Thread Tools
May 20, 2010, 08:56 PM
Registered User
oxxyfx's Avatar
Thread OP
Discussion

Future of the FPV


Well, not too distant future. I had the chance today to do something which can be considered ground-breaking - first ever in FPV.

To some people this might sound as science fiction but this is repeateable, proven by video and several tests we performed today.

I have flown my Mikrokopter M3-ME on 2.4Ghz (Futaba 9C transmitter with a spektrum module in the back and an AR7000 receiver with QuadroPPM module on the Mikrokopter) and received back flawless video on 2.4Ghz.

The credit goes to Rotoconcept Robotics from Mississauaga, Ontario, Canada for putting together a digital video transmitter which contrary to some beliefs in other threads - does not have any visible latency. The latency is less or close to 50ms which will still allow perfect piloting of any aircraft - including helikopters, airplanes and mikrokopters.

We at New Generation Hobbies are greatful to be able to present this product on behalf of Rotoconcept Robotics and thank them for the opportunity to represent and distribute their products in the future on the North American market.

Now, for the below video we had the following setup:

- Mikrokotper M3-ME
- RQ4 digital video transmitter set to 50mW on 2.4Ghz
- KX-550 NTSC camera

The receiver was inside of the office under a heavily shielded roof. We know this for fact because inside we cannot get a sattelite lock on any kind of GPS - it will not even find a single sattelite - that gives you the idea on how heavily the building is shielded.

The receiver was connected to a laptop and the video was recorded using Rotoconcept Robotics proprietary receiver, software and mpg2 encoder.

The RQ4 transmitter had a small 2.4Ghz antenna connected to it (less than 2dbi) and the RQ4 transmitter was mounted directly below the spektrum AR7000 receiver. We wanted it mounted this way - as close as possible to each other - so if there is any interference we would detect it.

The video receiver had a 6dbi onmi directional antenna connected to it. There are some frames dropped in the beginning of the video because I took off with the Mikrokopter already when we noticed that the receiver antenna was not properly plugged in. After that thas been corrected flying around the building and going higher up was flawless.

Here is the video so you can judge the quality for yourself. The video coming back from the transmitter is 720x480 (DV) resolution. After watching the video read further:

Mikrokopter FPV with RQ4 digital video transmitter (3 min 59 sec)


I didn't fly this FPV for 2 different reasons:

- we needed to make sure there is no interference on the 2 systems running on 2.4Ghz together
- this was in an industrial area zone new to me, and I do not wanted to loose my Mikrokopter by not knowing the area from the air.

On the video at some point across the street you see 4 huge antennas. These radiate 60W each on 700Mhz, 900Mhz, on 2.4GHz and some other frequencies - so if you think that there would have not been any interference factors think again.

With the receiver and recording laptop inside the building the 50mW transmitter gave us more than 150 meter range - NLOS - and we could not test any further because my LiPo died exactly where the video ends. At that point there were already 2 buildings in between the transmitter and receiver.

Next test will be somewhere outside with bigger distance between the TX and RX - perhaps we will put the RX inside of a closed wan or maybe my Pacifica will do it with closed doors and windows. I will also try to have the goggles connected to the receiver and try to fly this FPV to proove all the geeks wrong about latency on digital video transmitters - if a system is designed and put together properly.

Thanks again to Rotoconcept Robotics, this product will be soon on the market through New Generation Hobbies.

That's all folks for today.

PS. I just received an E-mail with some techincal data for those who are interested:

- automatic frequency shifting
- 200meter+ NLOS range
- 25Km+ LOS.
- 17 dBm average TX output,
- 20dBm (100mwatt) max
- 128 bit AES encryption
- <50msec to 100msc max latency
- TTL, RS232, Ethernet, USB ports onboard.

Since this is a computer - theoretically the possibilities for OSD, GPS and other things are limitless.
Last edited by oxxyfx; May 20, 2010 at 09:04 PM.
Sign up now
to remove ads between posts
May 20, 2010, 09:31 PM
DJI Supporter
Sid3ways's Avatar
Definitely looking forward to seeing more of this. That looks and sounds amazing. I know it is very early but what sort of price range are you expecting? 2X-3X the price of a normal rx/tx setup or 4x-5x the cost or more?
May 20, 2010, 09:36 PM
Registered User
some dudes are going to start eating their hats
the range is 30km @300mW manufacture rated, the price will hurt a little however

see http://digitalwirelessvideo.net/products.html
May 21, 2010, 01:06 AM
Suspended Account
Quote:
Originally Posted by jimbob00
some dudes are going to start eating their hats
the range is 30km @300mW manufacture rated, the price will hurt a little however

see http://digitalwirelessvideo.net/products.html
Not really, as stated you need a PC to decode the data, hence the PC is doing all the numbercrunching, not exactly something you'll be lugging to the field for some occasional flying.

Also the #1 issue with a digital video link still isn't tackled: freezing or blackout on loss of signal. I'm going to quote from my reply in another topic to give the proper perspective:

Quote:
You get freezes or total blackout if even part of the data is missing. As I mentioned in my previous reply such is not the case with an analog video link, there's a large grey area, starting with specks of noise, then the image getting noisier, occasional dropouts, more dropouts that last longer and longer and finally a total loss of signal. I usually head back if the dropouts last longer than required to control the plane.

To my knowledge there's no digital video protocol that handles this very well, as it cannot reconstruct part of the image due to lacking data. Hence it is a all or nothing situation. Which for FPV is just not suitable.

I'm not even going to go into the latency requirements for flying for example a FJ at >100mph just a few feet off the ground (or in trappy's case a Zephyr).
That being said I'd love to hear more about this system and the associated pricetag, as well as how they handle the above mentioned drawbacks.

Cheers,

Sander.
May 21, 2010, 01:09 AM
Registered User
Please, one Jim Drew is enough...
30km @ 300mW, did you computed what antenna gain you need to accomplish this ?
May 21, 2010, 01:22 AM
Suspended Account
Quote:
Originally Posted by oxxyfx
The credit goes to Rotoconcept Robotics from Mississauaga, Ontario, Canada for putting together a digital video transmitter which contrary to some beliefs in other threads - does not have any visible latency. The latency is less or close to 50ms which will still allow perfect piloting of any aircraft - including helikopters, airplanes and mikrokopters.
Well, their datasheets specify otherwise, <500msec and 100msec max. depending on frequency and configuration. Although that's fine for flying an Ez at altitude, I wouldn't want to be flying low with a FunJet with these values. Simply because latencies add up. A common RC Tx has between 20 and 50ms of latency, add to that the 100msec for this video link, another 50ms or so for the LCD display (goggles or screen) and we're well into 200msec already ...

Quote:
Originally Posted by oxxyfx
We at New Generation Hobbies are greatful to be able to present this product on behalf of Rotoconcept Robotics and thank them for the opportunity to represent and distribute their products in the future on the North American market.
We all love you for it, but it would be nice if the folks at Rotoconcept Robotics would be willing to participate in the discussion here. There's several folks (myself included) that have some valid arguments against a digital videolink. I would love to be proven wrong, but discussions on the forum usually end up going nowhere as the advocates for digital video cannot prove their claims. Rotoconcept Robotics should be in an ideal position to step forward and show what they've developed here and how this ties in with some of the concerns we (the seasoned FPV pilots) have.

Quote:
Originally Posted by oxxyfx
The receiver was connected to a laptop and the video was recorded using Rotoconcept Robotics proprietary receiver, software and mpg2 encoder.
That makes me cringe a little ...

Quote:
Originally Posted by oxxyfx
Next test will be somewhere outside with bigger distance between the TX and RX - perhaps we will put the RX inside of a closed wan or maybe my Pacifica will do it with closed doors and windows. I will also try to have the goggles connected to the receiver and try to fly this FPV to proove all the geeks wrong about latency on digital video transmitters - if a system is designed and put together properly.
There's nothing to be proven wrong about, simply because there *will* be latency, there's no such thing as real time video encoding, there *always* is latency. Whether that latency is something that will interfere with your ability to pilot the FPV plane is another story.

That latency will be governed mostly by the CPU horsepower dedicated to the encoding process, the codec/algorythm doing the work and the nature of the analog signal to be encoded. I think I do not need to illustrate further that an almost stationary image is less of a challenge for the encoder than a rapidly changing image where each individual frame is significantly different from the previous one.

So, if the folks at Rotoconcept Robotics are willing to participate in the discussion they're welcome to post here. I'm sure there's plenty of interest in hearing how they approach some of the genuine concerns uttered here.

