I made a promise to you that I would not directly comment about your posts. So far I have kept that promise but you have asked for help and I am willing to try.
Legend stock LG28 11" coil settings: You said you were using "factory Park M2" but you changed the tones to 60 (factory is 4), changed recovery speed to 1(slowest) of 10 (fastest), maxed out sensitivity 30 (factory is 25) and used the A (all targets accepted) discrimination pattern. Did you change the iron filter setting from factory default 8 which is one setting from maximum 9? Basically, Park M2 will hit any metal that is in the ground that is a millimeter in diameter or larger and a tenth of a gram or heavier since it is running with the highest weighting towards 39 kHz. So running recovery speed at 1 and iron filter at 8 with sensitivity on 30 is just going to magnify and lengthen the most predominant target signals in the ground which appears to be manmade iron with a little magnetite. You said the Legend basically was not hitting any of the freshly planted (?) 5" deep coins. That is not what I heard or saw.
These are the target IDs I saw in your Legend video for the coins: Quarter-31, 32, 50(correct ID) along with mostly numbers between 1 and 8. Dime- I saw a 43 twice which is a good ID, the rest were below 10. Nickel 18 to 28 which are really close to accurate 26 ID and 31/32 with lots of IDs below 10. Penny-(what kind zinc or copper?) I saw a couple of 29s with the rest of the IDs below 10.
Equinox 800 stock 11" coil settings: Park 1 Multi, 50 tones, recovery speed 1(slowest) of 6 (default is 5), maxed out sensitivity at 25 (default is 20), accepted all targets. Did you change the default iron bias setting of 6 to something lower? Park1 Multi has mid range weighting (somewhere around 20 kHz) so it won't hit the tiniest targets like Park 2 multi or the Legend's Park M2. However, the recovery speed on 1, maxed sensitivity and iron bias on maximum setting of 6 again like the Legend's settings will magnify and elongate the most predominant metallic signals in the ground. You said the 800 "Couldn't hit a one". That is not what I heard and saw.
These are the target IDs I saw in your Equinox 800 video for the coins: Quarter-15, 16, 18, 21, 22 (correct is 30) and there were lots of IDs below 0, Dime-all of the target IDs were below 0, Nickel: 14 (really close to the correct 12/13 ID), 16, 18, 27 with the rest being below 0, Penny 15, 18, 19(if it was a Zinc penny then really close to accurate ID of 20) and the rest were below 0.
Deus 1 with 9" round HF coil: Deep program with 14.4 kHz single frequency, 2 tone ferrous/non ferrous audio (why not use full tones like the Legend and Equinox????), disc on 5.5, sensitivity 90 of 99, iron volume 3 (similar to the Legend and 800), reactivity 2.......the settings are 0, 1, 2, 2.5, 3, 4 and 5 so a setting of 2 is mid range and not the absolute slowest like the Legend and 800 were set on, default silencer setting in Deep is -1 which is the minimum setting as opposed to the maximum settings that Legend and 800 may have been set on. So, even though I personally wouldn't use the Deep program for unmasking shallower targets in iron, it does do okay. At least having a reasonable recovery speed and silencer setting won't iron mask non ferrous targets as easily as the settings that the Legend and 800 were set to.
Deus 1's iron mineralization meter (for magnetite mineralization mostly) stayed on 1 or 2 bars of 12 most of the time and peaked at 5 a couple of times, so your soil's magnetite conditions are fairly mild. My Deus 1/ORX/Deus 2 mineralization meter stays on half full to full most of the time.
You said that Deus 1 hit the targets best.....part of that is due to it being in 2 tones in my opinion with a reasonable reactivity speed 2 and the minimum iron bias setting -1. Did it really?
These are the target IDs I saw in your Deus 1 video for the coins. Quarter non ferrous target IDs were 00 and 98, 99....mostly 00 which will often produce higher non ferrous responses. That is just how Deus 1 works. Correct quarter ID is 94. Most of the non 00 IDs were below 5. Dime non ferrous target IDs were 07 a few times with some 00, correct ID is 86 or so. Most of the non 00 IDs were below 5. Nickel non ferrous target IDs were 00 and 96 (correct ID is 63). Most of the non 00 IDs were below 5. Penny non ferrous IDs were 00, 35, 82 (correct for zinc) 94, and 99. Most of the non 00 target IDs were below 5.
For me in those conditions with freshly buried 5" deep US coin targets using the Legend or Equinox 800, I would set both on 2 tones, keep the recovery speed on a mid range number like 4 or 5 at least, lower the iron filter/iron bias setting to 0 or 1, lower the sensitivity to around default 20 or so, use the same discrimination settings you did with all targets accepted and try the M1 versus M2 frequency weightings and see which one does best. I would expect some funky target IDs and plenty of ferrous/non ferrous tones even with those settings but there would be some accurate ones too and you might be able to isolate the targets using the pinpoint function then scan again when you get the coils exactly centered over the targets and see if the IDs and amount of non ferrous tones improve.