Re: OpLink trouble
« Reply #60 on: July 29, 2016, 09:08:40 pm »
forgot to ask, what kind of range should I expect, also the one ground unit that is not shorted has a push button switch next to the antenna, would you happen to know what it's for?

hwh

  • *
  • 1018
Re: OpLink trouble
« Reply #61 on: July 30, 2016, 02:06:44 am »
...push button switch next to the antenna, would you happen to know what it's for?

Originally it was a "bind" button similar to the ones on regular RC transmitters/receivers.   Current (and future) firmware ignores it and you set the coordinator ID in GCS to "bind" the modems.

Re: OpLink trouble
« Reply #62 on: November 02, 2016, 12:03:55 am »
I get my oplink working with revo for telemetry. But PPM does not work until now.
My radio is HiTec Eclipse 7. Please help me out. Firmware is 15.09. BL is 4 on Oplink and 6 on Revo.

f5soh

  • *****
  • 4572
    • LibrePilot

karla

  • *****
  • 629
Re: OpLink trouble
« Reply #64 on: November 03, 2016, 08:21:13 am »
I have bought at least 4 sets of revo FC /oplink ground combo kits. two won't link at all, the others link and I get about 60db on the graph, but no matter what I do I can't get over a 2 to 3 foot range and it drops out.

I just did buy 4 set of Revo + OpLink just like alnrenny, from two different suppliers, I think... its hard to tell. These two seem different though since one has a black casing for the Revo with the pins color coded in Yellow, Red and Black on the board, a nice black case for the OPlink board, and the redundant 'bind' button is there but only 2 white ports (Main and Flexi),
http://www.thanksbuyer.com/openpiolot-cc3d-revolution-flight-controller-upgrade-version-integrating-oplink-revo-44119

the other supplier got blue casing for the Revo board, no casing for the OPlink and that has no 'bind' button on it but has 3 white ports (have no idea what the 3rd one is for).
http://www.thanksbuyer.com/openpiolot-cc3d-revolution-flight-controller-board-oplink-mini-cc3d-revo-transceiver-tx-rx-module-40847

These two suppliers I have noticed are distributed on all major platforms like
Banggood.com
eBay.com
DHgate.com
Aliexpress.com
Thanksbuyer.com
The platforms differs but products are the same.

I was very dissatisfied with the result of trying them out for signal strength.
dont know about dB but the Signal strength meter in the GCS show at best +10 (middle, pointing straight up). The distance was some 3-10 meters indoors.

My problems were all on the oplink end but I only have one revo so it's hard to draw any conclusion.  Most of the time I was running oplink to oplink tests or measuring the frequency on a spectrum analyzer.

Other than the two oplinks that were way off frequency (about 100 kHz) the other's frequencies were close enough to link and work after the antenna connector short was removed.

The antenna short is easy to check if you have a meter.  The oplink's center antenna connector pin to the outside shell should read open.  If it reads shorted or any resistance value it's a bad board.  Someone else who had a low power one read 6 ohms instead of a complete short. Anything other than open (infinite resistance) is bad.

Now, inspired by hwh, I just ran a continuity check with a multimeter between the middle pin of the antenna and any one of the surrounding 4. It should not be a short circuit here since then the antenna has no meaning.
. 4 out of 4 revo boards had no continuity - good
. 4 out of 4 OPlink boards did have continuity - bad, 0 or 2 ohms, full short circuit.

I feel good about finally getting an explanation for the poor range but also depressed by this.
I dont want to buy more OPlink boards.

To fix the issue: Maybe some tips on How to repair them? Would be great.

To solve the problem: since the issues with the hardware is widely documented in this forum.
Is it possible for LibrePilot to launch some kind of Product Quality Supplier (QSP) license and have a link to suppliers that sent in good boards? I cannot test board myself for the frequency spectrum mumbo but I really willing to help form such a program. This situation is not really good.

What do you think?

