Convert cheapo flashlights to run off a transformer

[smack2000]

Hi there,

I have a little project I'm getting into, and I have a wiring question. I know if there's anyone that can give good advice on what I'm trying to do, they're somewhere in these forums...

There's these cheapo 21 LED flashlights you can get on ebay amongst other places. They have a plastic battery caddy inside that holds 3 AAA's, and they have a tail click switch. I'm attempting to build a ring light out of them for my camera by utilizing about 20 of these flashlights in a custom enclosure I'll build out of PVC. In order to save weight and money, I want to run these off of AC power via some kind of wall-wart transformer. Can anyone give me any advice? I don't even know, from an electrical standpoint, if what I want to do is feasible...

I want to keep the tail switch in tact so I can choose which lights are 'globally' turned on and off with the main power

I don't care about drilling holes in the bodies to accommodate the wiring.

I don't know enough about electricity to know what the power requirements would be and how to do the math, or the proper way to wire all the lights together so they all get the appropriate power, and can be individually turned on or off by the tail switch without breaking the entire circuit.

If anyone can help me get started on this idea, or even tell me there's no way it will work, I would really appreciate it.

Thanks!

Scott McCoskery

 

[Mr Happy]

Well fortunately for you, it's not electrically very complicated, though the wiring will be a chore.

You will need to take individual pairs of wires back from each flashlight to a regulated DC supply of about 4 to 4.5 V that can supply up to 10 A (more about that in a moment). You will have to be very careful to keep the polarity right (that is, connect positive and negative wires to the right battery contacts inside the flashlights).

Three AAAs will give a working voltage of about 4 V, 4.5 V max, hence your transformer must do the same.

You can expect each flashlight to draw about 0.5 A (21 LEDs at 20 mA each with a bit extra for luck is 500 mA). That is also about as much as an AAA cell can comfortably supply, so it figures. Then 20 flashlights at 0.5 A each is 10 A.

Your 4.5 V 10 A supply is going to be tricky. It's not in any way a standard wall wart. You may be able to cannibalize an old PC power supply (they can typically do 5 V at humongous amps) but you will need to be a bit confident with electrical stuff to follow that path.

You also will want a regulated supply (it keeps the voltage at about 4.5 V whatever the load), so that you can run just a few flashlights or all of them without burning them out. Cheap standard wall warts are not always well regulated.

 

[mdocod]

I think the best bet is going to be to install a bucking regulator in each flashlight, there will be plenty of room since you won't have a battery pack in them. I would personally just use 350mA drivers for these so you don't have premature LED failures.

In all great honesty, there are much more ideal ways to accomplish what you are doing that will be similarly priced, substantially brighter, and have a much lower rate of failures crop up in the long haul.. Using bare high power emitters (like Cree XREs), buck style regulators, reliable toggle or rocker switches, and a 12V power supply..... The cheap LEDs used in those 21 LED lights are indeed fun (I have a few of em around here, some I have grown to love for what they are)... but these aren't going to make a good platform to build what you desire. (IMO)

 

[TMorita]

...
In all great honesty, there are much more ideal ways to accomplish what you are doing that will be similarly priced, substantially brighter, and have a much lower rate of failures crop up in the long haul.
...

I agree with this too.

It's kinda like....you decide you want a 250 horsepower hotrod, but you can't afford a 250 hp engine. But, you can get lawnmower engines very cheaply, so you decide to build an engine mount where you can mount 125 2hp lawnmower engines...

Toshi

 

[smack2000]

Thanks, everyone, for your replies.

First of all, I can buy a new ring flash for $400. So there is an upper limit to how much I'm willing to spend to build a DIY version. I could go out and buy one, but where's the fun in that? In fact, the goal is to build it as cheaply and easily as possible, and I'm sort of competing with a handful of other photographers that have attempted to build an effective ring light/flash for very little money. Most of the other solutions do not distribute light evenly around the ring and/or are enormous in size.

I am looking for ideas on how I can accomplish my goal without spending too much time and money. My budget is under $100. I own a mill and a lathe, etc; and have the machining expertise to probably build a $700 ring light given that budget for materials and unlimited time on my hands, but again, when I'm done, anybody with a soldering iron, a hacksaw, and some JB Weld should be able to replicate my project.

Mr. Happy: Thanks for the electrical advice. I thought it would be pretty straight forward as you've described. I have a couple of heavy-duty CC/CV DC Power supplies I use with my anodizing setup that should do the trick for testing purposes.

mdocod: You have some very valid points, and was expecting someone to leave a post just like your's. In fact, when I first started thinking about this, I looked through various components on theledlight.com. I quickly realized I'd be spending a lot of time designing this thing electrically -- granted it's probably a no brainer for you, but unfamiliar territory for me (I don't even know what a bucking regulator is), and those components add up quickly in price. I'd be interested to hear you expand on what you've suggested.

