Go to a  "printer friendly" view of this message which allow an easy print Printer-friendly copy Go to the page which allows you to send this topic link and a message to a friend Email this topic to a friend
Forums Lobby MASTER YOUR TOOLS - Hardware & Software Nikon Speedlights & Lighting topic #11919
View in linear mode

Subject: "rear curtain sync question" Previous topic | Next topic
scjames702 Basic MemberFri 09-Dec-05 03:30 PM
2 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
"rear curtain sync question"


Los Angeles, US
          

Greetings!

I have a D2X and SB-800 combination. I'd like to use rear curtain sync to more effectively imply motion. My question is this: When I set the D2X to "rear" and shoot, I get flash both at the start of the exposure and at the end. I think this is because the camera automatically selects slow sync mode in combination with rear curtain whenever the latter is selected. As "slow" fires at the start and "rear" at the end, I get both. Is that correct? More important, is there a way to stop the speedlight from firing at the start? Or is this not a problem? I also worry that the speedlight doesn't have the time to fully recycle before the second flash occurs.

Thanks,

Scott James

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

Replies to this topic
Subject Author Message Date ID
Reply message RE: rear curtain sync question
avm247 Moderator Awarded for high skills in documentary architecture and aviation photography Donor Ribbon. Awarded for his support to the Fundraising Campaign 2014
09th Dec 2005
1
Reply message RE: rear curtain sync question
scjames702
09th Dec 2005
2
Reply message RE: rear curtain sync question
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources.
09th Dec 2005
3
Reply message RE: rear curtain sync question
sixgun
09th Dec 2005
4
Reply message RE: rear curtain sync question
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources.
09th Dec 2005
5
Reply message RE: rear curtain sync question
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources.
13th Dec 2005
6
Reply message RE: rear curtain sync question
alm
14th Dec 2005
7
Reply message RE: rear curtain sync question
blw Moderator Awarded for his high level of expertise in various areas
15th Dec 2005
9
Reply message RE: rear curtain sync question
sixgun
20th Dec 2005
21
Reply message RE: rear curtain sync question
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources.
15th Dec 2005
8
Reply message RE: rear curtain sync question
alm
16th Dec 2005
10
     Reply message RE: rear curtain sync question
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources.
16th Dec 2005
11
          Reply message RE: rear curtain sync question
alm
16th Dec 2005
12
               Reply message RE: rear curtain sync question
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources.
17th Dec 2005
13
                    Reply message RE: rear curtain sync question
alm
17th Dec 2005
14
                         Reply message RE: rear curtain sync question
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources.
17th Dec 2005
15
                              Reply message RE: rear curtain sync question
alm
18th Dec 2005
16
                                   Reply message RE: rear curtain sync question
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources.
19th Dec 2005
17
                                        Reply message RE: rear curtain sync question
alm
19th Dec 2005
18
                                             Reply message RE: rear curtain sync question
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources.
20th Dec 2005
19
                                                  Reply message RE: rear curtain sync question
alm
20th Dec 2005
20
                                                  Reply message RE: rear curtain sync question
alm
21st Dec 2005
22
                                                  Reply message RE: rear curtain sync question
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources.
22nd Dec 2005
23
                                                  Reply message RE: rear curtain sync question
alm
23rd Dec 2005
24
                                                  Reply message RE: rear curtain sync question
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources.
23rd Dec 2005
25
                                                  Reply message RE: rear curtain sync question
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources.
04th Jan 2006
26

avm247 Moderator Awarded for high skills in documentary architecture and aviation photography Donor Ribbon. Awarded for his support to the Fundraising Campaign 2014 Charter MemberFri 09-Dec-05 04:15 PM
18608 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#1. "RE: rear curtain sync question"
In response to Reply # 0


Rancho Cordova, US
          

Scott, are you sure you aren't seeing a preflash? I know that with the older TTL systems Nikon used, the preflash was disabled but with I-TTL, the preflash is used to calculate the amount of flash exposure.

Anthony

The Moderator Page and My Gallery
"Eliminate the unnecessary and focus on the substance." (Tommy Ramone of the Ramones)

Anthony

The Moderator Page and My Gallery
The important things in life are simple; the simple things are hard.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

    
scjames702 Basic MemberFri 09-Dec-05 05:22 PM
2 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#2. "RE: rear curtain sync question"
In response to Reply # 1


Los Angeles, US
          

Hi Anthony,

Well, you may have just exposed my ignorance. I wasn't aware of the preflash function. Is there a preflash when the camera is set to front curtain sync (and perhaps not perceived distinctly b/c of its proximity to the moment of the flash itself)?

In any event, it sounds like you're telling me not to worry about.

Thanks for taking the time to respond.

Scott

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources. Charter MemberFri 09-Dec-05 05:43 PM
8508 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#3. "RE: rear curtain sync question"
In response to Reply # 0


Phoenix, US
          

Scott:

You are seeing the preflash on the SB800. This occurs in theTTL-BL-FP, TTL-FP and AA-FP modes so long as the flash head is at the 90 degree or -7 degree position. When tilted up for bounce flash, the preflash is cancelled.

When using the SB800 as master with one or more SB800s or SB600s as remotes, you will also see a preflash.

See page 36 of the SB800 user manual for a brief description of the "Monitor Preflash"

With a D2X and two or more SB800s, you are into the wonderful world of Nikon's CLS/AWL. I currently have eight SB800s used extensively on outdoor night shots.

I'm waiting patiently for Nikon to release their new SU-800 speedlight controller. While nobody seems to have their hands on one yet, we all suspect that it will eliminate the preflash problem when used to control wireless remotes. As you probably already know, the preflash will cause some models to blink and then when it fires again to capture the image, their eyes are closed.

The SB800 system is a superior design. Yes, the 125 page user manual is a bit dense because there is a lot to cover but the rewards are worth the effort to get through it. Stick with it.

