At 11:51 PM 14/04/2011, you wrote: >Round two: > >5mm LED manufacturers provide severe warnings about the dangers of >stressing the LED body during installation or in service. Probably because they don't want to be blamed if you crack the epoxy. ;-) >The industry standard arrangement is to have two separate mechanical >parts one per lead, with alignment and location being carried out by >the epoxy body prior to soldering and by the soldered leads plus body >post soldering. Yes. >Forces on the leads relative to the body pre-soldering >must be resisted by the epoxy. Post-soldering the PCB/leads/body >alignment must not place excessive stress on the body. During >soldering and post-soldering when the epoxy body or the leads entering >the body are at elevated temperature stress tolerance is even lower. Not sure about that.. the epoxy may be a bit softer and less brittle. >LEDs may have "stoppers" - small protrusions on the leads designed to >stand-ff the LED from the PCB by a millimetre or few or may be >stopperless. Best practise and manufacturers' assumption are that >stopperless LEDs will be spaced off the PCB by some other means BUT it >is near universal practise in low cost and / or low quality >equipment to use stopperless LEDs and to push them home against the >PCB surface as closely as possible prior to assembly. Yes. Although it is rather advisable to use BIG holes for this purpose, and perhaps a single-layer board (or pseudo-single layer- unplated holes with no top pad and a grossly oversize filleted bottom pad, for example). Otherwise there is no room for differential thermal expansion between the LED epoxy (large) and the PCB (rather small in the X-Y plane, and much larger in the Z plane!) ______ >Given: LEDs are mounted either > >A) With "stoppers" so that a mm or few of lead exists between PCB and >base of LED. > >B) Without stoppers and pushed home so base of LED is almost flush with PC= B. >("Almost" because leads have epoxy protrusion at lead to body join - >in practice this may be small to absent.) > >_____ > >If group "A" LEDs have "enough" force applied at right angles to body >and at 90 degrees to the plane that the two leads are in, the LED will >bend on its leads, pivoting mainly around a bend near the PCB surface. >Medium finger pressure sufficeth. Dropping a small PCB with say 10 >LEDs on it from 600 mm / 2 feet onto a wooden surface may cause >noticeable bending. Casual contact and "pushing around a work bench" >is not likely to cause noticeable bending. Maybe with only a mm or two. With ~10mm of lead, they bend rather easily IME. >Empirical tests, sample of few, brand "N" 5mm LED, 0.5mm square leads, >suggests that 2 Newton force (200 gram, 7 ounce) tends not to misalign >LED overly noticeably whereas 300 gram almost invariably does. > >Applying force along the plane the leads are in will only bend the LED >if a lead buckles - a very large force and liable to cause immediate >and obvious damage to the LED body. If you're lucky. _____ >Group "B" LEDs will not be visibly bent by any forces liable to be >experienced in reasonable handling. The LED body resting on the PCB >and/or the much greater stiffness of the very short leads (<< 1 mm) >will resist such forces. Where the LED body sits on the board a moment >arm is formed between the two leads to resist the moment arm formed >by the LED body. A force multiplier results with force applied to body >being multiplied by about the ratio of > > LED_length / interlead_spacing. > >Maybe 5:1 to 10: depending on various factors. > >So a 3 Newton / 300 gram force may result in a say 1.5 to 3 kg pull on >one LED lead. > >Where the LED body does not sit on the PCB and when force is along the >plane of the two leads one lead acts in compression and the other in >tension to resist rotation. Similar forces to the above result. > >Impulse forces from dropping may be higher. > >___________ > >In real world situations, LEDs mounted on stoppers may be bent to and >from with reasonable abandon without immediate evidence of damage. > >LEDs mounted on the PCB surface without stoppers may have forces well >in excess of that required to bend stoppered LEDs (as above) with no >apparent signs of damage. > >In past equipment, all or mainly hand soldered, I see a very small >percentage of LED failures with no apparent correlation with soldering >methods. > >__ > >Questions: > >How significant a real-world issue is LED mechanical installation >damage caused by rough handling.? > >If a stoppered LED is bent how much risk is there in rebending it? > >How bad is on-surface mounting compared to using stoppers in real >world installations. > >Are the manufacturers "gilding the Lily" (or leadening it) and is the >problems smaller than they suggest? > >How much effect do elevated post soldering stresses have? > >Any real world feedback? In quantity, I've used extended leads (held by a jig and hand soldered) with 3mm LEDs and soldering right down to the board with 5mm LEDs (both hand and wave soldered). Never used the leadframe standoffs- they've never been at a convenient height/tolerance and might constrain sourcing. No failure at all that I can recall on the 3mm LEDs and a few (no field failures that I can recall) on the 5mm LEDs. There may be some issue with solder pooling under the part when dip soldering with large holes (for ease of assembly). I can't recall a single mechanical failure of the LEDs stood off by 10mm or so, except maybe for the sample that our Golden retriever gnawed on for a while (she was subsequently involved in an incident involving contact with a moving car, and when the vet X-rayed her bruised body there were a couple of components clearly visible in her stomach!) She did recover and has outgrown both tendencies. >Best regards, Spehro Pefhany --"it's the network..." "The Journey is the rewar= d" speff@interlog.com Info for manufacturers: http://www.trexon.co= m Embedded software/hardware/analog Info for designers: http://www.speff.co= m --=20 http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .