EEVblog
HTML-код
- Опубликовано: 11 окт 2024
- Dave discovers a potential systemic manufacturing issue with Lifud LED light drivers.
A repair of two of Dave's failed Lifud LED studio panel light drivers.
Forum: www.eevblog.co...
EEVblog Main Web Site: www.eevblog.com
The 2nd EEVblog Channel: / eevblog2
Support the EEVblog through Patreon!
/ eevblog
Stuff I recommend:
kit.com/EEVblog/
Donate With Bitcoin & Other Crypto Currencies!
www.eevblog.co...
T-Shirts: teespring.com/s...
Likecoin - Coins for Likes: likecoin.pro/@...
💗 Likecoin - Coins for Likes: likecoin.pro/@...
I believe it is systemic to lead free solder. I work on Motorola products, and that's 75% of the issues. Usually on connectors between boards, then SMD's or even flat pack legs. The rest of the issues are switches, chip/trans failures, pots, knobs and buttons roughly in that order.
But, even so, in 1985 I worked on an old box of a 5 watt radio that was outputting less than a watt. I saw a bad solder joint, then just resoldered the whole board and bam, got all of the 5 watts back. That was 6 months into my first electronics job ever and the boss was amazed and I was somewhat proud of that fix.
Tricks of the trade
The micro cracks strike again!
My old truck had a faulty ABS module (over $1000+ to replace at the dealership!). After seeing lots of people claim its just faulty solder joints, i took the task to take it out, take it apart, and clean up the joints. Its a big pain since the unit is all sealed up with silicon, but the solder joint touch up did the job. Before the fix, every time i hit a bump, the abs light would come on. Just guessing the vibration + bad joint would cause the fault to trip. Who knows how many people were out $1000+ because of this.
Never saw someone so happy to find a cracked solder joint :P
Except maby myself lol
Great vid thanks! :)
MEAN WELL LED Power Supplies!!! Go for them in future. Can't go wrong. I've had good results from all the Mean Well LED power supplies I've used. Many years in with some of them, all still going strong. Great build quality.
I repaired similar thing in my LED power supply. Apart from cracked joins I had to add solder to main IC, because half of legs wasn't touched by wave at all. The satisfaction is always there, when you find that thing.
Lead-free solder is too brittle if a bit of thermal cycling between what, 18 degrees and 27, is enough to crack joints. Nice to hear you working the rabbit/herring connection.
2:25 Dave, if you'd allow me an advice: I always solder the end of multi stranded wires before placing them into termination blocks. (1) You never have to worry about oxidation, (2) You need less torque on the bolt for the same result (less mechanical stress for plastic parts) (3) Once the bolt is tightened, it makes a dent on the wire, that provides an extra security for slipping out like what we've just seen. In my lab, all wall sockets, wall plugs, light switches, extension leads are mounted like that.
Or you simply just ensure you screw it up tight enough.
I reckon it’s the lead-free solder Dave. I repair a lot of audio gear and the newer stuff with lead-free solder often exhibits these types of faults. Older gear with leaded solder tends to last a lot longer before these sorts of faults appear.
Yeah, same here. Old stuff soldered with leaded solder just keeps working. Lead free is just too brittle. I hate lead free solder.
Liquid Audio
deffinetly a big factor.
Desoldering lead-free stuff is brutal! Specially when they use tiny amounts!
Simon Tay Me too, hate lead-free solder!
Nice bandwagon. Mind if I hop on? Gods bless LEAD!
Interestingly in the old days, in television horizontal stages and switchmode power supplies with high pulse currents, localised heating causes the tin in leaded solder to migrate to the outside, leaving a 'leady' high-resistance joint that eventually fails. Used to be the one of the first thing to check in a dead telly... :)
This video helped me find a faulty solder joint on a cheap 12V plug pack. Thanks!