HBB in Phoenix, Arizona

HBB in Phoenix, Arizona
Nikonian Team Member

Photography is a journey with no conceivable destination.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

    
sixgun Registered since 26th May 2004Fri 09-Dec-05 08:41 PM
480 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#4. "RE: rear curtain sync question"
In response to Reply # 3


Austin, US
          

>You are seeing the preflash on the SB800. This occurs in theTTL-BL-FP, TTL-FP and AA-FP modes so long as the flash head is at the 90 degree or -7 degree position. When tilted up for bounce flash, the preflash is cancelled.<

Are you sure about the pre-flash being canceled. When I have my head straight up for bounce trying to capture motion there is a small but noticable delay from depressing the release to the flash firing in TTL or TTL BL and what seems to be a pre-flash (same noise/sight pattern), front sync, D-70. If I switch over to manual there is no such delay.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

        
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources. Charter MemberFri 09-Dec-05 10:38 PM
8508 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#5. "RE: rear curtain sync question"
In response to Reply # 4


Phoenix, US
          

Scott:

Sorry I cannot test an SB800 with a D70 as I sold the last one several weeks ago. I will put one on my D2X over the weekend and try all combinations. Since both cameras are CLS compatible, the results should be the same ... we think.

I'll get back to you.

HBB in Phoenix, Arizona

HBB in Phoenix, Arizona
Nikonian Team Member

Photography is a journey with no conceivable destination.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

            
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources. Charter MemberTue 13-Dec-05 03:56 PM
8508 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#6. "RE: rear curtain sync question"
In response to Reply # 5


Phoenix, US
          

Scott:

Sorry for the delay. My weekend disappeared at a Native American craft show here in Phoenix. I have a couple board meetings to get through and then I can get back to your question. I will run through all possible modes, etc. with the SB800 and see just where preflash does and does not occur.

Now that you have asked the question, I'm also curious. Thanks for your patience.

HBB in Phoenix, Arizona

HBB in Phoenix, Arizona
Nikonian Team Member

Photography is a journey with no conceivable destination.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

        
alm Registered since 29th Oct 2005Wed 14-Dec-05 05:03 AM
596 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#7. "RE: rear curtain sync question"
In response to Reply # 4


Zoetermeer, NL
          

I tested this recently for another thread (http://www.nikonians.org/cgi-bin/dcforum/dcboard.cgi?az=show_thread&forum=DCForumID7&om=15111#6). My D70 still fired the pre-flash with the head 90 degrees up, this was also reflected by the double lighting bolt icon on the SB-800's display (top left corner).

I believe that i-TTL always fires pre-flashes, and that the only way to disable this is use A/M mode, FV lock, or use TTL on a film camera (except the F6 maybe). I think that tilting the head up to disable pre-flash worked for the recent film camera's like the F100, the SB-800 manual contains some info on disabling pre-flashes in the section about SU-4 (non-CLS wireless flash) if I remember correctly.

Since the D70 is my only Nikon SLR camera, I can't test it on a D2X, so I'm still curious about HBB's results, but I'd be surprised if it worked differently.

Alson

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

        
blw Moderator Awarded for his high level of expertise in various areas Nikonian since 18th Jun 2004Thu 15-Dec-05 10:49 PM
26841 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profileClick to send message via AOL IM
#9. "RE: rear curtain sync question"
In response to Reply # 4


Richmond, US
          

On *digital* bodies, preflash is NEVER cancelled by tilt/rotate/bounce. This is only on film bodies (and I'm not sure about the F6 as it is both iTTL and film).

The ONLY way to get a dTTL or iTTL body not to preflash is to set it in A or M mode. TTL, TTL-BL and AA always use preflash.

_____
Brian... a bicoastal Nikonian and Team Member

My gallery is online. Comments and critique welcomed any time!

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

    
sixgun Registered since 26th May 2004Tue 20-Dec-05 10:40 PM
480 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#21. "RE: rear curtain sync question"
In response to Reply # 3


Austin, US
          

>I'm waiting patiently for Nikon to release their new SU-800 speedlight controller. While nobody seems to have their hands on one yet, we all suspect that it will eliminate the preflash problem when used to control wireless remotes. As you probably already know, the preflash will cause some models to blink and then when it fires again to capture the image, their eyes are closed.

<

Would it?

You would elimnate the visable pre-flashes that communicate from the Master unit to the remotes but wouldn't the remotes still have to do their pre-flash for the camera to meter them? I don't think were done with pre-flash yet.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources. Charter MemberThu 15-Dec-05 07:49 PM
8508 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#8. "RE: rear curtain sync question"
In response to Reply # 0


Phoenix, US
          

Scott:

Before I get too deep into the almost infinite number of SB800/D2X states, could you please tell me what the camera settings: Mode (A, M, P, S), Aperture, Shutter Speed, etc., lens and the speedlight settings (Mode, Head Position, etc.) you are using? It will help shorten the process a great deal.

Yes, with the D2X on a tripod in the A mode (Aperture Preferred) and the speedlight set to PF/TTL/BL/FP (PF = Pre-Flash = the two lightning bolt symbols) and "Rear Curtain" selected I also automatically get "Slow" along with it. The shutter speed selected by the D2X for this setup in my study with a Nikon AF-S 12-24 mm at 12 mm and F/8 is 1.3 seconds. And yes, I do get two flashes: one at the beginning of the 1.3 second exposure which is the preflash and a second one at the end which is for exposure while the shutter is open. I also get two flashes in every case with the head pointing straight up at the ceiling.

I get results identical to the above with the speedlight set to the FP/TTL/BL/FP, the FP/TTL/BL and the FP/AA/FP modes, with the head at ninety degrees and straight up positions also.

I think it is safe to assume that all preflashes are at a much reduced power and do not detract seriously from the SB800s ability to get the job done when the main flash is triggered.

The Nikon manuals "suggest" that if the two lightning bolts in the upper left corner of the SB800 monitor screen are on, you are going to get preflash. Some of them we can see. Others are invisible (at least to my eyes) in most situations. See page 13 of the SB800 manual and Page 112 of the D2X manual for a concise summary of the various SB800/D2X combinations.

