A fix for GPS loss during video recording

[HunterSK]

Credit deserved

I’d just like to give a shout out to Burlbark (a member on this board) for providing me with some invaluable tips and advice in helping me resolve a very annoying GPS loss issue.

Burlbark runs a drone repair business and I read his regular updates on his Facebook page (search for USdronerepair). He seems quite enthusiastic about this area of work, and is a regular contributor on this board. Over the past few weeks we’ve been exchanging some rather long chat sessions, which he freely offered as I’m in Australia and don’t really want to ship a drone across to the US, so I’m giving him a shout out here.


Background

My issue was loss of GPS lock when recording video. Having no GPS lock isn’t the end of the world as I prefer to shoot video while in ATTI mode.

However if the Phantom is unable to obtain at least 6 satellites while shooting video, in the event of loss of communication to the controller, this would result in automatic landing due to insufficient number of satellites received (as auto-home cannot work). That is unacceptable!

I own two Phantom 2 Vision+ (the original v2 units), both units are affected by a well-known loss of GPS lock when recording video. I’ve installed additional GPS antenna shielding, ferrite chokes, shielding of the GPS cable. All of this had some small measure of improvement but I was still losing at least 3 satellites, which is pretty critical because I may only receive 6-7 satellites in some areas.

I’ve spent several weekends experimenting with grounding the shielding, installing more shielding, ordering 3M EMI-specific shielding tape. All of this didn't resolve my issue, I was losing 3-5 satellites when recording video, and gaining them back when I stop recording video.

As I couldn’t measure what frequency, power and exact source of the EMI (I know it is likely the gimbal/camera assembly), it was like grasping at straws.


GPS cable contact issue - possible issue?

Well after several weeks of futile effort, I came across a post by Burlbark on the increasing the tension on the GPS cable’s side contacts. This involves removing the plug, then carefully bending the two side contacts on each pin (of the 4-pin cable), then reinstalling the plug. This increases pressure of the GPS pins when the plug is inserted onto the Phantom’s motherboard and makes better contact.

Result of testing

Based on some preliminary testing after five 15 mins flights today and recording video throughout, I saw loss of between zero and one GPS satellite while video recording. This is much better than 3-5 satellites loss than I normally encounter.

I’m not calling it a success quite yet, as I want to do another 10 flights in various conditions and solar Kp index environments before I can make a more confident conclusion.


Conclusion

Thanks to Burlbark for all his help thus far, much appreciated as I would have never guessed that the GPS cable contact on the main board would be the issue when recording video.

This issue does throw up a number of questions:

1. Assuming that it was poor contact of the GPS plug on the main board; what is the mechanism for RF coupling to occur between the camera and the GPS receiver and why is the effect is more pronounced when video recording is activated?

2. I’m curious to see what the RF spectrum looks like coming from the camera unit. Burlbark has posted an RF plot before, but I’d like to know which parts are noisy and the effect of video recording.

3. How much care did DJI put into EMC compliance when they designed this thing?

The Phantom 2 Vision+ (ver 3) has additional shielding and ferrites installed shortly after the community made the discovery on this board. Seems to me their solution was not borne out of detailed engineering analysis but perhaps a quick-fix?

I want tools that deliver capability, not toys.

 

[msinger]

This involves removing the plug, then carefully bending the two side contacts on each pin (of the 4-pin cable), then reinstalling the plug.

Got any tips on this step? Did you use needle-nose pliers?

 

[burlbark]

This involves removing the plug, then carefully bending the two side contacts on each pin (of the 4-pin cable), then reinstalling the plug.

Got any tips on this step? Did you use needle-nose pliers?

I love it when there are success stories.... :lol:

All you need to pull the plug apart and bend the contacts in just a little, is a small screwdriver. I started doing this to all Visions coming in for repair.

Once again in a vibrating platform this plug is not secured and it moves. Just like solid motor wires it begins to fatigue with time. I use a dab of low temp hotglue on the plug now.

I found this issue 5 months ago but did not actually realize what was happening. I was probing around with my Spectrum analyzer and some near field probes and saw a bunch of 1575 mhz being emitted near the plug. I thought it was residual. What it may have been was signal loss from the plug resulting in half the signal strength being emitted like an antenna from insufficient contact.

Jeremy James

 

[Brymel]

I had problems obtaining satellites for long periods of time here. Then I found and bought this - http://www.ebay.com/itm/151551733854?ss ... 1558.l2649
Improve GPS signal on Phantom 1 2 FC40 Vision and Vision +. Very well technically made shield. It has been working very well. I went from 3 to 5 satellites to 7 to 9. The most I have gotten was 13 at a different location. I not telling anyone to do this, just sharing my story.

 

[RichWest]

Please follow-up either way, this an interesting solution based on all that you have done to date. Experienced my first issue with loss of 3+ satellites while recording just yesterday. I had done some mods for improving FPV video feed speed and now having backed those off, I now have a max of 1 satellite loss when recording...

So grounding the GPS shield didn't help? Was the method a soldered lead to the shield? Where did you ground the wire?

 

[robinb]

I had problems obtaining satellites for long periods of time here. Then I found and bought this - http://www.ebay.com/itm/151551733854?ss ... 1558.l2649
Improve GPS signal on Phantom 1 2 FC40 Vision and Vision +. Very well technically made shield. It has been working very well. I went from 3 to 5 satellites to 7 to 9. The most I have gotten was 13 at a different location. I not telling anyone to do this, just sharing my story.

Sorry but 45 dollars shipping is a joke along with the price at 23 dollars.

I found the use of 0.0005 c bit of velcro at the back of the GPS connector has improved reception no end.
Probably bit like the bending of pins to make the contact much firmer.

 

[msinger]

So grounding the GPS shield didn't help? Was the method a soldered lead to the shield? Where did you ground the wire?

Grounding? I don't think that will solve anything.

You should do this and then tighten the contacts on the GPS plug (as described above). Both solve different issues.

 

[HunterSK]

Got any tips on this step? Did you use needle-nose pliers?

I used the smallest watch/jeweller's screw driver I had to press on the four indents to allow the GPS plug to be released.

Then I used the same screwdriver to bend the contacts on the connector. Each wire has a left and right connector, so I pushed both in gently. A better tool would be a small blunt instrument to avoid damaging the connector.

This is what I did:

 

[HunterSK]

So grounding the GPS shield didn't help? Was the method a soldered lead to the shield? Where did you ground the wire?

I used an Anderson quick disconnect, one connection only (I used two Black ones). Made up two cables of an appropriate length, and crimped both to their Anderson connectors as required.

I soldered one wire to the negative battery terminal on the motherboard. I assume this is common ground of the aircraft. The terminal is located directly underneath the WiFi Controller box. Because of the large current on the motherboard's battery connection, you need to crank up the soldering iron power as the motherboard's power ground pane around the battery connection is a substantial heatsink. My soldering iron is adjustable up to 60 watts and it was barely adequate.

I then soldered the second wire to the EMI GPS shield I installed. The 3M tinned-copper material is easy to solder to.

When assembling the Phantom, connect the ground wire together first.

Do a continuity test to verify low impedance from the GPS shielding (including GPS cable) to a ground connection of the Phantom. I measured less than 1 ohm.

Then connect the GPS cable.

After all of that, I found no evidence that grounding improved GPS performance or resolved my GPS drop out issue.

 

[RichWest]

So grounding the GPS shield didn't help? Was the method a soldered lead to the shield? Where did you ground the wire?

After all of that, I found no evidence that grounding improved GPS performance or resolved my GPS drop out issue.

Wow, great effort, sorry it wasn't a success. Thanks for laying out your process, I learned several things.

I want to retract me having success, was out flying this afternoon and was losing 3-4 satellites once recording started. Bummer...

 

[Sandman1962]

I'm sorry you are still having loss of GPS satellite lock while videoing.
I might suggest : if you have more shielding available, filling in the gaps a bit more, (I think the shielding needs to touch the rest of your shielding, and then be covered in your non conductive tape)
Perhaps running the GPS cable underneath the thick shielding that was originally installed by DJI.
If you were so inclined, the GPS antenna can be upgraded with a thicker antenna.

I've only been able to improve mine with the advice and tips found on this forum... But it
Has really improved my GPS performance , even while videoing.

Good luck!

 

[HunterSK]

Thanks Sandman1962. as you can see in my original post, the loss of GPS while video recording appears to be resolved by improving the contact on the GPS cable. This was in additional to the GPS shielding.

