The humbug doesn't approximate a linear filter. It generates a "fit" to the incoming data, and subtracts. Zero delay, zero phase shift once the fit is 'close enough'. Of course if the higher frequency noise (typically
60Hz+harmonics) isn't consistent, this approach fails miserably. Slow drift isn't a problem as the fitted function is continuously updated.** That post was under the name of "Ken Morrow" who then wanted a 60Hz notch filter that had low phase distortion. Posted on 12/2/14
.... Phil
On Dec 19, 2016, Frank Miles wrote (in article ):
While Digitimer doesn?t say, this appears to be modeled on a classical Adaptive Sidelobe Cancellation algorithm from the radar world. There is a huge literature. Cancellation can be done using analog hardware, but digital works far better.
Here is an example:
They use an antenna on the box as the sidelobe receiver, which may not be all that close to the sample under investigation. If I were implementing such a system, I?d use an extra ekg/eeg probe attached to the sample, but well away from the source of the signal of interest, to get the most accurate sample of the interference possible.
Joe Gwinn
No, this is not what I am doing at present.
But I do like to work with ELF's and always keen to learn more about overcoming the limitations of analog filters from those who have been kind enough to assist.
Kevin Foster
Based on the responses so far, I will try an elliptical LPF, or a two pole Bessel followed by the gyrator notch circuit that Jim kindly provided.
Here is a relevant article. See figure 11:
The Humbug looks great, but at $1,700 it is beyond my budget.
Kevin Foster
On Dec 20, 2016, Joseph Gwinn wrote (in article):
I had an afterthought - I recall that HP designed an EEG with a reference electrode attached to the patient?s leg, well away from the head. The human body is a leather bag filled with dilute salt water, so external interference has the same effect throughout the body, and can be removed by a differential input at each non-reference electrode. No phase shift required. I recall that HP published a big article in the HP Journal.
.A little searching dug up the article by H. Ronald Riggert in the June 1967 issue of the HP Journal. This article shows a driven-leg circuit, there the patient?s right leg is driven to be a virtual ground, mostly for safety, but also for the reduction in hum and noise.
US Patent 5,382,956 talks about the input IC that HP used in ECGs (EKGs) et al, which has a reference to an article describing the driven right leg circuit.
The September 1985 issue of the HP Journal also has a relevant but high-level article.
Joe Gwinn
The last ECG I had, I believe there were at least _7_ electrodes, all the way from chest down to abdomen, plus a couple on my sides close to the back.
I do recall that the sticky pads don't come off pleasantly from a hairy person such as I :-( ...Jim Thompson
I was in the hospital a few weeks back and must have had ten EKGs (ECGs are different). Each EKG needed its own set of electrodes, evidently, and none of the techs took off their sets or the last tech's. I was picking them off for a couple of days and it took a week to get all the stickum off.
They don't come off "cleanly" at all. They're covered in goo that laughs at water and soap. Alcohol will take it off but I would have had to take a bath in it to get all the goo off.
The full set is 12 electrodes.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required