Thanks Dave, just had a defective Lifud LED panel light and PSU on the bench where all components measured good but it would still not power on nor generate anything near the specified 55VDC output. You gave me the clue I needed. Despite using a microscope, I could not find any broken joint but it's up and running now that I have redone just all of them. 🤞
I might have spent around the equivalent of 100 AU-$ on a new power supply, and discarded one where all except the soldering was in best shape, if it wasn't for you. Thank you so much! 👍
This is why we love LEADED solder!
If you suspect a solder joint problem try probing around with an insulated poker on the top components. It often won't take much to make or break the connection and find the faulty part. Then just re-solder all its leads.
Great vid for a Saturday night! Thanks Dave!
Gah, that's pretty rough.
Potential for mass quantities of those getting borked in no time, and Dave might need to open up to fix them multiple times until most components have been resoldered.
Awesome footage. Dodgy solder joints can be quite troubling. I fixed a very old EZflash 3 in 1 by simply removing everything and putting everything back.
With the first cracked joint we can see that there was not much solder once you heated it, too thin and just cracks.
That hole was too big, and because it's single sided PCB, without plated holes, it haven't got filled properly.
Could be also because of oxidation on the components(stored for long time).
Very good repair video Dave!.
I had to repair a wine cooler and the amount of dry joints/cracks was remarkable.
I love these fix it videos!
after seeing your ceiling, i feel much much more comfortable with my light installations :D
I have found that large lead parts need a couple of seconds more in the wave solder machine to prevent a cold solder joint.
This was found in failed power supplies for gaming equipment at a manufacture i worked for several years ago,
Dave, you make electronic repairs look easy! I'd never have figured out that problem.
Great video Dave.take care mate...
17:00 "Is that because of this new fangled lead free rubbish" Yep, that is exactly the reason. Lead free solder is brittle and doesn't flow as well as leaded solder. Also their wave soldering is crap. Wave soldering never properly solders large high thermal mass components like power transistors and coils. Large components should always be hand soldered.
Ive lost count of the number of single sided lead free wave soldered boards just like yours that ive repaired just by resoldering cracked joints. On some power supply boards in LCD TVs that I've repaired, ive resoldered every joint with leaded solder because the wave soldering just didn't apply enough solder.
@Simon Tay
> Lead free solder is brittle and doesn't flow as well as leaded solder.
Bullshit. If your solder is brittle, then you're using the cheapest China stuff. Modern top shelf lead-free solder has much better properties than the best leaded solder alloys that can be bought nowadays, since vendors stopped producing the above average leaded stuff due to massively dropped demand.
> Also their wave soldering is crap.
That! Could also have happened with leaded solder but it is much more forgiving to manufacturing problems like that.
I only use leaded solder. I mean the existing lead free solder on the board. It cracks because it is brittle and there not being enough used. Failure rate for lead free solder is definitely higher than leaded solder. Leaded solder doesn't crack as easily.
Since the advent of lead-free solder, on anything like this, with only maybe 50 or 60 connections, if I suspect soldering issues I just reflow everything on the board with leaded solder. Takes maybe 5 to 10 minutes at most and problems solved. If I try to hunt for individual ones, I've wound up with something that works for another month or two, then I'm right back in there.
Not only is the lead-free a problem, but I'm finding that on through-hole stuff they're too skimpy with the solder. They're barely maintaining the physical connection, and vibration easily causes cracks.
Probably wise to go round and check the wiring on all the drivers now you've found that one.. If you can mess it up once you can mess it up twice.
Monkeh Also, make sure to follow the precise cutting lengths specified on the lid, or the wires will tend to come loose!
"we have a bona-fide failure! Beauty!" haha this crowd welcomes them!
Great video Dave...Thanks!
Quality repair vid. Thanks Dave
I would agree its most likely the solder that's the issue, nicely figured out.