While I have been using SB800s (eight of them) for months now, my curiosity has been aroused. I am going to dig out my oscilloscope, rig up an optical sensor and record a number of preflashes to see just what is going on and when. I likely will not get to this until sometime after the holidays but will report when I have something.

Once again, I must admit I do not quite understand everything I know about the SB800. As a forty-seven year survivor of the computer industry, I would love to see the logic flow chart for the SB800 and the D2X. When I recently downloaded and installed the D2X firmware updates (1.01), I noticed that they are almost five megabytes or so of code. There are a lot of things going on in this little black box!

I hope this bit of insight helps. If you have any other questions, let me know.

HBB in Phoenix, Arizona

HBB in Phoenix, Arizona
Nikonian Team Member

Photography is a journey with no conceivable destination.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

    
alm Registered since 29th Oct 2005Fri 16-Dec-05 03:35 AM
596 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#10. "RE: rear curtain sync question"
In response to Reply # 8


Zoetermeer, NL
          

Made curious by your suggestion and because I had the parts in a drawer, I built an extremely simple optical sensor: a light sensitive resistor in series with a regular resistor, a 5V lab supply, the sensor locally stabilized by some capacitors. I measure the voltage over the light sensitive resistor with an oscilloscope. The slow response of the resistor (the powersupply is completely stable, but it takes about 40ms for the resistor to completely regain its resistance after a flash burst) makes it hard to see high-frequency stuff like the communication between master and slave or the duration of a flash, but it's easy to distinguish the pre-flashes. I pointed my D70 with SB-800 in the direction of my test rig.

I only did a few quick tests because lack of time, but still some interesting observations: the time between the pre-flash and main flash was shorter if the ambient light level was higher. If I turned the fluorescent light 1m above my camera on, there was 23ms between pre-flash and main flash for TTL/TTL-BL. If I turned it off, it took 112ms. I used manual exposure with front curtain sync, so my conclusion is that it takes the camera more time to calculate the exposure, or it waits intentionally because it needs a larger burst (more time to charge the capacitor).

The delay between pre-flash and flash for AA mode was shorter than TTL/TTL-BL mode. The difference is about 20ms, independent of the ambient light. It was 2.4ms with overhead fluorescent tube, and 88ms without. This is obviously because the flash doesn't have to communicate with the camera, and because the math is easier (I believe AA just looks for 13/18% gray on average, without any advanced matrix metering).

I didn't find any differences in timing between different head positions (I tried straight, 7 degrees down, 90 degrees up).

I plan to do some more extensive tests and get some graphs soon if I succeed in connecting my oscilloscope to my computer (else I'll have to resort to the old photograph of the screen trick). Tests I planned are: more tests with different ambient light levels, tests with advanced wireless stuff (master with slave in different modes/groups to see what the communication looks like). Any other suggested tests?

Alson

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

        
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources. Charter MemberFri 16-Dec-05 04:36 PM
8508 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#11. "RE: rear curtain sync question"
In response to Reply # 10


Phoenix, US
          

Alson:

Beautiful piece of work! I'm delighted you found the time and the components to take our quest to this next level.

Short of taking an SB800 apart and examining the chip sets with a logic analyzer (No, I don't have one sitting around my office idle, do you?) I suspect the approach you are pioneering will have to do.

The thing we (and perhaps all reading this thread) are after is the encoding scheme used between the master on the camera and the remotes. I had assumed it would be sub-millisecond based on shutter lag and other factors relating to the SB800. You have neatly confirmed that.

I am assuming for the moment that the signalling is some form of pulse coded modulation, either frequency or amplitude. You will know this when you can expand the signals on your scope. I had a couple hours yesterday to experiment with a master and two remotes in various states. It is still not clear to me that the signalling is entirely with the visible preflash. I can "hide" remotes well out of line of sight from the master and they still fire. Yes, the room has an eight foot high, white ceiling and light colored walls that are reflective. Is there possibly some infrared component involved?

I have started compiling a list of the various modes (states) that can exist between the D2X and the SB800 in an effort to identify just how much (or little) data may be exchanged between them while in each state. It may be a manageable matrix that will quickly lead to a relatively small number of variables which are conveyed in "packets" from the master to the remotes. Obviously, one of the more interesting variables in each state relates to time intervals.

We know that the master must identify channel and group for starters. I suspect that the master broadcasts data (including the "do not fire" signal) for all three groups on the selected channel whether they are populated with remotes or not.

Beyond that we would like to know the content of each packet conveyed to the remotes: exposure compensation values; how these relate to flash duration; etc. It is easy to assume that the remote triggering sequence is initiated by the leading edge of the master flash when the shutter is open. But, how is this done when the master is "Off" during exposure and not firing? Is a time constant communicated to the remotes during the per-flash?

The quenching logic for shutting down the strobes is not immediately apparent, at least to me. Is it done entirely in the camera or are the camera and strobes somehow synched here also?

I could go on but must leave soon for a photo shoot at a school for homeless young people here in the Phoenix area. This is an annual event for me and is always very moving.

We are clearly on the same path with this thread and I appreciate your interest in the pursuit of knowledge. I look forward to your next post.

Regards,

HBB in Phoenix, Arizona

HBB in Phoenix, Arizona
Nikonian Team Member

Photography is a journey with no conceivable destination.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

            
alm Registered since 29th Oct 2005Fri 16-Dec-05 06:32 PM
596 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#12. "RE: rear curtain sync question"
In response to Reply # 11


Zoetermeer, NL
          

HBB,

I don't have a logic analyzer in my closet, nor do I have a spare SB-800 to take apart. I could take apart a SC-17 cable to analyze the hotshoe signals, but I believe the wireless protocol are more interesting (unless you want to built your own flash).

The communication between master and slave seems to consist of light pulses about 100-250us apart. I don't expect that pulse amplitude is used (since a flash can't really change its amplitude, and because it's highly dependent on obstacles). I expect that either pulse length or the time between pulses is used. The former is hard to measure with my current setup, the latter seems to vary somewhat (if I remember correctly there were two or three different pulse frequencies).