Mateusz

  • *
  • 808
Re: OpLink trouble
« Reply #65 on: November 03, 2016, 08:57:33 am »
It is very complicated. There are issues that conflict eachother:

  • Design files for Revo and Oplink are released for public under CC-BY-NC license. Which means Creative common (CC) - Attribution (BY) - Non commercial (NC). This implies that legally you are not allowed to produce Revo/Oplinks to sell them and make profit, you can only assemble board for hobby or own use.
  • Cloners don't care about Copyrights and make derivatives anyway, some better some worst and you can't control that externally.
  • It is ok to use compatible hardware, not derived. There are very nice compatible boards designed from scratch like FC Faclon which are high quality, or Sparky2 boards released fully as open-hardware but authorized supplier chargers 2x the amount that clone costs.

It's also by nature non-profit project, gathering people that like to tinker with hardware and software, more like DIY. On another hand, selling high quality hardware would require commercial steps which might not be the route LP guys want to take.

Best you can do is test hardware you buy, demand from seller full specs and return it if it's not what was advertised. Many suppliers or payment companies have customer/buyer protection and you are allowed to demand refund/return in case if item is broken. It's usually plenty of time to test it. Returning not functional items, eventually force clones to do the hardware properly.

I got my OPlink from Banggood and it works fine with 5 certified sparky2 boards, but I re-call that there was a discussion that some cloned OPlinks had their frequency a bit off (hwh might know more). You need software defined radio (SDR) dongle ($25-$30) to detect that as long as it supports waterfall spectrum AFAIK.

If you have power meter, then lucky you, that's expensive. But if you find short on your board, maybe check for short under magnification glass, sometimes there are solder joints. Maybe all your OPlink boards can be repaired.

Update:
By the way, OPLinks use standard HopeRF RFM22B 100mW 433MHz module, so it should be possible to get one of these
http://shop.top-electronics.eu/20dbm-transceiver-module-433mhz-smd-p-16847.html
and order PCB from OSHPark, and assemble it. Brian (https://forum.librepilot.org/index.php?action=profile;u=264) designed some Nano OPlink boards for his own use https://oshpark.com/profiles/webbbn
but ask him for BOM if you want to play in DIY project :)
« Last Edit: November 03, 2016, 09:10:34 am by Mateusz »

hwh

  • *
  • 1018
Re: OpLink trouble
« Reply #66 on: November 03, 2016, 09:54:10 am »
...To fix the issue: Maybe some tips on How to repair them? Would be great. ...
The shorted ones are shorted on an internal pcb layer so they can't be fixed in the obvious way.  There are two fixes:
  • remove the antenna connector, drill the hole the center pin went in bigger until the short goes away, and replace the connector with a pigtail or a connector that the center pin is insulated where it goes through the board so it doesn't touch the exposed inner ground layer.
  • ignore the connector, cut the trace between the connector and the rfm22b module, and solder a pigtail with connector directly to the rfm22b.  This one is fairly easy to do.  The picture below is an oplink I destroyed testing so it doesn't have the connector or rfm22b on it.  The red circle shows the trace to cut and the first pin on the rfm22b that the trace went to is where to attach the center conductor of the pigtail.

The ones that are off frequency f5soh has a change he's been working on that allows you to tune them back on frequency if they're not too far off.  I don't think it's been submitted or merged in yet.

...Is it possible for LibrePilot to launch some kind of Product Quality Supplier (QSP) license and have a link to suppliers that sent in good boards? ...
While I don't speak for the project (no one really speaks for the whole project)  I don't think any of us would really want to get involved in doing something like that.  Both from the standpoint of not wanting to do it and because as Mateusz said, the designs are a type of open source with a non-commercial clause and we don't own the rights to control it.

Almost forgot, the third 4 pin white connector is a debugging port used to debug the firmware.  A device called a STLink or a JLink is attached to it.

karla

  • *****
  • 629
Re: OpLink trouble
« Reply #67 on: November 03, 2016, 10:40:41 am »
Thanks a lot. The responsiveness and depth of knowledge here is just awesome.
As usual it will take me 1-2 weeks to digest it :)
K

karla

  • *****
  • 629
Re: OpLink trouble
« Reply #68 on: November 04, 2016, 12:36:10 pm »
  • remove the antenna connector, drill the hole the center pin went in bigger until the short goes away, and replace the connector with a pigtail or a connector that the center pin is insulated where it goes through the board so it doesn't touch the exposed inner ground layer.


hvh, If I go with the option to use the 5 pin original connector, suppose I get the center pin through the board without touching the exposed inner ground layer, How do I get it securely connected to the rfm22b module? By finding the 'trace' indicated on your picture? and then solder it to that? You have a picture? How did you do it?
 
