CFL ballast design, and using dead lamps for repair

I have a desk lamp, with a magnifier and a 12W T4 circline fluorescent bulb, that I use constantly for working on PCBs and electronics projects in general. I bought it several years ago and recently it started flickering and then died. I replaced the bulb, but still no joy, and after replacing the blown fuse and two damaged transistors, I found that the little transformer had an open winding.

So, I thought, a 60W equivalent CFL is actually about 11-13 watts, and the little circuit in them should work. I had a couple of broken or dead bulbs ready for recycling, so I opened the bases, cut the leads, and extracted the PCBs. After a few unsuccessful tries, I was able to get it to work and now my lamp is once again operational.

I found some schematics of the CFL driver boards here:

formatting link

Some of those circuits matched what I had almost exactly. It was a little difficult to follow the explanation of how they operate, but what was confusing is the four pins shown on the lamp itself, which is also how the bulbs are made. I assume they are the heaters that are usually activated

with a starter, but I did not find any continuity on those pins. The desk lamp only had one wire to each of two pins on the circline bulb, but in the CFLs all four wires were connected to different points on the PCB. It would only work when I shorted the connections that would have gone to the heaters, and it seems to work very well. It starts to light at 50 VAC and reaches full brightness at 100-120V, at which it draws about 100mA. That's close enough to 12 watts for me!

Here's a little clip of my repair project:

formatting link

This is good to know. Those little circuits in each CFL have a lot of good components, including a DIAC and high voltage transistors, and it's probably possible to use them to drive small fluorescent lamps for DIY projects or repair. I doubt they could be used for the 40W tubes, but a driver from a

100W CFL should work on a 20W tube such as are in many desk lamps and small kitchen fixtures.

Does anyone know if there is any problem with this?

Thanks,

Paul

PS: GO RAVENS!!!!!

Reply to
P E Schoen
Loading thread data ...

So, I thought, a 60W equivalent CFL is actually about 11-13 watts, and the little circuit in them should work. I had a couple of broken or dead bulbs ready for recycling, so I opened the bases, cut the leads, and extracted the PCBs. After a few unsuccessful tries, I was able to get it to work and now my lamp is once again operational.

I found some schematics of the CFL driver boards here:

formatting link

Some of those circuits matched what I had almost exactly. It was a little difficult to follow the explanation of how they operate, but what was confusing is the four pins shown on the lamp itself, which is also how the bulbs are made. I assume they are the heaters that are usually activated with a starter, but I did not find any continuity on those pins. The desk lamp only had one wire to each of two pins on the circline bulb, but in the CFLs all four wires were connected to different points on the PCB. It would only work when I shorted the connections that would have gone to the heaters, and it seems to work very well. It starts to light at 50 VAC and reaches full brightness at 100-120V, at which it draws about 100mA. That's close enough to 12 watts for me!

Here's a little clip of my repair project:

formatting link

This is good to know. Those little circuits in each CFL have a lot of good components, including a DIAC and high voltage transistors, and it's probably possible to use them to drive small fluorescent lamps for DIY projects or repair. I doubt they could be used for the 40W tubes, but a driver from a

100W CFL should work on a 20W tube such as are in many desk lamps and small kitchen fixtures.

Does anyone know if there is any problem with this?

Thanks,

Paul

PS: GO RAVENS!!!!!

+++++++++

so the original lamp was conventional ballast and 2 pin bimetal switch starter ?

Reply to
N_Cook

/open circuit

no. there is a small electronic driver - plays hell with FM radio :-)

IME the heat shield deteriorates allowing lamp heat to saturate the board and kill circuits. It is possible the line voltage to rail transformer may also be affected by such a failure of the shield. I replace the shield with a thin stove mat.

--
no sale of foxs
Reply to
Foxs Mercantile

Besides the electrolytic and bulb, the starter cap is very failure-prone. In circuits where the filaments are used, it's in series with the output, forming a series resonant tank; the low resistance of the cold filaments draws lots of current, starting the tube quickly. When the tube starts up, it appears in parallel with the cap, transforming the circuit into a good old series inductance ballast.

The kind with only two pins skips the heating step and allows the tank voltage and current to resonate even higher, until the tube breaks down cold-cathode style. Once ignited, ion bombardment keeps the filaments warm, keeping the reignition and operating voltages normal.

High voltage film caps are big and expensive, so understandably, they don't like to use them very much. Often, a poor green (polyester) type is found, which isn't even green anymore, but black from the abuse. Others may be burned through, having experienced too many starts (too much peak voltage) that self-healing has burned away most of the capacitance. The circuit then either tries oscillating too high (burning itself from switching loss) or oscillates lazily or latches (resonance no longer strong enough to draw enough current to provide sufficient feedback).

Tim

--
Deep Friar: a very philosophical monk. 
Website: http://seventransistorlabs.com 
 Click to see the full signature
Reply to
Tim Williams

Problem? Yes. It would have been more sensible to replace that ancient, flakey, mercury-filled toob with LEDs.

