32bit Analog VAC Readings

[jcottrill]

My team is using a BRX PLC for a project and we need to be able to read a 0-10VAC analog signal and get 32 bits of resolution on the reading.  I don't believe anything in the BRX offering will do this but I was wondering if any of you have used an remote modules (RS485, Ethernet, etc) to accomplish this.  I have done some searching but most of what I come up with is just 16 bit.  We aren't really sampling the signal, just taking a few measurements for comparisons.  The thought now is to use a digital multi-meter with an RS232 port and poll the data.  While this works it is also a 9" x 2" x 12" piece of equipment that will only be utilized to a fraction of its capabilities.  Are there better ways of doing this that would integrate into a panel better?

[BobO]

2.3 nanovolt resolution on a time varying signal. I are confoosed.

[plcnut]

That does sound rather peculiar. Is it a true RMS reading, or peak to peak, or

[jcottrill]

The guys are trying to test transformers to ensure their turns ration are within 1% of their name plate rating.  I believe for a valid measurement (and as plcnut mentioned) a TRMS reading is required.  After thinking about what they are doing and seeing your guy's comments I need to dig in a bit more to the test plan as something smells a bit off.  I think the measurements for the primary and secondary could be made with 16 bit resolution as long as the scale is set properly and still be able to catch the error needed.

If I truly can make it down to the 16 bit range does BRX have any modules that handles VAC and perform a TRMS reading?

[BobO]

If you are trying to determine to 1%, that's only 1 part in 100, so 16 bit is 600x better than that.

How fast do you need a reading? What frequency is the input? The built in analog updates at just over 1ms, but that isn't fast enough to give very good resolution. If your input is 60hz, you would only get about 15 samples per wave. I suspect that taking samples over many waves would give you enough data to make a really good estimate of true RMS. Easy to test. May be good enough.

Might also be able to full wave rectify and filter (with a cap) and then measure the DC. Would have to account for the diode drop. Might work. Not sure about accuracy, but it wouldn't be hard to test either.

Barring either of those, then no, BRX doesn't have a method to accurately measure RMS.

[plcnut]

If you go look at fluke volt meters, you will notice a significant price change for one that says "True RMS". I believe it is going to require a specialty instrument to do what you are asking.

[plcnut]

Maybe this would work?
https://www.omega.com/pptst/DP25-VRMS_CRMS.html
Edit: you would need the -S4 model to get the RS-485 communications (~$515).

[jcottrill]

Hey guys,

Thanks for the feedback and sanity checks.  It looks like the issue boils down to scale in order to detect a 1% difference in the resulting ratio of the primary voltage to the secondary.  These are high power current transformers and I think the lowest ratio is 100:1.  To keep the test voltages low and safe they are using an input voltage of 1VAC so the ideal expected output will be 10mVAC.  With an analog input scaled to 5V, this would put us right on the edge and in a couple corner cases results in a false pass.  It think this puts the BRX builtin analog out to be on the safe side.

I had stumbled across a few items similar to the Omega last night and will dig in a bit more.  I also have found some devices similar to the one listed below.  Not much cheaper than the DMM they are using now but the multiple inputs would allow them to get rid of some relay logic being used to share the meter.  Also, since we have a sinusoidal wave then the True RMS isn't really needed, just good old RMS.

https://www.icpdas-usa.com/m_7017rms.html