Thoughts on How to Raise SMD LEDs Above the Board

[fuzzoli]

I have a bunch of 4014 SMD LEDs which will be mounted on a PCB.   Does anyone have thoughts on how to raise them off the board?  They would not need to be raised much -- perhaps 1-2mm.

(So I'm hearing the question of why the heck anyone would want to raise parts off a board -- good question!  This isn't for an electronic device, but rather for a lighting system for a model ship kit.  The 4014 LEDs are just about the right size at 1/350 scale to simulate ceiling florescent lights.  I'm trying to figure out how to make the lights appear to hang from the ceiling).

Some thoughts I've had:

• Tiny bits of wire soldered to the pads first, then hold the LEDs in place while remelting the pad/wire bit.
• Tiny bits of brass sheeting.
• Put a thin, non-solderable wire between the pads to "suspend" the LED while soldering.  Once soldered, slide out the wire.
• Solder super thin magnet wire to the LEDs, then, instead of SMD pads, have thru-hole pads.  The magnet wire would simulate the chains used to hang the lighting.  <-- Um, yeah, not going to happen -- too many LEDs!

Any votes for the above?  Any additional ideas?

Thanks!

[Alex Eisenhut]

light pipes

[nctnico]

Put a resistor or capacitor underneath them.

[fuzzoli]

Put a resistor or capacitor underneath them.

Oooo, I like this one!  Putting an 0402 zero ohm resistor on-end should actually give me the exact height and effect I'm looking for.

[Nusa]

Put a resistor or capacitor underneath them.

Oooo, I like this one!  Putting an 0402 zero ohm resistor on-end should actually give me the exact height and effect I'm looking for.

But then you'd have to line up and keep separated two of them for each LED.

Probably easier to stack components in parallel. A very low-value cap or a very high-value resistor in that case.

[james_s]

Use enameled wire of the appropriate diameter, you should be able to solder a pair of wires to each LED.

[JVR]

To prevent killing yourself, there are SMD machine place-able risers on the market for this sort of thing, pre-pregged with paste on the "top" side.

[ajb]

To prevent killing yourself, there are SMD machine place-able risers on the market for this sort of thing, pre-pregged with paste on the "top" side.

That's neat!  Do you have a link to an example?  I haven't been able to google anything up.

[mariush]

Get a stripboard : https://www.amazon.com/ST3U-StripBoard-3U-Uncut-Strips-Sided-x/dp/B00C9NXP94
Cut chunks the length of the led ... use a sharp knife to cut the copper and separate the strip in two segments
tin the segments and insert a couple of thin wires through two holes :

The thickness is 1.6mm i think, so you should get the high you want... could add a layer of copper foil over the stripboard to thicken it further.



Sort of ... you'll probably have only two holes on the tiny board, maybe 3 max ... there's like 0.1" (2.54mm) between center of holes...

[zitt]

At the risk of stating the obvious; are you derating the current thru those LEDs?
Given your planning to "rise" them from the PCB; you'll be effectively removing the direct to PCB coupling for heat transfer and the only source of heatsinking will be the power connections.

Given you're putting this in a model; which I assume will be glued together... heatsinking should be considered.
Either derate the current or find a way of coupling the heat from the LED die to the PCB.

Assuming these are standard SMTs rated for 20mA max... I'd about half that current to feel comfortable with the heat sinking of just a solder pad.

[james_s]

If he wants them to look like miniature fluorescent fixtures and not carbon arc floodlights I think it's safe to say they will be substantially derated.

[JVR]

To prevent killing yourself, there are SMD machine place-able risers on the market for this sort of thing, pre-pregged with paste on the "top" side.

That's neat!  Do you have a link to an example?  I haven't been able to google anything up.

Sadly not, I remember an FAE showing us the stuff last year. I'm pretty sure it was Wurth who made them.

[Doctorandus_P]

I have loads of 1208 resistors, and sometimes I need a resistor that's not in E12.
Then I simply stack another resistor on top of it (or even a 3rd).
The 2nd resistor is automatically centered on the first by holding them with tweezers during soldering.

This is probably easier to hand-solder than trying to balance 0402 on their sides.

[fuzzoli]

These are all great ideas!  After examining more photos, it appears the lights weren't hung but rather fastened to the ceiling.  If you think about it, having lights that hang (and swing) on a ship probably isn't the best idea.  Therefore, based on the attached photo, I'm going to try Nusa's suggestion and stack one or two 20M resistors beneath each LED (1k 20M @$0.01344 is only $13.44 USD -- not bad for 1000 of something) <= That was for 0402's.  1k 0603's is only $2.75 USD -- Even Better!

@james_s  - yup, you're spot-on...the LEDs most definitely won't be running anywhere near full brightness. 

I'll post some pictures in a few days.

Thanks!

[james_s]

Yeah those are definitely rigidly attached.

I've seen fluorescent fixtures on ships that were suspended several feet below the ceiling but always on rigid pipes, you wouldn't want them hanging on chains for obvious reasons.

This sort of miniature modeling is something I've always found fascinating but never had the patience or creativity to do myself at that level. When I was in middleschool I had a shop teacher who used to build model ships and other vehicles, I was always impressed.

[DaJMasta]

In terms of expense, I'd use low value capacitors rather than high value resistors, and definitely mounted normally on the board instead of on end - will be much easier to actually install.  <1000pF capacitors can often be had very cheaply because they get used in commercial/industrial designs and then when they pop out as excess reels to hobbyists, no one wants to bother.  It's nice that at least with 0603 and smaller caps, they're also usually as tall as they are wide, whereas resistors tend to be flat, so it should take fewer to get the right height.  Unless you're LEDs are transmitting information at several Mbaud, a thousand pF of capacitance across the LED will make really no difference to performance.

[Bucataru]

I have a similar issue. How do I raise this UV sensor (GUVB-S11SD) by 3mm or even higher? Light pipes, either glass or plastic, aren't transparent to these wavelengths of light.

[Mechatrommer]

Some thoughts I've had:...

if its for small batch, you can do whatever fancy technique like that. for more mass production alike, i would make pcb adapter on top of mother/main board.. like this...



reflow and hand solder friendly, you can stack them easily 2,3 or more for more height... fiddling with smd resistor stack is not fun. if LED requires heat transfer, you can design pad underneath the LED and then vias right down reflowed to main board. ymmv.

[Nominal Animal]

I have a similar issue. How do I raise this UV sensor (GUVB-S11SD) by 3mm or even higher? Light pipes, either glass or plastic, aren't transparent to these wavelengths of light.

It's footprint is 3.5mm × 2.8mm, centered on two 1.3mm × 2.3mm pads separated by 1.7mm.  The center-to-center pad distance is exactly 3mm.

I'm only a know-nothing hobbyist, but I would convert the sensor from SMD to through-hole, with at least two male pins (at standard 0.1" = 2.54mm pitch) on the sensor side, and a female header on the board.  I know of three ways to do this. 

One is to solder a 90-degree two-pin male header (standard 0.1" = 2.54mm pitch) on the sensor directly.  It should be doable by hand.  If you use through holes on the board, depending on the pin length you use, you can get 3mm – 7mm raise.  Note that the pins will be off to one side, not centered with the sensor, and accurate placement can be a bit difficult.

The other is to create a set of tiny boards in a strip or panel, with the sensor on one side, and pads for an SMD header, connector, or wires on the other side, populate and solder them, and then cut each one out.  If the UV sensor pokes through the enclosure, the helper board can be used (perhaps with tiny screws) to connect the sensor to the enclosure, and just use wires to the board.  The board alone gives a 1.5mm - 1.8mm raise, but since it can be larger than the sensor (say 7mm square), you can then easily use e.g. single through-hole pins (one on each side of the sensor) to raise it as much as you can find long enough pins.  You might use three pins for stability, and heatshrink to cover the bare pins between the two boards.  Or, for more flexible placement, just use wires between the two boards.

The third is to make your own riser blocks, for example by gluing two layers of one-sided PCB material together, conductive surfaces outwards, in strips; or by casting copper strips in suitable non-acid curing electrically insulating rubber or silicone.  Both yield one long riser block you then cut (shear and sand) to the desired lengths.

For proper production, I think the second/middle choice is the most sensible one.  But again, I know nothing.