69 replies to this topic

[Dupsbear]

Working also on a microflash. Here some conclusions:

- IR leds consume to much energy for the amount of light they give
- use red leds with a wavelength as high as possible

And now, gone do some more tests!

Hello Owfotograaf,

I am not an electrical engineer, but do have some experience in electronics. Would you please explain your comment that IR LEDs consume too much energy?

An IR LED uses about 1/2 the voltage, but does use more current. A typical IR LED would drive at 50mA @ 1.5VDC voltage drop, where a visible red LED might use only 20mA @ 2.4VDC voltage drop. Would not the trade off of voltage to current keep, what we might call the battery usuage about the same?


These LEDs are going to be on for such a small amount of time and number of flashes, I wonder if worrying about battery drain is an issue.

I am not saying you are wrong, I want to know in case I decide to build one of these circuits for myself or a friend.

Many thanks,
Barry
www.bareports.com
www.keywordworkflow.com

[rtrski]

It might be more a matter of luminous efficiency. I've already noticed that efficiency for the LED's I've been looking at varies significantly with wavelength, with greens being the most efficient. But it doesn't seem to be linear at all, and is probably related to some range of effective wavelengths for each LED semiconductor type. Osram's for example ( http://catalog.osram.../00051743_0.pdf ) are more efficient for amber (redder, longer wavelength) than yellow, for the same spectral bandwidths about their center wavelengths...which is counterintutive in comparison to green being higher efficiency.

And of course if there's any water in the path, IR is going to attenuate far faster than any visible light, of course.

Edited by rtrski, 21 October 2011 - 07:00 AM.

[Dupsbear]

It might be more a matter of luminous efficiency. are more efficient for amber (redder, longer wavelength) than yellow, for the same spectral bandwidths about their center wavelengths

And of course if there's any water in the path, IR is going to attenuate far faster than any visible light, of course.

I tried an amber LED with my SB800 and could not get the flash to fire, but I am not sure of the output of the amber LED I used. Of course, red worked fine, but the IR LED seemed to work the best. To test the IR LED efficiency with your flash unit, without having to purchase and IR LED, you can use a TV or cable remote control.

Obviously, as you stated, the IR would attenuate much faster though water, which is why the Nikon CLS system doesn't work well underwater. I would think with fiber optics that we would not be shooting through water. I want my light sources and receivers to be as close to each other as possible.

Thanks for the reply,
Barry
www.bareports.com
www.keywordworkflow.com

[rtrski]

I tried an amber LED with my SB800 and could not get the flash to fire, but I am not sure of the output of the amber LED I used. Of course, red worked fine, but the IR LED seemed to work the best. To test the IR LED efficiency with your flash unit, without having to purchase and IR LED, you can use a TV or cable remote control.

Obviously, as you stated, the IR would attenuate much faster though water, which is why the Nikon CLS system doesn't work well underwater. I would think with fiber optics that we would not be shooting through water. I want my light sources and receivers to be as close to each other as possible.

Thanks for the reply,
Barry
www.bareports.com
www.keywordworkflow.com

Well, a TV remote has perhaps only 3V supply (2 AAA's are kind of typical these days) but definitely has room for all sorts of voltage regulation to supply a charge multiplier to a higher voltage, caps for storage to deliver high current for the miliseconds that the IR is flickering, etc. Not necessarily a good test of what I can deliver on my own.

And now that I think about it...the Inon z240 (and z2000) photodetectors in the latest versions are designed for 'wireless' photodetector response, with water in the path, and assuming the camera's main flash (which will be generally fairly white) as the trigger source. Put it all together and that strongly implies that the photodetector is optimized for shorter (bluer) wavelengths vs. longer, since they might assume the longer wavelengths have been attenuated before they get to the photodetector. So the yellow and amber ones I just ordered from Mouser.com might be going the wrong direction...I might be getting the best coupling if I did try a nice bright green LED after all. Strange that the SB800's work best for you with IR or redder...maybe they were only ever designed for fiber use?

Another comment re: your current and voltage drop info above: the net 'wattage' from each might be about the same, assuming the current draw is available. With the CR2016's specifically, they're struggling to output 100mA. So given the current draw limitations of the batteries, its easier to get the light out of the visible LEDs.

All the above (plus worries about fitting the circuit into the housing - on my Sony for example I had to put the circuit board slightly offcenter atop the hotshoe, as the hotshoe location isn't centered beneath the viewfinder hump allowance on the Ike housing) sort of explains why no one has offered a simple product like this yet. Different LED colors for different strobes, LEDs either on the circuit board (like mine, to look thru the clear wall) or pigtailed to extend up to the fiber bulkhead adaptors (as in escape's images), available circuit volume, even different hotshoe mounts (I guess Sony is the only unique one, everyone else is pretty standard)...there'd have to be a lot of configurations to assure it would work. Couple that with a low count customer pool tobegin with and you end up with something that stays DIY.