TMorita: I believe I've made my point about why I'm taking this approach. I *can* afford the 250HP hotrod, but this is about DIY on the cheap. If what mdocod is suggesting is both cheap (or equivalently priced) and easy, then by all means, I'll listen.


I will say that my initial draw to the 21 LED flashlights (other than the price) is the volume of LED's that would be used. An important aspect of a ring light is that it has even light disbursement all the way around the ring, and the more individual sources, the better. There will be a diffusion material (probably 1/8" white acrylic) an inch or two in front of the LED's to help with this.

Finally, what would it take to make this thing flash? Cameras (with sync ports) just send a small electrical signal to a standard flash to tell it when to fire, then there's some other circuitry that must regulate the length of the flash. The light would need to fire within 1/250th of a second (sync speed), and last for only about 1/1000th of a second. I know this would be impossible with, say, a tungsten light because it needs to 'warm up', but LED's seem pretty instantaneous to me. Thoughts?

Thanks again,

Scott

 

[Mr Happy]

As far as bucking regulators and Crees and stuff are concerned, that is adding cost and complexity. Everything needed to light up LEDs is already in a given flashlight, even the cheapest one. You just have to supply the right voltage. You can make things as complex as you wish, but why make them more complex than you need?

Making it flash? No, that doesn't sound right. The LEDs can be on permanently, like photo flood lights. That should help you too with modeling; you can work by eye to get angles and coverage right.

(Flashes only flash because it is impossible to make them be on permanently; there is no photographic reason to prefer flashes unless it is for stop motion photography, reducing heat on the subject, or something like that.)

Further thoughts: even 400 low power LEDs will have a lower light intensity than a ring flash, leading to longer exposures. The color temperature will be different too, so you will need to adjust for that. You may not get good color rendition across the whole color spectrum either.

 

[mdocod]

I don't know much about this light ring thing and what goals are far as lighting are concerned you are trying to achieve, but my gut tells me that you will also have problems with the quality of light from the cheap LEDs, these tend to be very bluish and lack output in a lot of the spectrum... modern high power emitters aren't perfect either, but they are a lot better. This reason alone is enough to consider other options...

My question is this: how many lumens does one of these rings need to be able to emit?

The other issue is.... you're limit is $100... 20 of those 21 LED lights you are talking about is going to cost $80-100+ by itself...

http://www.kaidomain.com/WEBUI/ProductDetail.aspx?TranID=2028
http://cgi.ebay.com/21-LED-small-3A...eZWD1VQQtrksidZp1638.m118.l1247QQcmdZViewItem
From there you either have to have 20 350mA driver boards and an unregulated DC power supply capable of delivering 7A, or a more expensive selectable output voltage bench power supply capable of delivering ~7A or more... You could cut costs by wiring pairs of the flashlights in series so you could get by on half the current (3A power supplies are about $80-100) (but you would loose half your control over that is on and off, would have to be turned on and off in pairs)... Either way I think you are looking at another $100 worth of stuff to power the darn things and maintain control over each one individually... if you were dealing with a situation where they were all going to be ON all the time, you could just wire the whole string up in series and run about 70V DC through it and call it a day. (could probably just use a large rectifiers and some big resistors and maybe a few caps to dampen the output)....

Ok.. So I'm wondering- If you have access to mains power, then do you really need this to be LED or lower power? what about a simple string of 120V light BULBS. Bulb Sockets and bulbs and PVC and wire and wire nuts and a bunch of 120V+ few amp rated rocker switches... You could get almost everything you could need for such a project in town at lighting supply stores or even at a place like Home Depot.... Though a lot of it could be online for cheaper.. This way you aren't even dealing with having to regulate this whole thing.

 

[smack2000]

Mr. Happy: Thanks for the response. I echo your thoughts exactly regarding building a 'new' light from the ground up with components. Like I said, for some people it's a no brainer, but for me, the fact that these flashlights already contain the appropriate circuitry is part of the appeal. But sometimes you find out it really is easier than you think, which is why I'm always open to new approaches.

I am familiar with the various photographic differences between flash and continuous light -- I take pictures for a living. A couple other advantages of flash -- no squinting, and mixing flash with ambient (dragging the shutter). Also you can front or rear curtain sync while dragging the shutter to make your motion blur be before or after the moving subject. Squinting is a big deal when your light source is on-camera as you typically want your subject looking directly into the lens. Anyways, the question was more out of curiosity than anything -- I don't really know if an LED, along with the appropriate circuitry, is capable of flashing within the necessary requirements of a photograph.

