Potting LEDs Under Pressure

[LED Boatguy]

Hey there: Got a project going that involves potting LEDs in an optically clear compound. For best results, they say to "pressure cast" the light fixture at 60 PSI while the potting compound hardens--I guess to minimize bubbles? Devices to be used are CREE XP-G and XP-E as well as some linear regulators. There will NOT be any electrolytic caps in the potting area.

Anyone see problems with this?

Thanks.

 

[Mark620]

Anyone see problems with this?

I do not see a problem with potting the topside of the led but encasing it is a problem. heat problem?? what about the heatsink? air flow across. need more info.

 

[LED Boatguy]

I do not see a problem with potting the topside of the led but encasing it is a problem. heat problem?? what about the heatsink? air flow across. need more info.

All electronics will be mounted to the back of the "box" and the "box" will be submerged in water. Of course I'll add thermal protection to ground out the enable pins if temps rise.

 

[East Tn Bowman]

All electronics will be mounted to the back of the "box" and the "box" will be submerged in water. Of course I'll add thermal protection to ground out the enable pins if temps rise.

curious, since i'm a new guy to this led stuff, what sorta application is this underwater box gonna be used for? and any ideas on how far out the light will go say in a lake or pond? all this stuff sounds cool to me just need some guidance, well a whole lot of guidance if i'm going to make this work

 

[spencer]

But from the back of the LED you NEED a thermal path to a heatsink or to the water.

 

[Mark620]

But from the back of the LED you NEED a thermal path to a heatsink or to the water.

If the "heat sink" is enclosed it is a thermal mass.

 

[LED Boatguy]

But from the back of the LED you NEED a thermal path to a heatsink or to the water.

There is a direct thermal path from the MCPCB to the case, and the case is submerged. Heat is not an issue. I'm more worried about 60 PSI and where it will squeeze the potting material.

 

[HarryN]

Potting the LED is a good idea, but just keep in mind that this will dramatically alter the optical characteristics. Normally, you want the encapsulant to have a higher RI than the encapsulant on the package, which is not that easy.

Not all LEDs can deal with the shear forces involved in a pressure potting operation, so I would double check with the supplier to make sure that the dome will not come off.

 

[SemiMan]

Pretty much a given you are kissing any warranty on your LEDS good bye. You will run into a few issues perhaps.

1) Light loss, change in lighting pattern, etc.

2) Possibility of premature failure due to bond wire stress due to thermal cycling as well as general thermal stress.. You have taken a soft dome that is meant to be pliable and essentially hardened it in place. That is a big no-no as far as the LED makers are concerned.

The LED makers have been investigating clear potting compounds as there is a need. The big thing for them is of course reliability. I am not aware of them approving any potting compounds, though I vaguely remember someone giving their blessing to a conformal coat that I believe was pliable.

Semiman

 

[saabluster]

Pretty much a given you are kissing any warranty on your LEDS good bye. You will run into a few issues perhaps.

1) Light loss, change in lighting pattern, etc.

2) Possibility of premature failure due to bond wire stress due to thermal cycling as well as general thermal stress.. You have taken a soft dome that is meant to be pliable and essentially hardened it in place. That is a big no-no as far as the LED makers are concerned.

The LED makers have been investigating clear potting compounds as there is a need. The big thing for them is of course reliability. I am not aware of them approving any potting compounds, though I vaguely remember someone giving their blessing to a conformal coat that I believe was pliable.

Semiman

The warranty and number 1 yes but the rest no. The silicone dome is not designed to move around. As a matter of fact you are specifically told not to do that. The bond wires are remarkably robust as well. I don't see a problem with what he is trying to do.

 

[SemiMan]

The warranty and number 1 yes but the rest no. The silicone dome is not designed to move around. As a matter of fact you are specifically told not to do that. The bond wires are remarkably robust as well. I don't see a problem with what he is trying to do.