[owfotograaf]

An IR LED uses about 1/2 the voltage, but does use more current. A typical IR LED would drive at 50mA @ 1.5VDC voltage drop, where a visible red LED might use only 20mA @ 2.4VDC voltage drop. Would not the trade off of voltage to current keep, what we might call the battery usuage about the same?

These LEDs are going to be on for such a small amount of time and number of flashes, I wonder if worrying about battery drain is an issue.

I have tested them, and have no good experience with them. The power consumption is about the same, your right. The specs also don't tell the mcd brightness.
If you want to work with CR.... cels, witch are small, 100mA is the maximum you can draw out, even only 2ms time, the tension drops at lot.

[Jesper64]

And now that I think about it...the Inon z240 (and z2000) photodetectors in the latest versions are designed for 'wireless' photodetector response, with water in the path, and assuming the camera's main flash (which will be generally fairly white) as the trigger source. Put it all together and that strongly implies that the photodetector is optimized for shorter (bluer) wavelengths vs. longer, since they might assume the longer wavelengths have been attenuated before they get to the photodetector. So the yellow and amber ones I just ordered from Mouser.com might be going the wrong direction...I might be getting the best coupling if I did try a nice bright green LED after all. Strange that the SB800's work best for you with IR or redder...maybe they were only ever designed for fiber use?

I don't think the Inons are necessarily optimised for shorter wavelengths of light. As they market the Clear Photo System for clear housings, which is basically a piece of film that is placed over the internal flash and it blocks out visible light and allows the IR from the flash to trigger the strobe.

[rtrski]

Jesper:

Thanks for the reassurance; that does make sense.

I've been able to put it all together and am not getting very reliable triggering, even with the ends of a 3mm lightpipe nicely polished, unless the LED to F.O. input alignment thru the Ikelite wall is nearly perfect. So....I'm back to the drawing board a little bit. It's just the triggering alignment between lighpipe input and LED thru the housing wall that's causing a bit too much loss at the moment. I already know that 1) if I can go back to two 2032's from the 2016's, the LED flash gets a lot brighter, and 2) I've got higher (about 2x) luminosity LED's on order than the white ones from Fry's (plastic encasing provides only a 20degree view angle) I'm using now. The new ones are 60deg view angle in orange-red, amber, and yellow options (Osram, Kingbright, and Avago). Between the two I think I'll be able to git-r-done, but the LED's won't arrive until after this weekend. Sigh....on to iteration 4. I think I'll have to shave down the plastic housing of the hotshoe adaptor a bit more to get the dual 2032 battery holder to fit the vertical space in the housing, and I'm getting awful close to out of space to shave it further already. Thanks for the extra-high T-grip style iISO hotshoe mount design, Sony/Minolta.

Edited by rtrski, 22 October 2011 - 05:41 AM.

[Dupsbear]

Jesper:

Thanks for the reassurance; that does make sense.

I've been able to put it all together and am not getting very reliable triggering, even with the ends of a 3mm lightpipe nicely polished, unless the LED to F.O. input alignment thru the Ikelite wall is nearly perfect. So....I'm back to the drawing board a little bit. It's just the triggering alignment between lighpipe input and LED thru the housing wall that's causing a bit too much loss at the moment.

I have some little reflectors, 12mm O.D., that slip over an LED and increase the forward intensity of light, like a flashlight. It would be interesting to know if these reflectors would help with your LED/housing/lightpipe interface. A reflector could also be put on the lightpipe, using the same principle of using a dish to capture more satellite signals. I have had the reflectors for many years and do not remember where I got them; I feel sure an internet search should find something that would work.

Barry
www.bareports.com
www.keywordworkflow.com

[rtrski]

We think alike...I'd already stolen one off of an old MINI-MAG flashlight (the bulb kind) but I think even with that, I'll still go the bigger battery and another circuit spin. The reflector is just some sort of 'metalized' plastic...I'm sure the coating isn't really chrome or anything, but I do admit I'm not sure how it will work in water. I could use it around the fiberoptic receptor outside the housing but it definitely won't fit around the LED in the housing (max available height above the circuit board is about 6mm).

My wife did buy me (promise not to laugh...too hard) some sequins at Hobby Lobby that are about a 15mm diameter...might be able to hole-punch one of those, split the resulting donut on one leg, and twist and glue into a bit of a cone shape.

[Dupsbear]

but it definitely won't fit around the LED in the housing (max available height above the circuit board is about 6mm).