I did another two flights today in a different location with a higher Kp index (was at 3 during the test), and GPS remained steady during video recording. I will do more testing over this weekend to be more conclusive and will report back in a week.

 

[DJIDrDan]

My new Vision plus has the same problem, DJI is going to replace it via Droneworld Cape Town even though I bought it from US Hobbies, Thanks to Anthony who contacted DJI in China.

 

[RichWest]

GPS cable contact issue - possible issue?

Well after several weeks of futile effort, I came across a post by Burlbark on the increasing the tension on the GPS cable’s side contacts. This involves removing the plug, then carefully bending the two side contacts on each pin (of the 4-pin cable), then reinstalling the plug. This increases pressure of the GPS pins when the plug is inserted onto the Phantom’s motherboard and makes better contact.

Result of testing

Based on some preliminary testing after five 15 mins flights today and recording video throughout, I saw loss of between zero and one GPS satellite while video recording. This is much better than 3-5 satellites loss than I normally encounter.

I’m not calling it a success quite yet, as I want to do another 10 flights in various conditions and solar Kp index environments before I can make a more confident conclusion.


Conclusion

1. Assuming that it was poor contact of the GPS plug on the main board; what is the mechanism for RF coupling to occur between the camera and the GPS receiver and why is the effect is more pronounced when video recording is activated?

2. I’m curious to see what the RF spectrum looks like coming from the camera unit. Burlbark has posted an RF plot before, but I’d like to know which parts are noisy and the effect of video recording.

3. How much care did DJI put into EMC compliance when they designed this thing? You think they did any at all?

The Phantom 2 Vision+ (ver 3) has additional shielding and ferrites installed shortly after the community made the discovery on this board. Seems to me their solution was not borne out of detailed engineering analysis but perhaps a quick-fix? Yes, I agree!

I want tools that deliver capability, not toys.

Confirming my improvements by making more contact pressure for the GPS plug pins. I removed all the pins from the plug body, I only bent in one side of the contact pad. I went a step further, solder the terminated cables to each pin to insure the best bond to the plug/pins (pins in picture, not to the board pins). I was able to obtain 11-12 Satellites, normally I would have seen a max of 7-8 at this particular location at 75ft Alt. My bird has never had over 10 satellites from the day received.

Still haven't done the GPS foil mod, as I was trying to narrow down the source of interference. I did lose Sat numbers when recording was turned on, basically dropping to 8 and then would slowly recover to 10. I have a clearance issue at the moment with an added heatsink to the wifi module for heat related issues.

Hunter I appreciate your efforts to isolate the root cause. Your testing allowed me to make a stepped improvement. I'm beginning to believe the interference is sourcing from any number of connectors rather than short-comings of the shielding. My thinking, the unit had no issues in the first few weeks as it was built by DJI. I believe the connections and/or the cables are the source for deteriorating services.

My addition of Stage 3, bird side, did negatively impact my GPS signal. I believe the mod'd Rx board, to accept the new u.fl connected cables, could be the current issue for my installation. I mangled two Rx boards and will now go for a third, better tooled to tackle the black glue per Tony's tip for hot air removal.

Just sharing my results...Thank you HunterSK!

 

[flyNfrank]

This involves removing the plug, then carefully bending the two side contacts on each pin (of the 4-pin cable), then reinstalling the plug.

Got any tips on this step? Did you use needle-nose pliers?

I love it when there are success stories.... :lol:

All you need to pull the plug apart and bend the contacts in just a little, is a small screwdriver. I started doing this to all Visions coming in for repair.

Once again in a vibrating platform this plug is not secured and it moves. Just like solid motor wires it begins to fatigue with time. I use a dab of low temp hotglue on the plug now.

I found this issue 5 months ago but did not actually realize what was happening. I was probing around with my Spectrum analyzer and some near field probes and saw a bunch of 1575 mhz being emitted near the plug. I thought it was residual. What it may have been was signal loss from the plug resulting in half the signal strength being emitted like an antenna from insufficient contact.

Jeremy James

So you found this problem 5 months ago, but yet you never mentioned anything about it. But then you decided to perform a fix on all the visions that came in for repair from that point forward. You say you used a analyzer and discovered current bleed-out.

