Voltage reducer schematic

Thread Starter

Lesliev

Joined Jan 13, 2021
19
Hello,
I am working on a project (Small portable table lamp) that uses a LIPO battery (3.7v-4.2v)to power an LED. The LED is a Cree XPg3 and requires 2.87v to give the lumen output I require. Currently I am using an Adafruit 4920 breakout to provide 3.3v and act as the driver for the LED. I then use a 1ohm 1W resistor to give me the required 2.87v. This was calculated from the Cree XPg3 datasheet. I am sure this is probably not the best way to set up the circuit but with limited electronics experience I used what was readily available and have built several prototypes all of which work pretty much as required. I recharge the LIPO via a USB powered charge base I have designed and containing a LIPO charging board that connects to the lamp base via detachable connections. So far all is working OK.

Now however I would like to go to the next step and build my own PCB to be incorporated into the lamp base that will produce the required 2.87v constant voltage from the LIPO recognising that when the LIPO voltage will vary over its charged lifetime and when it falls below a threshold it will stop operating until the LIPO is recharged. I have looked at the schematic for the Adafruit 4920 and that uses a TPS62827 IC and did think of using that schematic reworked to my needs to produce the result I am after. However two things I need to consider. The TPS62827 is not available anywhere I can find and the delivery time from Texas Instruments is 50-70 weeks. Secondly I am sure for my basic, simple purposes there is a less universal circuit than the very lovely Adafruit circuit and that will give me the 2.87 (or3.3v) in a simpler way and maybe using an IC like the 555 that is available and may work.

So to my question. Can anyone recommend a circuit/schematic I should look at with a view to using it or adapting it to my needs. Not wedded to the 555 but it is at least still readily available and from what I have read may well do what I need but I am sure there are many other IC that will do as good or better a job.

Sadly at this stage in my amateur learning process (I was an architect in my professional life and have no training at all in electronics) I am not capable of just starting with a datasheet and getting a working circuit and also I am sure this is a problem that has been solved elegantly countless times before by many much smarter than me.

Hope someone can help and advise. Just one final comment this lamp I am designing is not for commercial sale but when finally working will be gifts to my family but I benefit greatly because it is also a very engaging, engrossing exercise with huge rewards to my learning along the way. I am very much enjoying the journey.

Les
 

SamR

Joined Mar 19, 2019
3,617
I did not find the PDF for this specific LEDs electrical requirements but LEDs are current driven devices, not voltage. There should be a PDF listing the voltage and current specifications. Typically, the voltage can be within a range and there should be a max current rating. To keep the current below that maximum you use a current limiting resistor, not to drop the voltage but to limit the current. Depending on the LED and the brightness needed I like to run LEDs well below the max current, say ~50-75% of max current @ a voltage within its operating range for longevity. YMMV
 

Thread Starter

Lesliev

Joined Jan 13, 2021
19
Thanks Sam,
I am obviouly mistaken about the LED voltage. I thought as all LED have a forward voltage measurement I assumed it was the voltage that counts. Thank you for setting that straight. I have attached the datasheet I have for the LED I am using and hopefully that might help a little. I used the graphs on pages 17-19 to determine the required voltage. I have been working on the criteria of keeping the amperage to between 650mA and find currently I am running at around that current. The batery charge lasts about 2-2.5 hours but with some degradation of the brightness after about 1.5 hours.

Here are the calcs I used:
Calculations for 3.3v supply
3.3-2.87 = 0.43V
0.43V/650mA = ~0.66Ω Use 1ohm
0.43V*650mA = 0.41W+ 50% = ~0.6W
So a 1W resistor would cover it quite comfortably.

Regards
 

Attachments

Ian0

Joined Aug 7, 2020
3,783
Do not drive an LED with a constant voltage. The forward voltage drops as the temperature increases, that makes the current increase, it gets hotter, forward voltage drops further, current increases, Dead LED.
ALWAYS drive LEDs with constant current.
There is also a wide variation of forward voltage from part to part.
 

DickCappels

Joined Aug 21, 2008
7,997
While you have the resistor there you might want to measure the current. Illuminace is nearly a liner function of current if you can keep the junction cool, and this curve from the datasheet will give you an idea of how current and thus illuminance changes with the voltage across the LED for a fixed junction temperature.

1633689871431.png

The curve is not very sharp so maybe you can make do using only a resistor, but if I might make a suggestion, check the current (by measuring the voltage drop across the resistor) for each LED while using a freshly charged battery and adjust the value of the resistor to obtain the current you desire, and do this after the junction of LED gets up to a sable temperature. That way you just took care of the highest current. As the battery runs down there should be no noticeable color shift but the LED will run cooler. Note: A better way would be to drive the LED with a constant current source as @Ian0 suggests in the post above.

