How do i design a desk lighting circuit using 8 superbright LEDs?

I don't quite understand. Is this for a school project or are you trying to produce a LED desklamp product for consumer sale (as these already exist).

More information is needed.

Hi

thanks for the interest.I am designing this as an examination project for my Year 11 Engineering students. Therefore i have developed this idea because it allows them to work with many different materials and processes whilst incorporating some electronic work, all elements of the examination requirements.

Although i was a mechanical engineer before i started teaching my electronic knowledge

is poor. I want to give the pupils a finished project that not only looks well made and attractive but actual works to a reasonable standard. I believe that professionally designed lighting circuits using LED's can give out sufficient light and last for a great number of hours. What they include in their circuits is what i wish to include in this project.

However, because it is a school examination project we are not allowed to use mains electricity in the design.

I hope that answers your question and makes clear what i am hoping to achieve.

thanks for your interest.

I am unsure which country you are in (I am assuming your students are not in their 11th year of an engineering double or triple PHD).

So, some basic questions to help taylor this project to the ability's of your students and your available electrical facilities.

1) How old are the students?

2) How much mechanical expertise do they have (for example, are you getting them to also machine and build the lamp housing, LED supports, etc)?

3) How much electrical expertise do they have (at the moment I am assuming they have none and this is their first foray into the electrical world)?

4) How much electrical work are the students expected to do (are they expected to assemble basic printed circuit board lighting control kits or just learn a bit about voltage, current, series and parallel circuits by using a soldering iron and doing a bit of wiring)?

No use us offering you a solution that is totally impractical after-all.

My students are in the UK. Year 11 pupils are between 15-16 years old.

They will be machining the connectors, moulding a Casing, fabricating a lamp cover, making the PCB's for the given circuit and the LED holder, Lamp base. The only part which they will be given are the manufacturing Drawings and an adjustable Goose neck arm.

They can produce reasonable PCB's and do have a basic knowledge of basic electronics eg transistors. potential dividers, various sensors, 555 timers etc

They will be expected to make the PCD and practically wire up the circuit into their Lamp that they will have made, and explain their choice of materials and components.

i have produced a simple electrical circuit using 1 switch, 8 LED'S 1 Resistor 27ohms 3 x 1.5v batterries and a PCB with the LED's in parallel.

This circuit is still alight after 50 hours. However i would like to get an efficient and long lasting circuit which makes the whole project worth the pupils keeping it and to use at home.

Although the circuit is still alight i require a circuit that give out an ambient light to be able to read or work with.

The use of batteries is due to health and safety regulations in schools in the UK schools.

Hi Jackotalltrades

I have just thought if you are really interested in this project, i could email you the design for the lamp and my circuit direct to. Then you may understand what i am trying to achieve given the limitations of the pupils, the variety of processes and materials that i have included. I could send as a PDF file attachment in a couple of days (time difference and weekend allowing)

If this is of interest, then email me direct to:

brianmorse@hotmail.co.uk

Please include your return email address, if not, no offence will be taken

Now days there are a lot of microchips which integrate most of the circuitry necessary to produce a LED driving circuit that is efficient, cheap and easy to build. The main reason to go for one of these circuits over just using a resistor in series with a string of LED's is that losses are reduced and the LED's can operate down to a lower voltage (hence extended battery life).

Below is an example simple design and info (there is also plenty more on the internet).

http://electronicdesign.com/Articles/Index.cfm?AD=1&ArticleID=8968

Unfortunately I don't have any more designs on me . Come on, someone else out there on CR4 must have a link to a basic circuit for this where the parts are readily available locally.

Hello brianmorseteacher

Just going by your Username, presumably you intend to use this as a class exercise, in which you are the class teacher.

<"I have produced a basic circuit using 8 LED's connected in parallel with a 27 ohm resistor.">

Using a resistor in parallel with the 8 LED's is a wasteful use of electrochemical energy from the 3 AAA 1.5V cells which you evidently have in series.

As said earlier, by jack of all trades these desk lamps already are on sale, very cheaply, in many discount stores.

A wonderful piece of apparatus for the battery makers they are, because the cells do not last very long in continual use, and users find there is a need for regular replacement of the cells.

Advise further please, with

Kind Regards....

Hi

thanks for your interest.

Please see my reply to "Jack of all trades" as it explains in more detail what and why i have chosen this project. One reply spoke of PSU's this is as clear as mud to me. Electronics is not my field and when replying consider me as an idiot, so simple pieces of information or diagrams would be more helpful.

Further to previous answers, a typical super-duper-bright LED is going to need at least 50mA (and will probably need a heatsink). 3 off good alkaline AAAs will last about 8hrs @ (8 x 50 = ) 400mA (running down from initial 1.5V to 0.9V each).

