Upgrade: Prodmod LED Camera Light - Constant Brightness - Fixed Voltage - Flexible Battery
By using an $11 DC-DC boost circuit in the Prodmod LED Camera Light I am now able to power it with one, two, or three batteries and can use either rechargeable or alkaline and still maintain a fixed LED brightness. The brightness of the LEDs remains unchanged until the very last moment when the total battery voltage drops below 1V. The flexibility of the DC-DC boost circuit allows you to use either one or two batteries if I wish leaving room for other items in the case. It also allows you to position your mount stud somewhere else if it better suits your camera.
I tested a 5V circuit and a 3.3V and compiled the results. There is an effect on battery life but its not as bad as you might think.
The Background
If you have already made the original Prodmod DIY LED Camera Light you may have noticed that the light will start at full brightness and begin to slowly dim as the batteries are depleted. This is because the output voltage is not regulated. As the battery voltage reduces so does the current
through your LEDs. Not only does it emit less light, the color temperature can change a bit as well as the current drops.
To regulate a voltage you can typically use something like a LM317 voltage regulator. But you’ll need to supply it with more voltage than you want it to regulate. So if you want 3.6V you might need to give it 5 or 6V. Well, I don’t want to add more batteries to my light just to get a regulated supply. In fact I would rather use fewer batteries.
This is what a DC Boost circuit is for. It converts a low voltage high current supply like a battery and boosts it to a higher voltage yet lower current output. It consumes current from the batteries to get the voltage boost. So even though you output current is low, you input current from the battery remains high. This of it this way: Power=Voltage x Current. Assume the battery has a fixed amount of power, if you increase the voltage in this equation the current must drop - this is the output. But if the current is increased, the voltage must drop - this is the input. There is a calculation given by the Vpack manufacturer that helps explain this relationship. You can download the PDF.
The DC-Boost PCB
So for my first experiment I chose to use the VPack5.0V_AAA_1 from Bodhilabs. These guys are great, they are only $10.95 pre-assembled and free shipping! The circuit is mounted on a small PCB which is glued to a battery holder. It is the smallest one they have and its limited to 100mA output which is fine for this light.
Relocate the mount screw
I removed the PCB from battery holder and wired it up to my original 3LED light. Since I don’t need all three batteries anymore I removed one which made room for the PCB, and it also made it possible to move my mount screw. This is great if the upper right hand corner of the original mount stud was inconvenient.
Moving the screw to the second battery chamber worked great for my Lumix camera because now the LEDs are positions further out under the lens and there is less chance of casting a shadow.
After I removed the LEDs I went ahead and drilled two new holes in the second battery chamber. One for my Panasonic Lumix off to the side and another for my Nikon near the center. I then relocated the red positive terminal so I now have a circuit for only 2 batteries. you only need one battery, but since we have the space here I settled on using two.
Choose NEW Resistor
Make sure you CHANGE YOUR RESISTOR. You now have a 5V supply where you previously had only 4V or 3.6V. So you’ll need a higher value resistor to limit the LED current. I used 23ohms for my test with only 3LEDs because it was handy and I didn’t mind running the LEDs hot. But you should run your calculations again with 5V supply. This time you are actually sure that number wont change. So 5V-3.3V = 1.7V. And say you want 20mA per LED and have 3 LEDs. That’s a total current of 3×20=0.060A.
So 1.7V/0.060A = 28 ohms.
If you are using 4LEDs the result would be
1.7V/.080 = 22 ohms
Wire it up
You can go ahead and wire up the DC boost as shown - I rearranged direction of the LED leads because I though it would be easier, but you can put them the other way if you want. And you can surely fit another LED to make 4 total. I don’t have a photo of how the wires connect to the PCB. If you need help on that just add a comment below and I’ll add to the post. By the way you don’t have to put the PCB above the screw head like I did, you could put the PCB above the LED chamber, just make sure you protect the PCB from the exposed LED leads and resistor with some electrical tape.
To The Test
Ok, so now we have a steady output voltage, and best of all we get to use only one or two batteries, but how long will it last?I ran tests with one and two batteries for both the 5V and 3.3V version of the Vpack PCB. The batteries were rated for 1000mAh.
Here are the results.
5V Vpack
one AAA - 3LED - 22.4 ohm - 80mA = total run time 50 minutes continuous
two AAA - 3LED - 22.4ohm - 80mA = total run time 4 Hours continuous
3.3V Vpack
one AAA - 4LED - 4 ohm - 80mA = total run time 2.8 Hours continuous
two AAA - 4LED - 4 ohm - 80mA = total run time 6.8 Hours continuous!
Clearly the best choice here is the 3.3V Vpack. The reason it lasted longer is because it requires more energy for the Vpack to boost to 5V than it does to boost to 3.3V. Plus with a 5V supply your resistor (28 or 22 ohms) has to waste a bunch of the extra energy as heat. You’ve basically wasted energy twice.
**The hours above are worst case. The actual battery life will be longer than what is shown since the light is ower is turned on and off during normal use.
