Maxim's 1674 IC as a LED driver...

Status
Not open for further replies.

papasan

Enlightened
Joined
Mar 25, 2001
Messages
621
Location
Northern Virginia
Maxim\'s 1674 IC as a LED driver...

http://pdfserv.maxim-ic.com/arpdf/MAX1674-MAX1676.pdf

wanted to start a new thread about this chip because i'ld rather not clutter up the other threads...

i think it'ld be nice to compare this with the zetex and 757 offerings to see if one stood out as the best amoungst them...

this looks like a promising chip with a low outboard component count and small size for driving an LS or several 5mm LEDs...

i have a couple samples of this chip and i'ld like to put them to some use, i just wanted to get some input before i do my next digikey order...if i put this together it would only be my second cicruit building attempt, so i'm not a guru or anything...

question...
am i correct in thinking that for a minimum set-up this circuit needs a .1uF reference cap, an output cap (Kemet's tants) and a coil?...if so, very nice...
 
Re: Maxim\'s 1674 IC as a LED driver...

I got a couple samples of the MAX1674. The specs look very desireable for our applications.

I wasn't able to get it working using a normal soldering iron and my 20/20 eyes. That pin spacing of less than 2/100" is beyond my ability to work with. I might take a look at it sometime again with a good magnifying glass.
 
Re: Maxim\'s 1674 IC as a LED driver...

The big problem with this is making it work in a constant-current mode. If its low-battery comparator had a lower reference voltage, you could use it as a current sense amp (like I'm doing with a Linear Tech LT1308B). The LT1308B is also in a hobbyist-friendly SO8 package (0.50 inch lead spacing). And the 36V max output voltage (10V max input) of the LT1308B is better, as well. Downside is that it needs an external Schottky, and so is not quite as efficient. But I'm showing around 81% efficiency with a 2.5V input and two white LS lamps in series as a load.

You can see the schematic at http://nedkonz.dhs.org:8080/Ned/27
 
Re: Maxim\'s 1674 IC as a LED driver...

As I mentioned in the other thread, the main drawback is its size---difficult to solder.

Also, the spec sheets are a bit confusing, but they seem to imply that the MAX1675, with its 500mA current limit, is much more efficient than the 1674, with a 1000mA limit.

The 1675 is only rated at 280mA out with 2.4 volts in and 3.3 volts out---too low for practical use with an LS, but could certainly drive a few Nichia's.

The 1675 requires no external transistor or rectifier, but uses two additional capacitors, so it's doubtful its circuit board could be smaller than the Zetex's, nor any cheaper.
 
Re: Maxim\'s 1674 IC as a LED driver...

Why can't you get Luxeons? Future/Active shows the LXHL-NW98 (the white Star/O) in stock, at least; this makes for interesting projects...
 
Re: Maxim\'s 1674 IC as a LED driver...

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by papasan:

wanted to start a new thread about this chip because i'ld rather not clutter up the other threads...


question...
am i correct in thinking that for a minimum set-up this circuit needs a .1uF reference cap, an output cap (Kemet's tants) and a coil?...if so, very nice...
<HR></BLOCKQUOTE>


papasan,

The chip size is a killer. Who can solder this small chip? Who? They call it the uMax package. It is just so small.

Looks like it requires a 22uH inductor, a dual 47uF output caps, a 47uF input cap, it can be preset at 3.3 volts with a resistor, or use a resistor comparitor setup for variable voltage (which lowers efficiency.) A .1uf cap is used at the reference pin.

Perhaps electrical conductive epoxy, the 70% silver kind, might be the only way to go with this small chip. Has anyone tried this approach. Think, "NO SOLDERING!"

Wayne
 
Re: Maxim\'s 1674 IC as a LED driver...

nah, soldering is the way to go...try cleaning up epoxy if you make a mess =P...and the chip doesn't look to small to me, looks decently easy to solder...think 'solder wick' and 'flux'...and use the force...

i don't think the input cap is necessary...i would also be willing to bet the dual output cap can be replaced with a single kemet tantalum cap...at least for our needs...and for 3.3V you just jump FB to OUT, no resistor needed...

any thoughts on this?...
 
Re: Maxim\'s 1674 IC as a LED driver...

If you're really concerned about soldering these by hand (though many people have), you might be able to use solder paste and a toaster oven to solder with.

I've found that the 0.050" pitch parts are really simple to solder, and the 0.025" parts are a little more fussy, though not impossible. In the smaller packages, the biggest problem is holding them down (like the 0402 size resistors, for instance).
 
