Thread Tools
This thread is privately moderated by alicecooper, who may elect to delete unwanted replies.
Jul 12, 2021, 11:20 AM
Registered User
Thread OP
Build Log

A New Long Range Ground Station


It has been a long time coming, but finally I created a ground station that ticks nearly all of my boxes.
This is not a new idea, I have been creating these things for nearly a decade, this one being number six, I feel it's the last with current technology.
I decided for this model to just have a few key features, this is in an effort to both simplify its functional performance, and also to minimise setup time, enhance system reliability and condense the amount of overall equipment to a minimum.
From a functional standpoint it is fairly basic, having a diversity receiver that sits up on a tripod running a high performance directional antenna which can be steered to any part of the sky with a single joystick. The other diversity channel just uses an omni antenna for 360 degree performance up to around 4km.
The driver end of the system is a seat positioned console that incorporates the pan tilt joystick itself, an integrated audio system, a preview monitor, and enough battery power to run everything including external devices like my goggles. Really it couldn't be simpler.

I designed the mast head with the same mandate, to have everything fully integrated including the 360 degree pan servo, 90 degree tilt servo, receiver, associated electronics and antennas. A key part of the integration was to simplify connection between the mast head and the console to just a single plug lead.
I did this using a common VGA cable as these provide all the connections necessary for servo control, dual loss free video and screened audio connection to provide for both buffered AV outputs, including power and ground, and importantly quality versions of these cables are cheap and easily available in various lengths, the HD15 connections are quick and well retained and the connections themselves are very reliable. I enhanced this aspect by pairing up all unused wires (most XGA cables support 14 of the 15 plus independent group shield) for the power connection, and having two video feeds using a pair of the independently screened coaxial video wires I have a layer of redundancy in all but the servo signals.
I integrated a servo BEC into the pan tilt head unit meaning the 12V (3S) power fed up the link cable runs both the receiver and the servo's. This overcomes voltage drop and any potential coupling of servo noise into the video/audio path, a problem I encountered with a previous version. For this I also integrated a custom PWM filter for the servo signals, as the servo's were sensitive to ground path coupling (being trailing edge triggered digital servo's) this Schmitt Trigger filter removed all traces of jitter and twitch which plague many servo's when used in long cable runs.
The entire head unit assembly is fitted together with thumb screws and is quite easy to disassemble, but to minimise setup time in the field and reduce wear and tear on plugs connectors and fixtures I customised the carry case to allow the whole head assembly to simply drop into a receptacle, only requiring the omni antenna to be removed. The patch array antenna has an additional support frame to keep it's set position accurate, without that it tends to droop under it's own weight with just the semi rigid coax for support.
The head unit also utilises my own quick fit attachment system for the tripod mast, it basically uses a machine fit sleeve with a keyed slot to prevent weather cocking, and it just drops onto the pole and locks itself in place. It provides very quick and tool free attachment.

The carry case was one of my typical wooden constructions, using mostly regular case hardware, I had the benefit again of customising an exact fit and retention for everything, including the console itself, the mast head / antenna assembly and all required accessories, cables and additional antennas.
The construction is done with 12mm clear pine and 5mm marine ply, coated with urethane satin clear coat and internally designed to self lock all components into place when closed, so nothing needs to be tied in, just open the lid and freely lift the pieces out.

In use I really like only having to lug two cases out to the field along with the tripod and chair, with the plane in the other hand I make it in just two trips out to become ready to fly.
On the folding chair front I have also done some extra modifications, I basically made a second fold out locking table to hold the console on the left side, (the right table holds radio, phone etc) and I also made a custom goggle holder which is great, I might do a basic follow up thread on it as it is one of my favourite yet simplest projects. No more faffing around trying to prevent the sun from entering the goggs, or having them fall into the wet grass...

The console itself was built from a NOS Radio Shack slope front project case that has rattled around my parts store for over 20 years looking for just the right project... it turned out to be a perfect fit for this one.
The decision to go with a modified RC joystick for piloting the antenna direction was done for a few reasons. The biggest of them being manual control. Despite integrating a port for automated antenna tracking, I have always preferred to have full manual control at my fingertips, (used to be grabbing the mast and twisting..)and I find it often performs better at the very fringes where I can overshoot or undershoot the signal line to mitigate fresnel effects, as well as to utilise unique terrestrial nodes to improve the path at long distance in high elevation terrain, (also climbing mountains and following complex steep contours) sometimes the best signal is not actually a bullseye shot.
I also prefer the reliability of manual control, many previous tracker experiences turned into near disasters with fringe signals frequently losing the tenuous link long enough (<one second) to lose or corrupt the direction info and suddenly it goes rogue, then requiring the tracking system to be shut down and the mast dropped to manually steer the antenna enough to re locate the signal... almost never again for me..!