Cheers,

Sander.
May 21, 2010, 01:25 AM
Registered User
Daemon's Avatar
Fascinating. Could certainly be very useful for indoor FPV where multipathing is usually
a major problem.

I'd like to see
1. Actual FPV piloting to prove the latency is low and freezes are not fatal.
2. High motion video
3. LoS Range test
4. Minimum CPU specs for ground station computer
5. Price

ian
May 21, 2010, 01:27 AM
Registered User
wallaguest1's Avatar
some pictures of the hardware? thnks
May 21, 2010, 01:49 AM
I think my wheels fell off.
Thumb_number's Avatar
I'll stick to analog, this stuff is gonna be well outside my budget.

Very cool technology though, well done. I'm guessing in five years time it (digital links) will be all you will be able to get.

Subscribed, out of nothing more than pure interest.
May 21, 2010, 04:11 AM
Suspended Account
trappy's Avatar
Quote:
Originally Posted by jimbob00
some dudes are going to start eating their hats
the range is 30km @300mW manufacture rated, the price will hurt a little however

see http://digitalwirelessvideo.net/products.html
if they use WLan, I'll even eat my plane.

Well done Zoltan! This is the beginning of a new era. Some things we'd like to find out is price, weight and power consumption, though and is a PC required for operation?
May 21, 2010, 04:18 AM
FPV Junkie
tazdevil's Avatar
cool , stuff , waiting for the price for this OFDM video link

and for mk piloting for aerial photography this could be nice

keep me informed of the avaibility Zoltan

do you plane export for this product or only north America distribution ?

rgds
May 21, 2010, 04:52 AM
Registered User
PeteSchug's Avatar
In a quick glance I didn't see weight mentioned, but 5" x 5.5" x 1.5" for the tx is HUGE by my standards. I don't have I plane I could fly that in except the quarter scale Schweitzer 2-33 I'm building. Planning to try FPV aerotow, I've had lots of aerotows in the full sized 2-33 and really look forward to trying it with a large model. Progress is slow right now, but I am retiring at the end of the summer, so will probably get it going before winter. I will probably not go digital just for that trial.

My favorite video mount of all time is still my pico Tiger Moth (fly anywhere, intimidate no-one) with an all up weight of 9.5 oz. As fast as the TM flies I don't have to worry about latency, but I think the digital tx may outweigh my plane! I live in a population dense area in case you guys who live in the wide open spaces are wondering what I am babbling about with "fly anywhere, intimidate no-one.)

Sounds like a good first step. I also look forward to something that can transmit from a pure digital camera with no intermediate conversion. There are lots of very inexpensive, very small cameras that are pure digital that cannot be hooked up to current tx's. Stuff like the keychain cams. I really like small gear.

Pete
May 21, 2010, 05:34 AM
Registered User
Hey Guys;

My name's Zoltan as well (NOT the same as Zoltan @ NG)
I'm part of the team that developed the system. I'll do my best to answer all your questions here. I'm also an RC and real helicopter flyer and joined this forum well before I joined Rotoconcept.

The size of the production FPV model will be less than 4X2" including a VERY sturdy enclosure as we want it to survive "rough landings"
It is is actually different than the RQ4 listed on the website as that was designed for ground applications.

There are a lot of valid concerns posted here, such as latency and blackout, however we took all of that into consideration.

It is powered by an ARM9 CPU @ 1GHZ so there is plenty of kick for fast encryption. As throughput gets worse, resolution gets decreased automatically. The latency is just below 50msec and it adjusts picture quality to stick to that as much as possible. (spec on the site had a typo)

No I would still not put on a 150km/h funjet as I cannot make any guarantees on latency.

The minimum hardware req of the RX computer is very small, it runs fine at 30FPS 720X480 on a 1GHZ Via C7. We are working on a computer free TV out receiver, but that won't be anytime soon. (Latency, cost, etc....)

I don't have pricing information yet, however this is specifically designed for FPV not professional applications so we're doing our best to keep it affordable for everyone, but do keep in mind they're made in Canada in low volumes. We should have $$ info in the next few weeks.

As far as LOS range goes, we have done over 10KMS and considering the RX sensitivity is 98 dBm @2Bmps so it's enough to say it goes far. RSSI can be included on the screen as well. The OSD part is opensource and all the code/tools will be available soon. We are working on a drag&drop OSD designer so everyone can play with it. (If you're good at C for Linux and .NET for Windows and would like to contribute, let me know)

More videos to come.
May 21, 2010, 05:53 AM
Suspended Account
Hi Zoltan,

Welcome! Glad you could join our discussion here. As you may have noticed the topic of a digital video link for FPV is always cause for heated discussion.

