DJI FPV Range Testing - Pushing the system to the limits on different setups

[Music] so this is the dji fpv system in 2018 it revolutionized the fpv industry by providing a real-time hd fpv link that looks stunning in comparison to well everything else available on the market but is it any good for long range is that any good for long okay hold that thought looks like we might have a problem here so situation update i've flown out about 10 kilometers only to realize that yes that my transmitter battery is low and we're on the return home trip now only a couple of kilometers out but things are really starting to get dicey you can hear the radio clicking in the background with the super super low battery and i don't want to leave it in return to home because it's not a great location if this thing tries to auto land here so fingers crossed that we actually do get back home hey guys what's going on so it's been a minute since i've uploaded any range testing videos and the reason for that is because i've been super busy lately with this last lot of testing with the dji fpv system now as some of you may know this testing was requested by joshua bardwell who very kindly sent me out a pair of v2 dji goggles a couple of months back and these results been taking a little bit longer than i would have liked to get published mainly because i dumped the first dji wing into a muddy lake as you saw in the intro it was a silly mistake i was testing on 1200 milliwatts and got out to about the 10 kilometer mark only to realize that i hadn't charged the batteries in my t16 so i started getting transmitter battery low warnings turned around got back to about the 500 meter away from home mark and then completely lost the transmitter and ended up dumping it into a lake missed it by that much it was an expensive lesson to learn and in hindsight it's a pretty obvious thing you should have on your pre-flight checklist for long-range flying but that's my excuse for taking so long to get this testing out but it's here now and in tonight's episode we're going to be checking out some long range testing with the dji fpv system [Music] now there'll probably be a few people watching this and having a bit of a laugh because i said the words dji and long range in the same sentence and the reason that that's a bit of a joke is because the dji system has a hard cutoff of 13.3 kilometers and when you're talking about proper long range you're talking 30 40 50 kilometers and in comparison 13 kilometers is a bit laughable and it also raises the question if you already know the range limit is 13.3 kilometers why bother range testing it right and you would indeed be right so in order to make this testing useful i've approached it from a slightly different way so rather than just focusing on the all out max range which we already know i've tried to make this as useful as possible for guys looking to use the system for some mid-range mountain surfing or cinematic type stuff where you're out at like the five six kilometer mark or even as a rough comparison for the freestyle guys who are looking to compare how the system performs with different antenna and power combinations so some of the things we're going to look at in this episode are how does the stock out of the box setup perform dji marketing material claims 0.7 kilometers for the ce region and four kilometers for the fcc so can we achieve these results in real world testing what's the usable range versus the hard cutoff range so the dji system is pretty good at keeping the link alive but after a certain point it all just becomes this blocky mess of color so what's the soft cutoff versus the hard cut off we actually lose the link is there any range difference between 25 megabits and 50 megabits in the past there was but is that still the case now how much better does the range and image quality get with some good directional patches for example with the iflight crystal and this is especially important in situations where you're using low power like for racing or in the eu and finally which goggle antenna connectors are best for patches and why but wait and i'll ramble on a bit towards the end about why the 13 kilometer hard cut off kind of makes sense now as i mentioned earlier this testing wouldn't have been possible if it wasn't for jb getting in touch and sending over a pair of v2 goggles out of his own pocket so the least i can do in return is recommend that you go check out his channel using the link in the video description and i genuinely do appreciate the support and i also like the way that jb has been kind of reaching out and supporting other channels like mark spatz from uav tech and ian from mads tech to name a couple and kind of helping to gather the best information possible in order to make it as useful as he can for you guys so keep doing what you're doing jb now before we jump into some flight dvr let's quickly take a look at the test setup that i'm using for these range tests so on the goggle side it's pretty straightforward we've got a pair of v2 dji goggles with the stock standard on the antennas and as i mentioned later on in the video we'll also take a look at the performance we can get using the iflight crystal patch as well as with some larger high