P7 LED driven by Buck driver and 4 x Li-ion 18650 cells in series

Hi all

I am working on a rather complicated lighting setup that is to produce 4000lumens - and be used underwater.. This part of the post is all about the lighting so I'll leave the underwater part for later.

I have researched plenty, discussed this with a friend who is an electrical engineer, tried and tested a number of things and come up with the following design. (fulllllll long description with all calcs and justifications can be posted but I thought I'd keep it simple for now)

Plan is 1 x P7 700 - 900 lumen LED powered by 4 x 18650 Li-ion 3000mAh cells in series + protected by a buck driver that outputs 3A and has various types of protection (heat, voltage etc)

I finally set up the circuit using : 4 x Cells > switch > buck driver > 1 x LED
This powered the LED on but it was not blindingly bright like others I have purchased are. I can still look directly at it which is odd

I then tried adding LED's to see the affect. No change in light. All 4 LED's had the same utput which makes sense as I added them in parrallel.

My ultimate plan is to power 4 or 5 P7's using a battery made up of 3 sub batteries in parrallel each is [4 x 18650 3.7v, 3000mAh in series] + [4 x 18650 3.7v, 3000mAh in series] + [4 x 18650 3.7v, 3000mAh in series]
So I'd have 14.8v output and 32Ah / 32000mAh
Given that current draw should be 3A per LED continuous, I should get a long burn time

My 2 questions are :
1) What is the first thing to measure / check?
2) The buck driver has 3 output modes, I cannot switch between them, how does one do this? Using the latching switch I have yields on > off > on > off etc. Could this be the reason the light power is weak ie it's stuck on the first setting - LOW ?

I can post LOTS of information but want to see if there's something glaringly wrong first. Thanks as always

Technical info :

LED's:
SSC P7 700 - 900 lumen LED's presoldered to PCB

Li-ion cells:
Each cell is a Surefire Li-Ion 3.7V, 3000mAh

Buck driver:
- Current regulated at 3.0A
- Input voltage 5.5V-15V, (no more than 12V would be the best option for protection of LED board according to accurate test)
- Buck Voltage Current Regulated circuit
- Gives 1+ hour current regulated (3.0A) output with four Eneloop AA
- Gives 3+ hours current regulated (3.0A) output with two protected KD 18650
-Gives 90%~95% efficient output
- Measures 19mm in diameter, and 8mm in height
- Multi-Mode: Low(5%)/High(100%)/Middle(35%)
- Designed for SSC P7 LED for maximum output 300lumen-900lumens(C bin)
- Prewired, ready to use
http://www.kaidomain.com/Product/Details.S005296

Very possible that it could be in low mode. The modes should switch by turning the light on, then off and back on suddenly. Usually in flashlights with the light on all that is needed is a half press of the switch untill off then release switch for on.

Thanks a lot for the advice. It's a latching switch that's very robust as its IP68 but I'll try switching it quickly and kayaking contact. Really appreciate the help

An update which I hope attracts more attention and help : See photos below of the components
The Driver is wrapped up in electrical tape as it's the most delicate part and the attaching wires are thin. Please excuse the untidy cobling together - it's a very first go to check that the circuit works!
The main item to please look at is the switch. It's a £30+ IP68 rated switch to go underwater which is why I needed this particular one. I'm hoping it will work with the driver ie allow it to switch through the modes (low > high > med)

Thanks


The full circuit


The switch


The batteries showing the spec


A close up of the 4 P7 LED's on a 6mm thich aluminium heatsink. The proper one will be 15mm thich and cut to size
The LED's don't get hot at all - as they're not outing full brihgness yet..
Wired in parrallel - ** should these be in series? **

Edited

Edited

Have you tested your batteries they could be some of these. You should consider getting some high quality batteries to try and run that at 3 amps. Most batteries at 1 to 2 amps will work fine. But when you get to 3 amps thats where the higher quality batteries shine. Just a suggestion. From the test I have seen on this particular battery they do not preform well at above 2 amps. Their internal resistance is to high. I see your planning on running a 4s3p that will better cope with 3amps. But your test bed is only 4s.