I don't understand why it is when you found such an issue like you say, that you did mention to anyone? Especially the guy that works with you and is noted as being your partner. How is it all these visions get this repair would done to them and the guy you work with doesn't know anything about it at the beginning of December? When I pointed out a problem with the GPS Connector in December your partner had never heard anything about it prior to the posting. He offered his help and offered to take pics or anything else to help find a resolution to the issue. He also began replacing the stock gps module with a 4mm one at that time.

Seriously Jeremy, I don't think you have to throw smoke & mirrors at the other pilots in this forum to get them to pay you any attention. From what I have seen by pictures, you obviously do decent work and you appear to be dedicated in the work that you do. But there really is no need here for all the self proclaimed stuff you post. Point I'm trying to make is, had you posted this issue 5 months ago as I did during the 1st of December, you may have helped several pilots save themselves from a lot of grief. The thread I have been referring to is here. viewtopic.php?f=27&t=30912

 

[burlbark]

I love it when there are success stories.... :lol:

All you need to pull the plug apart and bend the contacts in just a little, is a small screwdriver. I started doing this to all Visions coming in for repair.

Once again in a vibrating platform this plug is not secured and it moves. Just like solid motor wires it begins to fatigue with time. I use a dab of low temp hotglue on the plug now.

I found this issue 5 months ago but did not actually realize what was happening. I was probing around with my Spectrum analyzer and some near field probes and saw a bunch of 1575 mhz being emitted near the plug. I thought it was residual. What it may have been was signal loss from the plug resulting in half the signal strength being emitted like an antenna from insufficient contact.

Jeremy James

So you found this problem 5 months ago, but yet you never mentioned anything about it. But then you decided to perform a fix on all the visions that came in for repair from that point forward. You say you used a analyzer and discovered current bleed-out.

I don't understand why it is when you found such an issue like you say, that you did mention to anyone? Especially the guy that works with you and is noted as being your partner. How is it all these visions get this repair would done to them and the guy you work with doesn't know anything about it at the beginning of December? When I pointed out a problem with the GPS Connector in December your partner had never heard anything about it prior to the posting. He offered his help and offered to take pics or anything else to help find a resolution to the issue. He also began replacing the stock gps module with a 4mm one at that time.

Seriously Jeremy, I don't think you have to throw smoke & mirrors at the other pilots in this forum to get them to pay you any attention. From what I have seen by pictures, you obviously do decent work and you appear to be dedicated in the work that you do. But there really is no need here for all the self proclaimed stuff you post. Point I'm trying to make is, had you posted this issue 5 months ago as I did during the 1st of December, you may have helped several pilots save themselves from a lot of grief. The thread I have been referring to is here. viewtopic.php?f=27&t=30912

Why do you always come off confrontational. Please go back and read what I said and quit trying to assume everything. If you are not using some scientific method you are allowing your own personal bias to cloud your judgement.

You want me to post my findings but you dont want to me to proclaim that I found them?... This is just.... I cant even be upset at you :lol:

Smoke and mirrors..... What are you talking about? I have been 100% transparent and have shared all of my findings implicitly. In November I posted on the plug bleed off. Sorry if you didnt see it... http://www.rcgroups.com/forums/showthre ... 2&page=549

There have been much more pressing issues and most people found that shielding the plug and gps receiver was of enough benefit.

"I do decent work" What is your evaluation based upon, DJI?

Come on man.... I noticed this issue 5 months ago and brought it to every ones attention that I knew....I wasnt even a forum member here when I found it.

How about we just continue to try and bring things to light and remove all the weaknesses that we find. Hani and I work together and try to provide everything we can to this community. I dont know why you would try to drag him into this? Hani and I have only been speaking for a couple months he was not aware of all my findings.

I am glad your zip tie fix is working for you and many others, there is allot to be aware of.

 

[flyNfrank]

Your approach towards me a few times was not pleasant and not necessary. Your posting in the link you provided talks of needing shielding and nothing about the actual issue at hand. It may have been in line for what was being discussed in that thread, during that time, but nothing else. Your recent talks about modifying the plug pins would apply however.

The pic you posted, what's up with that? Are you showing the mods you have put together? If so that's great. Now that you mention it, it is the motor wire mod where I have the biggest issue with you. To begin with, you are way off by putting effort into convincing those that are clueless that the motors are problem, and that doing such mod is the cure. You then have people considering doing the mod who have never even picked up a soldering iron much less attempt to do the mod for fear of losing their quad. And that, is wrong!

