Actual Measured Flash Durations of Small Speedlight Strobes

It’s been a while since I’ve written a technical article. But today I’m going to talk about flash durations. After a bit of wild searching on the internet, I couldn’t really find any good resources showing actual real-world measurements of flash durations of speedlights, especially t.1 measurements.

Canon 580EX at full power

Lights Tested

I decided I would measure the flash durations of various lights.
These include:

These speedlights above are have IGBT triggered power control which result in shorter durations for lower power discharges.
I also performed tests on these units:

  • Broncolor Flashman floor pack with Pulso 2 head (1800Ws symmetric)
  • Broncolor Impact 41 mono light
  • Godox 120 Ws
  • Alien Bee B400 160 Ws
  • Paul C Buff Einstein E640 (640 Ws), measured for constant color mode and action mode
  • “Mystery” eBay light 400 Ws

Here are the results of the real world measurements made. Later in the article, I’ll describe, how I performed the measurements.
I’ve added some results of a few mains powered strobes for reference.

The Results

Canon 580EX

Power μs s
1 4000.0 1/250
2 1088.0 1/919
4 484.0 1/2066
8 266.0 1/3759
16 166.0 1/6024
32 105.6 1/9470
64 71.6 1/13966
128 50.4 1/19841

Nikon SB-24

Power μs s
1 4320.0 1/231
2 1260.0 1/794
4 732.0 1/1366
8 292.0 1/3425
16 192.0 1/5208

Nikon SB-26

Power μs s
1 4080.0 1/245
2 720.0 1/1389
4 368.0 1/2717
8 204.0 1/4902
16 128.0 1/7813
32 92.0 1/10870
64 72.0 1/13889

Nikon SB-28

Power μs s
1 3780.0 1/265
2 1048.0 1/954
4 496.0 1/2016
8 276.0 1/3623
16 158.4 1/6313
32 100.8 1/9921
64 72.0 1/13889

Nikon SB-80DX

Power μs s
1 3880.0 1/258
2 864.0 1/1157
4 408.0 1/2451
8 222.6 1/4492
16 135.2 1/7396
32 90.4 1/11062
64 63.2 1/15823
128 45.6 1/21930

Yongnuo YN560

Power μs s
1 3200.0 1/313
2 736.0 1/1359
4 356.0 1/2809
8 202.0 1/4950
16 124.0 1/8065
32 79.2 1/12626
64 54.8 1/18248
128 43.4 1/23041

Cactus KF36

  • This is the same as Vivitar 285HV
Power μs s
1 3640.0 1/275
2 1504.0 1/665
4 636.0 1/1572
16 188.8 1/5297

Broncolor Flashman with Pulso 2 head (1600 Ws)

Power μs s
1 5680.0 1/176
2 5880.0 1/170
4 7220.0 1/139

I double and tripled checked this, the duration is really shorter at higher power!

Broncolor Impact 41 Monolight

Power μs s
1 4280.0 1/234
2 2500.0 1/400
4 1408.0 1/710

Godox (120 Ws)

Power μs s
1 2880.0 1/347
2 3480.0 1/287
4 3360.0 1/298
8 3520.0 1/284

Alien Bee B400 (160 Ws)

Power μs s
1 824.0 1/1214
2 928.0 1/1078
4 1032.0 1/969
8 1328.0 1/753
16 1120.0 1/893
32 1296.0 1/772

Paul C Buff Einstein E640 (640 Ws)

We measured this under 2 modes, Constant Color and Action mode.
Under constant color mode:

Power μs s
1 2560.0 1/391
2 984.0 1/1016
4 592.0 1/1689
8 472.0 1/2119
16 396.0 1/2525
32 336.0 1/2976
64 296.0 1/3378
128 268.0 1/3731
256 236.0 1/4237

Under action mode:

