Using a DC/DC converter to power the laser

If by output cap you mean the cap across the load, then no. It just filters the output.
If you're talking about the LM3410 output cap, then yes, it is extremelly important. It's pretty much the main cap on there.


The Fb voltage should never change. It should always be 0.191-0.193, depending on the IC. If it was less than this, the IC could be struggling, but if it's more, then you're doing something wrong..

It shouldn't really be 300mV.. Which version of the circuit did you make, anyway?
If you make some pictures, it could also help.

300mV is very strange
and yes the output cap does provide power through part of the cycle. without the output cap, you would just get a heap of high voltage spikes. the output cap sucks up voltage and provides power in between the spikes.
so if you are not using a large enough cap, your multimeter could be reading the top of the spikes instead of the average voltage, but 2.2uF should be plenty to smooth it out. maybe it isn't soldered properly. thats the only thing i could think of that would explain a 300mV FB voltage.

that or the chip is somehow damaged or faulty.

post pics of your board

I don't think the chip is screwed. The previous IC I tried had the same problems (before meeting an untimely open circuit death)

My circuit is set up with the standard boost. Originally, I tried it with the sepic design so that I could power reds from a single Li-ion in a small box. It had the same problem that I'm having now.

How sensitive are ceramic caps to heat? If its the caps I'll try the circuit without them and see if the problem is the same.

Ill stick some pics in here soon.

I have made around 15 of boost only circuits, and like five of them didn't work. There is a simple way to test the IC if it's shorted or if a pad is not soldered well.

But when it all tested well, it was usually a ceramic cap, that was a short. I don't know if i got bad caps, or if they are so sensitive to heat. I solder quite quickly, and usually this doesn't happen. But i've had to replace several input caps and a couple of times, the caps i have across the load.


I have not tested what happens with the FB voltage in this case tho. Maybe i should.

Interesting experience with the caps  

I'm 95% sure its not a short  , after all, the circuit does boost. I'd like to believe the problem lies solely in the ceramic caps seeing as that would be the quickest thing to fix.

I did some more testing with the driver. Using 2 white led's in series as a test load, the Vfb is about 300mv at 4.5 input and about 180mv (yay its deceptively in range  ) with 3v input. Yep it really is 300mv, it gets noticeable brighter from switching that voltage up.

You're using a boost only circuit, right? What's the Vf of those LEDs?

If the input voltage is higher, than the Vf of the load, the entire input voltage will get to the load, and the current won't be regulated.

I made this mistake once, on a dummy load.. The dummy did not have the same Vf as a blu ray diode. When i lowered the input voltage to 3V, all was well.


If you're using a boost only circuit, don't use more than 3V on the input, and make sure, that the Vf of the load is well over 3V.

IgorT said:

You're using a boost only circuit, right? What's the Vf of those LEDs?

If the input voltage is higher, than the Vf of the load, the entire input voltage will get to the load, and the current won't be regulated.

I made this mistake once, on a dummy load.. The dummy did not have the same Vf as a blu ray diode. When i lowered the input voltage to 3V, all was well.


If you're using a boost only circuit, don't use more than 3V on the input, and make sure, that the Vf of the load is well over 3V.

Click to expand...

Yep, its the straight boost circuit.

With 3v or less, it seems fine, with the FB voltage at 182mv to my meter.

If you don't mind, what happens to your circuit if the input voltage is higher than 3v, but still under the Vf of your load.

I'm still convinced that theres an underlying problem here, I don't see any reason why it shouldn't be able to boost say, 5v to 7v.  [smiley=angry.gif]

Hrmph it was the cap after all. I got my ceramic cap kit from "sure" last week and a 10uf did fine. I played around with the sepic too, but it seemed to struggle with higher currents, the max I could get with it was about 500-600ma, even with 30uf of ceramic goodness on the output.

To all who have tried these types of circuits, do your inductors get warm with use? It would be useful to be able to get an amp or so out of it for other led's and..... stuff.

And lastly, why do you all insist on ordering the leadless packages on the samples. Is this some kind of manly thing I'
m not in on

Hmm. Ok, i made some more of boost only constant current drivers with this IC. Quite a few of them didn't work at all. With some it really was the capacitor, that was shorted, but with others, i actually managed to kill the regulator.

After replacing it, the circuit worked fine again.. But it was frustrating because of the sheer number of them, and because i had to do it fast.. So i have not even checked what happens to the Vfb, if a regulator is damaged like that.

In most cases, when i tested a driver with a damaged regulator on a dummy load, nothing happened.

But there was one case, where the output current depended on the input voltage.. This reminded me of your problem. You said your Vfb changes with input voltage.. Well i guess i did manage to kill one of them in the same way as you....



About yours... It can't be fine at under 3V input voltage, but not fine at over 3V but still under the Vf of load....

If it works properly, nothing happens when the input voltage is between 3V and Vf of load.. With the boost only version of the circuit, the input voltage has to be over 2.5V, but under the Vf of the load, or under 5.5V, whichever is lower.

Meaning, since the Vf of your load is 6.5V, the input voltage has to be between 2.5 and 5.5V, and the output current HAS to be the SAME, during this entire input voltage range! Oh, and sometime, they even work flawlessly down to 2V..


If the current changes with the input voltage, then either your regulator is damaged, or soldered bad, or there is something wrong with your circuit design..


Again, post pictures of your circuit please. With and without the components.


And just because the previous one was doing the same, doesn't mean it's not a damaged regulator. I've been working with electronics for 10 years, and yet i managed to kill many of them simply because i was in a rush.


With each one, i put it on a dummy load, set the input voltage to 3V, and measure the output current. Then i vary the input voltage with the fine adjust pot on my PSU (changes the voltage by 1V), to make sure the output current stays the same. But one time, i was looking at my DMM and grabbed the wrong pot and turned it... The voltage went to 16V! :

I don't remember what exactly happened to that one, because i was in such a rush. Maybe that one started varying the output with the input voltage.. Not sure.

But i do think your regulator is damaged. Soldering it for too long can also damage it. Or there really is something wrong with your circuit..

How do they keep dying for you? I find that strange, which packages are you using? It bugs me how all the nice schematics woop and others have posted are for the llp version. Although I've killed 4 so far, the only one that I've found to be dead from the start was when I was soldering another component and didn't notice the shaft of the iron was resting on the IC... until I smelled burning plastic.

Anyway, I've put it down to the caps since everything worked great after replacing them, but I can't rule out that it wasn't the IC since I hooked it up to 12v by mistake (POOF) before I got the new caps.

Heres my schematic, sorry if it makes your eyes bleed  

http://img371.imageshack.us/my.php?image=sepicpcbyg4.jpg

My board is also made for the 6LLP version.. I had to make a board 13x9mm small!

Even my first board several pages ago was for the 6LLP version, even tho it was MUCH larger, but that's because only 6LLP samples were available back then. But i made the new board slightly different this time, to make hand soldering easier.


The reason so many of mine died was because i had to make around 20 drivers in a very short time - i was in a rush. Sometimes i shorted a pad, and destroyed the IC during removal. Other times, everything was soldered correctly, but the IC was dead. Then, sometimes the cap was shorted, but i thought it was the IC, so i removed it, which again sometimes resulted in a ripped out pad.. : And in one case, all was fine, until i turned the wrong knob on my PSU.


All of this never happened to me before, simply because i wasn't rushing.


So your circuit works now? Vary the input voltage (CAREFULLY!) between 2.5 and 5.5V with the driver connected to a dummy load through a DMM, to make sure the current stays the same during this input voltage range.

But if you did the same as me, and connected it to 12V, you really should replace it. It's probably damaged already, even if it does still work.

Wait.. You made a SEPIC? With two coils?

In that case, it's not boost only. It can both buck and boost, depending on what the load needs. The output current should then stay the same all the time, as long as the input voltage is between 2.5 - 5.5V, regardless of the load Vf...

With a boost only version, the output current (and Vfb) goes up when the input voltage is over the Vf of the load.


Did you use the 5 pin SOT package? I would so love to use that one! Soldering would be SO MUCH easier!

IgorT said:

Wait.. You made a SEPIC? With two coils?

In that case, it's not boost only. It can both buck and boost, depending on what the load needs. The output current should then stay the same all the time, as long as the input voltage is between 2.5 - 5.5V, regardless of the load Vf...

With a boost only version, the output current (and Vfb) goes up when the input voltage is over the Vf of the load.


Did you use the 5 pin SOT package? I would so love to use that one! Soldering would be SO MUCH easier!

Click to expand...

Yep thats a sepic. For a boost only (i was troubleshooting with this type for simplicity) I use the same board, but I remove L1 and put a bridge between the pads of the two inductors, as opposed to the cap.

And yes thats the 5 pin version. They are available as samples so .... yeah, why not!

The one I hooked up to 12v was dead by a long shot (hence poof?) Can't believe how picky these are.

Anyway, with it working, I guess its time to put it to lasers. What filter cap size/type do you recommend?

I put two 10uF ceramics directly on the diode. Before i also had two 47uF tantalums on the circuit output for the powerup spike, and a 1uF directly on the diode for the ripple, but it turned out not to be required, and i don't have room for that now anyway..

Even one 10uF should be enough, but i put two on there just in case...