What Difference Does 1khz Have On The Equinox From 5khz To 4khz?

[Jeff McClendon]

4 minutes ago, phrunt said:

Thanks Jeff, yea I'm not sure if that is switching frequencies or if it was something else the way it's called noise cancelling and channels.  I have no clue how it works but I do know the good old fashion frequency shift works pretty good 🙂

 

Somewhat cryptic explanation from the Equinox manual about Noise Cancel.....

Noise Cancel

Noise Cancel reduces the effect of environmental electromagnetic noise from sources such as power lines, mobile phone towers, and other metal detectors.

Detectors may become noisy due to electrical interference from power lines, electrical equipment or other detectors operating close by. The detector interprets this interference as inconsistent, erratic detections.

The Noise Cancel setting allows you to change the noise cancel channel. This slightly shifts the detector transmit frequency to be less responsive to the source of the noise.

Noise Cancel affects both the audible detection noise level and pinpointing performance.

The Noise Cancel setting has 19 channels with a range from -9 to 9 with a default setting of 0 for all Detect Mode Search Profiles.

[phrunt]

Ok, so noise cancel does shift frequency, well that's how it works then, great!  I'm very rusty on my memory of the manual, it's been a long time 😉

So that's how the Manticore noise cancel works too then, and in multi they must just shift the weighted frequencies to clear it up the same as you would in single frequency but just with more complex calculations.  I should have expected nothing less, that's the obvious way it works, it's just their reference to channels that makes it confusing, rather than frequencies.   I guess they just wanted to call each specific frequency a channel and in multi they'd have combinations of frequencies pre-set as channels and it tries these combinations.  Perhaps the Manticore goes a step further and doesn't have specific "channels" and auto adjusts on the fly without using preset channels, so zoning in one of the noisy frequencies the multi mode is weighted for and shifting that while leaving the quiet ones at higher frequencies for example in it's detect mode is weighted for alone until it has a clean combination.

So it's sitting there saying OK my weighted 18kHz is fine, no trouble there but my 5kHz is a bit noisy, so I'll try 4.9, 4.8, 4.7 etc until it gets a good one, then once it does it's ready to roll. 

The Manticore is significantly better with EMI than the Nox for some reason. 

[Digalicious]

12 minutes ago, Chase Goldman said:

Dig - none of us can prove or disprove any of this because we didn't design these rigs.  So if you want concrete answers about what's going on under the hood.  Ask the manufacturers directly.  Who knows, maybe they can answer your questions without giving away state secrets.

Those using the Equinox don't really care how ML made 4 khz quieter, they just know from experience it is, they can run higher gains than 5 khz, find decent targets at depth, and are not seeing any noticeable performance hit (even if there is a slight gain reduction).  Simple as that.

Well Chase, the manufacturers don't even tell us what SMF frequencies are being received and processed, so I seriously doubt Minelab would tell me what they're doing with that 4 khz mode 😁

Anyway, my point has always been that a detector can't know the difference between EMI and a target in the ground. So, if EMI is somehow filtered out (or whatever), then there is going to be a performance hit in another aspect. I don't think I'm alone in my curiosity to find out what exactly that "hit" is.

A good example of that hit, is Rattlehead's "Silver Slayer" program for the D2. It's pretty much impervious to EMI, but that's only because everything below silver / copper is notched out, and EMI typically ID's below copper / silver. So the EMI is gone, but the hit is, every target below copper and silver is gone as well.

[Jeff McClendon]

16 minutes ago, phrunt said:

Ok, so noise cancel does shift frequency, well that's how it works then, great!  I'm very rusty on my memory of the manual, it's been a long time 😉

So that's how the Manticore noise cancel works too then, and in multi they must just shift the weighted frequencies to clear it up the same as you would in single frequency but just with more complex calculations.  I should have expected nothing less, that's the obvious way it works, it's just their reference to channels that makes it confusing, rather than frequencies.   I guess they just wanted to call each specific frequency a channel and in multi they'd have combinations of frequencies pre-set as channels and it tries these combinations.  Perhaps the Manticore goes a step further and doesn't have specific "channels" and auto adjusts on the fly without using preset channels, so zoning in one of the noisy frequencies the multi mode is weighted for and shifting that while leaving the quiet ones at higher frequencies for example in it's detect mode is weighted for alone until it has a clean combination.

I respect the way that Minelab says Noise Cancel reduces the effects of EMI by using slight shifts of the transmit frequency to be less responsive to the effects of the noise.

Words like eliminate, stop, eradicate, prevent, filter out EMI are not included in their description of this function.

EMI is never "gone" unless it was not there in the first place. Whether it seems to be gone because I can't hear it due to the way I have setup my detector does not mean that EMI is gone.

[Chase Goldman]

2 hours ago, Digalicious said:

Anyway, my point has always been that a detector can't know the difference between EMI and a target in the ground. So, if EMI is somehow filtered out (or whatever), then there is going to be a performance hit in another aspect. I don't think I'm alone in my curiosity to find out what exactly that "hit" is.

I disagree with the absolute "can't know the difference" statement as there are probably sophisticated ways to use signal processing and machine learning (assessment) to determine with good certainty the nature of the magnetic field disturbance (i.e., target based, EMI based, or ground based) under normal search conditions and to dynamically change filtering.  I don't know of any specific detector that does this...yet, but I find it totally plausible.  Automatic ground balancing does this to a certain extent by continuously monitoring the level of ground mineralization, and executing a ground grab when signficant changes in mineralization are sensed.  But it can also be fooled by ferrous targets.

