National Semiconductor LM2621 Evaluation Board

[KenBar]

Has anyone changed the value of R F2?

I was going to drop it to 3.5 volts or so.
It is now just under 5 volts with a 50k resistor.

This is working so sweet now I almost hate to change it.

I was thinking of putting a trim pot on top of it...mabye with external pot??
Any ideas? Gadget

Just don't make it too hard.
===================================

====================================

 

[**DONOTDELETE**]

You lucky dog! You got yours already! I'm sooooooooo jealous. The mailman is beginning to wonder why I'm waiting at the mailbox every day when he comes by

Since I haven't gotten mine yet, I can only tell you the math I've done ... we're obviously both looking at making it into an LS driver.

I'm going to test mine at 3.2 volts to drive the LS at 320 mA. That calls for an R2 value of 237 ohms. Best I could do at Radio Shack was 220 ohms and 33 ohms, so I'll put the two 33 ohms in parallel and then add the 220 ohm in series to get 236.5 ohms... which should produce 3.19 volts and that's close enough for me.

R1 is valued at 150 ohms on the evaluation board and R2 is determined by the formula:

R2 = 150(Vout/1.24 - 1)

For 3.5 volts, R2 = 273 ohms

I have to leave the trimming pot info to Gadget.

Good luck & let us know how it goes!

Mark

 

[KenBar]

The good fairy sent me a bunch of the LS modules. I absolutely LOVE the amber. It smokes the Steca module I was using.

I was wondering....since I have a solid 5 volts now, if I put 2 of the LS in series, how it would do.

As the board is right now... I can get 500 ma out of it.

Also....I was wondering since the current value of RF2 is 50K, if I could just "tack" another value on top of it. That would be easier for me to solder on to it. The module is so tiny....I just hate to cobble it up more than I have to.

ALSO..
I would appreciate ideas on heat sinking.
Pics etc would be great.
I really hate it to look too cobblely.
I would prefer to buy something already made..chip cooling fins etc.

 

[**DONOTDELETE**]

If I get any of this wrong, I'm sure someone will correct me, so...

If you put 2 LS in series, the voltage will be cut in half, but the available mAmps will stay the same. So, you'd have 2.5 volts available to each LS, which could draw on 500 mA, but they won't use all that because at 2.5 volts the LS module will use MUCH less than 300 mA (if it starts at all).

If you put two LS modules in parallel, each gets the same voltage, but they divide the amperage between them. So they'd get 5 volts each, but only draw 250 mA.

I think.

If you "tack" another resistor onto the existing resistor, your resistor values will change a tiny bit at a time. I guess with the right value surface mount resistors, you could literally stack one on top of the other until you get the ohms you need. To determine what resistor values to use and how to connect them to get different values, see the thread in Batteries Included: Resistor values

I'll be using copper sheets as a heat sink, but you might be interested in the information on these threads:
Heat sinking led arrays
Silviron's conversion (see his pictures)

There's another post in CPF somewhere (I can't find it now) that had a neat round heat sink with an adhesive tape that looked like it would just fit an LS.

Mark

 

[**DONOTDELETE**]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by KenBar:
The good fairy sent me a bunch of the LS modules. I absolutely LOVE the amber. It smokes the Steca module I was using.

I was wondering....since I have a solid 5 volts now, if I put 2 of the LS in series, how it would do.

As the board is right now... I can get 500 ma out of it.

Also....I was wondering since the current value of RF2 is 50K, if I could just "tack" another value on top of it. That would be easier for me to solder on to it. The module is so tiny....I just hate to cobble it up more than I have to.

ALSO..
I would appreciate ideas on heat sinking<HR></BLOCKQUOTE>

the radio shack has a SCR power transister heat sync that looks very suitable for a single LS heat sync, it is 1/2 x3/4x1/2 inch and has 4 or 5 fins , its the good one they sell for power transisters its black in color, not the cheap one whihc is one hunk of bent aluminum, we used this sync for many things

 

[Quickbeam]

I used the "cheap hunk of bent aluminum"

RS heatsink but either should work - check the LS mod page on my site (below) for the part number. I liked this particular heatsink because I could drill it at the corners and mount the LS on it with screws.