So after a previous version did duty using two knobs, one to pan and one to tilt, I decided a stick would actually be better, as groping around blind would inevitably find the wrong knob and my hunt to improve the signal was not always as successful as it should have been (a polite summary here...)
The joystick I found with mechanically linked trims was rescued from a junk box vintage RC radio and i modified it's movements to not return to centre and to have some positive independent resistance. I did this using some custom made felt friction disks preloaded with a spring on each pivot and the final effect is perfect, the movement on either axis feels independent, panning movement doesn't interfere with tilting and vice versa.
In the remaining panel control layout I also included a switch to reduce the pan range from 360 degrees to 180 degrees as this gives a much more precise feel to longer range ventures with the general direction being setup to centre stick.
I deliberately used different knobs on all panel controls making finding the right one easy by feel alone with no lifting the goggles to see. It works well, I only need the unit to be set square in position and I can easily find and control everything with goggles down.

I added the preview monitor for a couple of reasons, one is to check GPS count and FC status during initialisation without needing to put the goggles on, and secondly, it is a handy fall back in case of any problem with the goggles or cables. I used to carry a spare set with battery just in case, I rarely needed them but better safe than sorry. This display is only 2.5" but the image is surprisingly good, I can easily identify OSD text and terrain and it is still visible in sunlight.
As I didn't want to degrade the video signal with more DA's I set this monitor up with a switch allowing it to utilise either one of the two AV signals sent directly from the receiver, this still allows use of two goggles or DVR etc, and has the ability to override the source signal in the event of any external malfunction, eg a lead failure or short. It works well, and has proven better than I anticipated.

For powering the system I made up my own custom battery pack using liion batteries sourced from many new DOA failed HP tablets (from work) otherwise binned, these 4Ah 1s flat cells are very high quality and offer long life, I stacked six in 3S2P configuration which gives me 8Ah @ 3S total and the final stack fits the available space with minimal clearance all round, almost the luckiest fit.
I use Skyzone Sky04X goggles and the AV port plugs into one of two 3.5mm 4p connectors on the back panel and as I fitted a custom balance charge port on that panel, I made a second optional (aviation connector) lead to power the goggles. In use I get around 7-8 hours run time as the goggles use very little energy in AV mode. I included a 1A aviation circuit breaker after measuring a total combined draw of 900mA with everything working. The best thing is, should I forget to charge it, there is easily several days worth of flying time for me before it needs recharging. I included a small LCD battery indicator to display state of charge while in use. The aviation balance port has become a standard option for lots of my gear and I use a custom jst lead to balance charge, but also the spare pins on this 7p connector also allow pwm input to provide alternative tracker control and to simultaneously power that gear if needed using another custom cable.
I also added a second aux power port to supply even more external devices if needed.

The audio signal needed to be setup in a special way as RC video transmitters often use only one channel to carry audio. The other channel consequently contains very high level static noise which makes combining both channels for mono impossible. I overcame the problem by making the receiver connection cables unique for the ports, one diversity AV out connects only the left channel, the other only connects the right, they can then be switched at the mast connection point and the unused channel is taken out of circuit. Once at the console the signal is split to the internal amplifier and both left and right channels of the 4p AV ports, allowing my goggles to live record both L/R AV channels together as mono, and this system works well with a variety of different transmitters. I also modified the tone level to roll off high frequency from the audio, as most systems have extremely sharp peaky high frequency microphones, rolling off above 2.5kHz makes the sound much more listenable...

So I have now flown several long duration flights with this new setup, some in very poor weather conditions, and it has passed my own demanding tests so far with flying colours. The controls are easy to use with my fairly free left hand (mode 2) and the servo positioning accuracy is a good match for the 14.4dB high gain antenna. It could easily resolve to double that gain, but it is very easy to peak the image quality whenever needed, at 180 degree setting a quick narrow sweep side to side immediately resolves the ideal position to point.
Having good quality audio is also a great asset. I can hear wind noise, servo action and motor speed, including prop freewheeling, the silence of a stall or roar of a dive adds a valuable sense to my flying and I now use it all the time.

The ready to fly setup time from putting everything on the ground is under two minutes which is way faster than my previous systems, and similarly packing up takes no time at all. The carry case although slightly bigger than my transmitter/goggle case is still fairly light, it is similarly water resistant (to wet ground) and is well balanced, on uneven ground, it doesn't try to tip back when open.