PS: GO NINERS!!!!!!!!!

--
John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    
 Click to see the full signature
Reply to
John Larkin

Many CFL PCBs have a diac - you can make a neat little pulsed bicycle light using a blocking oscillator inverter to charge a 47u electro. Everytime it reaches about 32V, you have the diac dump the charge into the LED bank of a cheap & nasty LED worklight.

My project was published in Elektor a year or two ago.

Reply to
Ian Field

"P E Schoen" schreef in bericht news:kdj0bd$6ut$ snipped-for-privacy@dont-email.me... I have a desk lamp, with a magnifier and a 12W T4 circline fluorescent bulb, that I use constantly for working on PCBs and electronics projects in general. I bought it several years ago and recently it started flickering and then died. I replaced the bulb, but still no joy, and after replacing the blown fuse and two damaged transistors, I found that the little transformer had an open winding.

So, I thought, a 60W equivalent CFL is actually about 11-13 watts, and the little circuit in them should work. I had a couple of broken or dead bulbs ready for recycling, so I opened the bases, cut the leads, and extracted the PCBs. After a few unsuccessful tries, I was able to get it to work and now my lamp is once again operational.

I found some schematics of the CFL driver boards here:

formatting link

Some of those circuits matched what I had almost exactly. It was a little difficult to follow the explanation of how they operate, but what was confusing is the four pins shown on the lamp itself, which is also how the bulbs are made. I assume they are the heaters that are usually activated with a starter, but I did not find any continuity on those pins. The desk lamp only had one wire to each of two pins on the circline bulb, but in the CFLs all four wires were connected to different points on the PCB. It would only work when I shorted the connections that would have gone to the heaters, and it seems to work very well. It starts to light at 50 VAC and reaches full brightness at 100-120V, at which it draws about 100mA. That's close enough to 12 watts for me!

Here's a little clip of my repair project:

formatting link

This is good to know. Those little circuits in each CFL have a lot of good components, including a DIAC and high voltage transistors, and it's probably possible to use them to drive small fluorescent lamps for DIY projects or repair. I doubt they could be used for the 40W tubes, but a driver from a

100W CFL should work on a 20W tube such as are in many desk lamps and small kitchen fixtures.

Does anyone know if there is any problem with this?

Thanks,

Paul

PS: GO RAVENS!!!!!

The traditional way of igniting TLs is first heating the heaters and then start it with a relative high voltage puls. But heating first is not necessary. The lamp will ignite if only the voltage of the ignition pulse is high enough. Nevertheless, the heaters cannot be omitted as they provide the electrons required for the current through the tube. Especially low power TLs are sometimes ignited this way. When used with low frequencies i.e. 50Hz or 60Hz, it is said to shorten the lifetime of the tube. The circuit I ever found in a handheld 8W TL lamp did work but gave a flickering light.

I also use a magnifier lamp but a bigger one containing a 22W circline. It came with an old-fashioned iron ballast and no starter. I had to start it by hand using the special switch. That is I had to push the button and keep it pushed for a 6s. During this time the heaters were on. Relieving the button ignited the lamp... Most of the times. But over time igniting failed more and more and the circlines seemed to worn out faster and faster. Investigating revealed the special switch to be worn out beyond repair. So I got a 20W CFL, got out the electronics, put them in a box and connected it to the circline. Works fine for several years now already.

petrus bitbyter

Reply to
petrus bitbyter

Don't forget the significant increase of efficiency running the tube at around 20kHz.

Frequency may also contribute to easier starting, although most CFLs seem to have some sort of setup to run power into the heaters.

More often than not its a capacitor to bleed some of the AC current via the heaters, some Ikea branded CFLs I got from a car boot sale long ago had resin dipped double pellet thermistors bridging the heaters across the tube.

Reply to
Ian Field

"P E Schoen"

It starts to light at 50 VAC and reaches full brightness at 100-120V, at which it draws about 100mA. That's close enough to 12 watts for me!

** You are mixing up "watts" and "VA".

CFL inverters, like most simple SMPS, have a PF of about 0.6 due to the peaky current wave shape drawn from the AC supply - so VA and watts numbers differ by about a 2:1 ratio.

You did use a wide band "true rms " meter to measure the AC current - didn't you ?

... Phil

Reply to
Phil Allison

FWIW, of the modest pile of CFLs I autopsied from my own house, 100% of the failures were from failed heaters.

This may be atypical. I get good service from CFLs running them without enclosures; the enclosed, base-up fail modes may be different.

--
Cheers, 
James Arthur
Reply to
dagmargoodboat

FWIW, of the modest pile of CFLs I autopsied from my own house, 100% of the failures were from failed heaters.

This may be atypical. I get good service from CFLs running them without enclosures; the enclosed, base-up fail modes may be different.

