7D wireless flash trigger can't expose

What about lowering the built in flash value, try -1 or less??

I did a couple of more tests and I've come to the conclusion that there's something definitely off with the 7D as the master. Using the 550EX in Group B causes a little bit of overexposure, although not as much as with the Metz 58 AF, except for once when the shot was really blown out.

Also, when Group A flash is turned off, the WB of the image shot with Group B flash is different (too bluish) and not consistent with the WB when Group A flash is turned on. Using the 580EX II as the master does not produce this WB oddity nor the overexposure from the Group B flash.

This definitely needs further looking into when I have a little more time for a more thorough testing.

The built in flash is set not to contribute to the illumination.

I tried it now with only one slave, and internal master.

I actually get the same amount of overexposure regardless of whether I set FEC to 0, -3 or +3. If I on the other hand change the aperture from f/5.6 to f/16 (also three steps), I get a noticeable underexposure.

If I instead put the 580 EX II on the camera, changing the aperture makes no difference at all to the exposure, which is as it should be, when using E-TTL.
On the other hand, when setting FEC as -3, 0 and then +3, I get underexposure, correct and then overexposure, when using an external master. The image taken with +3 is actually about as overexposed as they always get, when using the internal master.

Yet another interesting twist:
I used the internal as non-contributing master, one 580 EX II as a slave. I held the 580 EX II on top of the lens, with the command receiver window facing the internal flash on the 7D and the reflector of the 580 EX II facing the cat (who now is pretty tired of all flashing, which prevents her from sleeping!). So this arrangement is equivalent to having the 580 EX II directly in the hotshoe, but it's controlled indirectly from the internal master.

Now, believe it or not, now exposure is correct (slightly blown highlight in the white fur on the cat, but that's excusable) and it reacts properly to FEC as well. -3 gives underexposure and +3 overexposure, just as it should.

Summary: My 7D can't control slave flashes that are bounced into a 2.5 meter high white ceiling, but it can control them if they are aimed directly towards the target. It can also do it with a 580 EX II used as a master, and then it works in all configurations.

Update: I've now walked around in my house, carrying around one or more flashes (all 580 EX II) and shot images of walls, furniture, curtains and I don't know what else. Not the cat, she got enough of this and went out.
The funny thing is that occasionally it works. When it does, it sometimes works perfectly. I can return to a place were I couldn't get a decent bounced shot a few minutes ago and do really well. It even reacts reasonably to the FEC setting. Then, a little later, it's as it was before again. Everything from pretty much to disastrous overexposure, and no reaction to FEC at all.

It will be difficult to claim a warranty case here, if the tricky thing refuses to malfunction when I claim my case...

I suspect firmware. If it was the camera's metering system, it ought to fail just as miserably when using a 580 EX II as a master, as it's the camera metering the flash light in all cases.

Not a single time during several hundred images have the camera said "Busy", though. Somebody else had that kind of problems when triggering with the internal flash.

On my 7D ETTL, the is 2.66 to 3 stops underexposed using the in camera commander. I noticed this doing MA on my lenses, but blew it off as an adjustment I missed in my haste to MA. With the tests I've just done it doesn't matter if it's bounced or not. When a 430ex II or 580ex II is in the hot shoe, all works well. I chalk it up to firmware growing pains. The 7D is wonderful. I can live with the FE comp in the interim. I'm pretty sure the problem existed before I updated the firmware.

Ralph

I meant
=On my 7D the ETTL is 2.66 to 3 stops underexposed using the in camera commander=

Also, my 580ex II is inconsistent in master mode, in slave it's underexposed as the 430.

Interesting to see that other's cameras are haunted by about the same problem. As it sometimes goes away, but then returns again, it could be that they have some timing problem with the metering in the camera, when using the built-in wireless commander.
I've not seen it fail once when doing the old fashioned way, i.e. with a 580 EX II being the wireless master on the 7D, only when using the built-in.

Chuck Westfall next, maybe?

430ex II as slave, on board flash disabled, under expose 2.5 stops
430ex II as slave, on board flash enabled, better exposure
430ex II as slave, 580ex II in hot shoe as master, master disabled, under expose 2.5 stops
430ex II as slave, 580ex II in hot shoe as master, master enabled, better exposure
non-wireless mode, either 430 or 580 give good exposure in hot shoe

Your's is worse than mine, since mine seems to always do it correctly when I have the 580 EX II as a master.

Might it be that the exposure is relying on HSS shot (which the 7D controller cannot handle)?

No, it's not. Correct exposure for ambient light in the situations I tested was anything in the range 30" to 1/15 s (dark to pretty dim) with f/5.6 and ISO 500. I used 1/125 s and f/5.6, so ambient was quite a lot underexposed, when I took the photos of the cat. At 1/15 s there some of the ambient came in a bit, but not much.

Besides, the 580 EX II, when used as a master, wasn't set to HSS either.

So your idea is good, but the answer is no.

cL0d wrote in post #8896619
Might it be that the exposure is relying on HSS shot (which the 7D controller cannot handle)?

not hss here either. I'm going to do a side by side with my 40D to see how different the exposure is.

windpig wrote in post #8896239
430ex II as slave, on board flash disabled, under expose 2.5 stops
430ex II as slave, on board flash enabled, better exposure
430ex II as slave, 580ex II in hot shoe as master, master disabled, under expose 2.5 stops
430ex II as slave, 580ex II in hot shoe as master, master enabled, better exposure
non-wireless mode, either 430 or 580 give good exposure in hot shoe

OK, what that tells me is without a flash burst from the position of the camera it underexposed - whether that be the built-in or 580. Was the second light positioned to adequately light the subject, and strong enough? Are you sure the off-camera light is really firing, and what you are seeing isn't (2.5 stops under) just from ambient?

I am not trying to say you guys are wrong, but narrowing down the circumstances in which troubles are occuring. That is, I am wondering if the glitch has to do with channel selection or something else. I tried different combos with my 580 classic (channel A slave) and 580ex2 (channel B), and while the exposures weren't always perfect (nothing new for ettl) they were not off by more than a stop. What I didn't think to try was just the 580ex2 by itself... that flash has always been flaky for me.

Try changing your flashes zoom to match the focal length. I noticed you had it set at 24mm and you were shooting with a 300mm lens. I don't know if it would affect it that much but I noticed that when using my 430EX II off camera I still need to manually adjust the zoom on the flash otherwise I get really inconsistent results.

I would try posting this in the Flash section as I think somebody else will figure out the problem for you. I haven't mastered off camera flash yet and there are some really good people in there.

The 24 mm zoom was for the master, which was set not to contribute to the exposure anyway. The default zoom setting for the master is 24 mm, in order to cover the slaves as good as possible. But it's of no importance at all in my case, since when I use the 580 EX II as a master, everything works perfectly. It's when using the internal flash as a master I get problems, and that one can't zoom at all.

I considered posting in the flash forum, but I'm of the opinion that this is a camera problem, and wanted to hear if anyone else saw the same thing, and also inform those considering the 7D.

After a side by side quick test with my 40D it looks to be user error on my part. The two cameras seem to meter differently. In ETTL the 7D really nails the exposure a lot, but also fails. I'm going to chalk it up to shooting in ETTL. Would the best test involve using evaluative or average?