Judging from the SU-800 and the color of the window, my guess is that it's sensitive to IR, but since white light contains plenty of IR (try holding your hand in front of a flash when it fires), it also works with visible light. Using IR wouldn't really help anything regarding the line of sight thing: it doesn't reflect any better than white light (even worse depending on the color of the walls/ceiling). I think the sensor is just very sensitive to the pulses (if my theory is correct, it only needs to count pulses and the time between them). A flash burst, doesn't matter how short, is still quite powerful (remember the amount of power is constant for an electronic flash, only the duration changes). If someone uses flash behind you accross a large room, you can often still see the reflection.

The light sensitive resistor doesn't behave as resistive as I expected: it behaves more like a capacitor whose discharge rate depends on the amount of light. This means that I get a exp(-t/RC) curve after each flash burst, instead of a nice pulse. I tweaked the series resistor so that it charges as fast as possible. Because I'm more a CS than an EE person, I plan to convert the sum of discharge curves to pulses via math, instead of building a proper active trigger circuit.

There's both communication before and after the pre-flash. I'd expect something like:
before pre-flash: Channel 1: group A, TTL, group B, AA
after pre-flash: Channel 1: group A, fire at XX power (since the camera determines the exposure)

The signals that I can think of to send before the pre-flash:
Channel number, group, TTL mode, AA mode, manual mode 1/1-1/128, repeat mode, modeling light, FP sync on
After the pre-flash:
only the power level for TTL and the signal to fire (may be triggered by the main flash burst)

I don't think the master sends any signals to the slaves to stop (would be quite hard to see with all the flashes firing at full power), since there isn't any flash metering during the exposure with CLS. Either the flash determines it's own exposure (AA/manual mode), or the body calculated it after the pre-flash (i-TTL). I think the TTL flashes just get the amount of power in the second burst of short flashes after the pre-flash. I'm not sure how pre-flashes with AA work in wireless mode, it might be needed so the camera can see the effect of the AA flash to properly set the TTL flashes.

I can only test on my D70 (and I have one SB-800 and one SB-600). If you send me a D2X I'll be happy to test . I don't expect the lack of slaves to be a major problem, since communication from the master to slave seems rather one-way to me (except for the pre-flash in TTL mode, the slave doesn't seem to confirm anything, although I didn't measure a slave yet).

To be continued .

Alson

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

                
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources. Charter MemberSat 17-Dec-05 02:21 AM
8508 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#13. "RE: rear curtain sync question"
In response to Reply # 12


Phoenix, US
          

Alson:

Since I don't have a window seat at your oscilloscope, I must rely on you for interpretation of pulses. Thanks for your very articulate interpretations.

I concur that the coding is probably not amplitude modulation even though the pulses must have a rise time that is greater than zero. Digging deep into my prior years in the computer industry developing early computer based data communications signalling systems (circa 1963-1970), I recall that there were some encoding schemes based both on the pulse width and the interval between pulses already digitized. (Return to zero and non-return to zero comes to mind. Does this ring any bells with you?)

Given that the SB800 designers would want to convey the essential data in the minimal time, such a scheme might make sense here. This would require a fairly accurate clock running at a suitable frequency. This leads to the question of accuracy and precision of the clocks in the master SB800 and multiple remote SB800s. If the middle of the rise and decay times can be fairly accurately established and the clock is of sufficient precision, then the width of the pulses and the width between pulses can each be interpreted as data. When the middle of the rise time is reached, the start time for this pulse is noted. When the middle of the decay time is reached, the time is noted again and represents the end of the current pulse and the beginning of the next one. Subtracting the start time from the end time gives the interval for this pulse already digitized. The process repeats for the zero (on the Y axis) interval between pulses. This approach also seems to simplify the analog to digital conversion issues associated with other approaches.

In the above manner, no time across the signalling interval is wasted. The clock base must be fast enough and stable enough to accurately accomodate the inherent variations between SB800 units.

I suspect that all of this communication occurs in the interval between the depression of the camera shutter button beyond the half-way point initiating the exposure and the opening of the shutter. This seems to be what I observe when using rear curtain synch with multiple remote strobes and multiple seccond time exposures.

I recall reading the shutter lag time in the D70 and D2X somewhere but cannot remember where. These numbers will give us the interval we have to work with and the shortest of the two will be limiting as the SB800s work equally well on the D70 and the D2X. With this shortest interval and the amount of data required in hand, we can begin to estimate the number of pulses we could expect in this interval and relate this to the data transfer required.

All of our discussion to date implies that there is a fairly accurate clock (probably piezo-electric crystal) and a microcprocessor with a modest amount of memory running in each SB800. Presumably, the same clock is used by the microprocessor and the rise/decay time detectors. It is difficult to imagine much of this system running in an analog mode, agree?

I wish we could discuss this over a pint or three of your country's famous beers and could draw diagrams. Thanks again for your interest and efforts to unravel this piece of technology.

I'm off tonight to investigate some sites for my next set of night photos of Phoenix police officers. One possible site is on the side of one of the mountains that ring the city. We plan to use a wide angle lens at dusk with a six or eight second time exposure to capture a bit of the fading blue sky and sunset with the city lights coming up and then fire six or eight SB800s (rear curtain synch) to illuminate the patrol car and officers.

I look forward to your next exchange.

Regards,


HBB in Phoenix, Arizona

HBB in Phoenix, Arizona
Nikonian Team Member

Photography is a journey with no conceivable destination.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

                    
alm Registered since 29th Oct 2005Sat 17-Dec-05 04:41 AM
596 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#14. "RE: rear curtain sync question"
In response to Reply # 13


Zoetermeer, NL
          

HBB,

Here's your window seat: http://waalsdorp.nl/~alson/CLS-wireless-protocol/

I modified my first sensor so the capacitor discharges faster (makes it less sensitive, but more accurate in the time domain, where I really need the extra resolution). I made a second sensor today from a spare CD-ROM player that was killed by a shattered CD. The second sensor (sensor 1 in the above page) doesn't exhibit the weird capacitor behavior.

It just contains the result of a few measurements I did to try to find out the type of modulation.