I would agree with you if you were right ......but...... I would suggest contacting the LED manufacturers who will specifically tell you NOT to do this as the silicon dome is flexible for a reason ... in order to absorb thermal stress. If you put a hard coating around it, then you take away some of the ability to do this... that will be especially true if you cast under pressure. There is a difference between physically "moving" the dome, and allowing the dome to naturally expand and contract slightly with thermal cycling.

In automotive we have wanted to pot/conformal coat on top of the LEDs for years but do not for just this reason.

Semiman

 

[saabluster]

I would agree with you if you were right ......but...... I would suggest contacting the LED manufacturers who will specifically tell you NOT to do this as the silicon dome is flexible for a reason ... in order to absorb thermal stress. If you put a hard coating around it, then you take away some of the ability to do this... that will be especially true if you cast under pressure. There is a difference between physically "moving" the dome, and allowing the dome to naturally expand and contract slightly with thermal cycling.

In automotive we have wanted to pot/conformal coat on top of the LEDs for years but do not for just this reason.

Semiman

LED companies are just being conservative. If we listen to them we also can't run an XR-E over 1000mA. They can't account for the vast array of materials and methods used to do this potting so yeah it only makes sense for the official line to be no potting. I do a lot of casting and have many projects going with potted LEDs and nary a problem. I pressure cast as well. There is no significant difference to the LED whether or not you pressure cast or not. Like I say though it depends on the materials and methods.

And it's not like the LED die is in there jumping all around now is it? Keep in mind the LED die is already potted so if potting was bad than the manufactures are all screwing up! Like I said don't listen to everything people tell you. Test it yourself.

 

[SemiMan]

LED companies are just being conservative. If we listen to them we also can't run an XR-E over 1000mA. They can't account for the vast array of materials and methods used to do this potting so yeah it only makes sense for the official line to be no potting. I do a lot of casting and have many projects going with potted LEDs and nary a problem. I pressure cast as well. There is no significant difference to the LED whether or not you pressure cast or not. Like I say though it depends on the materials and methods.

And it's not like the LED die is in there jumping all around now is it? Keep in mind the LED die is already potted so if potting was bad than the manufactures are all screwing up! Like I said don't listen to everything people tell you. Test it yourself.

The LED manufacturers have been trying out many different potting/conformal compounds with hopes they could recommend one that did work. Unfortunately that is not the case. Perhaps in a limited temperature swing environment where life is not as much of an issue that would be fine. However, certainly for automotive which uses the full temperature range of the LED, that is not the case.

It is not a matter of the die "jumping around", it is a matter of thermal cycling pulling ever so slightly on the bond wires each and every time.

If this is a one off, go for it, you are not likely going to run into issues. If it is a narrow temp range, small run commercial, likely the same thing. If you intend to run any number of them over a wider range of temperatures, then buyer beware....

 

[saabluster]

The LED manufacturers have been trying out many different potting/conformal compounds with hopes they could recommend one that did work. Unfortunately that is not the case. Perhaps in a limited temperature swing environment where life is not as much of an issue that would be fine. However, certainly for automotive which uses the full temperature range of the LED, that is not the case.

It is not a matter of the die "jumping around", it is a matter of thermal cycling pulling ever so slightly on the bond wires each and every time.

If this is a one off, go for it, you are not likely going to run into issues. If it is a narrow temp range, small run commercial, likely the same thing. If you intend to run any number of them over a wider range of temperatures, then buyer beware....

You still did not address my last point about potting at the factory. So have you seen their testing data? Who did it? What did they try? I guarantee you it can be done and easy. What have you looked at potting in the past? Headlights?

 

[bshanahan14rulz]

I've ripped the dome off of an xp-e and had one of the two - bondwires stay intact. I've potted a domeless XR-E and it runs fine (isn't to say the epoxy I used worked fine, it didnt. Turned brown. But the LED still lights)

I'm thinking that if you, unlike me, actually looked for something that doesn't expand with heat and doesn't brown, you'd get pretty good results.