You'll need a bit of fancy power supply circuitry to get anywhere - connecting straight to the cells with a series resistor will have your users reaching for the mains light-switch within a couple of hours.

Plenty of folks here can design you a PSU (if they've got the time, & it may cost a few pints) - but I don't think you'll do better that about 3 or 4hrs at full brightness on one set of cells.

(Sorry, I've had a long day. My figures may be a bit off - but I don't think they're too far out).

Hi

thanks for the interest.

The circuit i have designed so far has been running for 18 hours continuously. However if you read my reply posted to \jack of all trades, you will see my restrictions in producing this project.

i hope that helps.

Can you give me the part no. (or supplier + order code) for the LEDs, please?

Hi JohnDG

If that is a serious question, then i will obtain them for school tomorrow and post them here.

Serious. Thanks.

Any design is going to need the information from the datasheet to get the best out of the system.

Hi John

I have given Jack of all trade more info if you care to my latest reply.

To answer you request:

Supplier: Rapid Electronics

Part: High Intensity white Led

Code N0. 55-1886

Currently my circuit is still alight after 50 hours, but whether it is still giving out enough ambient light to use for working by is unknown as i only have the circuit on a circuit board and not in a the lamp.

I hope this is useful to you.

Thanks, Brian(?),

For anyone else interested, this is as far as I've got in tracking down a datasheet: http://www.rapidonline.com/netalogue/specs/55-1886.pdf. It's a bit thin (one page), but it's a start.

From the sheet, it's rated 10,000mcd @20mA. V_F typ. = 3.4V @ I_F=20mA.

Your best to put all the LEDs in series so that they all get identical current.

Get a dual output 1 Watt DC/DC converter. (5V to ±15V but works down to 4.5V input)

http://www.farnell.com/datasheets/83605.pdf part code NMJ0515SC 5V to ±15V

Get your batteries from Farnell too: £30 for 100 16.5 Ah
http://uk.farnell.com/jsp/Electrical/Batteries,+Chargers+&+Holders/VARTA/4120210501/displayProduct.jsp?sku=5088124

A great source for LEDs is dotlight in Germany

http://www.dotlight.de/shop/index.php/cPath/149_155 95 euros for 500 white LEDs (they'll take a UK sterling credit card)

I have no idea why they sell them cheaper if you buy them with resistors suitable for driving individual LEDs from a 12V supply, just throw the resistors away.

These are 14 Candela 20 mA 3.2V LEDS. So if you put 9 in series with a 60 Ohm resistor you should draw 20 mA from your 30V supply. Better use a 62 Ohm resistor unless you want the kids to work out what to put in parallel with a 62 to make a 60.

Obviously you could put in a bigger resistor and reduce the number of LEDs to 8.

Oh, you should get about 80 hours out of the three D cells.

JohnDG, There are two types of super bright LEDs: these standard 20mA types and the 400 mA types designed for real lighting applications.

Hi Randall,

Thanks for your interest in my problem, i appreciate your advice.

however as i am an electronic moron most of it went over my head.

Does this converter increase the input voltage from 5v to an output voltage of 15v an therefore run for a longer time period?. I also thought that if i had 8 led,s series then each Led would dim in output brightness?

I did say i was an electronic novice. Can i be cheeky and ask if you could draw the circuit with the converter etc an email it direct to me? I am limited for space in the lamp base. I assume that i still use the 3 x 1.5v batteries (AAA type) or is there a different type of battery that you are suggesting i use?

my email address is "brianmorse@hotmail.co.uk"

If you are unable to help further thanks anyway and you have given me something to work on or to ask questions about.

A couple more thoughts.

As Sciesis pointed out: in your existing design you are sharing out about 40 mA (4.5 - 3.4 = 1.1V; I = 1.1/27 ) between 8 LEDs: that's 5 mA each so they're probably not nearly as bright as they could be.

The calculations I did above are for LEDs with a forward voltage drop of 3.2V (typical), not 3.4 V (which is what you've got).

Hi, Randall,

I'm a bit concerned that the spec. for the NMJ0515SC gives a minimum V_IN of 4.5V - the datasheet doesn't say what happens below that. As you know, the 'useful life' of an alkaline cell is taken down to 0.9V (end point voltage), which only gives 2.7V for the converter input. In fact, a typical discharge curve shows that they don't stay at 1.5V for very long at all.

(Unless you've already tried it, or data for low input voltage performance can be found) it would be a good idea to get hold of a converter & try it with a variable PSU (or 3 cells and some time!) & a dummy load. It might just pack up, or the efficiency may drop dramatically.

... one way around this (if it is a real problem!) would be to use a step-up converter. The LM2704, for instance, is not too expensive (£0.91 ea. for 25+ from RS). It'll give up to 20V out for 2.2V to 7V in. Not enough out for 8 LEDs in series, but 2 groups of 4 (with a bit of current-sharing resistance and V_OUT set to say 15V) would work. Makes the circuit messier tho', as it needs a choke & a couple of capacitors, plus a couple of resistors to set V_OUT.