Power μs s
1 2640.0 1/379
2 624.0 1/1603
4 296.0 1/3378
8 162.0 1/6173
16 131.2 1/7622
32 116.8 1/8562
64 132.8 1/7530
128 128.8 1/7764
256 137.6 1/7267

“Mystery” eBay light (400 Ws)

Power μs s
1 7120.0 1/140
2 7920.0 1/126
4 8000.0 1/125
8 8080.0 1/124
16 9440.0 1/106
32 10160.0 1/98

How I measured it

I built up a basic little circuit using a BPV11 phototransistor as the sensor. The output was then read by an oscilloscope and the t.1 interval determined by setting two cursors to correlate to the respective t.1 amplitudes.

The resistor was chosen so that the transistor always operated within its linear region, and collector current was within the maximum limits allowed. The device never reaches saturation. It measures light between 620 to 980 nm, peaking at 850nm and has a turn on and turn off response time between 5 to 6 microseconds each (plenty fast enough for our measurements, within 5% for the fastest measurements).
The t.1 (sometimes called t0.1) is measured as below:

Sample Discharges

I didn’t bother to save the waveforms for each test, it would have taken too long. But I did keep a couple for my personal reference. Here are some of them, just as an example to see the shape of light output of the 580EX at different power levels.
The vertical axis is linear to the irradiance at the sensor, so something that is twice as high means it was twice as bright (one stop). From one image to another in the plots below, they are not to scale.

Canon 580EX at full power

Canon 580EX at full power minus one third

Canon 580EX at 1/4 power

Canon 580EX at 1/64 power

Final Words

This article might be updated later as I get my hands on more gear to test and add to the result list.
Leave a comment if you have any questions, or perhaps you’ve picked up a mistake somewhere.

65 thoughts to “Actual Measured Flash Durations of Small Speedlight Strobes”

  1. Thanks for this info Andy! Was trying to research a reference that gave me figures on outputs on various strobes and to find such an awesome resource on the one page was a godsend!

  2. Hi,
    I would just like to say a big THANK YOU for all the effort you put in. I’m still confused as to why flash durations are different on every site (I’ve seen sites that say 580ex flash duration is 1/8xx, but also yours), but any input is helpful, albeit confusing.

  3. I enjoyed your writeup. I am considering doing the same and I have a few questions. First, I noticed that the time you measured was off significantly from what the manufacturer states, for example with the Alien Bees/Einstein strobes. What could account for this? I have no doubt the manufacturer would present their product in the best light, but I would also expect you to at least come close to one of their times. I also looked at the specs for the photo transistor you used. I noticed a on-off time stated in microseconds, and you specified nano seconds. Could that be the source of error in your measurements? Also, why not use a photo diode, which I heard is preferred because it is faster? Thanks

    1. Thanks Randy,
      I just used parts I had sitting around the shed. Photo diode would have been a little bit better.
      I don’t know why the manufacturer has different times, and to be honest, i didn’t really check all of them to compare, i only measured what i could get my hands on.
      I did do some research and found most people measure t0.1 and t0.5 based on comparison with the peak intensity based on what i did. This is somewhat deceptive as it doesn’t reflect the amount of lighting coming from the flash, as it does not take into account the flash curve shape, so t0.1 is NOT 90% of the light coming out. The measurement of 90% of the light coming out would involve measuring the area of the curve swept which perhaps might be what some other manufacturers are doing?
      But its true that some manufacturer’s don’t tell you how they measure their specs. I should post some links later to those that do for the readers’ reference.

  4. Hello Randy.
    The result was amazing. I want to know real flash duration nikon sb28. Its so important data.
    I am interested in YN560II type. Its new version of YN560. How is flash duration YN560II?
    I think it same. But YN560II has mult flash mode. Is it different condencer?
    If you have chance than you use it, I hope please check fulash duration YN560II.
    Thank you so much!!
    Best regards

    1. Thanks Yuki,
      I don’t actually have a YN560 II, but if i get my hands on one, I’ll be happy to test it and add it on here!