As far as pressure is concerned, XP has a tough dome. I've heard people describe some domes as gummy, but I wouldn't describe the XP as gummy. I'll be straight, I haven't personally felt the LEDs that others call "gummy" but I don't think they'd use that word if it was the same stuff as what's on the XP.

Semi, if what you are saying is true, that the encapsulant would thermally cycle and push on the dome enough to shear the bond wires, then wouldn't XR be a better bet, since if you push on the dome, it just pushes on the ring, and the bond wires are in a gel, giving it more flexibility.

Either way, encapsulation has many uses and if we can figure out a way to do it reliably (instead of just going with what the manufacturer says), then we rock! and if not, then we know for ourselves that it won't work, and won't have just taken the company's word for it.

 

[spencer]

I've heard people describe some domes as gummy, but I wouldn't describe the XP as gummy. I'll be straight, I haven't personally felt the LEDs that others call "gummy" but I don't think they'd use that word if it was the same stuff as what's on the XP.

I would say the SSC P4 and P7 have gummy domes. I don't know what the XP series has.

 

[jashhash]

Actually the biggest Problem I've experienced with something like this is that the dome of the LED gets pretty hot. Hot enough to melt or brown certain types of plastics. I would recommend using Clear silicone as a potting material since it is both flexible and can handle high temperatures.

 

[Steve K]

As a practicing EE who deals with potting materials on a regular basis, I would discourage this approach. Potting is one of the tougher challenges in packaging electronics.

It sounds like you are using the potting for mechanical protection, which means it needs to be pretty hard. This is typically a hazard to electronic components, especially the more brittle parts like ceramic caps. It might be less of an issue if the only electronics are LEDs.

The other trick with potting is getting it to adhere to everything. There are all sorts of high tech and funky processes that may be needed to get the potting to reliably stick to the housing. Problems don't usually happen right away, but take time and thermal cycles to show up.

Getting the bubbles out of the potting is an issue too. Drawing a vacuum on the device after potting does help.

If I was given a choice, I'd avoid potting unless there was a large cost involved.

Steve K.
(I'm in the middle of dealing with a potting issue on a sensor right now, and it's a giant PITA)

 

[SemiMan]

You still did not address my last point about potting at the factory. So have you seen their testing data? Who did it? What did they try? I guarantee you it can be done and easy. What have you looked at potting in the past? Headlights?

High powered LEDs are not potted at the factory. Low powered LEDs are, but the bond wire is actually pretty substantial compared to the current and w.r.t. to current/bond wire size for a high powered LED.

What have I potted.... just about everything in a transportation vehicle that can use a LED .... in high volumes. To that end, I do tend to get the ear of the LED companies and I know that many compounds have been tried and I have even suggested some.

You are guaranteeing it, but I can pretty much guess you have not done extensive testing over wide temperature ranges with extensive temp/on-off cycling and with long duration lumen maintenance testing and over a reasonably large sample size. Purely for test purposes, I have potted over many a high powered LED..... even did a custom job for a customer for an outdoor application even though the customer was warned it was a bad idea and it came without warranty. Guess what, several years later, the failure rate is starting to become an issue.

I suggest you pick up the phone and talk to Lumileds, Cree, Nichia, Osram, etc. and ask them what they thick of the idea. Certainly Lumileds and Osram have very significant automotive experience and hence potting experience and customers who would like to pot.

Semiman

 

[SemiMan]

Actually the biggest Problem I've experienced with something like this is that the dome of the LED gets pretty hot. Hot enough to melt or brown certain types of plastics. I would recommend using Clear silicone as a potting material since it is both flexible and can handle high temperatures.

This is another interesting point. Cree has released reliability data that shows a correlation to not only die/case temperature, but to ambient temperature. Need to check my notes, but for the most part, optically clear potting material would not be overly thermally conductive. On piece wise basis better than air, but air moves.

Semiman