Conclusion - It’s Not Perfect
With every advantages there has to be a few disadvantages, right? By using a DC boost circuit we gain a lot of flexibility with our input voltage battery type and voltage. We also gain a constant output voltage to help maintain constant brightness. But its not perfect. As the LEDs heat up their voltage drop changes slightly, and as the resistor heats up it too changes slightly. So even though your voltage is constant your current can still fluctuate. Lucky for us these parts do not get too hot so they wont vary all that much but it should be said that current fluctuations are still possible.
I am sorry I couldn’t show you a graph of output current over time, my low current section of my multimeter isn’t working right.
Next Upgrade
Add a dimming feature to your LED light. - coming soon!
Receive automatic updates - Don’t miss a thing!
Do you want to hear about new updates as they are posted? You should subscribe to my free RSS feed or have my posts delivered by email. I am now setup on FeedBurner. See below. Try it it’s free.
Sweet mod. Excellent, fun DIY project.
Check out http://www.youtube.com/watch?v=gTAqGKt64WM for some DIY circuitry.
Also, check out http://ledsupply.com/20mm.php for lenses & reflectors.
http://electronicdesign.com/Articles/Index.cfm?AD=1&ArticleID=1823 DIY charging circuit.
You could sell this at the MAKE store (http://store.makezine.com/).
At least submit this project at Makezine for some traffic.
P.s. Thanks for the heads up on Bodhilabs.
Many props, you da man!
Thanks for the feedback and the links. I definitely check make every now and then. I like the joule thief idea but I am worried it’s too bulky with that big toroid. You can make your own DC boost using a few parts that are smaller. You can find directions on ladyada site. She made the mintyboost. That one is still too big for me and personally I didn’t want to mess with soldering a circuit so I bought the bodhilabs one for $11. Its the smallest I found and it fits into the case nicely since its a rectangle.
I did submit my original article to makezine.com and they did publish a link. I received a great response from it. Maybe I will try to sell through their site too. Thanks for the charging circuit, i was looking for something like that.
[…] Video Test | Photo Sample | Battery Test | DIY
[…] Video Test | Photo Sample | Battery Test Upgrade
I’d like to see a project for a still camera, like my Minox B. Perhaps I’ll do it. Using the VPACK3.3V_AAA_1 (http://www.bodhilabs.com/vpack33aaa1.html) and five 560MV7C LEDs (http://www.buy-leds-online.com/560mw7c.html) in parallel, is there a need for a resistor? Seems the power/voltage/amperage specifications are a perfect match. Light output would be about 12 lumens, less than 1% that of a 100 watt light bulb (http://led.linear1.org/lumen.wiz) of 1700 lumens. However, considering the light bulb is radiating in all directions and the LED at only 60 deg., a comparison of 2%, it’s like a 2 watt spotlight. Does any of this make sense?
Hmm. I guess you are looking at the Vpack output limit of 100mA? The question is: Does the Vpack limit the output current to 100mA? or Is the 100mA current limit more like a recommended limit to keep the Vpack from overheating. My guess is that it will not simply stop at 100mA, if you try to pull more than 100mA the Vpack will try to provide it and burn out. Remember the Vpack is not a current limiting device- so you can pull whatever you want from it. Also, even though you put 5 LEDs in parralel the current isn’t split up evenly unless they are all the same Voltage Drop (Vf). And their current draw might increase over time as the temperature rises. But anyway, your question is more about the Vpack. So I will ask the experts and get back to you.
I checked with bodhilabs, and the short answer is the Vpack does NOT limit the current output. So you will draw more than 100mA if you are not using a resistor. That value was more to represent what a typical AAA battery can supply after energy is used to boost up the voltage.
For good measure I tested a Vpack using only a 27ohm resistor and no LEDs and it was drawing 150mA.
The other issue is the voltage drop Vf of the LEDs. You are assuming they are all 3.3V but they are probably not.
In general powering LEDs without a resistor is a bad idea. So my advice is Don’t Do it
If I understand LEDs correctly, according to the data sheet current draw is dependent on the supplied voltage. Is that how you’re limiting the current? So, would it be more efficient to modify the resistor on the PCB
(http://www.bodhilabs.com/vpackbare.html)
to achieve 3.3v? I think you hinted at this under your “To The Test” heading. Perhaps it’d be wise to reduce the voltage to 3.2v to eliminate the heating problem. At 3.2v the current is 20mA according to the 560MW7C data sheet; this would also keep the drain of 5 LEDs at 100mA, well within the limit of the VPack. However, I’d like as much light as possible, so 3.3v may be the best compromise.
Another question: If I used the flash sync of the Minox B camera to control power to the VPack, kept the voltage at 3.3, and set the shutter speed at 1/20 second, would the high current drain not be a problem? Of course this begs the question, will the camera handle the power at the flash sync? Or, I could use a switch to turn on the LEDs and take a camera meter reading, set the shutter speed, take the shot, and turn off the LEDs. Thus, useful for both video and stills.
I realize my purposes are different than yours. You want constant light for video and I want a tiny flash for low-light photos (and not to blind or annoy my subject).
