[Project] LED Phone Case

[kon.bick]

Hello everybody,

I'm new to this site and this forum seems to be perfect to introduce my project and get some guidance from you guys. Special thanks to Giorgos!

My aim is to make a LED Phone Case like you can see in the pictures below. My final goal is to get a slow breathing of all of the almost 30 LEDs as a nice effekt. Constraints are regarding the size since the Phone Case shouldn't be too thick.



Parts I want to use:

-The biggest problem was the power supply. It had to be small like a coin cell but at the same time have a big current supply to power all the LEDs. Thats why I ordered a smartphone battery for the BLU Dash 5.0, as it comes flat (4mm) and has up to 2000mAh and a current of up to 1A (that's a suggestion from reading a paper about smartphone batteries).
-For the LED's I thought of 1206 SMD LEDs.
-The circuit should be built with a 555 timer IC. However, I need to use a low power SMD 555 that works with 3.7V power supply that I get from the battery.
-SMD resistors and BJTs.

The circuit I want to use:



This is a standard 555 timer circuit that I want to modify for my project. But first I want to make sure that I understood the circuit correctly. Here is what I think how it works. Please correct me if I am wrong!

1. As initial voltage is low at pin 2, the trigger will turn on the output (pin 3) and current will flow. At first the current will flow into the capacity as its resistance is lower than the BJT's. However, a little later current is also flowing into the BJT and the current will increase as the capcity's resistance will increase gradually. This gives us a gradual breathing till the LED is bright.
1. As the voltage across the capacity is getting greater than 6 volts, the treshold will fire and turn off the output pin. Current is now flowing from the capacity into the 555 and into ground. This phase might be slightly shorter since only the capacity current has to flow into ground as to decharge the capacity.
3. As the voltage across the capcity is getting less than 3 volts, again it begins to turn output on and so on.

The modifications I want to make:



The 555 timer I have to use will have a lower output/sink current (50mA/10mA). Thus, I have to use BJTs to get enough current through my LEDs which need approx. 3.2V and 20mA. But here we come to my questions: Is it possible to power 28 LEDs with the battery and the 555 timer and how could I manage to do that. Below is my first draft but I need to get a better understanding of how to power so many LEDs and your guys' advice. How can I calculate the current properly e.g. what outvoltage will the ouput pin 3 give me?
Any advice or comment on my project? Any help is appreciated!

Thanks,
Kon

[kam]

Hello kon,
well, bad news. your circuit will not work as you would expect. The 555 output (even for the LMC555) will  be less than Vcc and each transistor connected as a voltage follower will oppose a 0.6 to 0.7 voltage drop. So the first LEDs will receive 2.7V, the 2nd leds 2V etc.

I'd rather recommend to go PWM for this application. It will be easier to design (especially with a microcontroller) and you will have much much higher efficiency (and battery life).

If you are not into microcontollers, then you have to make a better interface between the 555 and the LEDs, not with just one transistor. I suggest you use an OP-AMP interface (google, it is simple circuit with one opamp and no external components). THis way you can provide enough current for the LEDs.

As for the transistors, one is enough. At 3.7V the emitter of the transistor will have some 3.1V. Each led must have a resistor of about 22 ohms. Then you connect all the led-resistor networks in parallel at the emitter of the single transistor. Measure the total current so that it does not exceed the maximum limits of the transistor. If so, either use 2 transistors or use a bigger led resistor.

[kon.bick]

Hello kam, thanks for your quick answer!

Right, the 0.7V drop, that won't work it all. Was just a first try though. I would do it with a microcontroller but space is really limited to have it inside a phone case.

What you mentioned last makes perfect sense to me. One transistor would be enough (if he can handle ~0.5 or 0.6A) and hence I could have all my LEDs parallel with a tiny resistor for each. Then the circuit would work already?

I will look into OP-AMP, as you mentioned. But could you elaborate more on what you mean with: " then you have to make a better interface between the 555 and the LEDs, not with just one transistor". Here you say, more than one transistor is needed?

