Chip wars? dat2zip & lamba

Hmmmm ... it almost seems heretical to bring this up, so before I do, let it be known that I greatly respect both dat2zip and lambda for their wondrous contributions to the forum and its members!

Lambda's Ill Pill Kit or dat2zip's MADMAX-ADJ 1674?

Apples-to-apples this may not be, but if we could keep as many extraneous variables controlled as possible, how would these two chip kits compare in a head-to-head standoff?

Questions about them just keep rattling around in my brainpan (there's not much else up there to interfere) and it seems this might be a good place for anyone with both drivers/LS combinations to chime in and share their opinions and/or empirical data.

Since dat2zip's offering can vary output, we shouldn't consider that function here (although it sounds like a great feature) ... and we should assume each driver/LS pair is properly heatsinked and driven to maximum output.

In all fairness, we probably shouldn't give weight to the variance in pricing or the construction (DIY) differences between the kits, either (although those are important considerations when buying, maybe those things are better left to our matchless reviewers to deal with).

The flashlight housing ... MiniMag, Brinkman, custom ... likewise isn't worth getting into as long as its understood it has to provide sufficient cooling.

What really intrigues me is:
• How do the chips compare when driving a Luxeon from two, new, alkaline AA's?
• Which produces the brighter initial light?
• Which chipset stays brighter, longer?
• Which has the longest total run time (if you can't read newspaper by it at 2 ft., it's dead)?

Anyone know the answers?

--------------------

-Mark

Step on up everyone. Witness the very first CPF Chipset Smackdown.

Ding.

I'm don't know if lambda and dat2zip want to comment on each other's chips, but I've been trying to buy both, so I've looked at all the info for both. It would be hard to compare the two at maximum output, because I think the madmax max's out at around 600-700, but the ill pill, although it could probably go higher, because of the heatsinking, you probably wouldn't want to take it that high. Maybe it would be more fair to compare the Illuminator with the MadMax, then they'd could be run about the same. From all the info I've read, I think these are the answers, but I'm sure someone else can verify this

1) Illuminator starts out brighter
2) Again, illuminator starts out brighter
3) the madmax will stay brighter longer because it's regulated
4) i don't know the answer, but I think it's the madmax. I know on certain versions of the madmax, it has a moonmode, kinda like the ARC's, so after the batteries die it will still provide some light at a very low power. I don't think there's been an official runtime, but I'm assuming it could be as long as 5 hours of moon mode. I had a couple of other circuits that continued to burn for over 6-7 hours on dead batteries, but at that point the batteries are probably leaking so you wouldn't want to run it that low anyway.

there are runtime graphs for both somewhere. but if you're looking for comparison's with Lambda's pill, then the answer's are probably different. again, I don't know how hard the pill can be driven without overheating

Great points, InTheDark!

In retrospect, its probably more fair to compare the Ill Pill DC-DC Converter Kit with the MadMax-Adj 1674 since both are build-ups for DIY'ers. Since the Illuminator is a completed flashlight, to me it seems a less even comparison.

The other quick-draw flashlight showdown I'd like to see is between a flashlight dat2zip has modified using his MadMax against lambda's similar Illuminator. Would be best if both used the same donor flashlight (MiniMag or Brinkman).

Odds are even that I'll end up buying chipsets from both gents eventually, regardless how any comparisons turn out.

My runtime graphs are at:

Run Time Graphs

All designs have their pro's and con's. Depending on what tradeoffs if any were made and what features were optimized it would be rather tough to compare apples to apples.

Generally speaking, the MAXIM167X has both the switch element (NFET) and an output switch (PFET) which is more efficient than a DC/DC converter that uses an external diode and an internal switch element. (MAX1675 I have actually measured 94% efficiency on a regular basis)

My other observations are that converters that use transistor switches (LT1932, ZXSC300) have slower turn on/off characteristics than converters that use FETs as switches (LM2621, MAXIM ICs). The time the switch element is in transition from on to off or off to on the switch element burns power and is less efficient.

Another point to take into consideration is the switching frequency. Lower than 600kHz uses larger component size elements (capacitor, inductor). Switching higher than 2MHz requires multilayer PCBs for good grounding and switch times.

Flat out perfect regulation is not necessarily the best design for a flashlight.

Compare the current LS flashlights (Mod'd or off the shelf) to conventional flashlights. Does a conventional flashlight maintain constant brighness over the life of the battery? NO

Does the end user usually notice the ever so slight decrease of light over time. Not usually.

So, an ideal situation may actually be a poor converter regulator. The ZLT project is a good case in point. The ZLT if set to 400mA with fresh batteries will outrun the MADMAX regulated design.

The reason for this is a regulated driver will draw a constant power. In the above case that would be approx 3.2V @ 400mA or 1.28W. Disregarding that we are overdriving the LS specifications we now look at the power to the LS say 1/2 hour later. The ZLT will be driving less current while the a regulated driver would ideally be still putting out the same voltage and current. This constant power output is not ideal for the battery. As the battery voltage falls it is actually better to not tax the battery as much to get the most out of the battery. Constant power means that as the battery voltage falls the battery current increases.

Since the ZLT is now drawing less and less power over time the battery life is extended as compared to the fully regulated.


The MADMAX-ADJ is full SMT with 9 SMT components all on the same side including the 3mm trim pot in a 14mm circular board.

