PocketWizard delay time measurement

[richard_cochran]

<p>I've been using PocketWizard Plus II modules to trigger flashes for several years, but I've often wondered how much delay they add to the time required to fire the flash. In other words, how long is the time delay between the instant the camera sends the "fire" signal to the PocketWizard mounted on the hot shoe, and the instant the remote PocketWizard sends the "fire" signal to the flash it's attached to?</p>

<p>I've seen the question asked on various sites, and the answer seemed to be something along the lines of "not very long", or "not long enough to matter", or "you might want to slow the camera down by one shutter speed, especially if you're using slow studio flash units and camera with a fast sync speed".</p>

<p>Well, I recently got an oscilloscope, and it makes it easy to make a precise measurement of this sort of thing. I attached a probe on channel 1 (yellow) to monitor the flash sync circuit between the camera and the transmitting PocketWizard. I attached a probe on channel 2 (cyan) to monitor the firing circuit between the receiving PocketWizard and flash. The results are below:</p>

<p> <img src="http://rcochran.homedns.org/PocketWizardDelay/PocketWizardDelay.gif" alt="" width="320" height="234" /></p>

<p>I got nice clean traces, and very consistent results. You see the voltage on the yellow trace goes down, and then a short time later the voltage on the cyan trace goes down. Each square on the grid represents 100 μseconds. You can eyeball the difference at about 700 μseconds, but the cursors measured it at 692 μseconds. That's 0.000692 seconds. Or if you prefer, 1/1445 of a second.</p>

<p>If your shutter speed is 1/250 of a second, this represents about one sixth of the time your shutter is open. If your shutter speed is 1/125, the delay now represents about one twelfth of the time the shutter is open.</p>

<p>Whether this delay is significant depends on several things, among them the time required for your flash units to fire, and the time that your shutter remains fully open. So as a practical matter, I'm back with the idea that it's probably not enough to matter in most circumstances, but it's probably reasonable to slow the shutter down a third of a stop to allow for the delay. Especially if you're using a camera with a fast sync speed and ambient light is dim enough that you don't absolutely need the fastest possible shutter speed.</p>

<p>But now I've got a number to back that up. 692 microseconds.</p>

<p> </p>

[pete_s.]

<p>I made the same measurement a few years ago on my Tektronix but I have forgotten the exact number of microseconds I got. So thanks for refreshing my memory!</p>

<p>On a focal plane shutter a 1/250s exposure means that the time from the front curtain starts to open until the time the rear curtain starts to close is 1/250s or 4000 microseconds. But the x-sync signal isn't activated until the front curtain is fully open and when that happens depends on the mechanical speed of the shutter curtain. The speed of the shutters is always the same regardless of the what shutter speed you have selected but we don't know exactly how fast it is so the time the shutter is fully open is really unknown and could be different on different cameras.</p>

<p>However if you extend the shutter time to account for the delay of the pocketwizard you are back to the same timing you would have had without the pocketwizard. So if you use 1/200 instead of 1/250 the shutter is fully open an additional 1000 microseconds which more than the delay of the pocketwizard.</p>

<p> </p>

[pete_s.]

<p>An interesting test is to shorten the shutter speed until you get a black band, a partially exposed frame. That means that the flash isn't firing until<strong> after</strong> the rear curtain has begun to close. If you do some calculations you can now find out how long time the shutter is fully open.</p>

<p>If you take the flash duration in account you can then select a shutter speed that would allow for both the delay from the radio trigger as well as making sure that the flash has burned out before starting to close the rear curtain.</p>

[hjoseph7]

<p>Not sure if you could try this test with strobes that are being usedin a master slave configuration. What is the delay time from when the main strobe fires and the slaves pick up the signal which causes them to fire ? I always wondered about that...</p>

[frank_skomial]

<p>Nice LCD digital storage scope you have!<br>

These are unreasonably high priced, and holding.<br>

With my old analog 40 MHz CRT based Hitachi, cannot catch/store many events I would like to, but no justification to dump otherwise a good scope, yet.</p>

[rodeo_joe1]

<p>Whoopee! My cheap triggers beat the PWs. I use some Chinese made "iShoot" triggers that have a delay time of 600us, but only at one particular channel setting. I believe I posted those results on this forum somewhere several weeks ago, but I'll repeat them now:<br /> The delay varies between 600 and 700 microseconds depending on the channel selected. This variation is due to the modulation frequency changing with the channel. In fact the change in modulation frequency is all that differentiates the "channels" from each other. My measurements were taken using a Tektronix digital storage 'scope with a setup similar to what Richard describes above.</p>

<p>Needless to say, I use the channel that gives the shortest delay, but even 600 microseconds is a long time in synch-years. It means that I have to drop the shutter speed to 1/200th using a Nikon D700, rather than use its maximum possible X speed of 1/320th. A Canon 5D (mk I) needs to be set to 1/160th for reliably clean synching with the same setup. However it's no big deal really.</p>