Thanks,
Kon

[kam]

First, you should keep the leds underpowered. You have a lot of them. At 20ma your mini will become a flash light. Maybe 10ma per led is ok. For your reference, I designed once an illuminated doorbell for the US big boxes. It had 8 leds sharing some 40ma (5ma each led) and I had to add a dimming circuit because it was too bright for a night ight.

As for the transistor interface. If you directly interface a transistor to the rd network, the base current will interfere with the rc network. Either you have to make low impedance rc (which requires a huge capacitor) or you make a high impedance interface with opamp so that no current is drawn from the rc. But if you plan to go digital with microcontroller then do not worry about matching impedancies. The output of the uc with proper resistor to the transistor. Pwm driven ...

[kon.bick]

Thanks for the answer again! I updated the circuit schematic in the first post and understand now why I need to look into opamps. Because the base current will be way too weak and making it stronger by increasing the capcity, like you said, will result in my LEDs not breathing evenly. I will have to match impedances.

Thanks for your hint regarding the LED current. I will aim for a peak of 10ma / LED making it a total of 280mA.

[kam]

You're on the right track! 

[kon.bick]

Hey kam, learning more and more about opamps I found out how ideal this is for this circuit...the high input impedance is perfect for not interfering with the current going into the capacity. Moreover, it gives me the necessary gain and I can get the necessary current in the rd part.

Here is what my circuit looks like now.



I have a couple of questions again:

1) To further lower the current in the diodes (currently in red) I wanted to use an inverting closed loop opamp. However, the circuit didn't work at all when I implemented it. Now I have a gain that is equal to 1+(Rr/Re) and tried to make it even weaker...but I think inverting closed loop op amp is better to do that since my gain can be below 1...should it work?

2) I'm about to buy parts. For the DIS and CV pin I used 1uF to save the from any static (this is how it's normally done, isn't it?). What value should these capacities have when I build my circuit?
Right now I used LT1813 as opamp and I think it does the job I need it to do. I was a little limited in parts when I simulated this with LTSpice of course. But should I use another opamp?
Diodes and other capcities are settled and the 88Ohm resistors do a good job.

Thanks for your feedback once again!

Best,
Kon

[kam]

Here is a simpler idea to reduce the current. You can control the voltage with a zener. So the output of the 1:1 opamp goes through a resistor and then a zener diode (say 5V) keeps the voltage constant at the base of the transistor. This turns the transistor into a voltage regulator with 0.7V lower voltage at the emitter. Now you have 4.3V at the emitter. You can choose the proper resistors to drive each LED color. It is easy now since you have stable 4.3V

2) The capacitor will not save from the static. It will save from noise. I usually use 3 capacitors, one 100uF, one 1uG and one 0.1uF in parallel. Sometimes further decoupling is required but one 10uF capacitor for you is ok. As for the opamp, you requirements are quite simple so get a single supply rail to rail opamp that can handle the voltage that you want (15V would be ok).

[kon.bick]

Hey kam,

thanks for your feedback once again. I actually am facing a very different problem. As I plan to use just 3.7V as VCC I cant seem to get more than 2.8V ish output from the OP amp. Minus the V_BE drop on the bipolar transistor I just get around 2V over the diode which is not enough at all. I think I have to redesign everything...if it is even possible to get my desired result...?

[kam]

is the green channel across the led? if so, then the voltage is correct, the led voltage drop

[kon.bick]

Hey kam, sorry that I didn't point out what traces belong to what points at the circuit.



As you can see in green is the Output from the OP Amp. The problem here is that due to the inner circuit of the OP Amp I would only get around 2.8V output maximum. That means, I can't get a good amplified signal..I also added traces for voltage over LED and current through LED. I already changed from white LED to green LED as they would only need 2V instead of 3 to 3.2V (white). However, still the current is not at 10mA max.



Another idea was to setup this circuit, at least the astable multivibrator with analog elements. That would mean I could get a signal as high as 3.7 to work with. However, R5 is problematic as it is needed to reduce the base current and make a slow charging of the capacitor 3 possible. But also voltage drops...I built this circuit with green LEDs too, but here the current is way too low @R5 = 100k, and still too low @R5=1k.

If you have an idea, let me know.

Thanks,
Kon