Jeez. Without having opened them they look identical. That's not such good news is it.
2 have already shown to have been duds and won't charge. They are reading 0.2v using my good quality meter to test them. I got them in batches of 10 for £12 odd so they were very cheap - but from reading that they are "cheap" and not inexpensive.
These lights ar efor my wife so another reason not to let them fail! I'll read up and get some of the top quality ones. Thanks a L O T 007, information like this is invaluable.

I'm still charging the other cells and once I have enough charged I'll do a test with 4s3p and let you know if that makes a difference + see if the switch is the issue or once thre's enough power it enables the driver to change modes properly. I'll be back in a couple days

Apol for the triple post before, I couldn't delete the extras

Getting Redilasts... Thank you. Again.

iv use these drivers befor and havent had this problem, most have worked fine, even with a single LED

that said, the spec you've listed says no more then 12V for protection of the board, 4 cells will be over 14V fully charged, so may have damaged the driver

a bit of maths though
3A into 4 P7's = 0.75A per P7,
a P7 is acutally 4 dies in parralel, so 0.75A into 4 = 0.185A or 185 mA per die, so not particulary bright or hot, but i dont think you would be able to look directly at it.

also the driver is a fairly efficient buck regulator it will be pulling about 12watts (3.7V x 3A)
for 4 series cells (14V+) this will give a current draw of less then 1A, which id have thought thoses cells would be able to do.

for max output the LED's will need to be in series, however that might not be within the drivers or the cells capability. (current draw will increase over 3A)

to change modes on the driver, you need to switch off then on again quickly, i.e. less then a second.

id sugget starting again with just the one P7, just to simplify things
also test the total battery voltage, just incase sumthing is wrong i.e bad cell

if you think it might be the cells, a 12Vcar battery should be within the drivers limits and might proove the cells to be bad.


check the basics first, it shouldnt be too complicted to sort out.

Thanks b-bassett, I'll post all of the details when I get to work
I'm busy re-jogging everything to start from scratch and prove each item.
Great idea on the car battery, I've got one fully charged ready to go.

I'm still learning so your technical info was greatly appreciated.

I think I need to start right at the beginning which is : what is the best way of powering 4 or 5 P7's..
I could ask loads of questions but let me do tests and post results so we're working with facts
Thanks again

I would recommend you use TaskLed driver H6Flex, you get thermal and voltage protection, multi-mode,...
I know it is more expensive but you get quality product,...

I'm a little confused by what the goal here is.
Placing 4 P7 LEDs in parallel and driving them with 3A total isn't much brighter than putting two P7s in parallel and driving them with 3A total.
The P7 is spec'd to 2.8A, so would handle 3A with proper heatsinking, so one p7 driven at 3A would be almost as bright as driving all 4 in parallel with the same 3A.
(An underwater light should have plenty of cooling available during regular use so I wouldn't hesitate to drive at 3A if you want max brightness.)

Also, with 4 batteries in series, your max voltage would be 4 * 4.2V = 16.8V. Well above the max voltage of 15V. I am not familiar with your driver but that probably isn't a good idea.
It probably isn't a good idea to run any cheap driver near its max, definitely not over.

So this is what I would do. Test your batts and find two good ones. Try powering one P7 with two batts. Use the switch to cycle the modes. The High should be plenty bright.

After that works, add a second P7. Add in parallel if you are trying to spread the power between multiple emitters. Add in series if you are trying to increase brightness. I don't know how much voltage overhead your driver needs so it may not do full power with two P7 in series and only two batts. If you are running P7s in series, add another 18650 batt to get more voltage overhead. Now your maximum input V is 4.2V*3=12.6V. It will sag significantly when in use (down to ~3.6V per cell).

Keep us updated.

