[time-nuts] Re: Another reason to monitor line frequency :) - My AC measurement project & question

[Wilko Bulte]

Given that it sounds that these capacitors  are mains connected I would recommend using Y rated capacitors for safety reasons. That should prevent them from going short circuit, Y caps being designed to "fail open".

Or did I misunderstand the proposed circuit?

Wilko

> On 22 Jan 2022, at 03:06, Lux, Jim <jim at luxfamily.com> wrote:
> 
> On 1/21/22 4:43 PM, willl will wrote:
>> Hi everyone,
>> 
>> I have an recently finished project that also measuring AC waveform, full
>> description here:
>> https://github.com/will127534/RaspberryPiAtomicNixieClock/wiki
>> 
>> Basically using an AC transformer and Ti's ADC8681 @ 50Khz sampling rate.
>> 
>> This year I'm working on a earthquake sensor + AC mains monitor system (In
>> an earthquake prone area, AC mains frequency will fluctuate by power
>> generator and machine emergency stop like this one:
>> https://twitter.com/kuriuzu/status/1360602496821911553).
>> 
>> I want to improve AC measurement. Apart from the ADC sampling speed upgrade
>> (previously bottlenecked by the SPI connection to FPGA). I'm currently
>> debating about whether or not to bypass the transformer. How does the
>> distortion of an AC transformer impact the accuracy of mains waveform and
>> frequency? I'm not sure if it is worth it to go through the mains voltage
>> safety requirements.
> 
> You can use a capacitorsget your galvanic isolation, and a CR voltage divider with minimal waveform distortion. Pick a burden current (say, 1 mA) and for 120V line, you need 120k impedance at line frequency 60Hz
> 
> X = 1/(377*C) --> C = 1/(377 * 120E3)  = 22 nF
> 
> Say you want ~100:1 ratio? so 22 nF in series with 1.2k  (or 2.2uF) would do nicely. Then feed your high Z ADC with a couple 0.1 uF capacitors from the ends of the 1.2k.
> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com -- To unsubscribe send an email to time-nuts-leave at lists.febo.com
> To unsubscribe, go to and follow the instructions there.