Modulation: this seems to be a form of frequency modulation. I don't see any pattern in amplitudes in the measurements (in theory you could use amplitude modulation relative to the amplitude of the first pulse, but frequencies are less likely to be distorted in reflections). You are right that pulses don't have a zero rise/fall time, but so do my sensors . With a timebase of 10us/div, I can see oscillation from the first sensor. The pulse width of the first pulse seems to be about 300ns. To accurately measure the pulse width, the slaves would need a lot more resolution (AD convertor needs a samplerate of 10MS/s) than for measuring pulse width (which is in the order of 100 us). If I were the designer, I'd only use pulse width if I didn't have enough bandwidth with FM. And even if I didn't, it would be easier to just increase the base frequency.

Hardware is even easier: an analog trigger that triggers on sudden bursts of light, pin high on micro-controller, sensed at next clock tick. Works fine assuming the clock of the uC is fast enough (only need 100KHz for this according to my measurements).

Since the use of pulse width modulation seems unnecessary, and because it's hard for me to measure, I'm going to pretend it doesn't exist unless proven otherwise .

My analysis of the encoding of the channel number seems to suggest that a trinary (maybe more?) encoding with FM is used. I count 11 pulses in these graphs (these are only the before pre-flash pulses, since there's no pre-flash and no TTL metering in this case). Assuming a trinary FM encoding, this would allow for 310 = 16 bits of information, it may be even more for more groups. 2 bits for channel, 2 bits for group, a few bits for mode, 6 bits for power on manual, seems plenty of bandwidth to me.

Re shutter lag: dpreview.com may have shutter lag measurements (not sure if they reviewed the D2x). I believe bclaff posted a preliminary article about what happens with i-TTL (pre-flash before/after mirror up). He would expand it after he got information from people at Nikon. I think this was before the crash. I don't recall that he finished it, but I may have missed it. I don't think we need the shutter lag to calculate the max. amount of time, since it's easy enough with a scope to see where the pulses stop. I love digital storage oscilloscopes for single-shot events like this, I can't really fire my flash 50 times per second continuously without killing it, and I don't know any 50fps Nikon dSLR's.

I agree that there's obviously a uC onboard (even controlling the dot matrix would be very hard without a uC). Since all current computers are synchronous, this implies a fairly stable clock too.

It seems to make sense to assume FM for now (unless we find that not all information is communicated this way). I plan to write a program that analyzes the data from sensor 1, 'triggers' on the pulses, and outputs the interval between them. This can be converted to the trinary (or more?) encoding used by the SB-800, after I run it on a sufficiently diverse set of data.

I hope your shoot went well, the disadvantage of rear sync is that you have to keep your models stand still for 6-8 seconds until the flash fires.

Alson

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

                        
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources. Charter MemberSat 17-Dec-05 10:48 PM
8508 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#15. "RE: rear curtain sync question"
In response to Reply # 14


Phoenix, US
          

Alson:

Thanks much for the window seat. Great traces!

Your scope is obviously much newer than mine ... and with memory yet! I'm impressed. I will study them over the weekend and will be back with additiional questions and comments next week when I get a quiet moment.

I am assuming the variable vertical position of trace two is simply where you placed it on the screen. Correct?

I also assume that the time interval on the X axis is 10 us per small division, giving an elapsed time of approximately 250 us for the series of pulses.

When you say "trinary" coding do you mean three digit positions (columns) with a total of ten discrete symbols (e.g., 0 through 9) available for each position? This would give 59,049 combinations (3^10 = 59,049). I would interpret this as three positions of base ten.

Or, do you mean, three digit positions (columns) with a total of three discrete symbols (e.g., 0, 1, 2) available for each position? This would yield 27 combinations (3^3 = 27). To me, this would be three positions of base three.

I will study some more and get ack to you with more questions.

The shoot with the homeless children went well. I'm always amazed that they handle things as well as they do, given what some of them have been and are going through.

Yes, time exposures with rear curtain synch do require cooperative subjects. Officers in fixed poses work very well for this type of shoot. Hovering helicopters are another matter all together.

Thanks again for another great step.

Scott: If you are listening, thanks for getting this started!

HBB in Phoenix, Arizona

HBB in Phoenix, Arizona
Nikonian Team Member

Photography is a journey with no conceivable destination.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

                            
alm Registered since 29th Oct 2005Sun 18-Dec-05 12:08 AM
596 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#16. "RE: rear curtain sync question"
In response to Reply # 15


Zoetermeer, NL
          

HBB,

Vertical position is indeed where I placed them so you can see both traces. Absolute voltage isn't terribly interesting anyway, because for sensor 1 you would just see the voltage of the photodiode, and for sensor 2 you'd see the powersupply voltage I chose (5V). I might draw schematic of both sensors when they are 'finished', but the important point is that the lower the Y-value is, the higher the light level (I believe sensor 1's output (Y-axis) is logarithmic, I'm not sure about sensor 2).

'New' digital scopes (I believe mine is about 8 years old, but I bought it used a few years ago) really help in this case: they have memory so they can capture non-repeating events easily (you won't see a single shot event on a cathode ray tube unless you manage to trigger a camera at the same time), and you can transfer the data to a computer (these traces are screenshots from the computer program I use to transfer data from my scope via the serial port). The timebase setting is at the bottom of the images, 0.5 ms per large div, so 0.1 ms per small div. The 250us was for the interval between two pulses. The entire trace is 5ms (10 divs) long. Usually there would be a pre-flash after these pulses, then some more pulses (to transfer the results of i-TTL metering to the slaves I think), and then the real flash burst. But since I have turned off the flash on the master, and I set the slave to manual (and didn't have any actual slaves), you only see the initial communication.

By trinary I indeed mean three different tokens (eg. 0,1,2) per interval. This is from the measurements on the channel data (look at the two tables at the end of the page, I interpreted this as three tokens: ~77 = 0, ~114 = 1, ~185 = 2. I'm not sure why they count 01, 10, 11, 02 instead of 01, 02, 10, 11 or even simply 00, 01, 10, 11. I'm not sure about the trinary thing, since they may have even longer intervals for other tokens.

