Canon 600EX II-RT drops link- HELP!

[Phillyphoto]

Scenarios I have tried: 

Using Canon Transmitter ST-E3-RT as the master and (2) 600 EX II-RTs as slaves.

Using A 600EX as master and B 600EX as slave.

Using B 600EX as master and A 600EX as slave. 

I get the same result of the slave dropping link. The time varies. Sometimes it drops link in 4 mins, sometimes 10mins, sometimes 20 or more minutes. The only way to relink them is by turning everything off and back on.

All channels are the same. Yes, I have scanned for the best connection as well as every other channel and AUTO.

All IDs are the same. 

Not near a wifi-router or airport, I'm in a row home in Philadelphia. 

Using NiMh rechargeables and using freshly charged batteries for every test. Batteries are about 2 years old. 

I have spoken to 2 Canon service reps and neither of them has any idea what the problem is. I really don't have the money to spend on sending everything in for "repair". 

Any help is greatly appreciated, thanks!

[halocastle]

I don't know if you ever watch Bob Davis -- who is a Canon Explorer of Light -- YT videos.  He rave(ed/s) about Canon's RT system (which is a big reason I gave up on 3rd party solutions and went pure Canon) but then readily admits to using RadioPoppers (which operate at ~902-928 Mhz) as a backup for when the RT system, "...isn't quite adequate."

When I first saw this -- searching around for a fix after experiencing drops -- it kinda pissed me off.

ref: https://www.youtube.com/watch?v=IFnsLU6JV_E

P.S. Canon -- whatever that means -- knows about it.

[archiea]

Canon needs to do two things:

1) redesign the transmitter/flash so if link is lost, the user from the transmitter can reset them remotely… or have a setting in the flash that re-seeks connection every minute, 5 minutes, 10 mi utes, etc based on user input. This way, as wifi standards update, it will extend this feature tpinntge flash.

2) offer past canon flash users either an upgrade price or a priority on new refurb models.  This will at least allow users some reason for brand loyaty. These flashes are 5x the Price of the knockoffs… don’t make me think going knockoff was the smarter way. 

[halocastle]

#1 isn't possible.  If they could do that, then merely changing the channel or ID yourself would fix the issue.  But you can't.  I can purposely change the channel/ID on a slave, lose link, then change it to the correct setting(s) and it re-links.  I can walk too far away or introduce too many obstacles, lose link, and when I correct for that, it re-links.  But once the link DROPS, there's nothing I can do to get them to re-link -- no amount of channel/ID tinkering works.  I must turn radio off (on the slave) then back on.  It's as if a drop mini-bricks the flash.

I'd love to know the mechanism that's causing the drops.  Is it a frequency shift?  Like a machine designed to play only the black keys of a piano in a pitch black room, if it gets dumped...oops.  And neither the machine, nor the piano, "know" anything's wrong.

I believe Canon has too many IDs -- 10,000!  Other systems, which have IDs, have, at most, 100.  I doubt they're going to do anything firmware wise to fix this as the EX series must be sent to Canon for updates.  The EL series can be updated via the camera.  Maybe when the EL-5 comes out they will offer an update (ID reduction/greater freq spacing) just for the EL series.

I've been having great success with channel 15 and book ending my ID to 9999 or 0000.  This (guessing) limits frequency shift to just one direction.

And lastly, every trigger system in the 2.4 GHz band has issues.  Just spend a little time on a Godox forum!

[Mannies17]

I too found the most success with Ch 15 0000. 

Has anyone tried the RT noise flash sleeves they sell on B&H? People used to need them for the Pocket Wizard/580ex issue. Noticed someone suggested this on You Tube comments. Experimenting this week. I figured $12 was worth a shot for the time I’m wasting resetting these flashes WAY HIGH UP on light stands during an EVERY MOMENT COUNTS wedding. Before I lose even more money switching my whole lighting setup to Profoto. I am dumbfounded I have to say these things to Canon. 

[Steppy]

1) This is already implemented but appears to be broken. As holocastle mentions this does indeed work correctly if the link is lost when the RF signal is temporarily lost due to range issues. As soon as the slave loses the connection it immediately attempts to restore the connection (without delay). This is standard implementation for a wide range of connection types. Unfortunately, in the case we appear to be discussing here, and the issue I have indoors, both the master and slaves get screwed up and lock up completely. This reconnection system no longer works. This could be due to either a hardware issue, or more likely a firmware issue.