John.

I'm going all out of sync. here because I wrote my reply to your last comment before I read your latest......

Yes that's a much better idea for a commercial product, but, I think Brian and his students are going to have their work cut out doing the rest of the design (mechanical included). I suppose he could make some suggestions and let them decide, or he could divide up the design task, and, let them find out what collaborating (read fighting) in the real world is like.

I have in the past used DIP headers to take a surface mounted chip and make it into a DIP....if nothing else is available....

Hi, Andy,

I thought about DIP headers - trouble is it takes some very fancy soldering to hook up an SO23-5 to one. But then, I guess the kids have a lot better eyesight (& steadier hands?) than I.

The other approach I've been looking at uses SO23 to SIL adapters - titchy little PCB's with the SO23 pads, & tracks running down to 0.1" spaced pins. Best I can do so far is about £3.75 ea (the converter is less than £1!) - tho' maybe Brian could beg for an educational discount - or maybe even sponsorship? I've no idea what the budget for the project is - but being a school in the UK, I suspect it's ²/₃ of bugger-all.

I'm also still looking for a device in a 'sensible' package that will do the job.

John.

Brian,

I hope you can see that John is much better at this than I am. However, he probably doesn't realise (because he is so experienced) that understanding the design above would be beyond the capabilities of your students, so, they would just end up copying it without understanding it. I still think that starting with a fixed voltage and simple application of Ohms law together with the forward voltage drop of the LEDs would be the best teaching exercise.

But make sure that you take notice of his concerns about my first attempt.

Just one other idea: do the design constraints rule out the use of a mains to DC power supply to provide the initial DC voltage. Here's a 30 V supply for under £15 (if you buy more than 5). It is capable of supplying vastly more current than you require, so, if you look around you might find something cheaper. Certainly if you use a standard 300 mA 12V plug top and change the design to 9 LEDs (3 parallel strings of 3 series LEDs) You can reduce the cost. If you do this you will need to pay particular attention to ensuring that all three strings use the same current, so that the brightness of all LEDs is the same.

Very good point. I have to admit when I looked at the spec. I was thinking that 1.5 V cells always look like about 1.6 Volts with a meter, but, that's under no load. How about a 4.5 to 9V input device and 5 or 6 cells?

http://uk.farnell.com/jsp/Electrical/Power+Supplies+&+Converters/XP+POWER/JCA0205D03/displayProduct.jsp?sku=1289152&_requestid=67692

data sheet http://www.farnell.com/datasheets/94376.pdf

The initial set accuracy (±1%) and line and load regulation look much better too. Costs about the same but is a physically larger device. Let's hope he's not already committed to a design for the base.

Hi Randall

Thank you very much for the circuit drawing, when i returning to school after half term

i will seek to purchase the converter and try it out. I did not understand the comment or include a link field?.

Much appreciated

Brian

Hello again, brianmorseteacher

A further thought.

If you use rechargeable cells - NimH then they may be recharged many times, and save money in the long term.

Kind Regards....

Hi sparkstation

Good point, much appreciated

regards

Brian

Quick note: using NIMH will reduce your available voltage by about 20%. This may work in your favor and maybe not. the voltage drop across each LED theoretically would be 1.8 volts using my series parallel circuit as drawn. This is without using a resistor. You would have to try it to see if the luminance would be adequate.

Regards,

Keywalker

Keywalker you need to look at the forward voltage specs. of newer LEDs.

You are quite right. Sorry about that. It has been about 5 years since I retired and have not kept up with the latest innovations in LED technology. All of the LED's that I used in the design projects for the industry I was in were 2V Typ. forward voltage. This actually makes the project much easier using NIMH now. Three NIMH batteries in series equals 3.6 volts. The forward voltage for the RL5-W18030 Super-White LED (GaN/InGaN) <http://www.superbrightleds.com/specs/W18030.htm> is 3.4 volts with a maximum of 4 volts. Brian only needs to put the LED's in parallel and hook up to the series voltage source from the batteries. Any other additional circuitry will only degrade the efficient use of the available Amp Hours as I see it. Please correct me if I am wrong. No offense taken.

The budget has to be taken into account here (no pun intended), along with efficiency etc.

For 3 off NiMH cells you're looking at, say, £4.50. And then there's the charger problem - school budget couldn't possibly cover 1 charger per student, so what do you do? Send them off at the end of the project (having bought a couple of chargers to allow the students to recharge in a rota) with a lamp powered by soon-to-be dead rechargeables. They then have to find themselves chargers.