Had a Sub in a vehicles stock audio system not working, or sometimes working intermittently. A board mounted next to and connecting to the sub was my first spot to check, and after removing it, found the same problem, a cracked joint on a relay. Though, mine was much smaller, but after a quick re-flow of it and a few others just to be safe, popped it back in and it worked, good as new.
Wow, there is no way I'd have spotted that with my crappy eyes, probably not even WITH the microscope. Nice work, Dave-O!
Exactly, you would have swapped out the control unit. An hours labour? what is the unit worth.
That's a feature not a problem
Oh, it's the "please buy a new one" flicker feature...
"undocumented feature"
Hahaha, "working as intended"? :P
No it is disco mode
i get this flickering problem with a bunch of high bay 250w led fixtures. the problem is that they have poor design. using 12 36v cobs in 6s2p. when one or more led burns out the driver will start flickering the leds.
The most sophisticated method to check all solder joints simultaneously: smack it into the table repeatedly :)
If you cant bang it like that then the age-old technique is to whack each component with the back of screwdriver handle.
3:00 That tends to happen with stranded wires when they're twisted up and put into bare connectors, or even crimp connections. While it's easier to get them in there because the strands won't bend off as much when you're out of alignment, the issue is the twisted up wire end actually has a larger diameter than if the strands were all straight. Over time it can untwist, get stressed, etc. and then just fall right out!
4:35 that's a very effective way of killing an SMPS...
The first guy had the Line/Neutral backwards!
noticed that as well
Yeah, noticed that too. Even though it makes no difference because the incoming AC gets rectified to DC anyway, its not good practice.
Simon Tay hehe ac trips me out
That as well.
Simon Tay Only the live is usually fused though.
Looks like cold solder joints from the higher melting temperature of pure tin is the most likely root cause, and a manufacturing process that didn't get it up to full flow temperature. Looks like these assemblies were hand soldered vs wave or reflow.
ROHS is the dumbest idea spawned by politicians of the EU. And is having the opposite effect of more electronics thrown away prematurely. They didn't take into account the amount of toxic chemistry it takes to make integrated circuits, and the printed wiring boards. All they thought about was the lead being tossed into landfills. A better solution would have been a requirement to recycle all electronics.
At least the industry I worked in (avionics) was able to get a waiver on pure tin solder. Mainly because of the tin whisker problem.
I don't think this has anything to do with lead free solder. This problem has been around since well before ROHS was even thought of. This is a typical problem with using cheap single sided, non plated through PCBs. I first saw the problem in portable CRT televisions donkeys years ago.
The problem is always with heavier components (typically inductors/transformers) that are 2nd stage (hand inserted) just before the flow soldering process. As you will see in the video, the holes for these components are much larger that you would expect, for ease of component insertion, however this means that the only thing holding the component in is a very thin layer of solder between the lead of the component and the pad due to the reduced meniscus. Thermal cycling and vibration does the rest. If the PCB had PTH's then this wouldn't have been a problem!
Also, large heavy components such as coils, power transistors and caps should always be hand soldered. Wave soldering just doesn't apply enough solder to these because of the higher thermal mass.
On the cover 4mm lead length from outer insulation to stripped wires. By exceeding this yo are putting the cable outer insulation outside the clamp, or putting a stress on the leads as they 'S' bend between the end of th cable and the clamp.
As an FYI, If you're based in Australia, those led drivers are missing the RCM mark. Australia national deviation requires a higher glow wire test of 750C on the plastics, EU only require only 650C. Please be careful!
Every other RUclipsr when their LED lights fail: "Well no filming for me until this is fixed". Dave: "Ohh a chance for another repair!"
Free standing (large mass) components soldered to a one sided PCB with no thru-hole vias using lead-free solder.
Throw in some temperature cycling and heating on that nice big copper wire. I'm guessing that copper layer is pretty thin too.
Yep. What could possibly go wrong?
Good catch on the cold solder joints. I used to have to try and find those buggers without a microscope. Mind you my eyes were only 20 years old.