My wife did buy me (promise not to laugh...too hard) some sequins at Hobby Lobby that are about a 15mm diameter...might be able to hole-punch one of those, split the resulting donut on one leg, and twist and glue into a bit of a cone shape.

Thinking out of the box, why do the LEDs have to be above the camera? Why not place the LEDs in an area of your housing where there is plenty of room.

Also, here is a neat bulkhead that would give a better light transfer from inside to outside the housing. http://www.watershot...e-bulkhead.html and, as is available on my Anthis Nexus housing, http://www.anthisnex...;product_id=319. The Nexus part lens looks like it is made of glass and really transfers the light!

I have to admit, the sequin idea did bring a chuckle, but not a bad idea.

Barry
www.bareports.com
www.keywordworkflow.com

[rtrski]

Thinking out of the box, why do the LEDs have to be above the camera? Why not place the LEDs in an area of your housing where there is plenty of room....

Technically they don't have to be there, but this is an Ikelite housing, so I've got the clear plastic wall to use whereever it ends up, and I'm trying to avoid any mods to the housing itself (which would include removing the electrical bulkhead and replacing it with something to help with the light coupling). The left side gap between camera and housing is taken up by the existing electrical bulkhead connector - I can tuck it down there if it's being replaced with my circuit on the hotshoe instead. The right side is taken up by the stub end (larkshead loop and plastic clip) of my usual wrist strap, the bulk of which is removed when in-housing but if I can leave that part on it's just easier for converting between land and water use. And of course beneath the housing there's the big slab of the Ike electrical TTL circuitry which also acts as the camera's mount platform.

There's also a slight gap between where the top clamp handle rests in the locked position and the top ball mount plate (looking straight down at the top of the housing), providing a nice look-thru space that's also somewhat shielded from scattering forward, and the top ball mount plate is the perfect place to attach whatever I'm using to hold my fiberoptic instead since it's already fitted for a 1/4-20 screw thread. So having the LED's up there isn't really the biggest issue...it's fitting the batteries, frankly. Those I could try to tuck in elsewhere and 'wire' to the hotshoe mount, but its somehow much more satisfying to have a nice coherent part. Looks slightly less bomb-like in my carry-on luggage than a mess of wiring, too, if it's sort of more obviously productized to sit atop a hotshoe.

I've got a new circuit laid out that's a bit longer front to back and should put the LED's in a better position (per the gap described above) for me to couple, and with a new attempt at thinning the hotshoe mount I think I've gained about 1.2mm, which should be enough to let me slip in the dual-2032 battery clip vs. the dual 2016. I'll etch the board and re-solder everything this week once the new LED choices show up, after perhaps a quick breadboard to test between the 3 options. I might end up sacrificing the ball mount in the sense of not coming up with both a FO 'receiver' fitting and maintaining the attach point for the ball as well, but that's an easy enough price to pay as I can go back to having a focus lamp off one of my arms if I really need one.

[rtrski]

I think I might declare success this weekend. New circuit layout was fit-tested in the housing (paper cutout on the hotshoe mount with the battery holder taped on ), etched the circuit and soldered on all components last night, and was able to fire off the strobe just pointing the FO at the LEDs up to almost 5 inches away with no housing, and 3 inches away thru the housing wall. Didn't test lateral displacement that much but the new LEDs (I used 2 half-watt 'hyper orange' [617nm] Avago surface-mounts, in parallel even though I was the one recommending against that before ) are bright enough I can't comfortably look right at them when they fire, and they've got closer to a 120 deg radiation pattern than the domed 20degree one. So I'm pretty sure this time I can get a fitting atop the housing that will align the FO well enough for reliable triggering.

...And, I can still use the sequins to focus the pattern a little more if I need to.

Pool tests this weekend, followed by updated project shots including the DIY FO cable and mount, and I'll upload my final circuit image and BOM as well in case it can help anyone else. The new circuit is almost looking like a real product now, since I shaved down the half of the hotshoe adaptor I was using and then mounted the circuit board directly on it using its existing screw holes, and routed the wires from one to the other a bit more invisibly as well. Hopefully that'll calm the TSA dogs a tad when I try to get it thru security. I'll probably still gunk it all up with hotmelt glue as a bit of a potting agent against corrosion if nothing else though (I did tin the circuit before adding components as well.)

Thanks to all here for the comments and encouragement.

Edited by rtrski, 26 October 2011 - 06:37 AM.

[bvanant]

Next step is selling some.
Bill

[rtrski]

I don't think I can, at least not in the current incarnation. The fit for mine was kind of particular in terms of the circuit board having to be asymmetric on the hotshoe mount to line up best with the housing bump. Cameras are aligned in the Ike housings with the lens to port, not with the viewfinder position to bump. Then there's knowing how much vertical room you have on any given camera...etc. etc.

If I was serious about selling, I'd want to come up with a couple layer circuit board (throw some of the smaller components on the bottom, with the wiring to the hotshoe mount itself), be able to guarantee the battery holder would fit, and probably design in a socket for replacement of LEDs just in case someone manages to shoot fast enough in sequence they burn it out. (Replacement parts, cha-ching! kidding....)

And despite how long it took me in particular to iterate thru sizes and layouts....the circuit is pretty danged simple. Now if I could reverse-engineer TTL signalling and have the LED actually preflash/flash/squelch so one could set their Inon to sTTL...now THAT would sell.

But I will put up a full product log sometime to help others not repeat some of my mistakes and get it in the first or second iteration.

[rtrski]

Well, my LED trigger flash circuit project is finished, including the fiberoptic cable and mounting to both the z240 and the Ikelite housing.

Full photo log of the flash development is here. Best viewed as slideshow. Hit 'pause' at the bottom left, hit the 'show info' button at the top right to get the per-picture text info, and then continue thru images at your leisure.

http://www.flickr.co...716388071/show/

Also for those who want to try their own, I'm providing my bill of materials and circuit board layout and other comments in a powerpoint file attached here for download (Zipped). It includes links to extra details about DIY circuit board etching using a laser printer to make the mask and household chemicals as the etchant vs. anything fancy. Also makes some fit test suggestions to help others get this done in less steps than I did.

The fiberoptic DIY cable has been posted in the DIY subforum, here: http://wetpixel.com/...t...st&p=294502

Looking forward to giving this a real workout in Dominica in January!!

Attached Files

•  LED_Trigg_R_BOM_and_Layout.zip   680.96K   101 downloads

Edited by rtrski, 05 November 2011 - 11:18 AM.

[escape]

Very nice rtrski.

My led trigger for Nauticam housing ended up like this.
Led module separated(two IR LED 850nm = almost no visible light but it's triggering very well Inon's strobes).

[bvanant]

I don't think I can, at least not in the current incarnation. The fit for mine was kind of particular in terms of the circuit board having to be asymmetric on the hotshoe mount to line up best with the housing bump. Cameras are aligned in the Ike housings with the lens to port, not with the viewfinder position to bump. Then there's knowing how much vertical room you have on any given camera...etc. etc.

If I was serious about selling, I'd want to come up with a couple layer circuit board (throw some of the smaller components on the bottom, with the wiring to the hotshoe mount itself), be able to guarantee the battery holder would fit, and probably design in a socket for replacement of LEDs just in case someone manages to shoot fast enough in sequence they burn it out. (Replacement parts, cha-ching! kidding....)

And despite how long it took me in particular to iterate thru sizes and layouts....the circuit is pretty danged simple. Now if I could reverse-engineer TTL signalling and have the LED actually preflash/flash/squelch so one could set their Inon to sTTL...now THAT would sell.

But I will put up a full product log sometime to help others not repeat some of my mistakes and get it in the first or second iteration.

Isn't that (or the first part at least) what Mattias and crew did to get their TTL converters going? Now all you need is to do (clearly I am joking) is to reverse engineer their reverse engineering.

Bill

[rtrski]

Very nice rtrski.

My led trigger for Nauticam housing ended up like this.
Led module separated(two IR LED 850nm = almost no visible light but it's triggering very well Inon's strobes).

(pictures trimmed so as not to spam the forum linking them)

Thanks! I have to say your final product looks much cleaner than mine. Almost looks like you sliced the bottom off an old strobe for the mount, and then added the top mount?

Those battery holders are nice too...probably easier to swap out batteries than my top-spring clip. But I bet they'd be too high for me, as I can see a little space between them that would add to the overall thickness. Still, very nice, clean integration. I like the locking cord link to the LED's too (from a computer power supply MOLEX interconnect of some sort, right? )

[owfotograaf]

"Star Trek Enterprise"

My design, uP controlled, supply 2x CR2032, low battery warning, hot-shoe with spring contact,
goes to sleep so no battery consumption when not used (battery's can stay)
red LEDS, (2 cd), can be bent to the fibbers, those not needed can be cut off.

[rtrski]

"Star Trek Enterprise"

My design, uP controlled, supply 2x CR2032, low battery warning, hot-shoe with spring contact,
goes to sleep so no battery consumption when not used (battery's can stay)
red LEDS, (2 cd), can be bent to the fibbers, those not needed can be cut off.

VERY nice...you've clearly got access to some good circuit etching capability. And I like those low-profile battery holders as well. I hadn't found anything quite that small but I was searching for 2x stackers vs. using two single-cell holders side-by-side.

I'm guessing your chip is on the bottom side, and that's a double-side etched circuit board. Or maybe it's on the hotshoe sub-board?