I know color temp will be weird -- but what's normal, anyways? Even a regular camera flash is daylight balanced which does more harm than good when shooting in tungsten or flourescent light (in which case I use gels on the flash to match ambient color temp, and if I get enough output from this light, I'll use gels on it, too). I'm expecting a funky color temp, but I'm also expecting it to be consistent so I can compensate for it with RAW adjustments.

Thanks again! I'll wait around for some more action on this thread before I commit to trying my original idea -- but if and when I do build this thing, I will be shooting pics of the process for a DIY article on another site.

Regards,

Scott

 

[Mr Happy]

Just to mention though (to mdocod), that if you are using cheap off-the-shelf flashlights you don't need 350 mA driver boards or anything so complicated. You just need a rectified DC supply of about 4 to 5 V and enough amps. Scott mentioned a supply used for anodizing that sounds like it could be just the thing.

One thing that hardware hackers might have done in the old days is to custom wind a transformer secondary to the right specs on a 50 VA core and then rectify the output. Heavy, but robust and reliable.

Also (to Scott), I don't think you could make LEDs flash like a photographic flash. Those flashes have special xenon tubes in them. If you tried to put the same amount of electricity through LEDs they would probably blow up. Have you thought about a bunch of cheap slave flashes? (Obviously you have, but why isn't it a good option?)

 

[smack2000]

mdocod: I just read your post, and I'll need to chew on it a bit before I respond (and it's time to go home), but thanks for all the info!

Scott

 

[mdocod]

Just to mention though (to mdocod), that if you are using cheap off-the-shelf flashlights you don't need 350 mA driver boards or anything so complicated. You just need a rectified DC supply of about 4 to 5 V and enough amps. smack2000 mentioned a supply used for anodizing that sounds like it could be just the thing.

I guess my concern about this type of setup is that as you turn lights on the ring off the potential voltage available to the still-running LEDs may increase. I'm sure you are aware of how sensitive LEDs are to changed in input voltage, my gut tells me that to be able to reliably turn various lights in the ring on or off, you would need a more "controlled" power source. One that can "react" to the load and adjust on the fly. Like a Bench power supply.. or simply having the regulation at each LED or group of LEDs...

 

[Mr Happy]

I guess my concern about this type of setup is that as you turn lights on the ring off the potential voltage available to the still-running LEDs may increase. I'm sure you are aware of how sensitive LEDs are to changed in input voltage, my gut tells me that to be able to reliably turn various lights in the ring on or off, you would need a more "controlled" power source.

Yes, for bare LEDs this would be true. But Scott in the first post is talking about flashlights that run off 3 AAA cells. Now three brand new alkalines will deliver about 3 x 1.6 V = 4.8 V. When nearly exhausted they will deliver 3 x 1.0 V = 3.0 V (say). So the flashlight in question must contain sufficient regulation to cope with supply voltages over quite a wide range. Any supply that provides a voltage somewhere in that range should be OK I think. Maybe the LEDs are directly driven, but then they will be of a more robust kind that can cope with some latitude in the actual voltage, as long as it is not too high.

The supplies we are talking about in this thread will be (should be) fixed or regulated DC supplies. They will have a designed voltage of about 4.5 V at no load. When more lights are turned on the voltage will go down a little, but with a 10 A capable supply it will not go down a lot.

 

[mdocod]

the 3xAAA cheapo china lights have no regulation to speak of, they rely on the sag of the AAAs and internal resistance, some of them have a small value resistor in series and that's it. They start off very bright and overdriven and quickly diminish in output. I own a few.... here's some output graphs from similar lights:

http://www.flashlightreviews.com/reviews/amart_32led.htm
http://www.flashlightreviews.com/reviews/amart_28led.htm
http://www.flashlightreviews.com/reviews/i2tl_329.htm
http://www.flashlightreviews.com/reviews/muyan_19led.htm

and here's one very very similar to the OPs original plan (21LED 3xAAA)
http://www.flashlightreviews.com/reviews/i2tl_s21.htm
and another
http://www.flashlightreviews.com/reviews/muyan_21led.htm

As you can see these lights are not going to be able to control what they're getting power wise. On my 28LED light that's 3xAAA powered I have measured as high as 1.2A on fresh cells through the tailcap. that's ~43mA per LED, and these are probably not designed to handle anything better than 20mA. This is why back when these types of lights were more popular (before luxeon/cree/SSC etc because mainstream) we had tons of reports of premature failure on LEDs in these types of lights. Once a few burned out, the rest started going fast as the remaining LEDs would be more and more overdriven as more power was made available for them.