I'd be happy to get some feedback on these results. Everything except the measurements are obviously my interpretations, and can (and probably will) contain mistakes.

I didn't have time to do any new measurements today, but hope to write an analysis program (calculate interval between pulses so I don't have to measure them by hand) and gather some more data soon. If you have any special request for a certain measurements, eg. pre-flash and main flash, or a graph of a whole wireless TTL sequence with pre-flash with a larger timebase setting (you won't be able to recognize discrete pulses here), feel free to ask. It currently isn't very useful to look much closer at the pulses, since you see a nice second order differential equation (i.e. damped oscillation) because a piece of breadboard and some wires isn't really optimized for high-frequency work.

Glad that your shoot with the homeless children went well. Long exposures with helicopters seem especially problematic since a helicopter usually has lights that will show up on the photo no matter how much flash you use, and won't stay completely still for several seconds I imagine.

Alson

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

                                
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources. Charter MemberMon 19-Dec-05 04:43 PM
8508 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#17. "RE: rear curtain sync question"
In response to Reply # 16


Phoenix, US
          

Alson:

I had a few minutes this morning to look at the traces you posted on my smaller light table. I placed traces for channels two through four under trace one, one at a time. I then added traces two through four to trace one in pencil. I concur, this is where the channel signalling is done. Beyond approximately 2.5 ms, the traces are almost exactly congruent.

Out of curiosity, I also measured the vertical height of the last pulse appearing at approximately 2.5 ms on all four of the upper traces in millimeters:

Channel 1 = 32
Channel 2 = 19
Channel 3 = 11
Channel 4 = 41

While not precise, these are roughly in increments of ten. Dividing each of these by ten produces:

Channel 1 = 3
Channel 2 = 2
Channel 3 = 1
Channel 4 = 4

I recognize that the top and bottom traces are for the same signal using two different sensors you have assembled. There appears to be a correlation between pulse height and channel in the upper trace, agree? You will know better than I if this is significant or simply an artifact of your sensors. I also recognize that a single point does not make a curve and I await your computer analysis of the intervals.

Have you had a chance to capture pulses including signals for all three groups A, B and C? A first test would include identical data to all three groups on a given channel. Then we could identify the coding for that state. Next, try sending different data to each group to see if we can distinguish that coding.

I'm still a little puzzled concerning thet triggering of the image capturing flash when the master is set to off (---). If the remotes do not see this in visible light, they must be seeing it in infrared. The newly released SU-800 does not use a visible light preflash, but infrared. Current production SB800s appear to be completely compatible with the SU800. Ergo, they are happy with infrared and/or visible light for signalling. I recall that the Nikon web site indicates a 900 to 1,000 nanometer wave length for the infrared signalling. This is rougly twice the wavelength of the 550 nanometer green in the middle of the visible spectrum. With a fairly simple and inexpensive bandpass filter this seems to be fairly straighforward logic.

You're right, time exposures are useless for hovering helicopters. I do have the luxury of asking the pilots to turn off their navigation and other lights during these shoots which helps a lot by eliminating hot spots. Some time ago I posted a shot of one of the helicopters sitting on its landing pad with the rotor turning. I don't know if you saw it or not. With the bird on the ground, six seconds was perfect for bringing up distant runway and other lights followed by five SB8900s on light stands firing.

Thanks again for all your efforts in this very interesting endeavor. I look forward to your next post.

HBB in Phoenix, Arizona

HBB in Phoenix, Arizona
Nikonian Team Member

Photography is a journey with no conceivable destination.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

                                    
alm Registered since 29th Oct 2005Mon 19-Dec-05 11:54 PM
596 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#18. "RE: rear curtain sync question"
In response to Reply # 17


Zoetermeer, NL
          

HBB,

The reason why I discarded the option of AM is that the relative height of the other pulses (eg. compare the amplitude of the set of three pulses before the last one between channel 1 and channel 2) doesn't seem constant. Since I only changed the channel, I'd expect them to be constant. I wouldn't spend too much time looking at the absolute amplitude, since it's highly dependent on the distance between the sensor and the flash (which I didn't keep constant, nor is it in real life). I usually change the distance so the amplitude is as large as possible without getting out of range. It's good to know that the rest of the traces (after the first three pulses) are nearly identical, this confirms my suspicion that the channel is determined by the first three pulses.

While I agree that there's a correlation, I don't believe it's significant. I can do some identical tests to see if these differences are consistent (absolute differences won't ever be consistent since they change with the inverse square of the distance), but it's not really my priority at the moment. I plan to continue on the FM path first, and only if that fails (i.e. there's information that's not encoded in the interval), I'm going to look at other options. My other problem with AM is that as far as I know, the power of an electronic flash is constant, only the amplitude varies. Of course the sensor might do some integration so a wider pulse might appear higher, but I wouldn't trust it. If you look at the first trace below the '...' header (I should put an actual header there as soon as I think of one), you'll probably understant why I'm hesitant to say anything about amplitude or width from these traces (you're talking about signal bandwidth in the order of 10MHz, this is really beyond the capabilities of these sensors, only carefully constructed circuits have that kind of bandwidth).

I agree with your suggestion to start with the group encoding, and then the messages to that group (eg. look at AA, TTL, Manual, modelling licht, repeat flash), and then other options (eg. manual power or exposure compensation in AA mode). I didn't have the chance to do a lot of additional measurements or analysis yet. I've been trying to improve the sensor to decrease the oscillation, without much success, so I'll just have to live with it for now. I did add a few traces recently, including a picture of the setup, at your regular window seat .

Re triggering, as I wrote, I expect it to be triggered by IR, and the 'white' light from the SB-800 happens to contain IR. I don't think the SB-800 can change the spectrum that it sends. Most photodiodes (which I assume they use) are more sensitive to certain wavelengths. Plus the window in front of the sensor also provides some filtering. Since the wavelength doesn't contain much information, this doesn't have to be a very accurate filter.