It is a good idea to atleast check the current on each LED because they are all different.

Let us know if you need any ideas for simple LED drivers or you can have a look with a search engine.

On the subject of current, Please take SamR's advice in post #2. A little derating goes a long, long way in improving how long your LEDs will last. The maximum current for the white LEDs is 2 amps but if you run with 2 amps that would be 6 watts, most of which is heat, so minimizing your heatsink size and weight is another reason to not run the current too high.

This is a good time to invest in a thermometer to check the temperature of the case, and from that and knowing the power dissipation (watts = volts x amps) you can calculate the temperature of the die.
 

crutschow

Joined Mar 14, 2008
28,221
The resistor is wasting power and is not the best way to control the LED current.
Use a switching regulator designed to drive an LED which has a current (not voltage) regulated output.
That will efficiently keep the LED current constant independent of battery and LED voltage, and will also maximize the battery life.
Here's a search to get you started.
 

Thread Starter

Lesliev

Joined Jan 13, 2021
19
Thank you all for those great responses. Apologies for late response i have been wildflower watching in our outback of Western Australia.
So on reading the posts and with my limited understanding am i correct or even close when i say the advice is to concentrate on the current which i would like to be about 700mA at 2.87v but measure both current and fv to be sure and then use an LED driver circuit to produce that current. My battery is a Lipo (3.7v) 2600mA battery. So input voltage 3.2-4.2v depending on charge. Develop a circuit that produces the current required using an led driver ic maybe like this one
https://au.rs-online.com/web/p/led-driver-ics/1709376 thereby removing the need for a resistor.
Am i even close in my understanding?
Thank you Dickcappels for your offer re simple LED drivers i would greatly appreciate any suggestions from anyone.
one of my problems is when trying to breadboard a circuit i cant see how to insert an ic where the package is smd. Is there any way to do that with a socket or something?
thanks again to everyone.
 

DickCappels

Joined Aug 21, 2008
7,997
1634962988771.png

I don't know whether this will help, but Texas Instruments makes adapter boards so you can evaluate surface mount parts more easily without making a PCB for the surface mount parts.

1634963076243.png
I discourage the use of these plastic "breadboards" but this will illustrate how useful these adapter boards can be.

The adapter board come with pins that can be plugged into an IC socket.
1634963177085.png

I am not sure of the correct name for this kit but the instructions sheet is titled "DIP Adapter AVM".

There are similar boards (for which you need to get your own pins) and kits (pins included) available on eBay and similar sites.
1634963511393.png
 

click_here

Joined Sep 22, 2020
446
View attachment 250886

I don't know whether this will help, but Texas Instruments makes adapter boards so you can evaluate surface mount parts more easily without making a PCB for the surface mount parts.

View attachment 250887
I discourage the use of these plastic "breadboards" but this will illustrate how useful these adapter boards can be.

The adapter board come with pins that can be plugged into an IC socket.
View attachment 250888

I am not sure of the correct name for this kit but the instructions sheet is titled "DIP Adapter AVM".

There are similar boards (for which you need to get your own pins) and kits (pins included) available on eBay and similar sites.
View attachment 250889
Here is where the "art of electronics" recommends :)
http://www.proto-advantage.com/store/
 

Thread Starter

Lesliev

Joined Jan 13, 2021
19
Thank you all again.
I will look in to the adaptor boards from the sources supplied and TI.
i also looked at the circuit available from Crutschow named. LED 2-Transistor Low-Drop Constant-Current Driver.
https://www.electro-tech-online.com/articles/led-2-transistor-low-drop-constant-current-driver.937/

That circuit seems like a likely option using 2 transsistors but as most of the reading i have done recommends using pwm for an led driver i am not sure this circuit although it looks very simple and clever to me Is quite what i need. Love to hear anyones thoughts.
 

Thread Starter

Lesliev

Joined Jan 13, 2021
19
The adaptor boards recommended are certainly an option for prototyping with smt ic however their use still requires some soldering and for very small ic and a shaky handed solderer (like me!) that can be a challenge. As no one has suggested it i am assuming there are, not yet anyway, adaptor boards with sockets suitable for smt ic. I Believe many of the ic available use a range of standard packages so maybe someone will design (or already has) and make available a dip type adaptor board with a socket to suit standard packaging that would accept the smd ic and suitable for prototyping. Or is that just wishful thinking??
 
Top