So before I even make mention about people performing the mod, I want to point something out to you that you must not have thought about on modding the motor wires.

You put together a nice looking mod. Unfortunately the mod is band-aid and not a 100% cure. And because of this, what happens when you convince numerous people, like you have, to do this mod. And the eventually realize they need to install the correct ESC's. Now with the new ESC's they have new firmware that was not meant for modified motor wires. The spot in which the two wires have been spliced, of coarse has created a point of resistance. Regardless, they now are right back to where they started, which is ESC's and Motors combo that is mismatched for whatever reason. They possibly begin reporting their quads dropping from the sky all over again. And again they are involved in something they should never have to be. My opinion on this is there is no excuse for these guys to be in such a situation. And it's all a result of you misinforming from the get-go. I said all along to focus on the root of the problem. It had to be whatever drove the ESC's, which was the firmware. Tell other to wait and see how DJI would be taking care of it.

But if you just couldn't resist with your mod......You should have those with less abilities to send their quad to you to and let you do the mod. It doesn't take much at all to screw up by either not soldering correctly, forgetting to do something, or who knows what. Obviously you did not think about this. You always have to take in account who, or what type of person it might be that would be attempting the mod. Now the others that do know a little something, you could put together a "ready to go package" they could buy from you and then do the mod per your instructions.

Ok so I'm done.

 

[burlbark]

Your approach towards me a few times was not pleasant and not necessary. Your posting in the link you provided talks of needing shielding and nothing about the actual issue at hand. It may have been in line for what was being discussed in that thread, during that time, but nothing else. Your recent talks about modifying the plug pins would apply however.

The pic you posted, what's up with that? Are you showing the mods you have put together? If so that's great. Now that you mention it, it is the motor wire mod where I have the biggest issue with you. To begin with, you are way off by putting effort into convincing those that are clueless that the motors are problem, and that doing such mod is the cure. You then have people considering doing the mod who have never even picked up a soldering iron much less attempt to do the mod for fear of losing their quad. And that, is wrong!

So before I even make mention about people performing the mod, I want to point something out to you that you must not have thought about on modding the motor wires.

You put together a nice looking mod. Unfortunately the mod is band-aid and not a 100% cure. And because of this, what happens when you convince numerous people, like you have, to do this mod. And the eventually realize they need to install the correct ESC's. Now with the new ESC's they have new firmware that was not meant for modified motor wires. The spot in which the two wires have been spliced, of coarse has created a point of resistance. Regardless, they now are right back to where they started, which is ESC's and Motors combo that is mismatched for whatever reason. They possibly begin reporting their quads dropping from the sky all over again. And again they are involved in something they should never have to be. My opinion on this is there is no excuse for these guys to be in such a situation. And it's all a result of you misinforming from the get-go. I said all along to focus on the root of the problem. It had to be whatever drove the ESC's, which was the firmware. Tell other to wait and see how DJI would be taking care of it.

But if you just couldn't resist with your mod......You should have those with less abilities to send their quad to you to and let you do the mod. It doesn't take much at all to screw up by either not soldering correctly, forgetting to do something, or who knows what. Obviously you did not think about this. You always have to take in account who, or what type of person it might be that would be attempting the mod. Now the others that do know a little something, you could put together a "ready to go package" they could buy from you and then do the mod per your instructions.

Ok so I'm done.

There is no resistance at the motor splice at these current levels. In fact the cross sectional surface area in the wire is 25X smaller than the solder joint used to bind the wires. Heat is produced by resistance, with the amount of wire stripped of insulation and the amount of bonding area there is none at these current levels and the wire would need to reach 650f for the solder to give way. Its not going to happen from the joint, it would take catastrophic failure of the motor, esc or quad.

So now by showing a mod that can be done to improve the reliability of the quad you are saying that I am negligent..... This is unsightful thinking. Everyone can determine their own skill level. It is up to the user to determine this. With your line of thinking no mod should be posted because the public might not have the skill level.

You accused me of "smoke an mirrors" and here once again I have shared all that I can share.

Do you see how many times you have contradicted yourself in 2 posts? How about we both use our energy, that we plainly have plenty of, to further the advancement of these multirotors.