I think you are looking at the Forward Voltage vs. Current graph, right? 560MW7C.pdf
This is a guideline and each LED tends to vary in Forward Voltage- not only out of the box, but also
over time as its temperature changes. You can see that in the next graph of current vs temperature.
In practice an LED might start with Vf=3.6 for example, but after it heats up its forward voltage actually increases which then means it draws less current and gets dimmer.
Let’s say you did drive 4LEDs in parralel with only 3V. For each LED to acheive max brightness you would have to hand pick LEDs with exactly 3V Voltage drop. Then maybe it will work. - again check temperature effects over time.
If the LEDs are *not* the same Vf, the least Vf will be brightest, and if you dont reduce your output to that level you’ll damage that bright LED while the others are less bright.
This is where resistors are handy. To account for variations in Vf among a group of LEDs in parrallel you would select one resister per LED that is appropriate for each Vf. (my original circuit is not ideal because it assumes all Vf are equal)
For the record I did try driving all 4 LEDS in parrallel with the 3.3Vpack and no resistor. I don’t remember how much current they drew
but it was well over 100mA, So I quickly turned it off. This is probably becuase the lowest Vf of the bunch was 3V.
Imagine the equation: (assume 1 ohm resistor since I can’t divide by zero.) 3.3v-3=0.3v, 0.3v/1 ohm= 0.3A = 300mA!! So theoretically the 4 LEDS would be free to draw 75mA each. Not good.
I found this PDF
It does a really good job of explaining how to drive LEDs- specifically read the section on “Direct Control of Current”.
It states that you find the proper voltage for the LED using the V-I curve, to control the voltage across the LED it uses a
resistor. It supports your comment about how the power dissipated in the resistor is wasted energy.
It sounds like your goal is high efficiency and maximum brightness. If you read on in the PDF you’ll see that it concludes the best way to drive LEDs is with a constant current source. This does NOT require a resistor.
I think this is what you should focus on for your project.
Personally using a Vpack to stablize the output voltage and a resistor was good enough for now, but I plan to explore constant current in this blog in the next few weeks. You should go ahead and work on it now (dont wait for me;)
Now you also mentioned modifying the resistor on the Vpack PCB- yes this is the most efficient way to control voltage output and in turn current.
It has been specifically recommended to me by bodhilabs as a way to dim the LEDs so I encourage you to try it.
If you replace the Vpack resistor with a potentiometer you’ll end up creating a variable voltage power supply. (a dimmer I haven’t written about yet)
Then, to better understand how current relates to forward voltage I think the best thing is for you to use a multimeter to measure current,
and experiment driving one LED with 3v. Then slowly increase the voltage to 3.6V until you find what voltage that particular LED draws 20mA. Then hold the voltage steady and keep measuring current for a few minutes.
Does the current change? up or down? - I’d like to try this as well.
In the end it still sounds like you need a constant current LED driver.
For the flash sync idea.
I am not exactly sure but I can share this:
I was told by bodhilabs there is a way to pulse (switch on and off) the Vpack by lifting Pin 2 of the IC named “EN”
“Pin 2 of the IC, named EN, is currently tied to VDD which will permanently enable the circuit.”
If you lift this pin up you can turn the Vpack off and on.
So that might be one way to have the camera control when the Vpack is on. But I am not sure how to make it go from bright to really bright using that input.
Please try it out and let me know how if works. What I do know is that you should not drive the LED over the current limit for more than x seconds. You can find this out in the spec sheet under “pulse Current” which is 0.1 millisecond for this LED.
There are circuits specifically design to do this for Cell phone cameras. If you google it you’d probably find a LED flash circuit for your use.
check out this constant current circuit.
bad news is they have a 1000pc minimumand out of stock
but good reference
http://www.fairchildsemi.com/whats_new/led_drivers_fan5607.html
wait Digikey has it
http://parts.digikey.com/1/parts/571192-ic-driver-led-dc-dc-conv-16mlp-fan5607hmpx.html
only $1.26!!
i found a very good explaination you all should read at ladyada’s website
http://www.ladyada.net/learn/proj1/ledintro.html
it does a very good job of explaining the difference between a change in voltage through an LED alone as it relates to current and compares that to usnig a resistor in series.
Basically if you look at your V-I graph a 10% change in voltage results in a 100% change in current. But if you use an inline resistor, that 10% change in supplied voltage might only change the current by 5%. So it’s less sensitive to voltage changes. I know the Vpack is regulated, but this happens when Vf changes due to temperature as well.
So, the series resistor “tempers” the steep V/I curve of LEDs. Cool.
The “LED-driver considerations” (http://focus.ti.com/lit/an/slyt084/slyt084.pdf) and ladyada articles are educational.
I’ve ordered parts and will post progress (may be weeks before my next post).
MICRON EMS is a dependable Electronic manufacturing services company in bangalore. We are specializing in pcb assembly manufacturer, smt assembly manufacturer, smd assembly manufacturer, thro hole assembly, metal core pcb assembly, rohs pcb assembly, pcb jobwork, bga assembly, reflow soldering etc. Our goal is to consistently provide to our customers with products and services that meet or exceed mutually agreed upon specifications.