Using a DC/DC converter to power the laser

[phenol]

i guess it'd better suffice, as there isnt another option really without metallized vias. it would also work if a small copper or al slab is glued on top of LM3410.

 

[IgorT]

i guess it'd better suffice, as there isnt another option really without metallized vias. it would also work if a small copper or al slab is glued on top of LM3410.

I agree that more is definatelly better in this case.. Vias directly below the IC would transfer the heat away more efficiently, but i was thinking, if the upper ground plane would be made to have large exposed areas on both sides of the chip, pieces of metal could be soldered directly next to it..

This, with another piece of metal glued on top of it, like you said, should work ok.. (At least i hope so)


BTW: I just recieved the answer from a local National Semi supplier, that the Eval. board will not be available till the end of march..

So i guess Woops circuit is the best way to go. It shouldn't be too expensive, to etch a few of these circuits.. Now if only CoilCraft would answer me..

 

[woop]

metalised vias are not an option unless i get some made at express pcb or something. I don't think it should get too hot, it should only be switching <0.5A
I am going to get some solder paste from jaycar, make a board and see if i can solder it. I am still waiting on the inductors though. might try winding some

 

[IgorT]

Well, i just got some Fastron power inductors in different values, and the second 3410 sample package just arrived - the YSD..

Even better, i just recieved a confirmation from CoilCraft! They shipped my samples yesturday.. This is great! Seems it helped to translate the R&D part of my website for them...


So i'm just finishing the layout, and etching it later today.. Hope i get it working soon..

 

[IgorT]

I redrew the lm3410 sepec design in eagle. its 23.7mm high 19mm wide and single sided
still learning to use eagle. had to make a new package for the LLP6
what do you guys think?

Hi Woop!


I showed your PCB to my engineer, and this is his response:

- The connections for the inducors are a little smaller than in the original (less cooling, more resistance)
- Inductor L1 is a little further away from the IC (more inductivity and resistance)
- The IC connections are a little close to the central GND for the IC (testing necessary to make sure, there are no short circuits after soldering)

Still, it should work just fine, as long as the right elements are used.


He also had an idea for the vias under the IC.. He suggested using a double sided board, the lower side only GND, drilling some holes under the IC, putting some copper wire, or resistor legs in them, soldering them on the lower GND plane, and cutting them away on the upper plane, flush with the PCB.

During reflow, the vias would then get soldered by themselves to the upper plane.. This should work for cooling.. Unfortunatelly i just bought single sided plates, so i'll go with my idea, and hope for the best.




BTW: I still don't understand the dimensions you specified (23.7mm high 19mm wide)..  My AixiZ modules are 30.5mm long and 11.7mm wide. Do you have shorter modules? If i print the eagle pic of your PCB to only 16.5mm wide, it's still 34.6 long.. If i make it 19mm wide it's 40mm long. And in this size, the IC fits it's connections perfectly.

It still fits in my enclosure with enough space left for the batteries, so it's perfect..  I'm just trying to verify your results, because i don't understand, why mine is longer..

 

[IgorT]

Oh, does anyone know, if the 1N5817 Schottky diode would be good for this?

I have all the other elements in SMD form, except for the diode.. The ratings seem to match, the only difference is, that the forward voltage is 0.05V higher (0.45)..

 

[woop]

I showed your PCB to my engineer, and this is his response:

- The connections for the inducors are a little smaller than in the original (less cooling, more resistance)
- Inductor L1 is a little further away from the IC (more inductivity and resistance)
- The IC connections are a little close to the central GND for the IC (testing necessary to make sure, there are no short circuits after soldering)

Still, it should work just fine, as long as the right elements are used.

Glad to hear it. i was expecting more criticism
i need to change the pads anyway, because they are diferent to the MSS6132-472MLB inductor which is reccomended.
i guess i can swap L1 and C1. although they are so close anyway when you think of the scale of the thing. and the ic connections are just according to the datasheet. i will check it and maybe make the pads a little shorter

He also had an idea for the vias under the IC.. He suggested using a double sided board, the lower side only GND, drilling some holes under the IC, putting some copper wire, or resistor legs in them, soldering them on the lower GND plane, and cutting them away on the upper plane, flush with the PCB.

During reflow, the vias would then get soldered by themselves to the upper plane.. This should work for cooling.. Unfortunatelly i just bought single sided plates, so i'll go with my idea, and hope for the best.

i actually thought of that, but decided it would probably have more problems than it solved. like getting the wire to sit flush under the chip and not pushing it off the pcb. I donno. having a double sided pcb with the bottom layer all ground would help with shielding and slightly with heat dissipation even without vias under the chip, so i was going to use double layer anyway. i might try doing vias with solid copper wire.

BTW: I still don't understand the dimensions you specified (23.7mm high 19mm wide).. My AixiZ modules are 30.5mm long and 11.7mm wide. Do you have shorter modules? If i print the eagle pic of your PCB to only 16.5mm wide, it's still 34.6 long.. If i make it 19mm wide it's 40mm long. And in this size, the IC fits it's connections perfectly.

It still fits in my enclosure with enough space left for the batteries, so it's perfect.. I'm just trying to verify your results, because i don't understand, why mine is longer..

yup your measurements are better than mine. i made them inside eagle and somehow screwed it up.
my axiz module is 30mm long without the lens screw
the board i printed to scale is 37mm x 17.5mm (after measuring it with an actual ruler )

 

[woop]

I am going to have to redesign the pcb anyway to fit the proper inductors. we might as well use the layout in the demo board datasheet
http://www.national.com/an/AN/AN-1775.pdf
or just print it single layer with a bridge for the feedback wire

I hope you haven't etched it yet IgorT

 

[phenol]

Oh, does anyone know, if the 1N5817 Schottky diode would be good for this?

I have all the other elements in SMD form, except for the diode.. The ratings seem to match, the only difference is, that the forward voltage is 0.05V higher (0.45)..

im planning on using the same diode, 1n5819 rated for Vr=40V. i also have a bag /2-3kpcs/ of 16-Amp schottky rectifiers in TO220 but one would occuply like 1/2 of the total board space of the switcher.

 

[IgorT]

im planning on using the same diode, 1n5819 rated for Vr=40V. i also have a bag /2-3kpcs/ of 16-Amp schottky rectifiers in TO220 but one would occuply like 1/2 of the total board space of the switcher.

The 5819 has an even higher forward voltage (5.5V).. The 3410 datasheet specifies a diode with only 0.4Vf and it also says, that this diode can be the source of most of the efficiency losses, if not selected properly..

But if you think it should work, then great, at least i won't have to order a stupid diode from Farnell and wait another two weeks to get it...

I'll get both the 1N5817 and 19 and compare the results.


EDIT: My engineer says the 1N5817 should be perfect for this use, as it even has a slightly lower capacity, than the original.. The Vf difference shouldn't matter.. So i guess i'm all set.

 

[IgorT]

I hope you haven't etched it yet IgorT

Not yet, but i'm going to etch two now.. I don't have the correct inductors anyway, so they'll fit.

Oh, i see where the problem is now.. The inductor datasheet specifies only 2mm between the pads..
Oh well, looks like more drawing will be necessary.. And more label backing paper scratching.. But at least that means, the next batch will print even better...


On the other hand, i did just order some 4.7uH inductors from coilcraft, with a foot print, that fits your PCB perfectly...

The layout, the way it is in the datasheet would allow for different inductor sizes, so if you follow that, you shouldn't have any problems with any size inductors.

 

[IgorT]

I'm currently drawing the PCB the way it is in the original..

I'm just thinking of bringing the FB line onto the upper plane. I don't think there should be any problems, if i lead it up from the IC, under the coil, and back down to the resistor..


Any thoughts?

 

[woop]

ok I redrew the pcb. now it will fit the parts specified in the datasheet, I think.
this one is only 14 x 27.5mm
the second one is to fit big (by smd standards) diodes like the 1n5817

this design has a little less ground area for heatsinking so it is probably a good idea to glue a small heatsink to the chip or something (using a double layer board is probably a good idea just for more heat dissapation and shielding, just cover it while etching and ground it)
however it has lots of heatsinking for the inductors and they are nice and close to the chip
hopefully this one passes the engineer test

 

[woop]

here is a 600dpi bitmap suitable for printing to etch.
open it in paint, go to page setup and adjust to 15% size.
it should then print to scale

 

[woop]

I'm currently drawing the PCB the way it is in the original..

I'm just thinking of bringing the FB line onto the upper plane. I don't think there should be any problems, if i lead it up from the IC, under the coil, and back down to the resistor..


Any thoughts?

oh didn't see that post. i didn't do that in case the feedback wire picked up emf from the coil or track runing along side. its probably best to keep it under ground tracks with a bridge wire

 

[IgorT]

oh didn't see that post. i didn't do that in case the feedback wire picked up emf from the coil or track runing along side. its probably best to keep it under ground tracks with a bridge wire

You're right.. I'll lead it underneath..

I just finished drawing mine, came here, and saw you did yours too...

Now i have to decide which one to use..  Mine is pretty much the exact copy of the Evaluation board, only that i added some exposed GND plane on the left of the IC for cooling (the dog bone), for mounting a heat sink... I also added the elements, so their symbols will be etched into the copper..



Need to compare it to yours..


EDIT: Your new version is very very simple.. I like simple. Are the bitmaps mirrored already (for transfer) or not?

I also like the fact, that they would fit different size inductors..