I think Ian summed up some of the most pressing questions we have already, so let me just post them here once more:

We'd like to see the following:

1. Actual FPV piloting to prove the latency is low and freezes are not fatal.

This has been the thing that worries us FPV pilots most. With our current analog video links we have a good indication of when the link quality is less than optimal, with noise entering the image, dropouts etc. With a digital link there's no such warning, so how does your system handle this?

2. High motion video

As mentioned before, encoding an almost static video feed is entirely different from encoding some fast past FPV action. Where almost each frame contains a host of new information. How does your system handle this?

3. LoS Range test

Of course we'd all like to know how far it can be pushed in practice, numbers and free space loss/budget link calculations are all informative, but we're a tough crowd to please.

4. Minimum CPU specs for ground station computer

Will we need to be lugging a quad core notebook with 4GB of ram and a speedy harddisk to the field which will be eating through its battery in no time at all, or can we get away with a lowly Atom based netbook? And how about on the Tx end, does the Tx empty the battery faster than the brushless motor can? With a 1GHz CPU running I'm guessing it'll be upwards of 20W in power consumption?

5. Price

That's the million dollar question really, if all of the above seems to play out favorably will it be affordable enough?

Cheers,

Sander.
May 21, 2010, 06:06 AM
Registered User
oxxyfx's Avatar
Thread OP
Hi guys -

It is early morning here, and I have not woken up completly yet but here are some answers:

Sander First because you are my favorite Devils Advocate:

- Latency is 50ms and less. No matter what you read. No matter what you have calculated. It is still 50ms and less. I've seen this thing work hands on and I am confident to fly the Mikrokopter FPV with it - which test will be done in about 2 - 3 weeks after I finish my tax return. Government business comes first...

- Computer will be needed because the receiver connects to the computer at the moment. Decoding is not done by the laptop but by the receiver - the receiver is a computer based board with a very powerful processor inside. The laptop is merely used to display the data and do the recording - btw - the videos obtained this way are about 40-50% better quality then if you would run the signal through a composit video converter. I noticed that yesterday - the details on this video are amazing compared to a very similar recording I done a week ago and posted on vimeo here:

(10 min 28 sec)


Camera is the same, Mikrokopter is the same only the video downlink is analog in this case. Quality of the picture is not bad - but just compare the two.


- The transmitter on its own incorporates a processor over 600Mhz.

- Folks from Rotoconcept Robotics will participate in the discussion here later.

All the other quiestions:

- The weight and size of the transmitter will not allow you to mount this on a foamie to fly indoors. The 5" x 5.5" x 1.5" version incorporates an SD card reader and other connectors which will not be on the final FPV version, that will be smaller

- Lighter foamies will like EazyStar still not be able to carry this - however bigger planes like the cularis will have no problem flying it. I had no problem lifting the demo - which is heavier and larger then the final transmitter will be - with the 400mm quad.

- Price is not published yet - but if you look around other digital links available on the market are in the tens of thousands of dollar range - we would have not started this thread here if by all means this would not be in the range which will be accesible at least by some FPV fliers.

- One big accomplishment which I did not see anybody commenting on - that the two 2.4Ghz systems are used side by side - the RC link and the video system. That is unique in this perspective.

- This system with a pointed patch antenna or higher gain parabolic antenna will finally give those FPV cars a very usable range.

- One more additional thing - legality - this can be brought down to 10mW or less power output to the freqency range allowed by each country - hence it will be legal to use while 10mW power output still giving you miles of range - to fly to the end of the range of your EzUHF system...


Quick Reply
Message:

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion If This Guy Has a Future in Politics, then America Has no Future Highflight Life, The Universe, and Politics 15 Mar 13, 2010 08:46 AM
Discussion FPV Future is near? Snaydher FPV Talk 6 Dec 24, 2009 09:04 PM
Discussion my future FPV Tomeczek FPV Talk 2 Apr 15, 2009 11:30 AM
Yippee! Best future proof RC link with FPV in mind? JanBoothius FPV Talk 1 Jul 12, 2008 05:43 AM
Video future hobbies discount FPV setup wedge98 Electric Heli Talk 5 Oct 20, 2007 09:28 AM