gain patches on the air side we're using the ar pro wing as the test aircraft which has been fitted with the cadx air unit micro which should be similar in performance to the original dji air unit just with a slightly different camera i'm using the standard supplied dji antennas on the air unit and i've got them mounted out to the sides here to provide a good amount of separation between the two and i've just zip tied them to the foam hatch to keep them secure technically the proximity to the foam here might have a slight impact on the antennas but i feel like the effects are going to be negligible enough that we can just ignore them now in order to measure the actual max range of the dji system i've got a gps on board which will tell us how many kilometers out we are before we drop the video link but as many of you know the dji goggles don't record the on-screen osd information so when you play back the dvr later you've lost all that on-screen gps data now there's a number of ways that you can work around this and what i've chosen to do here is to add an additional analog fpv cam to the wing so this cam has been attached via the flight controller to a runcam onboard dvr module and this means we can record all the usual inapp osd information for example the distance from home and then we can overlay that on the hd dvr in post so that we can see the distance we're at when the link drops okay so with that out of the way let's get into some testing and the first thing i wanted to test was the stock standard out of the box setup so the stock on these on the goggles and the stock on these on the air unit now the dji marketing claims an expected range of 0.7 kilometers for the ce region and four kilometers for the fcc and these obviously refer to the 25 milliwatt power limit for the eu and they're probably referring to the 700 milliwatt power for fcc but we're going to test this one on the 1200 milliwatts fcc hack because i assume that's what most people are going to be using so i've got a couple of power levels to check here and this is usually the point in the video where i'd put the vtx onto the immersion rc power meter and check that the power output is correct but there's a problem with that so unfortunately i'm struggling to get any sensible power readings from the vtx using the immersion rc power meter so for example when i set the vtx to 25 milliwatts i only see three milliwatts on the power meter and yes i did make sure i wasn't in low power mode by arming the fc before i took the reading and i have a suspicion that this might be due to the way the system modulates the over-the-air signal so dji is sought to be using ofdm modulation for the rf signal between the air unit and the goggles and i'm far from an expert on how ofdm works but from a quick google you can see that rather than using a single frequency like a traditional vtx does it uses a whole bunch of sub carrier frequencies at equal spacing that you can see here which basically allows you to send more data in parallel lanes rather than in a single data lane like you'd get with a traditional analog vtx i don't know how this type of modulation is going to affect the way the power meter is trying to read the signal and i suspect we'd probably need a professional vna to properly measure this signal accurately and i don't have a cool 40 grand laying around to blow on a vna so we're just going to have to take dji's word for it that the power levels are correct the other thing that we want to check here before we head out for a fly is the tune on the stock antennas so let's get these on the analyzer and take a look okay so you can see we're looking at the reflected power here for the stock omni goggle antennas and you can see that we've got a dip here at 57 20 megahertz but it's not really that low at only negative 8.5 db really you want this to be as low as possible which means the antenna is reflecting back less power and instead you want that rf energy transmitted out into the air so this plot basically tells us that the stock goggle on these will be best at 57 20 megahertz and they aren't really that great overall now unfortunately i can't test the antennas on the air unit because they use an mmcx connector and i don't have an adapter to connect that to my analyzer that's probably enough stats for now so let's set this to 25 milliwatts and then head out for a fly all right stock antennas 25 milliwatts output power wire 50 megabits [Music] all right so there's a couple of things that we're interested in here one is obviously the maximum range so the point where the link drops completely and we're left looking at a blank screen but the other thing that i think will be more useful to people is the point where the image quality gets so bad and the latency starts to get so high that the link just isn't flyable anymore so basically the point where you'd probably turn back because it was just getting too hard to see any form of detail [Music] okay so we've just passed a 2.2 kilometer mark here and things are really starting to just become a blocky mess the latency isn't too bad at the moment definitely still flyable if you're doing some acro but you wouldn't want to have to avoid a tree or a hill or anything like that at this point