1. Setup of experiment # 14

This is the third test run between 2 CFL's: this time we use topnotch Osram Duluxstar models, both bought for 5.95.Euro:
the 21W/10 years rated model was bought about 6 months ago. It should survive 20000 switchings and has a luminous flow of 1230 lm. It should corespond to a classic 100W bulb. The brighter (and bigger) 24W/10 vears model was bought the 27th Apr.2011 and is a newer type. It too should sustain 20000 switchings, and has a luminous flow of 1500 lm. It is equivalent to a 120W classic bulb. 

See here the packaging of the 21W/1230lm lamp and here that of the 24W/1500lm lamp. Both lamps are rated A and are "Made in China".

Read here a German document on all Osram Dulux CFL's.

2. Start of experiment #14

Start  27 April  2011 15:34 UTC.
The webcam counters show the 21W/1500lm CFL at the right (blue sticker) and the 24/1500lm CFL at the left (yellow sticker). Due to the brightness of the 24W and the mandatory automatic settings of the webcam, the left counter display is a bit hazy when the lamps are ON. 

With a On period of 58 seconds and equivalent Off time, 20000 switchings correspond to 644.4 hours or practically 27 days of experiment duration. So let's watch if both lamps will survive the 24th May!

3. Problems with counting

I  first saw the 3rd June 11 a problem with the counters: normally the counters should increment by 1 when the lamps are switched OFF. At this date, they incremented at every action: ON or OFF. Left counter even sometimes jumps by more than one when lamp was switched on. On top of that both counters show a large difference. All this is probably caused by very rapid flicker which fools the light detectors. I rechecked today, the 25th June 11. The problem gets worse. Here a succession of counters:

ON        79016      85252

OFF      79017      85253

ON       79019      85253 -> 85256

OFF     79020      85257

ON      79022      85257 -> 85260

OFF    79025      85261

So the right lamps switches correctly when going OFF, but makes jumps of 3 when switched ON. The left one makes jumps (irregularly) when switched ON or OFF

As the experience started the 27th April, it runs now for about 70 days, which gives approx. 52138  POWER ON switching actions, well above the rated 20000 ! We will see how far this goes. For the moment, this experiment confirms the good behaviour of the OSRAM CFL's, compared to other brands. (the BMB team are not stooges of OSRAM; we are fully independent!

4. End of experiment 14

The 21W CFL (blue label, right on picture) dies the 14th Jul 2011 at about 16:00 local time; shown switching count is 120743.

The 24W model goes extinct 5 days later, on the 19th Jul 2011 at about 08:00 local time, with a switching count of 136877.

The switching counts  can not be taken as correct, as it was shown that flickering fooled the counters into higher totals.
So let us simply count the total switchings to the ON state from the number of days in use, rounded down; let us recall that the lamps are ON for 58 seconds and off for the same time. Which gives a switch to the ON state every 116 seconds:

CFL 21W:  from 28Apr to 13Jul = 77 days = 77*24*3600/116 = 57352 ; this amounts to 924 hours in the ON state.

CFL 24W: from 28Apr to 18Jul = 82 days = 82*24*3600/116 = 61076 = 984 hours in the ON state.

So both lamps exceed their rated switching withstand of 20000 by large margins of 287% and 305%.

It should be noted that the lifetime shown on the package is 10000 hours for both lamps, with 2.7 hours ON per day ( i.e. about 10 years lifetime). We could conclude that the BMB test bed accelerates aging by a factor of approx.10, compared to the "standard" operation on which the printed lifetime is based.

This experience again confirms the good quality of the OSRAM Duluxstar CFL's, compared to the competitors.

5. Background CFL

The closet containing the BMB test setup is illuminated by a small 5W CFL;  BMB started the 2th March 2009,  more than 2 years ago. