Which converter board for 5W LS ?

[BeeSee]

I am getting a McLux-PR head built with a 5W LS. I am undecided on which converter board (BB700/750 or other?) to use, and was hoping some of you experts could give me some recommendations.

I am not looking for maximum brightness, in fact I would like a good balance between brightness and run-time (1.5 - 2 hours on a set of 123's would be nice).

What run-times are you guys getting on a BB700 or 750?
Has anyone used a BB500 or DB700 with a 5W LS? Any feedback and suggestions would be appreciated. Thanks.

 

[CM]

I'd go with a BB666 (hmmmm, really really bad boy, like the devil /ubbthreads/images/graemlins/grin.gif ). You won't get the 1.5-2 hours runtime but it's probably the best compromise for brightness and staying power. A BB750 will probably get you less than an hour depending on the Vf of the LED. My BB "750" with W4W LED draws 1.3A on 2x123 which probably equates to 45 minutes on a set of 123's. However, it draws close to 2A (ouch!) on a Li-Ion cell. The BB's seem to be optimized for about 400mA-500mA. Much more above that and the inductor in the circuit starts saturating. What this means is that you need to keep the battery voltage as close to the Vf of the LED as possible or you will see drastic losses in efficiency.

Never tried running a 5W at 500mA, it would be interesting to see how much loss in brightness you get. Wish I had more time to play with the BB circuit. It's a nice design, very compact and has gotten a lot of mileage with the modders in this forum.

CM

 

[dat2zip]

CM,

One minor correction. At the high power the IC gets inefficient. The inductor is rated for 2A which is where the IC Ilimit is set to. The inductor doesn't saturate.

The poor IC has a 2A transistor in a 1/4W plastic package. We are pushing well past 1/2 watt of heat through the poor IC.

That's why in applications that pull more than 0.5A-1A input current over the battery input range the IC needs to be thermally heat sinked. More so that most IC applications since the package in by itself is only good for 0.25W (1/4W).

The diode drops almost as much power as the IC and combined together make the two parts that contribute most to the efficiency equation.

Wayne

 

[tvodrd]

CM,

You can cut/file and lap a small piece of copper (old penny?} or aluminum to sit against the IC and diode- they are very close to the same height off the board. Thermal-epoxy it in place. I've done it twice now, and it has worked ok.

Larry

 

[CM]

Wayne,

Thanks for the correction. I do have a question. I put another 10uF 16V X7r ceramic cap in parallel with the output cap and removed the 1608 inductor and put in a 3316 size (Coilcraft), same value of 3.3uH, higher current rating. I measured the LED voltage before and after and it went from 7V to 7.4V. I'm puzzled as to what's causing this. I don't believe it's the DCR of the two inductors. I am running off a Li-Ion in both cases. I repeated using 123's and got 7.4V using either inductor. So I presumed, probably incorrectly that the inductor is saturating. I need an expert opinion here and you're DA MAN. Also is there some kind of thermal cutoff on the IC? There's no mention of it in the data sheet but it behaves like there is a die temperature cutoff.

Larry,

I figure the IC's getting hot since in my 750ma 5W setup, I saw the light shut off when the light got hot. Cycling power turned it on only for a second before going out. I have since heat-sinked the IC and Shottky to the E-can using Arctic Alumina. Haven't tried the experiment again since I'm now paranoid. My McLux PR light is getting a much higher level of kid-glove treatment than my Surefires. I tend to baby this thing when I should get it all scratched up from heavy use /ubbthreads/images/graemlins/grin.gif

CM

 

[koala]

I don't know how good is the heat conductivity of the epoxy casing on the semiconductor. To further enhance cooling you can 'lap' the epoxy so that the heatsink element you are using will be nearer to the die.

The term 'lapping' is commonly used among overclockers, to sand and polish a heatsink surface to a mirror finish. The purpose is to provide a smooth and maximum contact between the CPU die and the heatsink. Other extremist take it further to lapping their CPU die /ubbthreads/images/graemlins/ooo.gif and they claim that it will be dropping the heat a few pinch.

I used to lap my SOT-8 microcontrollers so they are as thin as a credit card so I could build my own mobile SIM cards. It is not dangerous as long as you don't thin it too much. I am not sure how much heat conductivity you can achieve so try it at your own risk.

Another False and Useless cooling technique is to provide ample area of copper underneath an IC to act as heatsink. This is quite useful only if the underneath of the IC is applied with thermal compound. The leads of epoxy cased IC does not transfer much heat as they are connected by tiny thin gold wires to the die. The gold wires are thin and they do not conduct alot of heat. Again this technique is dissipating heat through the epoxy casing of the IC.

On the other hand, the above statement is not true for TO-220 sized semiconductors such as voltage regulator, mosfets and transistors. And also other IC with metal casing with is meant for heat dissipation.

Vince.

 

[Doug S]

[ QUOTE ]
nukiez said:
Another False and Useless cooling technique is.... The leads of epoxy cased IC does not transfer much heat as they are connected by tiny thin gold wires to the die. The gold wires are thin and they do not conduct alot of heat. Again this technique is dissipating heat through the epoxy casing of the IC.

On the other hand, the above statement is not true for TO-220 sized semiconductors such as voltage regulator, mosfets and transistors. And also other IC with metal casing with is meant for heat dissipation.

Vince.

[/ QUOTE ]

There are also some conventional looking SMT ICs where there has been enhancement of the thermal path from the die to one or more of the pins. This will usually be identified on the datasheet for the specific IC. In the case of the LT1618 which is the subject of this thread, I don't know if there is any such enhancement.

 

[CM]

I don't think there's any thermal enhancements to the LT1618. It's primarily low to moderate power applications. As the higher power LED's become more prevalent, or as we overdrive more, the IC itself needs to be looked at in terms of efficiency. My guess is that more parts will be required to drive to higher voltages and currents to the point where an external FET becomes beneficial. Anyone know if there's such an animal as a synchronous BOOST converter? I've seen many sync. bucks but haven't seen any boost. Not an expert in this area so this may be a stupid question.

 

[Doug S]

[ QUOTE ]
CM said:
I don't think there's any thermal enhancements to the LT1618. It's primarily low to moderate power applications. As the higher power LED's become more prevalent, or as we overdrive more, the IC itself needs to be looked at in terms of efficiency. My guess is that more parts will be required to drive to higher voltages and currents to the point where an external FET becomes beneficial. Anyone know if there's such an animal as a synchronous BOOST converter? I've seen many sync. bucks but haven't seen any boost. Not an expert in this area so this may be a stupid question.

[/ QUOTE ]

Yes! Both LT and Maxim have synchronous boost ICs with integrated Mosfets. They also offer synchronous boost controllers too.

 

[koala]

Wow, heat must be an issue to those chips!

Vince.

 

[CM]

Doug S,

I just saw the Maxim part this morning. Haven't had time to peruse the data sheet. I also saw the LT3402 (? it was a 3000 part) with the integrated MOSFET. Lots of great parts out there. Just need a good way to etch PCB's with fine pitch parts.

Vince,

Heat becomes an issue on the LT1618 (and similar switching controllers) when the input voltage starts dropping. What happens is that the internal switch must charge the inductor for a longer period of time as the battery voltage drops. The switch essentially shorts the inductor to ground and the longer the switch stays there, the more power the part dissipates. So the lower the battery voltage to the BB, the hotter the part gets. A more efficient switch (FET with low Rds(on)) improves things, hence my question about the synchronous converter above.

CM