Powergrid noise with 25Hz component?

[Karel]

Hi,

I received an ECG recording where all leads show strong (powergrid?)noise.
I have seen this kind of noise before but the base frequency componnet was always 50 Hz.
This recording however, has a strong 25 Hz component and I can't understand where this is coming from.
The recording was made on a remote location and I wasn't present so I don't have much info.
The ADC used is a delta-sigma type with a modulator frequency of 512 Khz and an output data rate of 1 KHz.
The system had a multiple-notch filter active with notches on 50 Hz, 100 Hz, 150 Hz, etc.

Any ideas where this 25 Hz came from?

[NiHaoMike]

Might there be some kind of load that's turning on every other cycle? Most likely it would be a solid state controlled heater of some sort.

[RoGeorge]

ECGs usually have adaptive filters to remove the 50Hz, something like this



so most probably you see "strong" 25Hz only because the adaptive filter was very good at removing the 50Hz, or maibe those 25Hz are in fact the adaptive filters' artifacts, IDK.

[TimFox]

Might there be some kind of load that's turning on every other cycle? Most likely it would be a solid state controlled heater of some sort.

I have seen interference at subharmonics of the power line:  in my case, a 10 Hz frequency with a 60 Hz line due to a soldering station with a zero-crossing switch in its servo connected to the same outlet strip as the device under test.  Resting in its holder, the thermal servo apparently cycled synchronously due to the zero-crossing switch.

[Vovk_Z]

or maibe those 25Hz are in fact the adaptive filters' artifacts, IDK.

Yes, it depends on how large is a relative level of those 25 Hz components on a TS spectrum. He shoved us absolute levels (uVolts) so it may be anything.

[Karel]

The multiple notch-filter is not adaptive. They are IIR filters with fixed notches at 50, 100, 150, etc. Hz.

[Benta]

Single-wave rectifier somewhere in your grid, perhaps?

[ejeffrey]

A common way I have seen power line subharmonic noise is actually from microphonic/vibration pickup from motors.  The most common AC induction motors spin at just under half the line frequency where the slip rate depends on load.  The clear sign of this is that the signal won't be at 25 Hz but at ~24.9 Hz.  This can be hard to see on a spectrum analyzer, but if you can amplify the signal enough to look at it on an oscilloscope then you can set the scope to "line" trigger.  If the 25 Hz signal is stationary it is a true subharmonic, if it is slowly scrolling to the right it is mechanical vibration from an induction motor.

To look at it on a spectrum analyzer you need to be careful.  The long term average of the line frequency is accurately 50 Hz, but there can be short term variations.  First zoom into the 50 Hz peak and see if the peak is split. There will always be an electrical 50 Hz component but if you have a second peak just under 50 Hz that could be a mechanical vibration. If you don't see a split peak measure the frequency of the peak to .01 Hz before switching to the 25 Hz peak and check if it is exactly half of the 50 Hz value.

[Siwastaja]

Single-wave rectifier somewhere in your grid, perhaps?

How would this cause a 25Hz component?

[TimFox]

I would think a zero-crossing switch in a servocontrol, as in my example above, would be more likely to produce a subharmonic component on the power line.
Mechanically, it would have to be a shaft running at half the line frequency, such as an 1800 rpm nominal motor on a 60 Hz line.

[ejeffrey]

1800/1500 RPM motors are pretty common especially in industrial settings but also things like HVAC blowers.

Another possibility is a VFD.  In principle a VFD can run at any frequency and do so while ramping up.  However, the ones I have seen in HVAC systems don't actually use continuous control range for steady state operation but run at certain fixed frequencies.  Common options are 25%, 50%, 75%, or 100% of line frequency and they cycle between those to track demand. A VFD set to 50% will generate 25 Hz power line noise and its harmonics directly.

[NiHaoMike]

Another possibility is a VFD.  In principle a VFD can run at any frequency and do so while ramping up.  However, the ones I have seen in HVAC systems don't actually use continuous control range for steady state operation but run at certain fixed frequencies.  Common options are 25%, 50%, 75%, or 100% of line frequency and they cycle between those to track demand.

There are some retrofit BLDC motors for HVAC equipment with individual inputs for different speeds but those are very uncommon. Pretty much any remotely modern HVAC system that has variable speed motors is going to have very fine steps in what speeds are available.

[ejeffrey]

So I'm not sure exactly how common it is, and I try to avoid learning any more than I have to, but we have a few systems, a chilled water system and a air handling system that has some VFDs.  In manual mode I can set the VFD to whatever frequency I want, but the way the system is programmed overall it actually runs at a few speeds.  For instance, 90% of the time the circulation pumps for the chilled water loop run at 30 Hz, but under some conditions it will bump up to 60 Hz.  It's also possible (likely actually) that our systems were programmed by idiots, so maybe this isn't normal, but it can happen.

[salihkanber]

It seems that low f Harmonics is a common thing in ecg or eeg measurements;

https://openbci.com/forum/index.php?p=/discussion/2472/25-hz-noise

[RoGeorge]

At a search for 25 Hz mains sub-harmonics also found:

A generator with a worn bearing can produce sub-harmonics. Instead of simply rotating, the shaft rolls around in the bearing.

and this:
https://www.erlphase.com/downloads/application_notes/Understanding_Sub_Harmonics.pdf