Cheers,
- Eddy
It's not the years, it's the mileage...
The reason for the flickering.. You said it was the active/live that had come lose from the terminal block. Well judging from the color then yeah your right. But i can guarantee you the receptacle on the other end of the cable isnt polarized. So the flicering would have been the because the live was feeding voltage into the caps and there is enough ground potential with the return so close to the terminal block. Its just caps charging up and dumping the power.
I hope your next video is addressing the wiring situation in your ceiling...
Lead free solder just sucks. thank you ROHS for destoying electronics, and therefore creating more junk and more polluting!
Jostling electronic devices is a valid diagnostic test; I learned from EEVblog hahaha XD
For academic history preservation, you should duplicate all those decks to an off-continent similar museum along with photos of any written or printed labeling. Do the same for them.
"u feel lucky....punk....." hahaha LOL! now i'm gonna have to go to my old vcr tapes and watch some of the dirty harry movies! i must have seen them 100 times, still love them!
I think after the first couple of dodgy joints, I wouldn't have bothered looking for bad ones, I would have just run across all the through holes with proper leaded solder.
Damn lead free crap, I swear it's the cause of 90% of failures.
It's much easier and faster to reflow everything instead of searching for cracks.
True! But it's also nice to know what the fault actually was.
Unless a proper stranded connector is used. In this case, the power supply used a really crappy spring clamp type connector, which is allowed for use with stranded connectors, its just not good at it. You are correct though, when only a screw is present, such as in electrical distribution boxes on the rails or some styles of breaker just use a screw, only solid wire can be used.
Same screw clamps in just about every breaker box that i've come across, designed to be used for solid copper. Maybe they're trying to accomodate solid AL wire with the springs too, but I'd at least try to tin anything stranded, even with leadfree solder, before trusting it under a clamp.
These connectors are fine with stranded wire as long as you don't twist it before you put it in. The issue is when you twist up the ends of the wire it has a larger diameter than if you leave the strands all straight. Eventually it can just fall right out, especially if the wire is bending coming right out of the connector into insulation bundling multiple wires together.
Lead free solder sucks for copper plumbing too. The first time I used it about 10 years back it was some wonky Cu-Se-Bi mix. Took me 4 attempts to sweat a half-inch copper pipe on to a brass valve. Previously I'd had about a 1% failure rate with with 60/40 in. I've since changed solders, and keep my 60/40 for copper drain lines, where it is still legal here.
18:00 So, "how ya doin" = bad or dodgy. Right?
Gotta love dry joints, when i repair vintage TVs i always see them
Those leaf clamps...unscrew a way and push a small screwdriver in the hole to align the leaf, allows for easy entry and grippier connection - need 'priming'
When I see a single slightly sketchy solder joint on a circuit board, I just reflow every joint on the board with leaded solder and fresh flux. Even if a joint is not broken now, it will be broken soon and you'll have to fix it again unless you improve them all at once.
In old days we had similar problem with our color CRT TV and its remote would stop working and service guy came and just resoldered the joins to fix it. Later he got fed up and changed the whole TV under warranty. Throughout the video I could not digest that it would simply be a disconnected main wire to cause flicker.
Davo, would you recommend soldering the copper on the ends of those mains wires to give the clamps more to squash down and overall conductor density? Once again, another good presentation... 10/10
This why you use ferrules on cables.
Or just suitable screw terminals. No need for a ferrule if the terminal is suited to fine stranded wire.
ferrules also look better. The flair can be a nuisance for closely spaced connectors but in this case would be ideal. IMO opinion it is the better way to terminate a connection.
uptherewhere Ferrules with flairs won't fit mechanically in the space between the cable clamp and the wire clamps. Simple ferrules might fit but are against manufacturer specifications as printed on the lid.
What?! No bootlaces? The outrage!
Tradie Trev my 1st thought. Bloody amateur elec-chicken.
Nice wiring neutral in the active terminal and active in the neutral.