Re: Maxim\'s 1674 IC as a LED driver...

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by papasan:

...wanted to start a new thread about this chip because i'ld rather not clutter up the other threads...

this looks like a promising chip with a low outboard component count and small size for driving an LS or several 5mm LEDs...


Originally posted by Duggg:
[QB]
Also, the spec sheets are a bit confusing, but they seem to imply that the MAX1675, with its 500mA current limit, is much more efficient than the 1674, with a 1000mA limit.
<HR></BLOCKQUOTE>

It's an intersting chip but wouldn't be the optimum for driving a Luxeon Star. I made a calculation of what a output current we could expect from this switcher on various input voltages. See here for more details: Converting Minimag (2AA version) (Page 10)

Looking at the expected input current during the duty cycle I think one certainly needs the input capacitor - at least for all batteries smaller than a D cell. Using the small SMD ceramic capacitors the circuit board could still be very small.
 
Re: Maxim\'s 1674 IC as a LED driver...

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by bikeNomad:
[QB
I've found that the 0.050" pitch parts are really simple to solder, and the 0.025" parts are a little more fussy, though not impossible. In the smaller packages, the biggest problem is holding them down (like the 0402 size resistors, for instance).[/QB]<HR></BLOCKQUOTE>


What kind of soldering tip do you use, and where can it be purchased. Surely the soldering tip I have must be too big?

Wayne
 
Re: Maxim\'s 1674 IC as a LED driver...

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Wayne Johnson:

What kind of soldering tip do you use, and where can it be purchased. Surely the soldering tip I have must be too big?

Wayne
<HR></BLOCKQUOTE>

I use a
Metcal iron that I got on eBay. They have a number of smaller tips; I often use a STTC-126 tip, which is a 1/64" sharp tip, bent 30 degrees. But this isn't a cheap iron ($150 or so used), and the tips (used) run around $9 on eBay.

My old iron, a Weller WTCPN, was also capable of doing this kind of work but had a bigger tip.

One strategy is to use a tip like Metcal's "mini hoof", which holds solder on its flat surface, and wipe the solder across the leads.

And you should use some magnification and good light.
 
Re: Maxim\'s 1674 IC as a LED driver...

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by bikeNomad:
I use a
Metcal iron that I got on eBay. They have a number of smaller tips; I often use a STTC-126 tip, which is a 1/64" sharp tip, bent 30 degrees. But this isn't a cheap iron ($150 or so used), and the tips (used) run around $9 on eBay.

My old iron, a Weller WTCPN, was also capable of doing this kind of work but had a bigger tip.

One strategy is to use a tip like Metcal's "mini hoof", which holds solder on its flat surface, and wipe the solder across the leads.

And you should use some magnification and good light.
<HR></BLOCKQUOTE>


bikeNomad,

Okay, thanks, that is a help. I have some second hand soldering equipment from SCE. I will have to check the make and model. I have a variable temperature soldering station with a desolderer vacuum, and I also have a desoldering station, with a vacuum, which has a real nice trigger action gun.

I will see if I can get a similar tip to the one you use. If not, I suppose I can try to make something up for it.

Wayne
 
Re: Maxim\'s 1674 IC as a LED driver...

got some time to build one of these today...when set to 3.3V out (as i had it) it only requires four components...probably would want to set it down to 3.0 or 3.1, in that case it would need two more components - two resistors...efficencey may suffer a little this way, dunno...

tested with a white bat-wing luxeon star, 1800mAH NiMHs...the board components where a maxim 1674 IC, a panasonic tantalum .1uF cap, kemet 150uF tantalum output cap, and the hand-wound coil from the zetex thread...perhaps a more specifically made coil would work out better, any thoughts?...

** updated with correct info **

Code:

not too bad, just slightly worse than the zetex 3x0...hard to tell from only one run (and a very over-driven one at that) but it seems to regulate better than the zetex circuit...

one odd thing, at 150 minutes when i stoped the run the light was flickering at about 6 or 8hz or so...obviously the last readings were off because i tested the cells and they red ~.95V each unloaded...might be handy as a warning that you need to change batteries but have no ideas why it happened...maybe because the battery power was so low that the it took quite awhile for the coil to charge up to the required voltage...
 
Re: Maxim\'s 1674 IC as a LED driver...

Wow, those are very impressive results! Regulation is excellent.

But how did you manage to solder that little thing? Did you etch your own board, or did you use one of those SurfBoards?

Did the Max1675 get warm after being driven at that high current?