But for today's detectors we are instructed to set up specific initial conditions so that basic assumptions can be made (or directly measured) to execute semi-automatic noise cancel (frequency shift/filtering) and ground balance routines.  In the former, the detector "brains" either know (via a sensor) or assume the detectrorist is following instructions and the coil is NOT moving so any magnetic field imbalances must be due to EMI.  Similarly, during ground balance, the machine "knows" or "assumes" you are swinging or pumping the coil over target free ground, so any magnetic field imbalances must be due to ground conductivity.  Either way, we trust the designer not to significantly degrade performance while going about the process of finding a quiet channel or setting the ground reference point, as instructed.

Be curious all you want.  But continuing to ask forum members here "why" or "how much" performance is affected by these EMI and ground noise cancellation routines is just unverifiable speculation precisely because we are not privvy to the specific design details.  

I you want to get down and dirty into this type of technical detail, I suggest visiting Geotech's (Carl Moreland's) excellent detector design focused forum.  https://www.geotech1.com/forums/forum

Carl is a forum member here.  Has written a number of books on detector design that are referenced in Steve's detector resources list.

Cheers

[Digalicious]

12 minutes ago, Chase Goldman said:

 

Be curious all you want.  But continuing to ask forum members here "why" or "how much" performance is affected by these EMI and ground noise cancellation routines is just unverifiable speculation precisely because we are not privvy to the specific design details.  

 

Yes, knowing the technicalities as to what Minelab is doing with EMI mitigation in 4khz mode would be ideal. However, not having those technical details, doesn't by necessity, mean the effects of such are "unverifiable". More specifically, I suggest said verification could be achieved by comparing the  detection characteristics between 5 khz and the different algorithm in 4 khz. 

 

[glacialgold]

1. I have nothing to add here. Still learning, and this is some advanced degree discussion in this thread.
2. Very, very informative to read all the different perspectives.
3. Super-entertaining witnessing what I'd call the "big dogs" duke it out! Yeah, it can get testy at times, but it's fun seeing people energized and pushing the limits.

[Chase Goldman]

1 hour ago, Digalicious said:

Yes, knowing the technicalities as to what Minelab is doing with EMI mitigation in 4khz mode would be ideal. However, not having those technical details, doesn't by necessity, mean the effects of such are "unverifiable". More specifically, I suggest said verification could be achieved by comparing the  detection characteristics between 5 khz and the different algorithm in 4 khz. 

 

So I did that with some quick and dirty air testing with my Nox 800 and the 6-inch coil (happened to be the coil I had bolted on, but it also provided a quieter test bed as I was able to crank sensitivity up to about 23 for all frequencies in my wi-fi saturated neighborhood).  Ran a zinc penny, modern nickel, clad dime, and clad quarter across the coil (air test) at 4 khz, 5 khz, and the dime and quarter across the coil at 20 khz.   The noise situation was such that I was able to run sat noise cancellations on all frequencies to get quiet channels on all 3 frequencies at Sensitivity 23. 

After running this test, I'm convinced that:

1) There is no signficant differences in air test detection depth between 4 khz and 5 khz for each of the targets. 

2) The dime and quarter at 20 khz saw a loss of about 1/2 to 1 inches air depth vs. the 4 and 5 khz air test runs.

Conclusion:  "Quiet" 4 khz on the Nox does not come at the obvious expense of performance versus "Noisy" 5 khz and at 20 khz, all other settings and factors being equal, you take a performance hit on high conductive target depth vs. 4/5 khz, which is not surprising.   

As the resident doubting Thomas here, Dig, I know you are not going to take my word for it and will want video proof.  Well, I leave it to someone else to go to the trouble of convincing you with that objective evidence.  In the meantime, I welcome you or anyone else to share or post results of similar 4 to 5 to 20 khz testing and would definitely be interested in any results that differ from mine.  But for now, I've convinced myself that 4 and 5 khz are basically equivalent in depth performance and I've seen 4 khz excel in EMI situations that shut down 5 khz, and to me, that's all that matters.

Regards.

[EL NINO77]

I think it's visible,, in this test...,,where the 4khz setting works best on a 1 silver dollar coin...

  The tests were done by my colleague Jack... on his test field..

[BigSkyGuy]

Quote

I really don’t know why 4 kHz is less interfered with here in the USA than 5 kHz. Maybe the 60 Hz USA and 50 Hz difference in the UK and most of Europe and Asia has something to do with it.

At the risk of stirring a hornets nest I will throw in my theory as to why 5 kHz is a problem but not 4 kHz. Power lines seem to be a big source of EMI, yet the transmission frequency, as Jeff mentions, is 60 Hz, two orders of magnitude lower than metals detectors. However, I have read that there is a power line carrier communication (PLCC) signal that is transmitted along with the power that is transmitted in the kHz range to as low as 5kH. This is why meter readers are no longer required. The power useage in your home is sent via PLCC. PLCC is also used for home security systems, and home control and automation.

I have used my DetectorPro UW, which operates at 2.4kHz, no problem right under power lines.