It's Dave's Disco Lab.
I have some test equipment i made my self, the live goes threw a cheap rocker switch and the unit it self is very low power, when turned off the backlight on it flickers every couple of minutes due to capacitive coupling
I used to tin wire ends with solder until someone told me it wasn't safe, so I bought a six sided ferrule crimp tool and now whenever I'm wiring anything, I always use ferrules.
Used to have solder joints go bad all the time on TV flyback transistors. Joints on components which thermally cycle a lot are worst, but there might be something with high peak currents too. Solution - use a plated-through board!
Nice repair and nice find! You fixed both using only a soldering iron and a screwdriver 😛
It looked like the solder was originally about 0.05mm thick which pretty much wouldn’t survive a sneeze!
5:09... Another problem that I can see is the bare wire strands are not all parallel and twisted together to get a good grip by the Blue Connector clamp.
I like your way of thinking.
This is a serious problem. Solder cracks and cracks in vias due to thermal expansions and this happens due to the bad selection of PCB raw material and matching PCB soldering process. For example a TG150 rated PCB material is better than TG135. The higher the number, the lower the Z axis expansion. We went down the rabbit hole for months to find out this problem.
6:45 he's happy about a broken PSU, that's Dave :-)
The tap test, always revealing.
praestant8 tell that to your gf
Thanks Dave 👍👍
The joys of lead free solder - now you know why the aviation industry is still allowed to use leaded solder.
To those whinging about it not being up to USA code. Very good as this is in Australia.
you guys make your homes out of wood and foam ....... your literally living in a fire hazard
Fouzai alaa. Exactly. Houses in europe, Australia and the UK are made properly with solid thick brick walls with caverty insulation, not thin walls made of plywood and plasterboad with no insulation.
KeeDx3 Oi mate! Don’t argue with these fine gentlemen! It’s their right as Englishmen to bitch about how shit America is, an’ how great the homeland is! ...and don’t mention the BBC! I hear they go bonkers at the mere mention!
walls are made by 2x4 or 4x6 depending, im sure if you wanted you could make em thicker. I know id like a brick castle style house if i had the money though
Kebabrullerne There, there, lad... let it out. Now tell him how uneducated and uncultured he is! You’ll feel much better!
Some ferrules might help with the screw connectors.
Hey Dave, proposal: a video about different solder leads and their pros and cons?
Seems like the HASL surface finish of the PCB somehow did not match up well with the wave soldering? It seems like the tin on the pads did not melt all that well on large components and form a good joint... just made it out good enough to survive for a bit.
Maybe the shipping case was damaged in transit or dropped? That would be more likely considering these are from a relatively trusted manufacturer. All the cracked joints were on high-mass components, so I'd suspect blunt impact caused the initial damage; thermal cycling just finished the job.
Every country has its own ways to do electric installations. But at least in Finland, you are supposed to put crimped sleeves on stranded wire that is connected to a screw header, yes it has the metal between the screw and the wire but still, at least you should have tinned the ends, an installation like that of course would actually need solid core cables, since it's not movable, but I'm pretty sure you know that already.
You should use pins for conneting multi strand conductor to screw terminal, in Europe that is the norm.
A friend has a broken Philips pc monitor. I wonder if I should give it another day in the lab looking at the solder joints...
That's quite a lot of power points and extensions you've got up there... :-o
They install a drop power point for every light
And how else are you gong to power 9 panel lights?
EEVblog It's fine as long as you trust the extensions.
it's less than 700W for all the light and extensions are rated for 2000 or 3000 W
Understood, just curious what the most likey yet still unlikely failure modes are besides overloading. They're typically temporary but it's something a lot of people do.
The comment thread starting with "EEEEKS" puts a finer point on it. A current path (dead short, some sort of contact resistance, etc.) will dissipate four times the energy at 220V versus 110V. While electrical codes vary, it seems setups like that (hidden in an enclosed space) would be even more risky with 220V.