The badboy converter I have yet to release does regulate current and will drive the 1W up to 420mA (2AA setup) or 750mA with 4V-6V input. It too has all SMT components on a 14mm circulate board.

I have another converter board that is based on the MAXIM167X series and it has the moon mode feature built in. Regulation cutoff is currently set to 1.6V and it throttles down to 50mA when in moon mode providing another 5-6 hours of safety light. This version will not damage or destroy rechargeables NICD or NIMH batteries (assuming you don't run it in moon mode too long).

Yes, Yes, No, No to upcoming questions in that exact order.

Ya, what he said.

The Illuminator circuit was designed for reliability and simplicity; one coil, one IC. When constructed it's size and shape is such that it will replace the bulb socket in AA Brinkmann, MiniMag and other small flashlights. The IL was designed with a fixed ouput that makes trade offs between, brightness, run time, and resulting Luxeon temperatures.

Aside from that, the only benifit it would have is some inherent "idiot proofing" as it is very forgiving compared to regular circuits. You can run it no load, shorted to ground (never recommended, but usually survives) and you can connect/disconnect Luxeons while the circuit is running with out any harm to circuit or Luxeon.

It was targeted at the beginer to average experimenter, with some forgiveness for mistakes made, and simple construction.

Dat2Zip's circuits certainly are more complex, more efficient, and allow more precise control over the output than the IL circuit.

To me you're trying to compare apples to oranges, or perhaps a cesna to the space shuttle.

Thanks for the replies! You gotta know I'd be happier 'n hog in a holler with either one ... more so with several of both!

Sure, from a design perspective they're radically different. From the user's perspective, they both modify a MiniMag into a better light generating system. I've read wonderful things about each one ... always in their own threads ... but never a word comparing the two. That seems unusual to old Joe Schmo (that'd be me), who's read comparisons here at CPF on just about every flashlight known to man.

Now I'm going to wonder about it until I can acquire and build one of each.

In the meantime, thanks for the input.

Great topic! I've had the same questions. OK, here are some more. What IC does the Lambda circuit use? Is the output (therefore brightness,runtime,etc.) the same for the LI and IllPill? In a nutshell, what is the difference between the Badboy circuit and the MAXIM167X? (When would you use one over the other?) And finally (for now) Dat2zip, How long does your Maxim167x circuit with moon mode typically run on 2 new AA Alks before going into moon mode?

sflate,

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>In a nutshell, what is the difference between the Badboy circuit and the MAXIM167X? (When would you use one over the other?) <HR></BLOCKQUOTE>

These are tough questions. The fact is there are mutiple variables that affect the runtime. Here's a partial list of several of the main variables.

1) Batttery - NICD, NIMH, Alkaline, Lithium.

2) Brightness setting (output Current)

3) Luxeon Star LED


MAXIMUM CONSTANT BRIGHTNESS:
Either the badboy or the MADMAX are about the same with the MAXIM 1674 squeeking out a few extra minutes of runtime in the continouus operating mode. Configuration: Badboy - 500mA drive, MAXIM1674 - Adjust Vout to generate 500mA of current draw.


MAXIMUM BRIGHTNESS:
It is possible to configure badboy to output almost 1A of output current into the LS. Whether any 2AA batteries can actually drive this converter has not been tested. Testing was done with a bench supply.

MAXIMUM RUNTIME: This would be achieved best with the MAXIM1675 running wide open delivering a starting current of 250-300mA with a fresh set of batteries. This plus Lithium batteries would generate the longest run times of all combinations starting at 94% efficiency and dropping to 85% at near dead battery condition.

The LS plays an important role in determining runtime. A hot LS is the lowest possible voltage at a fixed current as compared to other LS at the same current.

2.9V White LS @ 350mA = 1.015W Load
3.1V White LS @ 350mA = 1.085W Load
3.5V White LS @ 350mA = 1.225W Load

If you had the luxury of sorting thru many LSs to pick from always try to use the lowest one of the lot. This will produce a lower power Load.


Let's see if I can wing some specs off the top of head.
NOTES: 8 uMAX = 8 mSOIC

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR> How long does your Maxim167x circuit with moon mode typically run on 2 new AA Alks before going into moon mode? <HR></BLOCKQUOTE>

The setup I've been doing on Revision 6 boards (MAX1674) have been to configure them to run wide open so the run times should be similar to the run time graphs of rev5 I have posted on my web site. The only difference is thet it should level off at 50mA and then coast for another 4-5 hours at this reduced light level.


-Wayne

Wow, a few months ago we couldn't buy such boost converters so I started making my own circuits and pc boards. I'm still a beginner but designing and building such circuits is a fascinating hobby.

But now all of us have the possibility to buy a good or excellent boost converter - and even better we now can choose between different solutions which are all affordable !

I don't think we can exactly compare the various solutions as dat2zip an Lamda already have posted. Beneath the technical specifications it's also much a matter of personal preference which one you like more.

Just one example for what I said: If you want to have some kind of general purpose boost converter dat2zip's high sophisticated MADMAX1674-ADJ should be your choice. And if you only want to make a plain and fast Minimag mod Lambda's pill seems the way to go.

Dat2zip:
Thank you that you are showing us all the data and information about your circuits! I already learned a lot of you and out of all these information!

Lambda:
It would be very interesting to get some more information about your circuits as eg. which chips you use and of course also input/output/efficiency data. Btw, if you want to keep it as your business secret I do not mind! It's just a question!