--
Cheers, 
James Arthur 
 Click to see the full signature
Reply to
hifi-tek

** Most CFLs have PTC thermistors that increase current flow to the heater for the first few seconds - makes the lamp start quickly, particularly in cool weather conditions.

It also burns out the heater in a few hundred or so cycles.

Claims for life spans of 6 or 8 thousand hours are based on continuous use in a cool environment - plus lotsa good luck.

** Recently, a neighbour had an 18W CFL explode and bits go all over the kitchen - it was base up in a non ventilated fitting.

... Phil

Reply to
Phil Allison

of the failures were from failed heaters.

without enclosures; the enclosed, base-up fail modes may be different.

I don't know if any do fail specifically due to the start cap -- as you note, the filament tends to fail first. I haven't measured any of the caps to see the capacitance loss, but the color change is reason enough.

EPCOS makes capacitors specifically for this service; their graphs of peak voltage vs. capacitance are reminiscient of the Write-Only Memory's "insertions vs. pins remaining" graph. ;-o

Base-up tends to cook the capacitor; better ones don't mind so much. My parents have a GE "long life" model on their front porch, an enclosed (vented enough that bugs find their way in, but not so much that they find their way out before dying and making a pile...), base-up fixture.

I installed them something like four years ago; the other outdoor lights (upright, even) have since failed, but the front porch, surprisingly, still works.

I think. Come to think of it, I didn't look closely last time I was by; the front porch may've been replaced, most likely under pressure, the other "low priority targets" simply being "left in the dark"... :-)

Tim

--
Deep Friar: a very philosophical monk. 
Website: http://seventransistorlabs.com
Reply to
Tim Williams

I measured the heaters on one of the new circline lamps and they appear to be about 3 ohms each. So in series, using a starter, they would draw a high current surge and then probably much lower as they become nearly incandescent. Of course the ballast would limit the current to some much

lower value. I figure that an ordinary magnetic ballast for 12W would drop about 30 volts at 100 mA during normal operation, so probably about 400 mA into the heaters.

But since this lamp had only two wire zipcord going from the electronic ballast to the lamp, it obviously did not use the heaters with the original circuit. So I just shorted the connections for the heaters at the board, and it works perfectly well. I can't see where the 6 ohms of filament resistance would make much difference to a circuit designed for 100 mA, and I don't

know why the CFLs use the heaters at all, or why they should burn out. It does appear that the heaters of the one dead CFL are open, although I thought I had continuity through one when I first measured it. One of the CFLs actually got broken when I hit it with a flyswatter as I waged war on the stinkbugs that love to circle the lamp. It seemed to continue to glow for a while, which may have been the heaters.

The defunct circline lamp shows discoloration on one end and that heater is open, while the other is OK. I don't know how it can burn out if there is only one wire connected to it. Perhaps it just vaporizes because of the current flowing from it into the fluorescent plasma? I don't really know the details of operation.

When I was still in High School I made a little multivibrator circuit that fed a 12V transformer and I connected the 120V winding through a capacitor to a small fluorescent lamp in a camping lantern. It originally used a big high voltage battery (maybe 90 volts and maybe two of them), and maybe some sort of vibrator as in old tube type car radios to get the AC voltage for the lamp. Will it work on just DC? My conversion worked on a 12V motorcycle battery, and it was instant start, whereas the original had a starter button. I think I still have that lantern somewhere. I should find it and look inside.

I also have an old lantern that originally used two big doorbell batteries (remember them?), and I replaced them with a NiCad battery that was made of plexiglass so you could see the plates and electrolyte inside and possibly even service it. My father worked for a company that made exotic batteries for the military and I think this came from there, probably 50 years ago.

Paul

Reply to
P E Schoen

In article , petrus bitbyter writes

In other words, used like this, they are not heaters but cathodes. Presumably this is what a 'cold cathode' tube is.

--
 (\_/) 
(='.'=) 
 Click to see the full signature
Reply to
Mike Tomlinson

"Mike Tomlinson" schreef in bericht news:UjRdfqBzen$ snipped-for-privacy@jasper.org.uk...

Sure. During the classic startup you can consider the heaters to be direct heated cathodes. Once the light is on they are pretty hot cold cathodes :)

petrus bitbyter

Reply to
petrus bitbyter

Fortunately, I have not had this problem. The CFLs in my bathroom and bedroom have lasted two to three years, and are turned on and off several times a day.

Reply to
William Sommerwerck

Must've been a pretty big explosion for bits to go everywhere from a non-ventilated fitting!

Reply to
Ian Field

Two or three years? I have incandescents that last a lot longer than that.

Reply to
krw

On the last CFL I autopsied; the electrolytic read a realistic ESR on the Peak Atlas ESR70, but zero capacitance.

When I double checked with the DSE ESR meter it showed O/C.

On closer inspection, there was a tiny ring of corrosion "fur" round one of the legs.

Reply to
Ian Field

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.