Bad solder joints is the most common problem I have come across in my experience with led lights with separate drivers.
i just fixed an old novelty strobe light from 1971 (oct 26, according to the stamp on the PCB) and the 3 pins on the mount for the flash bulb were dangling in the breeze a bit. it might have taken a drop or 99 since the Beatles broke up... and also, the neon bulb that was in there (i believe as some sort of current limiter, not the same as the flash bulb) had to be put to pasture and replaced. one thing that i found interesting was that the neon bulb that i replaced it with arced a bit inside for about 2 or 3 seconds on it's maiden voyage... it was an EICO neon bulb from i presume /at least/ the 1950's, if that helps demystify the situation. also, i left all the old original caps, for now. the strobe light is working hunky dory so far and the neon bulb is orange-ly glowing happily - though, it's inside the enclosure and you can't really enjoy its warm orange glow. might have to install a mini wireless webcam focused right on it inside the enclosure just to enjoy it. nah, probably not... cheers.
Loose solder joints are very rarely the main cause of the flicker it is usually a component failure
In Austria it is required to use a "cable shoe" on this type of Connection. I am not sure if the english term is realy calble shoe. It is a small diamter copper tube around the calbe. The reason for this is to prevent the connection from getting loose due to changes in temperature.
It's hard to say for sure without a close look/feel/sniff, but from the dross and other debris still stuck to the mask it looks to me these have been dip or drag soldered (rather than wave) and they were a bit light-on with the (probably water-clean) flux in the machine. There has been some hand rework, though .
Disco Mode!
He needs multi-coloured ones which flash randomly, and a "magic smoke" machine under the desk with a high power disco laser.
Jim Griffiths LOL
Strobes and the works!!
The smoke machine was the cherry on top lmao
The Epilepsy Test is not a failure, it's a "feature."
Single sided PCB's with large pin/hole clearance for automatic placement of heavy component are prone to fail.
Vibration and thermal stress will eventually lead to the results you have experienced.
That was common knowledge when I started as a B.Sc.EE in 1978.
Forgotten knowledge???
Regards from Denmark
oakey
Hi Dave, the first one you removed with the loose wire, is wired the wrong way around according to the markings on the case and on the PCB...
Makes no difference.
At the point you found the flickering driver to STILL be defective I'd have just redone all of the through-hole stuff. Not worth the hassle of trying to find some super hard to see cracks.
wonder is all the hand trimming on that thing may have caused the too little amount of solder to crack.
(combined with the repeated heating cooling action on the device in use.)
I will never tire of dave referencing LMFAO :D
Ohhhhh hahahaha. Im new to the channel. So this has happened before eh. Haha
You should invite the manufacturer to join this discussion. Could get interesting.
I'd be motivated to trim up the leads on the bottom of that board.
Useful video
You don't really get those small terminal blocks with a cage (at least in the Degson range). At least the leaf spring is better than just the head of the screw pressing down as one gets on most of those cr@ppy choc blocks
dave, check your "dead" unit pcb again. i wonder if you have a soilder spil over. check at the bottom at 08:56
Maybe don't have them so close together or on top of each other might help with thermal cycling? That mess up there was killing my OCD.
Of course it’s blindingly obvious, but perhaps you should put a safety caption telling any beginners - to disconnect the mains supply before playing with, or re-terminating the input lead. Same with your flicker/arcing test.
Some punters just don’t think at the same time they’re using their hands.
A bit like walking and chewing gum,..!
I don't do this enough to warrant an expensive microscope but my hand held Pro-Scope 50X zoom is great for this sort of thing.
I once repaired a big kitchen mixer. The main choke broke off just before the solder joint. So it was in it's hole, looking all fine. No cracked solder or anything either.
I found it by accident before I went to measuring the non obvious.
How does that even happen...
You should have put bootlace crimp terminal on the wire strand ends....