I doubt I'll ever need to build another one again, although time may change that.. never say never..
Sign up now
to remove ads between posts
Jul 19, 2021, 07:32 AM
Registered User
That's real nice however I didn't see a stand for a sun umbrella over the chair - perhaps MkII?
You would have to be pleased with that piece of kit. Good tall mast to get clear of local clutter and away from any system noise, short antenna runs to the diversity rx and wide and narrow beam width antennas. Drool.
By the time you're out a few km and abiding by the 400 foot altitude limit (of course) the steering would be mainly azimuth only I would think. I have often wondered if a narrow beam antenna could be tracked by using a slow oscillation of the horizontal axis and detecting the -3dB points either side of the peak gain. It may only require a few degrees of movement and the angular rate of change of the craft is not too fast. It sounds feasible but then I have no experience on what sort of level changes you have to expect when tracking something along the horizon.
It's great to see and I hope it serves you well.

Cheers,
David
Jul 19, 2021, 08:38 AM
Registered User
Thread OP
Hi David, and thanks.
I do have a system like you mentioned that sweeps one antenna 360 degrees and points the main receiver antenna in the direction of the signal peak. it is the most reliable as far as trackers go but it tends to wear servo's out pretty fast from sweeping non stop for hours on end.
I had considered doing a stepper drive upgrade but it's faded from use being only single axis.

Good pick on the sunshade too, I have recently buried a dedicated pipe in the lawn that takes a large sun umbrella, and I fitted a spigot to the umbrella's peak allowing the head unit to use that instead of the tripod, it probably gives a good groundplane effect.. but it's winter here and an hour in the sun is pretty nice, it's still fairly low at midday.
The terrain here is pretty mountainous and maintaining sub 400'means some pretty solid climbs in virtually any direction, but interestingly the vertical azimuth is always much less sensitive to track than the horizontal at long range. I tailored the vertical servo range to just give -10 and +90 degrees movement and even at the signal fringes I can sweep a good 30 degrees vertically compared to about 10 degrees horizontally to peak the signal, it's a strange effect. Running very low power vtx's at closer proximity makes the vertical aspect much more sensitive above horizontal, demanding both share the same movement ratio.

In use I'm finding that 14db pagoda array to be one of the best 5.8 antenna's yet, it's pattern is very progressive and node free and it gives a fairly forgiving uniformity to tracking manually. I just give it a nudge once any static is noticed and it cleans the image up for quite some time, occasionally I will hunt around to overcome fresnel zone decay, and interestingly again, tilting upwards seems to show the biggest temporary improvement until I move beyond it and can drop back down.

I frequently use this setup for soaring, including with a few tiny ultra light fpv gliders and find the range and performance is great for the ones running low power transmitters (25mW), I can sometimes hang up there for hours on just one cell and get amazing distances. (thankfully as there is no RTH...)
As mentioned I used to use some older systems with pan adjustment only, then others with two axis knobs (pictured) to track the aircraft when doing this manually, and it frequently became a panic session trying to relocate the lost signal, the stick gimbal system has really fixed that. The larger unit pictured supports three different automated tracking systems that are pretty well outdated now, for me.
Here's a few pics of some of those previous versions, dating way back to 2014...

Cheers. AC
Jul 20, 2021, 05:58 AM
Registered User
Most interesting.
I'm sure you're right and a servo that has to constantly oscillate about a peak will have a lot of wear. The stepper solution would be a good option.
Have you tried helical antennas? They are inherently quite broad band unlike something like a patch which can be fairly high Q and hard to get right on frequency without at least a scalar analyzer and a return loss bridge to tune it. I often wonder how consistent some antenna designs are, especially when they use etched FR4 material without tuning as the dielectric constant varies from sheet to sheet directly impacting tuned dimensions.
I think the 14dB gain antenna is a good compromise between gain and beamwidth.
You've certainly made some nice stuff over the years and it's a pleasure to read good technical details about them.