DP

 

[**DONOTDELETE**]

Actually doug the one you used at least has some size to it, the cheap hunk of aluminum i was speaking of is not even colored, and its only purpose is to get the pinpoint of heat off a tiney area like the power transisters have.

i think with the LS because the base (if it isnt removed) is the heat sync, so to get better than that somone would need to transfer that heat OUT or into the air better, something with multiple fins ,or a wide width, for short bursts just a large absorbing hunk of metal.

then in plastic housings, to try and keep the base itself off of the plastic in that configuration, as it would surely melt the plastic without something at least between it.

the ultimate would be a hunk of metal that was the head, refleter or the flashlight body.

by the way doug i like the simple well laid out design of the site.

 

[Go Go Gadget Flashlight]

Ken, I got my board in the mail today, so soon I will be able to retro a trim pot and then show you how I did it.

I may get to it today, but can't guarantee as I am supposed to be be running down job prospects and corresponding with employers. The wife will be checking up on me when she gets home.

 

[KenBar]

Great!!!
Mabye I am easy to please, but the LS and the LM2621 are just way too much fun. I think anyone can do this mod. It will please the "Iwannasearchlight" people and the night light people also.

I think I am going to go ahead and order a few more to have around. When you remove the banana jack, the board is tiny...able to be stuffed anywhere.

 

[**DONOTDELETE**]

KenBar: My only beef with the LM2621 is its output being limited to 500 mA. Even so, there are a bunch of projects this will work great with!

I CAN'T STAND NOT HAVING ONE MUCH LONGER! The postman didn't bring mine again today

Most of the other circuits I'm working on will be more flexible and cost 1/4 to 1/2 as much as the LM2621, but had I known about the this evaluation board before I started buying all these other parts I doubt I would have started the project in the first place ... I think $20 each is a super deal for a circuit you don't have to build!

I'm sure several will be in my future ... if I ever get to play around with one.

Gadget: I'm very interested, too, in the variable resistor approach and can't wait to read how your testing goes.

Mark

 

[KenBar]

Originally posted by Luffokc:
[QB]KenBar: My only beef with the LM2621 is its output being limited to 500 mA. Even so, there are a bunch of projects this will work great with!

I CAN'T STAND NOT HAVING ONE MUCH LONGER! The postman didn't bring mine again today

=================================
Sorry re postman...

Quote from doc...
"By using a schottky diode and an inductor rated for higher currents, load currents up to 1 amp can be realized"

I have some inductors from Coilcraft standing by...if I can just figure out how to solder the buggers and get them to fit. They are of course a tad bigger. I have gotten a few SMT circuits to work but it is nerve shattering for me. What I have found is that cold solder joints, even partial, are death to things working correctly.

ALSO...the LN2621 comes with a GREAT CD....tons of toys

======================

The contents of the CD are structured as a set of web pages.

From the "home page", other web pages will guide you to product datasheets, application notes, articles, and design software for power management products:
> Switching Regulators (including Simple Switchers)
> Linear Low-Dropout Regulators
> Switched-Capacitor Converters
> Lithium Charging & Protection products
> Power Supply Supervisors
> References

Mabye the postman will come tomorrow...I know what it is to wait....believe me

ArcLS LOL

 

[e=mc²]

Where may I ask are you ordering the boards from? Seems like an excellent solution.

Regards

Ed

 

[**DONOTDELETE**]

jeff1500 found this for us. Great find, too!

Evaluation board

 

[**DONOTDELETE**]

Minor stuff...

Re: Calculating resistor values... you probably already know but... make sure of your units. Luffokc, you were talking 237 Ohms, but you probably meant 237 kOhms... If you didn't, you should. ;-) Several reasons, but mainly -- too low of values (i.e. 237 vs 237 k) and you wind up conducting lots of current through them, comparatively... you don't want to waste the battery juice, even if we're only talking, say, 21 mA... you'd much rather have it conduct 21 uA -- it's a reference voltage you're setting, not a driver.

