Read through your complete thread above again in detail. My takes -
The issues you reported on your first machine sounded pretty serious. You were right to get it replaced.
As I look closer at the second machine I can summarize the issues as follows:
The double press to get back to the detect screen from "settings" following a noise cancel. You also described lock-ups and freezes, which does not actually seem to be the case as you can readily get the screen to "unlock" with a second button push. Lock-ups typically mean the screen is unresponsive to any button pushes and are cause for concern, so use of the terms "freeze" or "lock-up" may be overstating the actual "issue". I now know how to make the double press issue repeat. See below.
Chatter when you lay the machine on the ground during target recovery (this can be "expected" during certain circumstances - I will explain).
Multiple Noise cancel attempts resulting in different channel numbers on back-to-back attempts - again this can be explained and is not unexpected.
First of all - I will say that it is hard to diagnose your issues merely by reading words - multiple interpretations of terminology can lead to false conclusions by those of us who can't see what is really happening first hand. I may be totally off base because all I can go by is your description of the issues. That being said, I will continue with my thoughts.
You stated that ML repair only replaced the control head. Why did ML think only the control head was the issue? What you may be experiencing with your first unit (and possibly the second and third) is some sort of control head damage caused by the coil. Unlikely, but possible. Swapping coils after you have experienced the issue may not be conclusive because after the coil causes the damage, it may look like the problem is independent of the coil because the damage to the control head has already been done. Make sure ML repair has an airtight case for not replacing your whole detector, coil included.
Assuming your coil is not the issue - the issue you are describing about multiple button presses appears to be somewhat normal behavior. I will say that I normally exit the settings menu by pushing on the detect mode select push button rather than the detect/pinpoint toggle button. The manual allows for either. I prefer not to unnecessarily cycle the detect/pinpoint button because it is probably the most used button on the detector and is probably most subject to wearing out over the detector life time, I suppose. Regardless, I figured out how to get the behavior you describe to repeatedly happen and I know how to make it not happen.
When you do a noise cancel - you will see the progress bar fill up the outer circle of the display that shows the discrimination pattern and graphic vdi. If you try to go back into detect mode while that progress bar is still showing, then you will have to press the button again to get back to the normal screen. If, however, you allow that progress bar to disappear which it does on its own a few seconds after completing the noise cancel, then you can go right back into detect mode on single button press. This is apparently normal behavior. Try it yourself and see. I also realized after doing this, that I usually go right into doing an auto ground balance after the noise cancel and the machine does the same thing when pressing the settings toggle to go to the next setting adjusment (ground balance is after noise cancel). As long as the noise cancel progress bar is showing, then I have to press the settings toggle button twice. If I just wait a sec and let the bargraph disappear, then I only have to press the settings button once to advance to ground balance. Bottom Line - Just wait a second and let the noise cancel bargraph disappear and you can save yourself an extra button push.
Chattiness when setting the machine down. If you set the machine down, it is best to set it down such that the bottom of the coil is face down on the ground rather than tilted on its side and or back with the bottom of the coil facing up into the open air. It then acts like a radio antenna and can more easily pick up interference. The other thing that can happen is if you are wearing a cell phone, and you set the detector down for target recovery, then if your cell phone gets near the control head then you can also cause EMI from the phone being picked up by the control head (not the coil). So keep your cell phone away from the control head or shielded from the control head by your body (by keeping it in a pocket or pouch away from the detector swing arm side). HTH. Don't repeatedly noise cancel to make this chattiness go away. Only noise cancel if you are getting chattiness while you are swinging the detector (even though you do the noise cancel with coil stationary).
Why do you get different noise cancel channel numbers when you do repeated noise cancels? First of all, noise cancel is not really a foolproof and powerful tool. You should use it, but don't expect miracles as far as noise cancellation is concerned. Super noisy sites might require more drastic action than noise cancel such as reducing sensitivity and even going to single frequency. As far as the different numbers are concerned, that is not unexpected. Noise cancel chooses what it thinks is the quietest of 19 different channels from -9 to +9 (0 is included). If there is not a lot noise, chances are that more than one channel may be quiet. In that case, noise cancel will probably randomly pick any one of the several quietest channels and if you run the noise cancel again, it will pick another quiet channel. This doesn't mean noise cancel is randomly guessing, it just means that there are several equivalently viable quiet channels to choose from. On the other hand, you may encounter a really bad EMI site where none of the 19 channels are really quiet and in that case you need to take the additional, less ideal measures I described above (lower sensitivity, go to a single frequency - higher single operating frequencies like 15, 20, and 40 khz tend to be quieter but they also affect depth because higher frequencies do not penetrate into the ground as much as lower frequencies - no free lunches - detecting is all about tradeoffs).
So what you are describing appears to primarily be expected behavior. If you are having extreme lockups or constantly random, worsening noise or erratic screen behavior, then, yes, you should send it back again, unfortunately.
HTH