Alternatively What do you mean by 'solder a pigtail with connector directly to the rfm22b'? What pins on the rfm22b and what thing to solder on there? Picture?

Sorry to ask so much, nu hurry to answer, I will be glad whenever, since this will go nowhere at the moment.
Best
K

hwh

  • *
  • 1018
Re: OpLink trouble
« Reply #69 on: November 04, 2016, 06:13:41 pm »
I actually hate those tiny connectors so I never actually did one using option one.  You're right, as long as the center pin hole has been drilled out enough to remove the short you can just put a tiny jumper wire to connect the new connector center pin to the spot I circled in red.

For option two you can use a "pigtail" wire like the attached picture.  A lot of the oplinks come with one like this or you can buy them on various sites.  Just cut off the small connector and solder the wire to the board.  The braid goes to any convenient ground like the 4 pins on the connector.  The center conductor goes to the rectangular pad to the left of my red circle.  Even with the rfm22b module on the board there's enough pad exposed to solder to.

I've misplaced the one I repaired so I can't post a finished picture.

karla

  • *****
  • 629
Re: OpLink trouble
« Reply #70 on: November 05, 2016, 06:29:06 am »
I just wanted to test if this will work and soldered an antenna straight on to the rf board.
It did! Signal meter now at maximum and stable and no drop outs.
Will do a nicer and more permanent fix.
Thanks a lot!
K

hwh

  • *
  • 1018
Re: OpLink trouble
« Reply #71 on: November 05, 2016, 07:42:08 am »
You're welcome  :)

Your pictures make a good example of the fix.

karla

  • *****
  • 629
Re: OpLink trouble
« Reply #72 on: November 19, 2016, 01:22:08 pm »
Now I have done several repairs using the other method, drilling a the center hole larger and re-solder the 5 pin connector. The reason this is preferred is because its impossible to fit the casing on to the OPlink board after you have soldered a cable directly on to the rf board. I have fixed 4 boards with this method now.
Will try to upload the pictures.
It fixes the short circuit but other issues remains...
« Last Edit: November 19, 2016, 01:35:10 pm by karla »

karla

  • *****
  • 629
Re: OpLink trouble
« Reply #73 on: November 19, 2016, 02:22:42 pm »
So fixing the short cut on the Oplink makes a much stronger signal on the Link Meter and before I could only get 10 m before failsafe it is dramatically improved.
Good.
However taking the quad to the field I discovered that there are generally a good signal strength but interrupted by sudden, short drop out of connection, that will will pick up again to a high level signal, like in 1-2 seconds.
I think this has been reported all over this forum already.
The video clip of the Link Meter I have is from flying around fpv with a Revo board on the quad and a Oplink on the ground. I assume the problem is all on the Oplink side but I do not know. I tried both with a Bluetooth connection to the GCS Mac and with a USB cable to the Mac and it makes no difference better or worse the Link strength or the dropouts. BT seems not to be the culprit here.
The video clip shows that even before taking off the ground and just standing still 10meters away there are dramatic dropouts of the signal strength. Then flying never further than 150 meters and 45m high the signal is pretty constant but the dropouts are there intermittently.
Now I noticed nothing of this while flying but looking at the capture file I see these scary lost signals. If flying later on away say 300 meters or more then it seems 100% sure failsafe will happen and my quad will crash to the ground.
I don't like that at all.

So, is this a case of a bad radio frequency match between the Revo and the Oplink?
How to fix some kind of tuning? I don't have a radio spectrum analyzer and would not know how to use it.
Are the radio modules on the Revo better than on the Oplinks? can they be swopped?

Will try to upload the video link...
« Last Edit: November 19, 2016, 02:42:51 pm by karla »

Re: OpLink trouble
« Reply #74 on: November 20, 2016, 12:44:54 am »
Some clone OpLinks use a bad RF board.  The bad RF board is off frequency by more than one channel IIRC.  Bad will only talk to another bad.

User HWH has the most experience with this.  I recall hearing that someone had some test code that allowed you to retune these bad boards.