I did see the picture of the stationary helicopter, this clearly worked fine, but dynamic pictures (although the turning rotor already helps) also have their advantages. I guess you just need to have your speedlights reproduce some more (exponential growth assuming unlimited growth) so you can illuminate the mountains on the background with flash too .

Alson

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

                                        
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources. Charter MemberTue 20-Dec-05 01:54 AM
8508 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#19. "RE: rear curtain sync question"
In response to Reply # 18


Phoenix, US
          

Alson:

Thanks for the update, the new traces and a shot of the setup.

There is an electronic supply store very close to me that provides all the components for a DeVry Institute (electronics training) facility right across the street. If it would help, I can stop in and explore their supply of photo diodes and related detectors. If they have one that will make your job easier, let me know and I will pick it up and forward it to you.

I suspect that your work on the SB-800 will lead to other opportunities with other units. We might as well be properly equipped. The Tektronix 220 is a nice unit. I looked at them recently in a store here and wondered if I should acquire one. I have resurrected my 4 X 5 view camera form its twenty-five year sleep and need to check the shutter timings. Yes, I could take the lenses/shutters to a camera repair shop near here and have them do it although that is not nearly as much fun.

I am amazed at how semiconductor technology has reduced the size of these units. My scope is so old I am embarrassed to tell you ... it was built long before the memory units appeared.

The more I think about the new SU-800 and your observations, the more I agree with you: the communication and triggering is IR. This suggests that the SU-800 was in the works when the SB-800s were designed. Nice "systems oriented" thinking.

I remember seeing a film clip of Dr Harold Edgerton, inventor of the strobe light while at MIT. He constructed what appeared to be an enormous unit with a reflector a couple feet in diameter and an enormous bank of capacitors. An assistant held a sheet of newsprint a few inches in front of it. When the unit fired, the sheet of newsprint burst into flames.

For the moment, I am content with eight SB800s and the SU-800 ... if it ever arrives. I have a number of other night shots planned in coming weeks. Now that the police department is beginning to recognize possibilities, imaginations are working overtime. I'm willing as long as they are. As long as my supply of lost socks and dryer lint holds out, I can provide nourishment for the reproductive process.

I'll let you know what I discover about sensors at the electronics store tomorrow.

Regards,

HBB in Phoenix, Arizona

HBB in Phoenix, Arizona
Nikonian Team Member

Photography is a journey with no conceivable destination.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

                                            
alm Registered since 29th Oct 2005Tue 20-Dec-05 02:49 AM
596 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#20. "RE: rear curtain sync question"
In response to Reply # 19


Zoetermeer, NL
          

I believe the current setup works fine as long as we only have to measure the intervals, the current traces are OK for distinguishing peaks. If I'd want a better resolution, I'd probably look at commercial optical probes designed for debugging remote controls or fiber signalling, but they are expensive and this is free.

The TDS220 is basically a digital equivalent of recent cathode ray oscilloscopes, with a few digital goodies like memory and on-screen measurements/cursors/math (user interface is still pretty analog-like, eg. real knobs instead of push buttons). It doesn't have all the advantages of the expensive digital scopes (>$4k), like a large memory that can store very long pulse trains (I believe it can store only 1000 or 2500 points), nor does it have a very low noise floor, but is relatively affordable (I think I paid 650 euro's for a used model with communications/math module that can do some stuff like fast fourier transform to mimic a spectrum analyzer), and I don't do very complex stuff, so I'm happy with it. The lunchbox size is of course a nice bonus. There's nothing wrong with old analog scopes (most analog work isn't high-frequency anyway), but they just can't capture single shot events like this very well.

Alson

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

                                            
alm Registered since 29th Oct 2005Wed 21-Dec-05 08:27 AM
596 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#22. "RE: rear curtain sync question"
In response to Reply # 19


Zoetermeer, NL
          

Update: I added some stuff to the webpage. I now tried different groups. Although I see some sort of pattern, the representation of C (296 148 in C and A,C, 74 146 in B,C and A,B,C) confuses me. This is clearly not a simple one token/set of tokens per group, but something more complex. I see three tokens in the group encoding: 74, 148, 294 (in contrast with the tokens for the channel: 77, 114, 186), but haven't figured out the meaning yet. It might help to test more combinations (combinations of ---, M, AA, TTL), but this gives 64 combinations (7 of which I already tested), plus AA and TTL introduce the complications of pre-flashes, so I was really hoping to figure the group encoding out before trying this so I only have to try each mode for one group. They might use some form of compression (eg. instead of A M/128, B M/128, C M/128 you can use A M/128, B same, C same).

I'm probably going to try harder to find a pattern in this later when I have more time, but I wouldn't mind help with it either .

Alson

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

                                            
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources. Charter MemberThu 22-Dec-05 03:11 PM
8508 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#23. "RE: rear curtain sync question"
In response to Reply # 22


Phoenix, US
          

Alson:

Sorry for the delay. I have been temporarily diverted. Once through the weekend I will have more time to get back to this.

I keep coming back to the following state which I use frequently for my night shots of police officers:

SB800 on camera in master mode.

Camera set to six second time exposure with rear curtain synch for master and any remote SB800s

SB800 on camera set to --- (Off)

I push the shutter, the preflashes fire, six seconds elapse and the remotes all fire while the shutter is still open and the master does not.

Question: How do the remotes know when to fire since there is no visible flash from the master to trigger on. I doubt that the SB800s are capable of accurately synching with individual, internal timers that, by definition, are running asynchronously even if the camera transmits the time exposure interval which I doubt. I suspect that the signal to fire for the exposure is emitted in infrared by the master SB800.

There is a large infrared window on the front of the SB800 below the strobe tube window with a pair of lens-like devices within that are used for focus assist.

If your sensor is sensative to infrared at 900 to 1000 nanometers could you set your SB800 up as described above with a several second time exposure with rear curtain synch? Then ignore the visible preflash and capture the proposed infrared exposure trigger from the infrared window?

