
Location: Phoenix, AZ

Please be positive and constructive with your questions and comments.