Reducing CT Noise

Post Reply
bd4c
Posts: 40
Joined: Mon Oct 12, 2015 9:00 pm

Reducing CT Noise

Post by bd4c »

I have some CTs attached to both the EFIS and also EIS. The sample rate seems to be around .5s (30hz), more or less. Unfortunately, this means the numbers jump around +/- 3 amps (more or less) all the time. This is quite noisy and distracting. Is there a way to reduce the noise (does not need to be an FFT, but anything would be helpful at this point). Both the EFIS and EIS show the same noisy behavior.
GRT_Jeff
Posts: 802
Joined: Tue Dec 11, 2012 12:11 am

Re: Reducing CT Noise

Post by GRT_Jeff »

If they are directly connected to the analog inputs on the EFIS, they have an adjustable digital filter (DF). Go to SET MENU, Display Unit Maintenance, Analog Inputs, and change the "Analog Detail" to "Raw data and filtering". This will show a DF number on each line. You can then use the knob to adjust each DF value. Higher values apply more filtering. The HXr analog inputs sample 10 times per second with no filtering by default.

The EIS does not have any adjustable filtering, and the EFIS does not have a way to apply filtering to data it receives from the EIS. The EIS aux inputs sample around 5 times per second.
bd4c
Posts: 40
Joined: Mon Oct 12, 2015 9:00 pm

Re: Reducing CT Noise

Post by bd4c »

Thanks Jeff.

Is there a relationship between the settings under the DF and accuracy? I got the jitter eliminated and set the offsets to zero out the CTs when not loaded (these CTs are measuring the alternator/generators), but I somewhat doubt the values toward the high end of the reading. For instance, zero indicated at no load, and then showing 58A when more or less fully loaded. However, the motor-generator is designed to supply 50A peak, and I suspect the load is only in the 40-45A range. I'll verify actuals (somehow) later, but the suspicion is enough to prompt the question for now.

Also, is there a way to flip the indicated polarity, without flipping the CT itself?
GRT_Jeff
Posts: 802
Joined: Tue Dec 11, 2012 12:11 am

Re: Reducing CT Noise

Post by GRT_Jeff »

The DF is just a simple low-pass filter in software. It doesn't change the way the input is sensed. If there are a lot of noise spikes in one direction then the filter would bias in that direction.

The sensor input can be flipped if you switch the Forward/Reverse (+/-) sensing and then recompute the offset. For example our 100A sensor in 100A mode uses Forward sensing with scale factor 164 and offset 313. If you switch to Reverse sensing with offset 341 then you will get the opposite value for the same input voltage.
Post Reply