Part 107 UAS pilot reported a flyaway that was due to an unknown aircraft issue or possible signal loss. The UAS was recovered with minor damage.

Narrative

DJI M30T Flight Log and Crash DRAFT Day 0 (XA00 - XC00); Person A written Day 1My Home Point for this training mission was the west side of the horse trailer parking lot at Location A. There was only one other car in the parking lot all morning. The weather was very good. Clear blue skies; 42 degrees; and just a small amount of wind around 5-10mph. My plan was to look for subjects in a forested area to determine the best height and sun angle for RGB visual detection. My subject was about 100 yards away from the home point up and over a small hill and in some trees. I checked my drone for proper arm extension and lock and propeller condition. All was good. I checked the controller preflight screen for height and distance and changed the height in case I needed to fly higher than 400' AGL to go up a mountain. I set it for 600. I glanced at the low battery and signal loss behaviors and did not notice my signal loss was set to hover. I knew; having been up here last time; that there was RC signal loss behind some of the close-by hills but I made a mental note to keep an eye on the RC signal and just fly higher if it became a problem - as I did last time It was nice enough to not have to be inside the vehicle at all so I got my chair out and sat in the shade of the vehicle so I could see the screen clearly and brightly.After getting the coordinates of my lost subject and dropping a pin on the controller; I launched the drone to my first search altitude which was 150' AGL. I documented last time that this was sufficient height to clear the nearest hill and trees and maintain visual line of sight. I was going to flew to the pin using the wide camera. I was able to see the subject almost immediately upon leaving the parking lot. He was under a tree but had on a brightly-colored jacket and was not shaded by the trees. He was a little off from where the pin was - about 20 feet - so I created a second pin for his actual location. I settled on a distance away from the subject to make my circles. I switched to the zoom (zoomed out) to see the subject clearly but also to be able to see some of the surrounding area; I guessed the circle had a radius of about 100'. I flew around the subject manually one time and determined that except for when he was blocked by trees; he was very visible from all angles. It was actually better when the drone was NOT directly between him and the sun because the sun made the brightly-colored jacket very bright and almost washed out. I was able to mostly correct this however by decreasing the exposure to make everything darker but it was interesting to observe that with the camera in auto he was too bright. I then tried the Point of Interest (POI) feature and flew around the subject a couple of times to see how this worked and it worked very nicely. I was manually adjusting the speed of the circle by using the side movement stick. I noted that the actual POI center point was NOT where the pin was unless the pin and the center of the camera were the exact same spot when I activated the POI. At least this is what I thought based on my observations of the drone behavior when I initiated the POI. I flew both clockwise and counterclockwise a couple of times to get a feel for everything and making observation of the surrounding area and what I was seeing. I was not paying attention to the RC signal after the first five minutes or so because it never dropped below full bars. The subject switched to camo clothing as I brought the drone home to change batteries.Everything was going very smoothly up to this point. I hot swapped the batteries and flew back to the coords to begin searching. Again I set my altitude to be 150' AGL from the home point. I could not see the subject at all this time with wide; zoom or thermal. I flew to within about 100' of the pin and circled while I tried different camera settings to no avail. The subject began moving to different spots and I was finally able to pick him out. I had to zoomin a couple of steps to consistently see him well - maybe 5X. I had to add a new pin because the subject has relocated to a different set of trees. I had trouble adding a new pin for an undetermined reason. I also was not able to figure out how to make the POI start at his exact location - it kept wanting to fly around a point that was about 10 feet to the left of where he was standing. I gave up so that I could focus on determining the best angle. I made at least one circle and I'm pretty sure that the center point of the POI was not holding steady. I was only able to determine that at any angle where the subject was not blocked by tress we equally effective at being able to see him. As I was trying to confirm this by continuing to circle using the right stick pushed full left the drone took off and I lost RC. It is very possible that I was changing positions slightly of where I stood but I guess the radio signal was getting blocked by the small hill to the south of where I was standing. Panic set in. The subject confirmed that the drone had moved rapidly to the south and was quickly out of his sight as well. I tried everything I could think of to get RC back but I was never able to re-acquire. I finally checked my setting for RC loss and found it set to Hover. Based on the fact that the drone was one-half to three quarters of a mile from home I don't feel it went into hove mode very quickly but this is just a guess because I was not paying attention to the RC strength and when I actually completely lost the RC signal. Better one half mile that one or two miles AND it looked like - or I was strongly praying - that the controller had the final location of the drone. In my panic and confused state and not able to think clearly I put my faith in the controller and began walking toward the drones last plotted location that it was able to communicate to the controller which - without RC - I felt could be a fair bit off of actual. I had enough sense to lock my vehicle and get the important things inside before taking off. The subject had radioed that he would be leaving his position to walk in the direction he saw the drone last heading. I caught up to him and continued to follow the controller screen. It look bad as we were heading right for a much higher hill in the distance. I was convinced the drone had crashed into the rocky peak. As we got closer it looked like the drone had flown around the peak and was on the other side. We tried to hike up and over the peak but it got steep so we worked our way around the east side to a saddle and the map said I was close and the drone was a little more south and west. I got to where the controller and drone were on top of each other and took a break. During the break looked up and saw the drone directly in front of me about 10' away and partially buried in a patch of crusty hard snow in an 20' diameter area clear of trees. I realized at this moment that I did not bring my phone and could not take any pictures. CRAP! The left rear propeller was completely buried in snow down to the ground but all the other quadrants look pretty good and especially the camera. The camera was up high on the snow and appeared untouched which it still looks pristine! I carefully lifted the drone out of the snow and set it on a clear spot while I considered what to do next. Also did more damage inspection. The right front propeller had the only other damage of having just the tip broken off. The left rear major damage was one propeller broke off (and recovered from the snow bank) and the other prop on this motor had a very small piece broken off near the tip but otherwise was attached to the motor.I did not have a camera or my phone so I was not able to take any pictures of the crash in situ.I am going to guess that the drone flew to this spot and began it's hover. When the batteries got low it tried to land. It got to it's landing height and didn't land because the bottom vision sensor did not know how to interpret the snow and it just continued to fly until the battery reached critical low and it did a force landing in the snow and for some reason it was not a very controlled landing. The ground and snow were not perfectly level but my opinion is that it should have been able to land gently while also level and not made this significant of an impact. It landed ass first and both batteries were ejected but still inside the drone - one almost fully inside and the other half slid out. This was battery pair set 2. Upon testing it has one solid light and one flashing light. I would have expected only one flashing light. For comparison; battery set 1 which was the first used also has the same battery light flashes.I powered up the drone this morning and all systems checked out. I ran the cameras through their paces and they all worked fine.I will be initiating a repair ticket today.After checking the flight log for the crash flight it turns out that after I launched and flew to the subject; and hovered trying to pin his location and having trouble with the pin; the drone made about a half of a rotation in the manual POI mode before flying off into a much larger loop and losing RC so... I am not sure if the loss of RC was the root cause (probably) or the fact that I didn't set the POI correctly due to the pinning issue or some other function for setting POI's.I did not have a sterile cockpit due to having radio comms with lost subject. There ended up being non-essential comms when the flyaway initiated.HP elevation 8090MSL Crash elevation 8200MSL per Google Earth - does not include forest height.Just shipped M30T off for repairs. Day 0 @XD45hrsThere were several things I didn't do once I realized I lost comms with the drone; and that it was set to hover for lost RC rather than RTH. I could see that I may have the crash or hover location. I considered that; when the batteries got low; the drone may initiate RTH and come home and I also may re-acquire the RC. I waited around the area for a while (5 - 10 minutes) and maybe even long enough that I knew the batteries must have gotten to the low point and if it was coming home; I should have seen it by now. At some point I decided the drone was most likely not coming back so I started off in the direction I saw on the RC. I did not take food; water; phone; or GPS; but I did have the RC in a harness on my chest. Many scenarios were playing out in my mind as we got closer to the Last Known Position (LKP). All I cared about was not losing a good drone. I was hopeful that there would not be extensive damage and that maybe it was able to initiate a controlled landing near the LKP before losing power. I just finished recovering the corrupted video of the drone's last flight and it clearly shows the drone flying away and ultimately flying into trees. It never went into hover mode after I lost control. I didn't feel the damage supported a high speed crash into trees but it is clear to me from the video that this is what happened. It also explains why the batteries were not more depleted. It should have stopped and hovered at about a minute before crashing because it should have detected the loss of RC but did not. I will be looking at the flight log more closely... Day 3 Having thought about this and having slept on it I am now entertaining the possibility that this was at least partially a flyaway situation and a failure of the drone to detect loss of RC and initiate the hover.I am trying to find a way to convert the downloaded flight log file from the drone into readable data. When you download about 8 months worth of flight logs (because you have never done this before) you get a .DAT file that is 3.5 Gigs in size. There is software that will convert .DAT files but nothing for files over a gig in size. What I would specifically be looking for is the timing around the loss of RC - if it was detected - and the crash. I looked at the data from the controller that showed the GPS breadcrumb