  5. Hello Andy.
    Thank you for your reply. I need fast flash duration for my photo. Becase Im skateboard photographer. The lumedyne action pack is 1/3000 @ 200W(T.1). Its best, but price is expnesive…
    If YN560 II is same flash duration as YN560, I will buy four and all flash power are 1/4. Its same flash duration as action pack.
    Please do the test when you hage free time.
    You are so nice flash master guy!!
    Thank you so much!!
    Best regards.

  6. Andy,
    Thanks for the work you’ve done. I was just wondering, can you use the graphs above to calculate the area under the graph and get the total power output in Watt*Seconds for the 580EX? I’m interested in comparing the amount of power to studio strobes, but I can only find hearsay on the internet. For example, “580 EX is 40 w*s at full power” or “580 EX is in the 60-80 w*s range”. But which is correct?!
    Thanks again!

    1. Hi Tim,
      Measuring the area under the graph won’t give you the watt-second (Joule) value. Watt seconds (energy) of a flash is usually not calculated by measuring the amount of light emitted, but by the storage capacity of the capacitor which is dumped into the flash tube. Different flash tubes may have different efficiencies, so 600Ws in one brand may likely be different to 600Ws in another.
      The best way to measure the watt-seconds of a speedlight is to open up the flash, measure the capacitance of the main capacitor (or capacitor bank), and then measure the terminal voltage when fully recharged. (this is hazardous, don’t try this unless you know what you’re doing)
      energy = 0.5 * capacitance * voltage ^ 2
      What you’re talking about is lumen-seconds which is actual light output, some manufacturers publish this, some don’t.
      Measurement is a little bit more difficult and is more than just measuring a curve under the graph as you need to take into account the shape of the output pattern etc, and working out a way to “measure all the light”.

  7. Hi Andy,
    Thanks for all this work too! I have a simple question, it’s probably really obvious to you.
    It’s great tohave the flash duration comparisons. I also need to be able to compare the power output of each flash. I am assuming the poster output/ amount of light is more on an einstein light at 1/128 of its strength than let’s say a speedlight 580Ex at 1/128 it’s strength. Is there a simple way of calculating this? Thanks for your help. Joschi

    1. Hi Joschi,
      Thanks for your comments.
      I don’t have actual light output comparison from different models of flash units. Its something that could have been measured to some extent but I haven’t done it in the past.
      The compare real world usable results, is to just use a light meter with a standard reflector on the unit and compare results that way. Or most manufacturers post guide numbers in their datasheets (with respective modifiers), which can also be used as comparison.
      e.g A 580EX speedlight at 105mm zoom has GN 58 while a Einstein 640 with a Retro Laser Reflector has GN 289 which is around 2.5 stops more light. (Ref)

  8. Thanks for your great work.
    The information was very interesting and useful for me.
    I wonder you would have information about the flash duration of macro flashes, because my mail work is dental photography (clinic en surgery).
    Thank you again.

    1. Thanks Gabriel.
      No sorry, I don’t have any information on macro style flashes, simply because I don’t own one or have access to one.

  9. Great work Andy, thanks very much for the info. The mind boggles as to why Canon cannot provide this sort of data for their own flashes. Your methods are way over my head but the data is exactly what I needed to know to figure out how much power I can get out of my 580ex II and my Einsteins when married with my fujifilm x100s leaf shutter. Time now to play with killing the sun with flash.

  10. Really good work. Thank you. Underlines how slow big studio units can be – the biggest I worked with was a Strobe Equipment 5000-joule Swimming Pool. When the capacitor (the unit was the size of a medium fridge on wheels) dumped its charge, it sounded like a small explosion.

  11. Very useful work Mr. Gock.
    It would be helpful to add either the joules (watt-second) or the measured f-stop (at a fixed distance, ISO and with a specified or common reflector / modifier ) or calculated Guide Number) for the different power levels.

