7D wireless flash trigger can't expose

windpig wrote in post #8900178
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?

Funny you should mention this. When I shoot with the flash metering mode set to Average, the resulting image is horribly overexposed. This is with the internal flash of the 7D as the master. No such overexposure occurs when the 580EX II is used as the master.

In my case, when it fails, it fails. In the situations where it overexposes, it overexposes regardless of whether flash metering is evaluative or average.
I also tried all other metering modes, for ambient, which shouldn't make any difference. They didn't.

Seems to be a systematic fault in Canon 7D's internal wireless controller. I've approached Canon's support with this, but no reply yet.

I see now that it's definitely not only my 7D, but indeed a fault in several of them. Probably in more than we hear about, as using one or more slave flashes, controlled by Canon's wireless system, isn't something every 7D user ever does.

Another report of the same problem

hmm i just tried this one myself....same result overexposure...its like the external flash doesnt know the ISO youve set the camera at...lowering the iso lessens the overexposure but the frame is still not exposed properly............

goddarnit....i decided right before posting that i should redo a shoot....different subject different location now the exposures are bang on..maybe sometimes you get a lil over or a lil under but very very minor differences...

im stumped....ill try again tomorrow and see if there are any differences....this might be a metering thing like suggested....

It seems you see the same thing as I do. Very often overexposure, but then suddenly it works for one single or a couple of images. Then back to overexposure again.

It looks like when you bounce the flash, the camera's E-TTL system isn't metering the flash illumination correctly, so it sends full power to the flash units that act as slaves. You can control exposure, but you have to use ISO and/or aperture, because nothing else works. FEC will not work, of course, since if the camera can't see the light from the pre-flash, it will always think it's too dark.

The obvious explanation is that it takes too long for bounced light to reach the subject, so the camera reads the illumination too early, and thinks that there's too little light, hence full power. Straight on, the light hits the target faster, and is metered correctly. But hey, how long time does it take for light to travel, say, three meters extra, for the bounce?

10 nanoseconds.

The camera's processors typically have clock cycles in that range, for one single instruction. Could there really be a timing issue with this? Could it be that the more powerful pre-flash from the 580 EX II, when used as a master, is able to trig the slaves some nanoseconds earlier, and that's the whole difference?

WHAT THE!!

i ate dinner, turned off the camera, flash etc....after dinner I tried it at the spot where i was having problems and now the exposure is great.....

maybe theres some sort of flash registration happening the first few times you use it? LOL

just kidding....this camera is really going to make us all relearn what we already know hehehe....

not a bad thing, just bad initially when we are still learning all these functions and behavior and spend hours being confused....theres a logical reason for all this.. we just havent unraveled it im guessing.
please note though that i am not making controlled test locations, the sofa moved a bunch..wasnt sitting in the exact same location but very very close...

This all supports my assumption that they have some kind of timing problem. It's on the verge of working, so sometimes it does, sometimes not. Mine also works, now and then. Without changing anything.

Makes it so difficult to proove your point if you take it back to the dealer, as it most likely works in the shop, of course.

I don''t see how the light could be "taking the light too long" (186,000 miles/sec is fast), but instead I agree that it is being fooled by soemthing. One thing that I tried was to set an "A" flash up shooting through an umbrella, and then a "B" flash pointed directly at the subject. The result: since I had the A:B ratio set at 1:1, the shot was way blown out. However, that makes sense in the context of canon flash: the A and B channels were set to the same power, with the metering preference for "A" (at least as far as I understand E-TTL2).

If you could retry your tests with BOTH flashes set to group A, you might very well get much better results. Alternatively, change the flash control from using a ratio (A:B) to just both (A+B). Also, the DISTANCE the flashes are from the subject matters as well - if "A" is much farther than "B", then "B" will be much stranger than you want.

For the halibut I emailed chuck gardner to see his take on all this

Well, 10 nanoseconds is the same as the clock cycle of a 100 MHz processor. 100 MHz isn't too fast these days. I don't know about the 7D, but the three years older 1D Mark III has 40 and 48 MHz RISC-architecture CPUs. Since they are RISC-based, they execute an instruction every clock cycle, normally.

So it's not that 10 nanoseconds of flight time for the light is far different to the timing the camera is working with. On the contrary, it's within the same order of magnitude.

Over to the flash system.
There is no "metering preference" for the A group. If you have one A and one B flash, both aimed at the same subject, but one 2 meters away and the other 5 meters away, the E-TTL II system will meter them separately. If you set a 1:1 ratio, the wireless flash commander will send two different power ratios to the A and B slaves, to make them contribute the same to the illumination of the subject.
The same goes of course for one direct and one bounced.