With a voltage divider calling for a 200K and a 270K, those resistors shouldn't affect efficiency much at all, and will get the LS current down, and extend the run time and maybe even increase efficiency. Could you do such a test?
 
Re: Maxim\'s 1674 IC as a LED driver...

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by papasan:

not too bad, an average efficiencey of 87.0% ...
<HR></BLOCKQUOTE>
Not too bad?
confused.gif
- I think it is a great result for a first prototype
smile.gif
!

How did you manage to get samples of this chip? I ordered some directly from Maxim on their web page more than four weeks ago (shortly before you opened this thread) and got a message that they are outstanding. So I'm still waiting ...
frown.gif


A bit irritating is your last measurement and your calculation: Input current 203mA at 0.42 volt and output only 8mA at 2.64 volt. There must be a big loss somewhere on this low input voltage. According my calculation this does not give an efficiency of 86.82%
grin.gif
?
 
Re: Maxim\'s 1674 IC as a LED driver...

soldering it wasn't too hard...the zetex chip is a little easier, this is probably the next finer step...th elegs are a bit longer so that makes it a little easier...basically i soldered it like i soldered all the rest of the SMD stuff -- paint the board in flux, place all the components and then let it sit and dry a little...the flux gets somewhat tacky and helps keep the components in place...after i solder all the pins down i've usually made a mess with solder bridges and whatnot so i go back over it with a small solder wick and clean it up...

yeah, i plan on lowering the voltage with the resistors and trying another test...i'll also put an input cap on and run it...the board didn't seem to get very warm, but the LS was sure hot to the touch...

i etched the board...it's about 1/2" x 3/4" rectangular...because of the lower pin count and the wire resistor it seems the zetex circuit can be shrunk down to a smaller size...but, just like the zetex circuit, the coil and the kemet cap are the largest pieces by far...

ahh, remuen, you are absolutely correct...i got some bad cells in my exel spreadsheet...oops...here's a better one...

Code:

that drops the average to 81.75%...irk...

i think the flashing at the end was the battery death causing the coil to charge slower...since the coil is in series with the driver circuit instead of parallel like the zetex circuit the chip will shut-down during the low-side of the input voltage wave...i guess an input cap is more important on this one...
 
Re: Maxim\'s 1674 IC as a LED driver...

Hey papasan,

Way to go! Your results look good to me. I also have a sample MAX1674. Sometime I hope to get to it. My only problem is too many projects to do.

Very encouraging. I do have a PCB board designed for the MAX1674, I just have not printed it out so I can etch the board. Now I just gotta get to it!
grin.gif
grin.gif


I received a few sample coiltronic inductors, and when I used one in the MAX757 design, it upped the efficiency by 6%. Something you might want to consider. Also hand wound ferrite toroid inductors seem to have lower DSR. ??

Wayne www.elektrolumens.com
 
Re: Maxim\'s 1674 IC as a LED driver...

if i have my theory down right, the DCR of the coil is basically just figured by how what guage and how long the wire is in the coil...since these hand-wound coils that duggg came up with use 24 or 22 guage wire and they only go around the magnet 2.5 times then i would think the DCR is lower than anything you would find commercially...another good thing about this circuit is that it specs a 22uH coil, so even the commercial ones should be lower DCR and be able to handle more current...i was kind of trepidatious about making my own coils also, but it's real easy and much cheaper than buying them pre-made...

here's a pic of my layout...not everything is on here, the input is the bottom right, the coil goes between the bottom right and top right...the ground is on the bottom and output top left...the unlabeled bar in the middle is where the resistors will go when i get some...right now i just have a short piece of wire jumping this to the output for the 3.3V pre-set voltage...

maxim1674.gif
 
Re: Maxim\'s 1674 IC as a LED driver...

papasan,

I have an extra 22uH inductor from Coiltronics. The resistance is .06 if I remember correctly. You can have it if you want it. I'd be interested in seeing if/how it affects the efficiency of your design?

Email me if you want it. It is for free!

Wayne www.elektrolumens.com
 
Re: Maxim\'s 1674 IC as a LED driver...

There's a science to winding your own coils.

First, you have to select a core material that has the "right" permeability. If the permeability is too low, the coil will require a lot of turns. If it's too high, it saturates and starts to act more like a resistor than an inductor.

Next, as Papasan pointed out, the DCR is directly proportional to the length of the wire, so the fewer the turns, the better. The Zetex coil has a DCR around 0.006 ohm.
 
Status
Not open for further replies.
Top