I agree the rechargeables would provide a better long-term solution, and much lower circuit complexity than my approach, as they also have a very nice flat plateau on the discharge curves - they stay up 'til they're nearly spent. I think a bit of series resistance (one resistor for all 8 in parallel) would be a good idea though, as the initial voltage of 3 fully-charged cells could be ≈ 4.2V, and the LED max. current limit of 70mA may be exceeded, particularly if one happened to be ill-matched, and with no current-sharing resistors.

The ideal would be a reasonably simple circuit which could use either rechargeables or disposables, providing as near as possible optimal performance with either, and at a minimum unit outlay. A typical "request" from the sales-team! ("This is what our customers are demanding! We've already sold it, now go design it!" ).

(BTW - the LED's you have cited seem very similar to the ones Brian selected, see my datasheet link, post #16).

Yes, and not only the budget but also the fact that he can not use mains. He would have to use mains to charge the NIMH batteries. What I was thinking of was allowing his students to charge their own batteries at home but this, upon reflection, is not a practical idea either. Most of them would probably not want to do this. This brings us to the point where I don't believe we can improve upon his original circuit.

Regards,

Keywalker

Hi Bill

Thanks for the interest and advice, but as i say remember i am an electronics moron in comparison to you professionals. I'm in the UK so whether i can find info on National semiconductors and if i did, i don't think i would know which LED driver would suit my purpose. I know nothing about the use of LED drivers, how they work or where to connect them.

regards

Brian

If he is getting 50 hours right now (say 40 if he ups the brightness slightly) this is weeks and weeks of normal usage!!! If the user used it for 2 hours a night (unlikely) the batteries will last 20 nights.

A better bet would be to add a connector for a charger, to allow trickle charging of the batteries when the light is not being used!!! Then you could have high mileage AAs at say 2700 ma.....

The charger plug is a good suggestion Andy. I would also suggest changing R1 from a fixed resistor to a 10 ohm potentiometer to vary the brightness of the LEDs. Often very bright lighting fatigues the eyes so some form of brightness adjustment would be welcome.

The mains PSU is a good idea, if it's acceptable in the terms of the spec. for the project. It depends on the interpretation of Brian's statement in the OP, viz.: "I cannot use any mains voltage circuits". I suspect we'll have to wait 'til after half-term for the answer.

" ... changing R1 from a fixed resistor to a 10 ohm potentiometer ... " is not such a good idea (assuming you're talking about my #31 circuit).

The LED current (I_LED) is set by

I_LED = V_REF/R_CS, where V_REF is the converter reference voltage, fixed at 1.237V, and R_CS is the value of the current sense resistor (R1).

You will see that 10Ω would try to drive the LEDs at over 120mA, and as the pot was turned down, it would get very much higher!

Even changing your pot to a more reasonable value (and including some fixed resistance) isn't an optimal solution, as to keep the current through the two sets of four LEDs equal, you need R1 = R2.

I agree a brightness control would be a nice touch (though I don't really think it's necessary here), but that isn't the way to do it.

My bad! I was referring to Keywalker (in #15) and his circuit. In my hasty typing I also failed to enter another zero which would make the potentiometer a 100 ohm value rather than the stated 10 ohm unit.

Recently my grandson and I modified his 12 LED miners style headlamp which was overly bright washing out the colors of his Spiderman comic books with it's blue hued intensity. We used a potentiometer between the battery and the bulbs which he selected because of it's small size and < 1Kohm rating (a guess). However it seems to work fine and he's happily reading in bed long after "lights out" as commanded by his mother (we do enjoy being conspiratorial now and then).

Hi John

I thought i would clarify the point about mains voltage use. Now under health and safety regulations i would be legally responsible for any accident/ fault in the project if this happened in school or out of school in the future. Because of the legal reasons it would not be the school who could be sued but the individual teacher. Also we would not trust pupils to wire up items from a mains 240v supply. In schools now unless you are trained and certified you are not permitted to put a plug on any electrical item, rediculous i know, but that is the legal requirements.

An additional point is that our (pennies) budget would not allow the inclusion of any PSU's, no money in schools for the pupils projects.

I would like thank all the people who have contributed their time and knowledge to helping me. It is often forgotten that teacher's need help in schools especially with electronic projects for pupils. It is a misconception that in school teachers know all the answers.

regards

brian

Look here for cheap (£6) PSUs:

http://cpc.farnell.com/jsp/search/browse.jsp;jsessionid=1PYXG0SI1YNQVQFIAEXZLTQ?N=411+1002139+500001+114130&_requestid=116856

If you used these 24V supplies you could do two strings of 6 LEDs.

Select On Test (SOT) resistors to get exactly 20 mA in each string.

....but not forgetting a smaller fixed resistor that will allow just the maximum current that the LEDs will handle, in series with the Pot., otherwise "DISASTER"!!!