The new SU800 using strictly infrared has been reported in this forum (Deven: Dec 15)with less range than the SB800 using visible light. This may be a clue for our exploration as I would expect the beam spread would be a bit greater and energy levels in infrared would be somewhat less than in visible. Is it possible that the signalling from the SB800 master to the remotes is a combination of visible and infrared? While we agree for the moment that bandwidth does not seem to be a problem, things may be more complicated than we think. Perhaps the SB800 is signalling in both visible and infrared in parallel instead of the initially presumed serial mode?

Thanks for your recent update. I will print it and add it to my file for examination once beyond the weekend events.

Question to others following this thread: Does anybody have an SB800 that has been trashed through some mishap that you would be willing to donate to the cause? If so, let us know as getting a look inside one of these little marvels would be a great help.
Thanks.

Alson, thanks for your persistence.

HBB in Phoenix, Arizona

HBB in Phoenix, Arizona
Nikonian Team Member

Photography is a journey with no conceivable destination.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

                                            
alm Registered since 29th Oct 2005Fri 23-Dec-05 07:47 AM
596 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#24. "RE: rear curtain sync question"
In response to Reply # 23


Zoetermeer, NL
          

HBB,

The delay is no problem, I understand that you have other things to do.

My sensor does seem sensitive to IR (it reacted to light of the remote that I use to trigger my D70, not the official ML-3 though), however, since the diode in the remote is a lot weaker than a flash, I had to press the diode to my photodiode. The photodiode actually seems more sensitive to IR than to blue, because I can't see any reaction to the blue LED that the remote also contains. I have no way of knowing how sensitive it is to different colors since I don't have the equipment to generate light with a constant intensity but with different colors, and I haven't been able to find a datasheet for the photodiode. I believe wavelength of the laser used for reading a CD is 780nm, so the photodiode is probably sensitive to light around 800nm, which includes (near) IR.

I tried the situation you described (D70 set to manual exposure, 2 seconds, f/4, rear sync, SB-800 on-camera set to M ---, A M 1/128, B ---, C ---, channel 3). I used an SB-600 as slave (channel 3, group A). The master flashes when I press the shutter (the slave doesn't), after 2 seconds both the slave and the master flash (suggesting that the slave is triggered by a white flash from the master). The fact that the slave doesn't flash until at the end of the exposure suggests that one-way communication is used (except the pre-flash, which I turned off by using manual mode). With the slave in TTL mode, the behavior was identical, except that the slave produced a pre-flash before the exposure. Could you try the same scenario indoors where you have a better view of your master flash (with only one or even zero slaves)? I suspect the master flash is obscured by all the slaves in your case.

It's possible that the SB-800 behaves differently on a D2X than on a D70, but I find this unlikely (since both use the same CLS system, it would make no sense to use different triggering). The fact that the slave flashes don't flash at all except for the pre-flash suggests that the master doesn't adjust its behavior depending on the type of slave (i.e. no different triggering for SB-600 and SB-800).

I wouldn't mind trying to measure this behavior, but from what I see it seems identical to the behavior I measured with first curtain sync: first the master communicates the modes and other parameters to the groups, then all groups fire an individual pre-flash if in TTL mode, then the master sends the results back to the slaves, then the shutter opens, the master gives a trigger flash, all slaves (and the master if not set to ---) fire.

I agree that if there was no flash from the master, it would be quite unlikely that all slaves would measure the exposure time individually (although not impossible if you allow for some slight margin), especially since you can use bulb mode (which behaved identical to the 2s exposure by the way).

Alson

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

                                            
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources. Charter MemberFri 23-Dec-05 10:37 PM
8508 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#25. "RE: rear curtain sync question"
In response to Reply # 24


Phoenix, US
          

Alson:

I will be setting up a D2X and a number of SB800s early next week to confirm that the D70 and D2X behave identically with the SB800s. I'm still puzzled about rear curtain synch with time exposures.

One of the buttons (right) on my Microsoft optical trackball died recently and I had to replace it. Naturally, I took it apart. This is the unit with two or three lights inside a tranaparent red shell holding the ball. This red light is reflected off a red ball with a reflective pattern below the transparent surface and two or more photosensors to capture movement.

Since this whole thing appears to operate in or near IR, I will happily forward it to you if you think it will help or if you would like to add it to your component drawer.

I just received a call that my SU-800 unit has arrived so I am going to dash off and get it. More later.





HBB in Phoenix, Arizona

HBB in Phoenix, Arizona
Nikonian Team Member

Photography is a journey with no conceivable destination.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

                                            
HBB Moderator Hal is an expert in several areas, including CLS Awarded for his excellent article contributions to the Resources. Charter MemberWed 04-Jan-06 10:01 PM
8508 posts Share on Facebook Share on Twitter Share on Linkedin    Click to send email to this author Click to send private message to this authorClick to view this author's profile
#26. "RE: rear curtain sync question"
In response to Reply # 24


Phoenix, US
          

Alson:

If you have time, I would be interested in seeing a trace from the Master SB800 providing different exposure compensations for each of the three groups, A, B and C. If the compensation factors could all be one stop apart it may help decode the signal. It shouldn't matter which channel you use, as we think we understand that piece of the puzzle.

I will have some time following Monday of next week to get back to this a bit. At that time, I will run a series of time exposure/rear curtain synch tests to identify the shortest interval between the pre and final flashes that I can detect. I'll let you know what I learn.

Thanks,

HBB in Phoenix, Arizona

HBB in Phoenix, Arizona
Nikonian Team Member

Photography is a journey with no conceivable destination.

  

Alert Printer-friendly copy | Reply | Reply with quote | Top

Forums Lobby MASTER YOUR TOOLS - Hardware & Software Nikon Speedlights & Lighting topic #11919 Previous topic | Next topic


Take the Nikonians Tour and learn more about being a Nikonian Wiki /FAQ /Help Listen to our MP3 photography radio channels Find anything on Nikon and imaging technology - fast!

Copyright © Nikonians 2000, 2014
All Rights Reserved

Nikonians®, NikoScope® and NikoniansAcademy™ are trademarks owned by Nikonians.org.
Nikon®, Nikonos® and Nikkor® are registered trademarks of Nikon Corporation.