the land is basically just one big dark green mass at this stage so let's mark that 2.2 kilometers down as our flyable range and continue on to link drops there's no way you could fly in this it's just a stuttery mess oh there it goes okay so on 25 milliwatts with the stock antennas we made it to 2.2 kilometers for the flyable range and 4.2 kilometers before the link dropped completely which obviously exceeds dji's marketing claim of a 0.7 kilometer range by a significant amount which is refreshing to see it also means we're pretty much guaranteed to make it past that four kilometer mark on the fcc mode since we can already make it past four even on 25 milliwatts but before we crank up the power and test that theory there's one more thing that i wanted to check so you might have noticed that that last flight was done on the 50 megabit setting and some time ago there was a known range difference between the 25 and 50 megabits rates so i'm keen to see if we get the same 4.2 kilometers range on 25 megabits if we leave the power set to 25 milliwatts so let's quickly head out and fly that and then we'll move on to the high power test all right dji goggles on stock antennas 25 milliwatts output power 25 megabits [Music] yep and the bit rate's starting to get up there now any second it's going to go yep and there it goes okay so 3.2 kilometers for the flyable range and 5.1 kilometers before the link dropped on 25 milliwatts and 25 megabits which is an interesting result because it's almost a full kilometer further than we got with the 50 megabit setting which is a bit weird because from what i understand it used to be the other way around so 50 megabits was the one that used to go further rather than 25 megabits and there could be a couple of things that could explain this first of all dji could be using different rf modulation parameters between the 50 megabit and 25 megabit modes which could have an impact on the range and second we know that 50 megabits uses a wider bandwidth on the spectrum than 25 which is why you only get three channels when you switch to 50 megabits mode and this could put us in a different part of the antenna tune which could then have a negative effect on the range but i'm probably getting a bit carried away with speculation now so moving on let's crank up the power to 1200 milliwatts for the fcc range check on the stock setup and see how far it'll go on 25 megabits [Music] okay so there's a couple things to note on this particular flight you might have noticed that i'm testing in a different location for this one which is not ideal because almost all the other tests in this video are at my normal long range spot but when i rocked up today they were spraying pesticide and had the whole area blocked off so i'm at the back up location for this one off to the left here you can see the muddy lake that ended up in the intro and the other thing to note is somehow i forgot to copy across the subtitle file for this first chunk of the flight so i've only got the delay and bitrate stats out past about six kilometers onwards [Music] [Music] [Music] okay so we're out at about 10 kilometers right now on the stock goggle antennas at 1200 milliwatts and although there's been a few sporadic drops in quality the link is still hanging in there really well i mean it doesn't look amazing but this is definitely still flyable in my opinion i'm keen to see what happens when we get past the magic 13 kilometer number though so let's keep pushing on [Music] okay there it is 13 kilometers and we've still got the link on the stock antennas [Music] uh yep it's starting to roll off pretty badly now i'd call that the flyable limit oh and there it goes all right i've got to say i'm pretty impressed with that fair enough the picture quality wasn't exactly amazing we were able to maintain a flyable link all the way out to the hard limit of the system at 13.3 kilometers on the stock out of the box setup okay let's quickly check the 50 megabit rate on 1200 milliwatts to see if we get less range again all right this one's stock antennas on 1.2 watt 50 megabits [Music] okay so once again we've got slightly less range on 50 megabits than what we got on 25 with both the flyable and max range coming in at about nine kilometers and this is consistent with our last test so it seems like 25 megabits is the rate that we want for maximum possible range whereas 50 megabits will give us a higher starting bit rate which will means we'll have a higher bit rate for longer into the flight and thus better image quality at the expense of max range now speaking of image quality you would have noticed that the bit rate drops off pretty hard after about the six kilometer mark using the stock omnis and if you're actually planning to use the dji system for some mid-range flying you'd probably be looking at upgrading the antennas to something a little bit better and really this isn't limited to just the mid-range guys so even if you're doing freestyle with dji which is kind of where the system suited best in my opinion then you're still going to benefit from some higher gain antennas for better penetration and latency so let's see what kind of performance we can get out of the system using the iflight crystal patch and