<p>Harry, what do you mean when you say strobes used in master/slave configuration? Is that using radio triggers or optical slaving?<br /> Optical slaves will be almost instant. There'll be a tiny delay due to the rise-time (response time) of the sensor used and its associated circuitry, but it'll be nowhere near as long as that of a radio trigger. I'd hazard a guess that optical triggering introduces a delay of no more than a few microseconds, which is totally insignificant.</p>

<p>So-called "wireless" systems using pre-flash optical signalling are a different matter. Those introduce a considerable delay<em> before</em> they even let the shutter fire!</p>

[rodeo_joe1]

<p>P.S. This quote from the "features" section of the Plus II on PW's webpage:</p>

<blockquote>

<p>Ultra-fast microprocessors enable trigger-response times as short as 1/2000th of a second allowing for reliable sync speeds of 1/250th for focal plane shutters and 1/500th for leaf shutters.</p>

</blockquote>

<p>Now I'm not saying that PW are stretching the truth, but 1/2000th of a second is 500 microseconds - not 700. Perhaps if they took the darned microprocessor out and used dedicated circuitry they could get the response time down a bit further!</p>

<p>[Gosub Trigger; case of "trigger" {bin} = false; Loop until "trigger" = true; Endsub; Gosub Fire.......... ]</p>

 

[richard_cochran]

<p>Thanks for the comments. I'm glad I'm not the only one who has been obsessive enough to measure this kind of thing. As for Harry's comment about (I presume) optical slaves, I think they're significantly faster than radio triggers, but I should measure them when I get a chance. </p>

<p>Since I only have three PW units, and I sometimes use more than two lights in a studio, I often end up using a combination of PWs and optical slaves, where the PW on the camera fires one light, and that light fires others via optical slaves. With that situation, I have the cascaded delay of the PWs plus optical slaves. To make it worse, I'm normally using my not-so-fast monolights with their long flash duration. Saving grace is that, in the studio, I've got so little ambient light that there's no harm in slowing the shutter down to 1/100 or even 1/80 or so. I've done exposure tests, somewhat like Pete describes, where I look at an evenly lit background, start with a very slow shutter speed, and speed it up until I can see the uneven light caused by the shutter curtain closing while the flash is still putting out light. In my worst case situations, I need 1/100 to capture all of the useful light from my strobes.</p>

<p>Next step, hook up a couple of silicon photodiodes to watch the delays between various slaved lights, and to track the intensities of the lights at various power settings. I wish I had more time to play with this stuff.</p>

[ian_casement]

<p>Hey Richard, neat test.</p>

<p>I like individual test results rather than relying on manufacturers figures. The PlusII's have always been quoted at 1/1800s which compared to Cybersyncs, Quantum FreeXwire and the new ControlTL are much slower.</p>

<p>It would be neat to try a variety of these to obtain direct comparisons.</p>

[pete_s.]

<p>OK, I redid my test. These are the Pocketwizard Plus 2 EU version. The delay is the same no matter what channel is used. These triggers are slightly faster than Richards triggers which I think is because the radio frequency is higher. 628 microseconds.</p>

<p> </p><div></div>

[pete_s.]

<p>And here is another test where I added an optical trigger. I kept the PWs in the loop because I didn't want a slave flash to affect the optical trigger. So it's flash => optical trigger => PW transmitter => PW receiver.</p>

<p>I found the optical trigger takes longer to trigger a slave when the flash it receives is weaker. I think this is because it has to reach some point of saturation before it triggers the slave.</p>

<p>The longest delay I managed to get while the slave still worked was 56 microseconds.</p><div></div>

[pete_s.]

<p>If you set the PW Plus 2 to Transmit Only Mode by pressing the Test button while powering on the delay is now shorter. 556 microseconds or 1/1800s if you like. This is PW to PW only, no optical triggers.</p>

<p>I also tested triggering through the different ports or hotshoe but it makes no difference for the PW.</p><div></div>

[pete_s.]

<p>Haha, just to make this test "complete" I tried another set of PW triggers. These are from 2006 or possibly 2007 (the one tested above are from 2011).</p>

<p>Delay is 692 microseconds and Transmit Only Mode makes no difference. So I guess that LPA Design have tweaked both the hardware and the firmware on newer sets.</p>

<p> </p><div></div>

[richard_cochran]

<p>My PWs are from about 2003 or so, if I remember right. And I tried transmit only mode and found it makes no difference. So it sounds like our measurements are consistent at 692 microseconds. It's interesting that newer ones can be a bit faster, though the difference isn't likely to matter in practice.</p>

[parv]

Thanks everybody for sharing your tests.