  12. Thanks for taking the time to do this! Can you please explain the charts? For example in the 580EX chart it says that at full power you have 4000.0 μs and 1/250 s. What does this mean? Thanks!

  13. To Don,
    4000μs=1/250 second=0.004 second=4ms
    This implies a full power discharge of 580EX is 4ms. (by t.1 standard) If you don’t know what t.1/t.5 is, find it at wiki.

  14. Hi Andy,
    Great article! I found your website when I was looking on Google for a fast visible light photo transistor. I was looking, as wanted to do the same experiment as you have done. Of course I still wanted to do the experiment, so found a BPv11 on CPC and tried this out. I am using the same circuit and 6v supply and have a 580EXII and an YN-560 to initially test. The problem I seem to have is one of replication. I do not seem to be able to get the same speeds as you. I know it s difficult without sitting in front of the setup up, but do you have any pointers that could be the reason I can replicate. Could I have something wrong on the scope? How did you set the trigger point above ambient or darken the room. How far away was the strobe? I get a flat top on the slope is this right?
    Thanks for any help

    1. Hi Doug,
      You shouldn’t be getting a “flat top” curve, this means you are getting incorrect results. This can be caused by several reasons.
      Are you using a 68R resistor, or something in similar vicinity?
      Don’t fire the flash straight at the transistor, just in another direction, and let it detect the bounce. Set the trigger above ambient measurement. Too much intensity can case it to “flat top” at the approx the battery voltage.
      Good luck.

    1. Hi Andy,
      It’s not actually flipped around, I remembering measuring and that is the way it came out. It is quote common for “full capacitor discharge” type flash units to exhibit this type of correlation.

  15. Thanks for posting the scope shots. It really gives a lot of insight as to how the flash works and how they regulate power. Looks like they fire towards the same peak power each time but regulate output power by timing when to squelch it. Which makes a lot of sense as I’d always thought the physics of those strobe lamps required them to fire at a certain fixed voltage level, meaning you can’t regulate power by varying applied voltage.
    So it falls out naturally that the less output power the strobe has, the shorter the duration has to be to squelch it down to that power. Its a good trick to know and understand. When you want to “freeze” action that you’re struggling to capture clearly, increase ISO, decrease power. Awesome.

  16. very interesting. I’m looking to buy a battery/head or studio flash (all I have now are speedlites) SPECIFICALLY for Hypersync (ODS, Supersync). So I’m looking for SLOW flashes. It looks like many of the speedlites have t.1 as slow as many studio lights! This is confusing to me, as everyone has said that studio lights are much slower (generally, excepting for “action”-specific setups), and so work much better for Hypersync (Pocket Wizard says this too– and they only test lights like the Quadra, not speedlites).

    1. Hi,
      IF you’re looking for “slow flashes”, i.e flashes with long flash durations, its best to look for older model units. Many modern units use IGBT based cut off, which means it has similar short durations like speed lights. For modern flashes still available, Alien Bees are of the long duration type, and some of the Elinchrom ones as well.

  17. Read in particular what Paul Buff has to say about freezing motion, which is what I care about, rather than total energy or color temperature. For water drops, etc, I can boost the power, etc. and deal with white balance on my own. What I want is is 1/millionth of a second (ideal) and your curves don’t seem to help -or- I don’t know how to interpret them. I get Paul Buff’s curves — too bad most mfg don’t want you to know how bad they are.

  18. Hi Andy, excellent article! Would you have by any chance the 600EX-RT data? Or where can I locate it? The Canon Speedlite manual doesn’t have it! thanks a lot!

  19. Andy, thank you for the information! What do you know regarding HSS? me and a friend are trying to figure out how a flash achieves HSS. Is it that the flash pulses or blinks light many times? Or is it a continuous burst of light? I thought I remember reading one time that the flash pulses because of the amount of light needed to expose at high shutter speeds would generate to much heat and damage the flash, so hence more bursts or pulses of light at lower outputs. Hope this makes sense and you can shed some light on the subject, yes, pun intended!