And, Kenbar, you'll probably find it in another section, but... resistors in // lower the value, resistors in series raise the value. If you tack another SMD over and existing one, I suppose it would work, but you may have a hard time coming up w/ your desired values. Anything you add to the 50k will reduce it's value. Tack a 5M Ohm resistor atop the 50k, and you've just made a dandy 49.5k Ohm. If that's the route you really want to go, then calc is simple:

1/R1 + 1/R2 = 1/Rp

Invert the values and add them together, and invert the sum --- voila, your // value. Similarly, if you want to figure what R2 should be, take the desired Rp, invert it, and subtract the (inverted) value of R1 -- and invert the result.

Instead, review the page that the caver-light did... ignore the spelling errors (grin) and you'll see that he completely removed one of the bais resistors, drilled a hole in the board, and soldered a standard wire resistor in place, using a standard up-ended mounting posture.

 

[**DONOTDELETE**]

D'Oh! One more thing...

The efficiency of the board's voltage conversion varies by current load and values of the voltage source. You might get it to work for a single 1.5 v cell (which degrades down to 0.7 v) but the efficiency will be a little less than if you used 3v and let it degrade down to 1.4v. Larger, but hey, you suck every last mA from the batteries that way.

Also, while Maxim has some boost/buck converters, I don't remember what the LM2621 does in that configuration... I don't know what it would do if the output voltage wound up being lower than, or crossing the input voltage. I don't think it's a SEPIC board.

---

I asked the guy about multiple boards (discount) or the possibility of buying the bare boards (cheaper) but he said that he only vends them from NS, and that he doesn't have access to any of the "stuff".

I don't doubt that we could get the cost down to at least 1/2 of its current cost, or, if we buy the components (and boards), down to a few bucks each. My salty/ fevered imagination tells me < $5 each. Plus, w/out the jacks, we could shorten the overall board length by at least 1/2 ".

 

[Quickbeam]

Well, I just placed an order for one... My wife's gonna kill me.....

I'm real interested in seeing if you can replace the 50k resistor with a small pot. Wow! A dimmable, efficient 2 cell LS light! I can't wait!

DP

 

[**DONOTDELETE**]

Marcus - Thanks for catching my mistake.

You are absolutely correct that the resistor values should be in kOhms. Dyslexia struck, I suppose.

Also good of you to point out it's not a SEPIC board, which means limiting Vin to two cells. Linear's 1308b can be configured SEPIC and yes, several of Maxim's offerings can be, too.

Thanks for your posts! Ingenious of you to inquire about quanity purchasing, etc., by the way.

Another point worth mentioning (although I still await the product & it's documentation, so this needs to be checked) I don't believe these National Semi boards have short protection ... wiring the input wrong may be a $20 mistake.

Mark

 

[KenBar]

What does the following mean in easy english please?
From extra sheet with LS module:

1) Use DC current source power supply, NOT voltage source.
2) Set current to 350 mA, forward voltage will be 2.5-4.0 volts.

I have a general idea what they mean, but I don't want to fry one of these LS.
I am not sure what CURRENT SOURCE POWER SUPPLY means....

Thanks

 

[Mike]

A current source power supply is one where you can set the current output to your desired level instead of the typical variable supply which lets you set the voltage.

 

[Go Go Gadget Flashlight]

Ken,
what they are saying is what Gransee has said, the LS's are current driven, not voltage driven. For a given current input, the voltage will differ as much as 10% (I think that's about the % Peter found).

So, in building a "source" for them, current needs to be limited, not voltage. It's pretty much the same as regular LEDs.

This is why we need to use a variable pot for the NatSemi board. The LS's vary so much, you can't just calculate the resistor values needed. You have to measure the current thru the LED as you vary the resistance. Then, you can measure the resistance of the pot and replace it with a fixed resistance.

If you have more than one LS, you can't just swap them out and assume the current is the same. Odds are, it won't be.