How long time from press to start of exposure in bulb-mode on EOS 600D/T3i?

Question

I want to capture a signal on an oscilloscope screen, but I do not know how long time the interesting signal is seen beforehand. So my thinking is that I use bulb-mode on the camera and control the remote shutter from my Arduino. Time will be about 200ms to 2 seconds, but I would like to stop as early as possible to avoid exposure of unrelated signals on the oscilloscope.

My question is how much time I can expect from press to exposure starts and how long time after release exposure ends. Are the numbers somewhat stable? Does it help or hurt if I start by having live-view active?

Answer 1

The published spec from Canon for the EOS Rebel T3i/600D is 90 milliseconds between the time the shutter button is fully pressed and the shutter begins to open. That's based on a specific set of circumstances and camera settings, though. Thing such as autofocus, metering, etc. can add to that time if they are enabled and the camera has difficulty accomplishing that task. The shutter transit time of each shutter curtain is approximately 3 millisecond, give or take a millisecond.

In the broader view, you're probably expecting more precision from a device than it is capable of delivering. Cameras designed to take artistic, documentary, historical, etc. photographs do not need to be precise enough for lab grade experiments. To make them so would add significant cost (on the order of several magnitudes, not several percentage points) without giving any additional benefit for all who are using the camera for what it was designed to do.

Answer 2

I don't know, but if you have an Arduino you can measure that. Just make a small setup with your camera shooting a LED lit up by the Arduino, and have the Arduino

1. trigger the shutter
2. after some delay, flash the LED

Increase the delay until the LED lits up in the pictures(*). Using mirror lock-up and non-liveview is likely the way to the shortest delay, with of course manual focus.

Perform a similar experiment for the end of the exposure: increase the delay until the LED in no longer on in the pictures.

I would expect the numbers to be reasonably stable, if your switch is reasonably good (an old clunky relay would create more variation than the camera itself)

(*) you can also use a dichotomic search to get at the result faster.

Edit: strike the LED... you can shoot your oscilloscope, after connecting it to an output pin of the Arduino. And if you generate a ramp on that pin you can deduce the delay from the measured voltage and won't need many tries...