FB

just wondering what type of illumination you are actually looking for? 4-5 P7's running at full whack will be a lot of light and heat, perhaps a little too much?

its usually best practice to match the number of LED's in series to the number of lithium cells in series, tends to make the electronics and calculations simpler.

that said, there isnt anything stopping you from using a different setup. just takes a little more planning and understanding of the circuits.

so... for 4-5 P7's running at 3A a piece. thats anywhere from 45 to 60 Watts, thats a lot!

your planned battery would be about 14.8V @9Ah so 118 Watts

best case scenario, you've got about 2 hours at full whack.

for these types of wattage, your going to need a good quality driver. the one youv'e got will definatly do 2 led's in series and i think it will do 3, but i wouldnt push it further than that.

another option would be to use a 7135 based linear driver, it would require a bit of electrickery to be able to use it with batteries over 6V, but is quite simple and reliable.

as a note, for 5 P7's in series you'll either need to use a boost driver, or to change your battery setup.




also, if you edit those extra posts of yours, you can delete all the pics and text, would make reading this much easier.

Tidra said:

I would recommend you use TaskLed driver H6Flex, you get thermal and voltage protection, multi-mode,...
I know it is more expensive but you get quality product,...

Click to expand...

+1 probably the best option in my opinion for 4-P7"s. Can handle 21v input with 93% efficiency.

Hi all. Here's an update for you. I have started right from the beginning which I hope will enable me to provide factual test results and therefore you to help me further

Background:
I wish to build a DIY high power underwater lighting solution for video and photography.
The driver to build them DIY is because I like building things and it's a project + the commercially available ones are way out of my price range!

I then researched as much as possible on CPF and the internet and came up wit hthe following requirements and solutions after ruling many others out..

Original requirements were:
1) Provide circa 3500 lumens of light at full power for circa 110 minutes (2 consecutive dives)
- I selected LED's given their higher light output to power and heat ratios over Halogen (worst) and HID (great but expensive). Provided by 3, 4 or 5 P7 LED's - depending on the output of each. I'm looking for a total lumen output and the best way of providing that

2) Provide adequate battery power to supply the LED's in 1)
-Still to be confirmed for hte final product but initially I bought the biggest Li-ion batteries I could find, 3000mAh 3.7V chinese ones. I am still testing these but am likely to stump up the cash for high quality ones, they're just 10x the price

3) Provide a waterproof switching mechanism
- Provided by an IP68 on>off>on latching switch. This is confirmed tested and working

4) Provide the power in a safe way given that the batteries of choice are Li-Ion batteries!
- I selected Li-ion batteries given their high power output to size / weight ratio. I have done tests with the batteries now and feel much more confident with them. Confirmed working

5) Provide a battery charger that charges the battery packs as quickly as is safely possible. Under 4 hours max
- Provided by a battery charger, one for each light that has 12V, 3A output. From the USA. Good quality.

6) A driver that would power the LED's with bult in protection
- Given that the P7 LED's require 3A to achieve max output (I think), I got a driver that looked to be of good, not best, quality. See specs. http://www.kaidomain.com/Product/Details.S005296
It is tested working. It now switches between power modes correctly.
Still to test : The number of LED's in series it can power. From the suggestions in this post, 2 is a good number, 3 maybe.
Given that I need up to 5 LED's (mainly because I have 5 available per light lead, 10 total ) I am looking at testing with the plan of getting a "TaskLed driver H6Flex" Tnx Tidra