I like your idea of a ready to run package, but I dont know if the folks can handle mounting the esc's. The grounding wire is wrapped around the garbage low temp insulation and can melt into the power lead easily. Just food for thought for those planning on replacing the esc's.

 

[burlbark]

Your approach towards me a few times was not pleasant and not necessary. Your posting in the link you provided talks of needing shielding and nothing about the actual issue at hand. It may have been in line for what was being discussed in that thread, during that time, but nothing else. Your recent talks about modifying the plug pins would apply however.

The pic you posted, what's up with that? Are you showing the mods you have put together? If so that's great. Now that you mention it, it is the motor wire mod where I have the biggest issue with you. To begin with, you are way off by putting effort into convincing those that are clueless that the motors are problem, and that doing such mod is the cure. You then have people considering doing the mod who have never even picked up a soldering iron much less attempt to do the mod for fear of losing their quad. And that, is wrong!

So before I even make mention about people performing the mod, I want to point something out to you that you must not have thought about on modding the motor wires.

You put together a nice looking mod. Unfortunately the mod is band-aid and not a 100% cure. And because of this, what happens when you convince numerous people, like you have, to do this mod. And the eventually realize they need to install the correct ESC's. Now with the new ESC's they have new firmware that was not meant for modified motor wires. The spot in which the two wires have been spliced, of coarse has created a point of resistance. Regardless, they now are right back to where they started, which is ESC's and Motors combo that is mismatched for whatever reason. They possibly begin reporting their quads dropping from the sky all over again. And again they are involved in something they should never have to be. My opinion on this is there is no excuse for these guys to be in such a situation. And it's all a result of you misinforming from the get-go. I said all along to focus on the root of the problem. It had to be whatever drove the ESC's, which was the firmware. Tell other to wait and see how DJI would be taking care of it.

But if you just couldn't resist with your mod......You should have those with less abilities to send their quad to you to and let you do the mod. It doesn't take much at all to screw up by either not soldering correctly, forgetting to do something, or who knows what. Obviously you did not think about this. You always have to take in account who, or what type of person it might be that would be attempting the mod. Now the others that do know a little something, you could put together a "ready to go package" they could buy from you and then do the mod per your instructions.

Ok so I'm done.

There is no resistance at the motor splice at these current levels. In fact the cross sectional surface area in the wire is 25X smaller than the solder joint used to bind the wires. Heat is produced by resistance, with the amount of wire stripped of insulation and the amount of bonding area there is none at these current levels and the wire would need to reach 650f for the solder to give way. Its not going to happen from the joint, it would take catastrophic failure of the motor, esc or quad.

So now by showing a mod that can be done to improve the reliability of the quad you are saying that I am negligent..... This is unsightful thinking. Everyone can determine their own skill level. It is up to the user to determine this. With your line of thinking no mod should be posted because the public might not have the skill level.

You accused me of "smoke an mirrors" and here once again I have shared all that I can share.

Do you see how many times you have contradicted yourself in 2 posts? How about we both use our energy, that we plainly have plenty of, to further the advancement of these multirotors.

I like your idea of a ready to run package, but I dont know if the folks can handle mounting the esc's. The grounding wire is wrapped around the garbage low temp insulation and can melt into the power lead easily. Just food for thought for those planning on replacing the esc's. Besides I cant offer anything here I am not a vendor.

 

[flyNfrank]

I did not continue mentioning the smoke and mirrors because I thought it was clear enough. You said on the GPS plug that you discovered 5 months ago there was a problem. But yet you didn't want to mention it until after it came up in my posting at the beginning of December.

You also said that since discovering it, you have been doing a fix on all the visions that come into your shop. I pointed out how it was odd that your own partner didn't even know about these quads being fixed in house. I shared a link where if you looked on page 9 on December 9th he was just learning about such a problem. Otherwise the man would have said how you guys had been addressing the problem.

You also then mentioned how this area of the GPS Plug needing to be shielded. You say you do place a small amount of hot glue, but nowhere do you show anything you have been doing to address the shielded issue.

Overall it simply gets to the point in which people begin asking themselves if Jeremy knows what he is talking about, or knows what he doing, ect.

I can get along with you. I just ask that you be considerate to those that could be less inclined, and in a desperate situation.