omnis so with this setup we should hopefully see an improvement in range on our 25 milliwatts test which will tell us if this patch is good for anyone using dji on low power for example in the eu or for racing and we should also see a significant improvement in bit rate and image quality on the high power range test before we check those two theories though let's put the iflat antennas onto the analyzer and check the tune now you can ignore this first dip here to negative 17 db because that's at 54 37 megahertz so unfortunately it's too low for us to use that leaves us with this upper dip here at 58 20 megahertz so around the frequency of channel 5 in 25 megabits or close to channel 3 and 50. this dip isn't particularly low at negative 12.5 db but one thing this plot doesn't show is the antenna gain which is up at 9 db for the iflight patch so a good jump from the stock on these moving on to the iflight omnis that come with the kit we've got a really nice dip here at 57 34 megahertz all the way down to negative 21 db so these seem like they should be a lot better than the stock on these if you're using a channel somewhere near that frequency so both the patch and omnis look like they should be an improvement so let's head out and put that to the test now in the interest of time i'm going to run both the 25 milliwatts and the 1200 milliwatts dvrs side by side here which should also help to show the difference in image quality and bitrate between the two power settings [Music] okay so things are starting to get pretty messy here now on 25 milliwatts and the bit rate's consistently dropping below one megabit so i'm going to call it there for the flyable range [Music] all right the latency values are really starting to get pretty crazy here now for 25 milliwatts getting way above 100 milliseconds of delay and there goes 25 milliwatts at 9.2 kilometers with 1200 milliwatts still holding strong at about 10 megabits at this stage although we're just starting to consistently dip below 10 megabits here at nine and a half kilometers which is still pretty good since we were already at about three megabits at this point with the stock antennas [Music] all right we're out past that magic 13 kilometer mark so we should see this drop off any second now and there it goes okay so the iflight crystal comes in at 6.8 kilometers flyable range and 9.2 kilometers max range on the 25 milliwatts setting which is a notable improvement over the dji antennas on 1200 milliwatts we managed to keep the bit rate up above 10 megabits out until about the 9.5 kilometer mark and the link was 100 flyable all the way out to the hard cut off at 13.3 kilometers so it's fairly obvious the iflight patch has a decent improvement over the stock setup so for guys that are running low power for example if you're racing or if you're actually following the rules and you live in the eu this is going to be a decent improvement over the stock dji antennas something else i want to check with the iflight crystal is a quick test to make sure that we're actually connecting it to the two best antenna outputs on the goggles to make sure that we're getting the best performance out of it now the v2 goggles use a slightly different configuration from the v1s and you can see from this diagram that mads tech has put together we've got two outputs that transmit and two that receive gail kramer claims that you get the best performance when you connect the patches to the bottom two ports which makes sense because then you'd have one patch transmitting and one receiving rather than just having them both on receive dji uses a bi-directional link so it stands to reason that you want high gain on both tx and rx so that you get the best performance in both directions so let's quickly verify that claim by running two range tests side by side one with the patches connected to the two bottom ports so tx and rx and the other one with it connected both on rx and i'm going to try and kill two birds with one stone with this test by running the system at a 500 milliwatt power output which should be a useful data point for comparison with the sharkbite system that i'll put to the test in a later video okay so on the left we've got the patches on tx and rx and on the right they're both on rx so keep an eye on the bitrate and obviously the maximum range [Music] all right so with both patches on tx we've dropped the link at 10.4 kilometers whereas we're still holding in there with patches on tx and rx [Music] okay so as expected the patches seem to work best on the two bottom connectors which is tx and rx and that managed to take us all the way out to 13.3 on 500 milliwatts so the iflight crystal is looking like a pretty good performer even at medium power levels but you do drop a little bit of image quality out towards the edge of range but what if you want the absolute best image quality all the way out to the hard limit what if you don't want to drop below 10 megabits out past 9 kilometers what if you don't care about your upper back or what if you think neck braces are a trendy look well in that case these are the go now this is the triple feed patch array antenna two in fact and this is my go to antenna for all my analog range testing and will fairly easily take