To answer some of the questions from the above suggestions:
1) I've reverted to 3 cells in series so it'll be 3s4p (3 batteris in series x 4 packs of each)
This provides 11.1V at 3.7 each.. Max is still in safe territory (3 x 4.2 = 12.6). The driver supports up to 15.5V
2) I'll decide on batteries and driver once I've tested what I have thoroughly. Your suggestions are top of the list in terms of purchase.. Just need to make sure because I've already sunk all my $ in to these parts. I thought I'd figurerd it all out Couldn't post till I had something useful to post!
3) I understand the calculations about the A flowing to each die on the P7. Thanks for this!
4) Understand how series and parrallel will affect brightness. Explains why series yields max brightness
5) Got the driver switching by touching the batteyr wire on and off. The switch also worked but it wasn't as consistent as touching the wire. I think this works though as it definitely showed a difference between the brightness levels TaskLed driver H6Flex is the plan, just need to confirm what I have won't work first tho
6) used a car battery to do the above tests. What a great idea! I have an 110Ah 12V battery so plenty power there. Will use this to test the lights till I have a god configuration and then look at the batteries.
7) Got a refund from the battery manufacturer on eBay. Well, half my money back but that was reasonable to me as I will use these to test. Thanks for this useful info!! Redilast are the plan but they're very pricey I know you get what you pay for. But I need so many that it'll be $200+

Notes:
I have a bicycle LED powered light that I think is a single P7 that is SUPER BRIGHT. IT is definitely not a CREE as the LED is much larger - like a P7. There's No way I can look at it directly - yes it has a reflector which focuses the light but it's blinding.. It's powered by 4 Li-ion cells, I think in parrallel but maybe 2s2p. I have taken the battery pack apart but it was a few weeks back. I'll check again. I mention this as whatever the configuration is it works VERY well and I'd like to replicate it

FB - I tried doing what you suggested. I took a 12V car battery instead of 2 Li-ions and ran it with the switch > driver > 1 P7 and it works well but it's not blindingly bright as I expected. I did get the driver to change mode and I believe I cycled through the modes. There are 3 modes. - Multi-Mode: Low(5%)/High(100%)/Middle(35%)
Driver details are : http://www.kaidomain.com/Product/Details.S005296
It says 90-95% efficiency - assume it's 90!
Is that what you would expect?

I think it's looking like I need a Taskled driver H6Flex Any recommended supplier / link?

-> Back to basics, thanks all and I'll post factual test results of various options this weekend..

About the driver - This is from a user who has run one like mine successfully :

o.k. back to basic,...

lets assume that SSC P7 driven with 3A output 900Lum and at 1A output 300Lum,... I know it is slightly more at 3A, but this is just math,...

if you wire led's in series then all the led's get the same output driver current through them,... lets say that the output current of the driver is 3A and you have 3 led's = every led is driven by 3A = 3*900Lum= 2700Lum
if you wire led's in parallel, then the output current from the driver is divided by number of led's,... lets say that the output current of the driver is 3A and you have 3 led's = every led is driven by 1A = 3*300Lum= 900Lum
this is how it works,... of course in practice there a few minor deviations but that's the way it is,....

Ok I understand now, really! Eurica
It explainswhy LED's are wired in series to get max output
I think I'm not getting enough current through my driver which is why the LED's are dim..

I'll test using the car battery and let you know the results in a table with number of LED's, series / parrallel, brightness etc. The one thing I haven't tested is current A output of the driver with my meter. I need to find the manual as the leads need to be plugged in differently to test this.. Once I know the current that's flowing throughthe LED's I will have factual data and can determine what output level they're working at.

I wasn't aware of the directly proportional relationship between the lumen output and current - although it makes perfect sense now. Like most things electrical.. The tech specs only show the max lumens and I've found it hard to fill in the gaps. Have a great weekend, I'll be back asap..

HaloclineDiving said:

I wasn't aware of the directly proportional relationship between the lumen output and current - although it makes perfect sense now. Like most things electrical.. The tech specs only show the max lumens and I've found it hard to fill in the gaps. Have a great weekend, I'll be back asap..

Click to expand...

Side note: If you want to figure out the output more precisely you should be aware that there's a property in LEDs called droop where efficiency falls at higher currents, hence the relationship won't be entirely proportional. If you look up the datasheets you'll be able to get a graph showing output vs current showing this, e.g. for a P7 at 2.8 A the output is only ~1.75 times the output at 1.4 A, instead of double.