So it's not that the ratio setting controls the power level of the flashes directly. It does indirectly, but just to achieve a certain level of illumination. This is clearly described in Canons more detailed description of how their flash system works.

You can rest assured that during the more than 500 pictures I took, trying to figure out if I did something wrong, I have tested having all flashes in the same group, in different groups, used A:B as well as A+B+C as well as 7D:s own mode of just consider all external flashes being one large group (literally the A+B+C with the internal not contributing) and so on.

No difference at all. That's not the issue.

I've also verified that the slaves do change form manual mode to E-TTL and back. No problem there either. Unfortunately, you can't see the power commands to the flashes when using E-TTL, so I can only confirm that the camera sends the correct power levels when the slaves are in manual mode. That works fine.
Besides, if all these different group settings and so on actually would make a difference, that doesn't explain why it works fine as long as I use a 580 EX II as the master flash, but not with the same ratio (A:B) setting using the internal flash. Or why it works (in many cases, although not all) with the slaves aimed directly towards the subject.

The problem is definitely in the camera. Let's now just hope this will filter through Canon's organization quickly enough, so that they can kill two birds with one stone, now when they are busting the ghosts out of the 7D images anyway. Better for them to issue one update than two.

OK, Anders. I don't want to get into a contest with you, but clock cycles of the CPU have very little to do with how long the firmware waits. For one, as you mentioned, it is a RISC processor, so the timing of a single instruction is not the determinant. The timeout is likely a multiple of the clock cycle, but that multiplier is more than 1. As you also said, it works with a 580 on camera, so timing of the preflash off the ceiling isn't the issue. Also, when I reproduced the issue, my A flash shooting through the umbrella was 4 feet away pointed directly, and the B flash was also pointed directly at the subject from 2 feet. Also, I did not mention that the slaves were not firing reliably because I originally didn't have the sensors pointing directly at the 7d. I confirmed this by firing a test pulse by hitting the picture style button while in the flash controller menu. After rotating the flash receivers so they had a clearer line of sight, they fired reliably.

The cause of what's going on is very much in the firmware of the wireless flash controller, which is why I asked a couple times for what settings you had there. Look, I know you're frustrated and me playing devil's advocate doesn't help. It very well could be timing, or that the onboard flash is too weak to be a controller unless absolute line of sight is maintained. I am not saying nothing is wrong, I am trying to help work through the permutations. If you'd prefer, I can unsubscribe and move on.

No, not at all. I didn't want to scare you away, just explain my thoughts. I also wanted to share my math considerations, where I found that there are at least not a factor of thousands or so between the time it takes light to travel compared to what timing the camera has when it measures the flash light.

You have seen my settings by now, haven't you? A couple of posts after the first one.

Sorry if I offended you. It's not always that easy not having English as a native language.

You may have noticed that more users are seeing the same thing now, as it's being reported on various forums.

what did you guys expect. Nikon had this technology down for years and Canon just decided to add it in 2009 for the first time.. There will be problems... May take Canon another 10 years and 20 EOS bodies later to fix it.

If you sold you STE2 transmitter, Get it back!

Just want to post some positive experience. After reading this post I wandered around the house with my 580 EX II in tow and have not seen any failures as described here. Also, all my previous "normal" shots with a mix of 430's and 580, I did not see any significant ETTL mixup with the 7D.

now, humor me ok.....

to set it up so that the onboard doesnt flash you choose the wireless mode where the icon is a single external flash right?

i dont have my camera now as im at work but could it be that the onboard flash is actually firing so we get double the amount of light we should be getting? thought of it just now while at work LOL. maybe you can try the tests again but this time put a gobo between your onboard flash and the subject and see if illumination lessens.

also there was a guy saying that the icon for onboard flash not firing is the icon of an onboard flash with no lines going out of the front....which i never saw in my menu selection....its just ext : onboard, external only, external+onboard (if memory serves me right). no such thing as an onboard flash icon on and onboard flash icon off. am i correct?

OK, I've done a little more testing and I have an interesting tidbit to add. While using my 70-200/2.8IS I could not reproduce the issue. However, I popped my 28-135 on the camera, and voila - I got major overexposure. I was shooting at relatively close distance - maybe 5 or 6 feet from the subject. I am starting to wonder if the issue is with lenses that don't report distance to the camera bounced from a relatively close distance?