Mains hum

I do not quite subscribe it.

The ferrite beads do not attenuate the 50 or 60 Hz hum from the mains.

Do you have a reference electrode on the patient?

Since these beads do not attenuate 50/60Hz, you probably had a problem with mains modulated interference being rectified in the input circuits of the amplifier. You need to add some RF filtering at the input, before any non-linear elements like transistors or op-amps.

Arie

That is true, but we do not have contacts to the low impedance source. The electrodes can have impedances of tens of kohm, with mismatch between the electrodes. We do not hit contact spikes through the skin.

A notch filter is not a good idea, as much of the interesting frequency components are in the mains frequency range. The ECG signal is spiky by its very nature, and narrow band filters will spoil the information looked for.

Didn't the original post say that the problem occurred with battery power? Sounds like you've found PART of the solution, not all of it. Ferrite beads don't do much at power-line frequency.

The body is still low impedance with respect to the amplifier inputs. They use paste and a large contact area to minimize skin resistance. Given the high input impedance of the amp, the mismatch in impedance is of little consequence.

You seem to be confusing a notch (band reject) with a narrow band-pass filter. An adequately narrow band reject filter would have virtually no impact on the signal of interest. It only need be wide enough to accommodate the normal variations in mains frequency.

I'd like to see what the output looks like with various inputs, all contacts shorted, resistors between contacts, etc. If the problem really has disappeared, I expect it will return and not by removing the ferrite beads.

Leaving aside that the problem has now been resolved, I would have thought Tauno was right in what he said. The wanted and unwanted components of the waveform are just too close together to target one without degrading the other to some meaningful extent.

Ricky wrote: ==========

** Shame you have never looked at the phase and amplitude curve of such a filter. ** About 0.1Hz is all the AC supply varies by.

...... Phil

I think that when OP tried battery power he was still examining the output on a mains connected scope? Interesting would be what is seen with a battery powered scope and the whole system free of any connection to mains ground.

No sane amount of ferrite beads will have effect at 50Hz, much more likely some tens/hundreds of kHz from a mains SMPS in the vicinity is getting in. The ECG amp will probably be heavily low pass filtered somewhere along the chain so the HF interference won't be seen at the output but it's 100Hz modulation rate will be?

piglet

snipped-for-privacy@hotmail.com wrote: =======================

** That is my contention too, posted 2 days ago.

The ECG amp will probably be heavily low pass filtered

** Huh ? 100Hz ?

Think is is just simple capacitive coupled injection of the AC supply voltage wave. The bane of many poorly shielded electric guitars .

...... Phil

Yes, but later on the OP said he solved the problem with ferrite beads. He hasn't said how many beads he used but it seems unlikely he used enough ferrite to attenuate mains frequencies directly so the theory now is that he instead attenuated some higher frequency pickup that was itself power line modulated?

piglet

snipped-for-privacy@hotmail.com wrote: ========================

** Should have tried Hippy beads or Rosary beads. ** Nor how much weed he smoke daily.

** I can see what looks like 50Hz in that slow ECG trace.

Try counting the wiggles.

..... Phil

With the timebase unhelpfully blurred out I didn't before count the wiggles to see if 50Hz or 100Hz but tried just now using the heart rate as a guide and agree with you - it looks like it is 50Hz direct mains pickup. Thanks Phil.

piglet

snipped-for-privacy@hotmail.com wrote: ========================

** From the OP's first post.

" This *is* mains interference as it works out at exactly mains frequency. "

...... Phil

Phil Allison snipped-for-privacy@gmail.com wrote in news: snipped-for-privacy@googlegroups.com:

Yep... practically wordlwide the electrical grids sport near zero variance with regard to frequency.

Right!

A narrow notch rings as badly as a narrow peak when fed with a spiky signal.

We did try all kinds of filtering already half a century ago, and the basic mathematics, physics and electronics about the filter has not changed.

snipped-for-privacy@decadence.org wrote: ================================

** No it does not.

** The AC supply has a fair amount of 3rd and 5th harmonics that are often more audible that the fundamental. But SFA 2nd. Magnetic hum fields radiated from transformers is mostly 3rd to.

..... Phil

Second harmonic is the acoustic (magnetostriction) output from transformers, and the light from fluorescent fixtures (or used to be; ballasts are more frequency-rich nowadays). Ripple voltage on a fullwave rectifier, also.

If you apply enough gain, you'll see lots of sneaky signal inputs. Microphonic tubes are /were a major concern, and that's why old HiFi preamps were on a different chassis than the power stages.

That's simply not true. You can design a filter to satisfy the requirements. It's not like the signal of interest has a critical component at 50/60 Hz. Look at the waveform, or better yet, look at the frequency spectrum. There's nothing important at that frequency. Tauno also spoke of it incorrectly, referring to "narrow band filters". What is called for is narrow band reject which is not at all the same thing and will not disturb the signal overall, unless there is a critical component at the same exact frequency, which is not the case.