Cheers,
David
Jul 20, 2021, 08:31 AM
Registered User
Thread OP
Thanks again David, and likewise, always good to have someone to bounce it around with..
I have done quite a lot with helicals and a few other various designs over the years. Most of my proper long range gear was built around 1.3gHz and I still run it with a few planes, but the new LORA based long range control systems got my attention so I was drawn back to 5.8gHz for a revisit to push those range limits again.
I made many different helicals, some of which I still use with goggles but in doing so I always found there was a pronounced crossing point where gain would give way to loss of efficiency, usually beyond 5-6 turns and although the pattern became narrower with a higher turn count the overall performance would flatten and then diminish as the signal output would drop at a progressively higher rate than the gain benefit would yield. That then took a toll on the remaining pattern efficiency also.
Increasing the coil's final coupling impedance would fend it off slightly but at the expense of a mismatch loss and any matching devices I tried at the feed point would always negatively disrupt the tune and take it back to an equal or more lossy transfer, even at exactly 50 ohms, at least that was my experience. Ideally a transformer that could maintain efficiency beyond 6gHz could probably overcome some of it but sourcing core material and matching both sides is beyond my capability and I still think the other losses from such a long coil still cost more than they give, even with a perfect time domain stacking the waves from either end, at some point some of the energy will just radiate back out through the unavoidable inductance.
Circular wireless managed to mitigate the problem to a point by changing the winding spacing towards the rear (reflector) which I know helped (I do the same), but it still seemed to have the same limitations, which it looks like they pushed as far as they could with their 8 turn version.
I only have limited test gear for 6gHz, which is just a spectrum analyser, a SWR meter and one of the most useful things, an analogue signal level meter that I made specifically for field testing.
With most of my own antennas I usually use that signal meter for final tuning (peaking), I have a large clear paddock outside my workshop and I set up a couple of tripods way out in the open and at night, around 300m apart with a transmitter at one end and the signal meter at the other, I can reduce the transmit power from 10mW down to fractions of a milliwatt using attenuators which gives me excellent resolution for tuning down into the noise floor, and it works well for pattern measurement, node shape, position and overall efficiency as well as for final tuning.
I can also swap between different antennas at either end to compare transmitting and receiving performance and do lots of comparative measurements.
It's frequently a very fiddly and fussy process which takes a lot of time (and walking), but the results are often worth the effort, although there is always that same trade off between gain and pattern shape. It's one thing to have a laser narrow beam with immense gain, but flying with it is usually no fun at all until the distance out is immense, I prefer a more predictable pattern that can still work at moderate distance, even with diversity system between a 12dB antenna and a 30dB dish it is still risky beyond 10km.
That patch array antenna that I used above seems to have the best of both worlds. I measured it's gain to at least match it's claimed 14.4dB and it is almost node free. I noticed also the PCB backing material is very thin at around 0.8mm (fibreglass reinforced), probably FR4, but half the usual thickness which no doubt helps. I haven't made any PCB based antennas so I don't know much about the different materials or their effects at gHz frequencies.
I did also make a few dish antennas some years ago, experimenting with agile focus systems, pattern directors and stubs, all good fun stuff but again performance was usually very fussy, often the pattern could yield intense nulls in an unpredictable way and have nodes that could almost match the prime beam, usually making tracking very difficult and unpredictable.
I have often made all sorts of unique antennas from different crazy ideas, usually based on HF designs, but one standout was an experimental semi hemispherical linear antenna that has a dual convex waveguide , really just another idea to see what would happen... lol, and it actually worked surprisingly well, not what I expected, but it has a very useful pattern.

It's one of the unique attractions of the FPV corner of the hobby, being able to play around with antenna's and associated gear, but unfortunately not many people are as interested in the DIY side of it, probably because the stuff is now relatively cheap to just go out and buy and have good performance..

Cheers, AC.
Last edited by alicecooper; Jul 20, 2021 at 10:18 AM.
Jul 21, 2021, 09:04 AM
Registered User
Again a most interesting discussion. You have certainly covered some ground (literally) with your antenna experiments in your back paddock. Once your antennas are electrically optimized the analog signal meter would be invaluable for measuring in the field. I have sometimes resorted to finding something like an RSSI or AGC signal on receivers but often it's so logarithmic that you need a significant signal change to register any DC output change so not good for fine tuning.
I think with helicals a 7 or 8 turn is about optimum. I guess you could stack them (as per patch antenna arrays) but getting the delays right would be a black art. I think helicals naturally pan out about 140 ohms so need a matching stub to get down to 50 ohms. Even if you have all the gear for doing this it doesn't tell you anything about the pickup pattern or how many lobes there are so again the analog meter is important. Having a good set of attenuators for 6GHz is very handy provided signal is not leaking out from other areas. I think to do any transforming at 6GHz you're looking at printed couplers on Rogers pcb material.
I guess there have to be limits as to how far you go. While a dish gives excellent gain it's a pain to keep on track and not that portable. I think what you have is an excellent approach while still maintaining good portability and relative simplicity.
I've gone totally the other way - fully autonomous and fixed on-board recording. No long range control, no long range video and actually a fairly high redundancy level. Perhaps not as exciting but cheap enough to write off if something went wrong.

Cheers,
David


Quick Reply
Message:
Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion Project Covert Ops: Building a covert long range ground station IBCrazy FPV Talk 7 Sep 20, 2012 04:00 PM
Discussion 12ch Radio relay for wireless long range UHF ground station i3dm FPV Talk 0 Mar 22, 2012 03:20 AM
Discussion Extreme Long Range Ground Stations questforstars UAV - Unmanned Aerial Vehicles 5 Jun 20, 2011 02:54 PM