    1. Hi Ross,
      Sorry for the late reply. HSS is achieved by multiple flash pulses (very fast blinking).
      Standard non HSS mode just fires the flash in a single pulse of light.
      At high shutter speeds above the cameras sync speed, the camera’s shutter is never completely open (only a smaller slit exposes the shutter at any one time), so they pulse the flash as the slit moves over the sensor/film.
      Try having a read of this link which might explain it better:

  20. Many thanks for your hard work. Its a pleasure to see a real-world measurement of flash duration and has been very useful in evaluating an industrial measurement task.
    I will be keeping an eye out to see if you get around to documenting the Canon 600Ex-Rt. For now I’ll use the figures for the 580.

  21. Thank you so much for this article. I’m just a bit confused as I thought the more powerfull the shot the slower it would be but the Alien Bees 400 are faster on full power than 1/32. How’s that? I’m checking LEDs can replace my speedlights your tables are of great use.

  22. Vasco,
    It has to do with the strategy Alien Bees (but not the Einstein from the same company) and most monolights use to reduce energy output. In short they reduce the voltage of the current flowing from the capacitors through the flash tube and that lengthens the duration of the energy pulse.

    1. Hi James,
      If you used 4x 580EX units, and set them to 1/4 power each, then yes, you’ll have a duration of around 1/2000th sec with the light output equivalent to what would come out of a single unit at full power.

  23. Great research. It’d be wonderful to see a single table, aligned by light output. For real world application, I care what f-stop and what flash duration I have. Whether it’s 1/4 or 1/16 of full power is not evident in my end result, but sharpness is.

  24. thanks for your detailed explanation.
    i have one more question, what is the flash duartion for ttl on camera flash for nikon d7100?is it same as faster as strobes?

  25. I can’t thank you enough for this research!
    Based on your numbers, I have purchased 3 Yongbuo YN560 flashes. They are incredible!
    Here is a sample of the clean, crisp motion-stopping exposures they give;
    Again, you have provided a great service to so many photographers!

  26. Interesting article Andy. I just took a look at Yongnuo flash and shocked about the result bcuz the Yongnuo itself claims the Flash Duration of this model and even newer models is 1/200~1/20000s and your results shows it’s much much faster and it’s fantastic. What do you think about their slower speed claim?

  27. Hi Andy, many thanks for this article. I’m new to flash photography, and have asked many so-called experts “when different power settings are selected on a speedlite, is it a change in flash duration or light intensity that determines the exposure that the flash gives?” I take it from your article above that it is the area under the curve prior to 10% decay that determines the exposure, and that the peak on the y axis is the same for all power settings on a given speedlite? Could you please confirm or otherwise my understanding of this? Thanks, Mel.

  28. Hi Andy,
    Thank your for taking the time to measure the flash durations of all the strobes. And also for sharing it online with all of us. Much appreciated!
    I imagine that this thread was published a few years ago, but still I thought I ask you, if there is anyway to identify the Mystery Chinese strobe unit please.
    Any pictures of the unit? Model number perhaps? Online retailer who may still sell them?
    I was looking for a strobe with a near constant, long flash duration and that unit you tested may just be a possibility for me.
    I thank you again!

  29. Great job!!!
    Perhaps you have already conducted measurements in flash Canon 600 RT ?
    Publish your results on this model.
    Thank you very much in advance!

  30. Hi,
    Do you have any idea for ETTL2 when a Canon camera samples the preflash light output relative to the center pin drop?

    1. Hi,
      I’ve never tried ETTL before but I can see where it might pose a problem. As the pre-flash is often a lower intensity, you could try setting your DSO trigger somewhere in between the two peaks.

Leave a Reply

Your email address will not be published. Required fields are marked *