[Steppy]

The number of IDs is irrelevant. They are almost certainly just numbers transmitted as part of the message to identify which master and slaves belong together in one set. I'm working on a device to connect existing 2.4GHz system which uses a 4 byte ID in the same way. The value 9999 would probably be transmitted as 270F hex and reducing the number of IDs would simply reduce the size of the message by 1 or 2 bytes. This would have no effect on performance or the reliability of the connection.

As you mentioned in a previous post Canon provides 15 channels whereas the 2.4GHz WiFi band doesn't allow this number of frequencies anywhere in the world. The system I mentioned above also provides 15 channels but these are closer together than the WiFi channels. This is not actually very relevant because the bandwidth of the channels is actually wider than their separation so they overlap anyway. Incidentally, the IDs of the system I'm working on are not number in order of their frequency but appear to be completely random - or at least I can't identify any logic to their order. I think we can assume that Canon has not used the same channel separations as for WiFi and the channel numbers are probably not comparable, if not even completely different.

[Steppy]

Every WiFi device both transmits and receives as do the flash master AND slave units. In my case I've identified the source of the interference to be 2 Apple HomePod Minis. If I pull the power on these two the flashes remain connected for as long as I've monitored them.

[Steppy]

I've continued my testing and the results may (or may not 😉 ) be of interest to others.

All Canon units are the first versions.

Firstly, I tested with all 15 channels and a wide range of different IDs but found no difference at all.

Master: ST-E3-RT
Slaves: 3x 600EX-RT
Slaves lose connection, master LED remains GREEN, apparently indicating that the connection loss has not been detected.
Connection can only be restored after recycling power on BOTH master and slaves (as halocastle reported it may be possible to just cycle the RF link but I've not tested this). Sequence is unimportant (slaves before master or master before slaves).

Master: ST-E3-RT
Slaves: 3x 600EX-RT + 1x YONGNUO YN600EX-RT II
Canon slaves lose connection. Yongnuo slave remains connected and continues to flash when triggered by master.
When YN600EX-RT II turned off and on again it fails to reconnect.

Master: ST-E3-RT
Slaves: 3x 600EX-RT + 1x YONGNUO YN600EX-RT II
Master initially turned OFF. All 4 slaves turned ON. Wait ca. 10 minutes.
Turn on master.
Canon slaves do not connect.
Yongnuo slave connects and flashes when triggered by master. 

Master: ST-E3-RT
Slaves: 3x 600EX-RT + 1x YONGNUO YN600EX-RT II
Master turned ON. All 4 slaves turned OFF. Wait ca. 10 minutes.
Turn ON all 4 slaves.
NONE of the slaves connect to the master.

Master: ST-E3-RT
Slaves: 3x 600EX-RT + 1x YONGNUO YN600EX-RT II + 4x Phottix Laso TTL Receivers
All Canon slaves lose the connection to the master. The Yongnuo and Phottix slaves remain connected and trigger when fired by the master.
If I only use 2 of the Phottix devices initially they remain connected as above, but if I turn on the remaining Phottix devices after the Canon slaves have been dropped they cannot connect to the master. 

Previous tests indicate the same behaviour when using a Canon 600EX-RT as a master. This is to be expected because the ST-E3-RT is probably just a simplified version of the flash, and probably uses very similar firmware and electronics for the link functionality.

Master: YONGNUO YN600EX-RT II
Slaves: 3x 600EX-RT + 4x Phottix Laso TTL Receivers
All Canon slaves lose the connection to the Yongnuo master.
Phottix slaves remain connected and trigger when fired by the Yongnuo master.
The Canon slaves reconnect to the master after cycling the power to the slaves only, with no further action required on the Yongnuo master.
If I only use 2 of the Phottix devices initially they remain connected as above, if I turn on the remaining Phottix devices after the Canon slaves have been dropped they RECONNECT to the Yongnuo master. 

[halocastle]

I don't believe the number of IDs is, "irrelevant."  Strong RF interference can corrupt these values and large ranges have less tolerance.

This is why you must toggle radio to fix the issue.  Otherwise, I could just change the ID by one then back again and it'd re-link (or the channel).  It doesn't -- a drop is FUBAR.  And using ID 9999 -- the top of the range -- (and channel 15) has fixed my drops entirely (fingers crossed).

All of this is rather moot, as I provided before (emphasis is mine):

Part 15 of the FCC Rules...(2) this device must accept any interference received, including interference that may cause undesired operation.

[Steppy]

You are of course free to believe what you wish 😉
The FCC rule you quoted simply describes good design practice for any interface. This is almost certainly what Canon fully intended, but it would appear that they have simply failed in some way. Yongnuo and Phoenix have managed to do a better job. It's very similar to a camera firmware issue except we cannot simply download and install a new firmware version (in contrast to the Yongnuo flashes!).