Teardown of el-cheapo fluorescent energy-saving lamp

[FenderBender]

Yeah I guess that's what happens when you put some cheapass parts on a cheapass PCB in a cheapass design. For pete's sake, if you are going to use electrolytics in ANYTHING, please use ones that actually have a distinguishable name on them. Or better yet, don't use them at all, if possible.

Grr. Bean counters.

[mariush]

Here's two CFL lamps that are sold here in Romania for about 2-3$ a piece....  20-22w lamps.

No surface mounted parts on their back sides so only front pics.

The first has about a third of the white tube black so I guess that's what failed... on the second I guess a diode is busted or the capacitor.. I don't trust them enough to waste my time trying to fix them.

...

Oh yeah... as I write this, I remembered posting on another forum another lamp I opened that got my attention with the Pchicon capacitor...thought it's a lame attempt to fake Nichicon. But at least the design seems a bit more complex and I think that's even a thermal protection near the transformer.

[Simon]

What do you expect from a product made by the cheapest supplier using the cheapest parts, assembled by the lowest cost poorly trained labour using the cheapest equipment. Amazing that most work when you buy them, as I often have faulty ones in a big batch, so I prefer to pay more for the better quality ones.

agreed, don't complain if you buy rubbish

[Mechatrommer]

What do you expect from a product made by the cheapest supplier using the cheapest parts, assembled by the lowest cost poorly trained labour using the cheapest equipment. Amazing that most work when you buy them, as I often have faulty ones in a big batch, so I prefer to pay more for the better quality ones.

agreed, don't complain if you buy rubbish

so I prefer to pay more for the better quality ones.

what is better quality brand? i have bunch of philips flou's that died, almost the same circuit as posted above. now i have in stock many type of brand, each time i replace, i will write the date on it, so i can check which one will last longer than others. your recommendation may speed this process up dramatically.

what is better quality brand? i have bunch of philips flou's that died, almost the same circuit as posted above. now i have in stock many type of brand, each time i replace, i will write the date on it, so i can check which one will last longer than others. your recommendation may speed this process up dramatically.

[Simon]

most CFL's in supermarkets are rubbish. I tend to try and buy online the ones that are 3500-5500K and cost a bit more and hopefully better built as they are for the more concious consumer

[grumpydoc]

agreed, don't complain if you buy rubbish

Sometimes there isn't much choice - the 3x3W GU10's - which are the only ones which approach a 50W halogen in brightness - were, for a long time, only available from one supplier in the UK.

Even if people want to buy better quality how does the average person know - most won't go online for quality - maybe to get a cheaper price, but probably not for quality and even if they did there's no guarantee that a "brand name" is any better than "no-name rubbish" when it's taken apart.

[Balaur]

I've actually repaired one of my bulbs. And it was a Philips, no less.

One of the ceramic capacitors failed in a very obvious way (hole in the package). I've replaced that one and still works after two years.

BTW, if you ever need HV transistors (ex: for Nixie clocks), CFL bulbs are a very good source of them.

Cheers,
Dan

[BravoV]

Someone from Czech Republic did an intensive teardown & documenting the circuit at varous types -> http://www.pavouk.org/hw/lamp/en_index.html

Worth a place in your bookmark imo.

[RJSC]

I like to fix them, not so much for the cost, but for the joy that saving something from going to the trash (landfill, incinerated/etc) gives me.
This wasteful society that makes things designed to fail in order to sell more disgusts me.

I've fixed a dozen and I've never seen the electrolytic capacitor die!
Its usually the tube died, or much more frequently, a very crappy ?polyester? green capacitor, supposedly 1000V that sits between the two filaments that shorts out!
It's function is to pass current through the 2 filaments during the heating phase, and then, the frequency changes, the tube presents a lower impedance and current starts to flow trough the tube and not trough the filaments and capacitor.
When it fails shorted, the current is always flowing trough the capacitor and filaments, and if it's left on, it leads to the blackening of the end of the tubes because of filament burn out.

These are the capacitors I was talking about:


You guys blaming the poor electrolytics... It's not always them!
Do you believe me if I tell you I have fixed a laptop motherboard with a shorted ceramic capacitor? Who had thought, right?
I've carefully dis-soldered every electrolytic from the bus and the sort was sill there!
Then I remembered of the ceramics! bingo!
Usually on old CRT TV I've fixed, its a resistor or a ceramic capacitor damaged, more rarely it was an electrolytic,
It goes against what I've read everywhere: "electrolytics are the fist to fail", but it's my personal experience.


Please take some time to watch this, there is proof that manufacturers of incandescent bulbs agreed to intentionally shorten their lifespan.
As for the CFLs life, I wouldn't be surprised if something like this is going on with CFL's as well.

[Mechatrommer]

we (i) need some info how actually the circuit works. any idea/link?

[RJSC]

The link above (http://www.pavouk.org/hw/lamp/en_index.html) with the teardown also explains how the circuit works, in a slightly broken english 

Electrical construction

Princip of function we explain on a LUXAR 11W lamp. Circuit contains supply section, which includes interference suppressor L2, fuse F1, bridge rectifier from 1N4007 diodes and filtering capacitor C4. Starting section includes D1, C2, R6 and diac. D2, D3, R1, R3 have protect function. Other parts have normal operation function.

Luxar 11W
Lamp start

R6, C2 and DIAC mades first pulse to base of transistor Q2 and cause his opening. After start is this section blocked by diode D1. After every opening of Q2 is discharged C2. There is not possible to collect enough energy for reopening of diac. Next are transistors excitated over very small transformer TR1. It consists of ferrite ring with three windings (5 to 10 coils). Now are filaments powered over capacitor C3 from voltage rises from resonant circuit from L1, TR1, C3 and C6. Than the tube lights up is resonation frequency specified by capacity of C3, because he has much lower capacity than C6. In this moment is voltage on a C3 over 600V in a relation to used tube. During start is peak collector current about 3 to 5 times bigger than during normal operation. When the tube is damaged, there are hazard of transistor destroying.
Normal operation

When the gas is ionisated in a pipe, C3 will be practically shorted and thanks to this frequency goes down and changer is now drived only by C6 and changer generates much lower voltage but enough to keep the light on. In a normal situation, when transistor opens, that current to TR1 increasing until his core is saturated and next his feedback to base drop away and transistor closes. Now opens second transistor which is excitated by reversly connected windind of TR1 and all process repeats.

[wkb]

These are the capacitors I was talking about:

These look very familiar... 

[RJSC]

These look very familiar... 

Please elaborate a bit... I'm not getting it.
Where else are they famous for failing?

[wkb]

These look very familiar... 

Please elaborate a bit... I'm not getting it.
Where else are they famous for failing?

These are the same caps I found in the broken lamp that started this whole thread.

[RJSC]

These are the same caps I found in the broken lamp that started this whole thread.

I haven't got time to read the whole thread in detail, so, excuse me if you already fixed it, but...

If you still have the lamp, can you test the capacitor between the filament poles for a short?

[westfw]

Interesting.  I wouldn't have suspected the mylar caps (what's to do wrong?), but I've definitely seen swollen and even burst electrolytics as well.

There are CFL reference designs out there, especially from the vendors offering ICs for circuits more complex than the cheapest.  (there ARE dimmable CFLs, for instance.)  IFR has a bunch of products: http://www.irf.com/technical-info/refdesigns/lightingkits.html

People who dismiss CFLs for their "manufacturing expense" haven't thought too much about what goes into manufacturing incandesant bulbs, I think.  I mean: tungsten wire doesn't exactly grow on trees!

[wkb]

These are the same caps I found in the broken lamp that started this whole thread.

I haven't got time to read the whole thread in detail, so, excuse me if you already fixed it, but...

If you still have the lamp, can you test the capacitor between the filament poles for a short?

Sorry, it went to the recycling bin.  I still have the 2nd one, unopened.  Plan to take that to the shop I bought it from and do some whining.  Found a carton of a smaller version of the same brand, and they promise 5 year / 5000 hours....