[edit in]
the other issue is that these cheap LEDs are really only about 30 lumens per watt (give or take)... I just don't see it being worth it. If you take and drive those 21LED lights at a level that could be used reliably long-term, you would only be getting about 30-40 lumens per flashlight, or about 700 combined possible lumens, about the equivalent of a typical household 50W bulb. For photography purposes I would want to have a few thousand lumens at my disposal if I was going to go through all this trouble. A ring of Solux 35W MR16 halogens would be my personal choice. 10,000+ lumens.

 

[Mr Happy]

I just don't see it being worth it.

Yes, regarding that last edit, I was kind of thinking the same thing. If you have 4.5 V at 10 A, that's only 45 W. Even with a really good luminous efficiency that's a feeble amount of light for photographic purposes. I suspect even a simple flash on a compact camera has an output of 1 kW to 10 kW or more! (That's 100,000's of lux)

Here's an interesting paper on the subject of xenon flash tubes and LEDs with some advice on how to use LEDs for photography. But it doesn't look cheap or simple!
http://www.cap-xx.com/news/CAP-XX_LED_Flash_and_Xenon_Comparison.pdf (PDF warning)

(Interesting flashlight review links by the way. I've always passed over that kind of light before thinking that for $20 - $30 I wouldn't waste my money. But if I see any of them in Big Lots or someplace for about $5 I might pick one up.)

 

[Illum]

Your 4.5 V 10 A supply is going to be tricky.

is it possible to do series parallel configurations with 4V inputs? parallel wire batches of three in series and connect the whole smash on a battery to solve the "nonstandard wall wort" issue?

I'm pretty sure you could find 12V 500ma-7amp battery chargers fairly easily

 

[smack2000]

Well alright -- I think you guys have successully talked me out of pursuing this.

The more I think about it, the more I think I want this to flash as opposed to being a continuous light. I think if I build a continuous light with the output I want, I'll be blinding my subject, not to mention closing their retinas which is always looks bad. The cap-xx pdf is what really talked me out of it -- reminding me of the fact that my camera sensor cares about cumulative light output over the duration of the flash or exposure (in the case of a continuous light), and most of what I will do with a ring flash (high-fashion style portraiture) requires me to freeze my subject with alot of sharpness -- not something you can do with an exposure time of less than 1/60 -- probably a requirement of a continuous light given the lower output. That's not to say the light itself wouldn't be useful, just not useful enough for what I want to do.

Thanks to all for an enlightening discussion -- now on to my next idea

The main downfall of the other DIY ring flashes is the light distribution around the ring -- most designs try to bounce the light from a single high power flash all around the interior of the ring by making it highly reflective and very big. So I turned my searching towards 'cheapo flashes' instead of cheapo LED lights. Mr Happy gave me the idea -- I've only thought of using an expensive flash that I already own (i.e. a Canon 580 EXII which is over $400) as the other DIY'ers have, but cheap small flashes -- I could use a few of them if I could find them small and cheap enough...

Here's what I found:

Holga 120 Mini Flash ($9):
http://cgi.ebay.com/100-BRAND-NEW-H...ryZ48549QQssPageNameZWDVWQQrdZ1QQcmdZViewItem

Samyang Mini Slave Flash($10)
http://cgi.ebay.com/New-Samyang-Min...ryZ30086QQssPageNameZWDVWQQrdZ1QQcmdZViewItem

Now I'd say my new focus is to do this as cheaply as possible, as there are already some cheap ring flashes made with equivalent electronics in the $60-$150 range (however most are designed for macro photography and small-barrel lenses).

I'm thinking of using 1 of the Holga's and 3 of the Samyang's in a similar PVC/Acrylic enclosure.

I could take the holga and connect it to the camera's sync port with a $10 Sync to Hotshoe adapter. The holga would then trigger the Samyang's via their optical slave. PVC cost is negligible, and maybe $10 for a sheet of white acrylic. Again, I know I could just buy a ring flash (and I have rented the Canon ring flash for a couple of shoots) but this is more of a challenge to produce and effective DIY ring flash for very little cash.

I am still interested in applying LED technology to photography, but I don't think this is the project for it... Maybe an LED softbox...

Thanks all,

Scott

 

[TorchBoy]

... not to mention closing their retinas which is always looks bad.

It certainly sounds bad. And painful, whatever it might be.

New thread for the reconceived project?