me out to about the 40 kilometer mark with a one watt vtx on the wing so i'm keen to see what the performance difference looks like on dji with these since they've got an even higher gain than the iflight patches at 14db and they look great in terms of the tune as well obviously the form factor isn't anywhere near as nice as the iflight patch and with the additional weight it pretty much drags the goggles off your face as you're wearing them so this really is a function over form thing and not something i'd want to wear on a regular basis but at the risk of spinal injuries let's head out and test this patch on 1200 milliwatts all right dji on two triple feed patches and two stock omnis 1.2 watt 50 megabits [Music] okay so with the iflight crystal we had flyable video all the way out to the 13 kilometer cut off but we really started to see a pretty substantial drop in the bit rate at about nine and a half kilometers so with this flight we want to keep an eye on the bitrate as well as the image quality to see if triple feeds can do any better all right so we're out past the nine and a half kilometer mark which is where we saw the bit rate start to drop right down with the iflight patch and as you can see we're still sitting in the 30s with the bitrate so still up above 25 megabits and we've also still got a beautiful clear hd image without any blocky compression which is kind of what you'd expect when you've got two 14 db dinner plates strapped to your head but it's fairly impressive nonetheless [Music] [Music] [Music] okay so form factor aside the performance you get with two triple feed arrays is pretty damn incredible we were just out at 13 kilometers with a bit rate of 25 megabits and perfect hd image quality with no blocky compression breakup whatsoever which is pretty damn impressive when you consider the amount of data that's been sent over the air reliably from that far out it also answers one of my own personal questions which was can you make it past that 13 kilometer hard limit if you throw some really high gain antennas at it and the answer is no no matter how much antenna gain you throw at it the hard limit really is a hard limit which leads me on to a bit of speculation as to why that might be the case now when i first heard about this hard limit the reasoning that was given to me at the time was that the limitation was due to the speed of light rf waves behave the same as light so what they were referring to was the speed that the rf signal can be transmitted from the air unit to the goggles now the speed of light is pretty damn fast like 300 kilometers per second type fast so naturally when i heard this i was pretty skeptical on how this could be the limiting factor of the range of the system at 13 kilometers it only takes 43 microseconds to transmit an rf signal from the vtx to the goggles and fair enough that doesn't take into account all the data required to make up a full hd frame that's just a simple step response but still that's stupid fast you think of the occusync protocol timings with that type we'd be getting black screens every time we flew behind a tree or a building which obviously we don't from what i understand of the occusync protocol it uses acknowledgment of frames and can also re-request dropped frames as the link degrades and obviously the protocol allows some level of packet loss which is the blocky mess that we see at the edge of range but what i expect is happening here is the goggle firmware has some sort of threshold where it will only allow a certain amount of dropped packets within a certain time period before it bails out and completely drops the link at a frame rate of 120 fps the goggles only have 8.3 milliseconds between each frame that they can use to receive all the packets to construct the next frame decompress the h264 compression and then render it on the screen anyway this is all just speculation without being able to see dji's code and that's one of the downsides of a proprietary system and we'll probably never know but hey 13 kilometers ain't bad for a high quality hd fpv feed all right wrapping up with the stock antennas we made it to 5.1 kilometers on 25 milliwatts and 13.3 kilometers on 1200 milliwatts and we noticed a slightly better range between the 25 megabits versus 50 megabits using the iflight crystal that improved to 9.2 kilometers on 25 milliwatts and 13.3 kilometers on 500 milliwatts and again 13.3 on 1200 milliwatts with a much more flyable video in the last few kilometers before the hard limit and finally using the triple feed arrays we made it out to 13.3 again but we had beautiful crystal clear 25 megabit plus performance all the way out to the very end that's going to do it for this one guys next up i'm planning to test the sharkbite hd system and see where the range limits are with that and i'm also keen to see if we get any improvements to image quality at the edge of range due to the fixed latency and bit rate that the sharkbite system uses until then thanks for watching guys see you in the next one

Channel: Wezley Varty

Views: 34,794

Keywords: DJI, DJI FPV, iflight crystal hd patch, range test, dji range test, long range fpv

Id: TSk1RPB-3T0

Length: 28min 26sec (1706 seconds)

Published: Sat Sep 11 2021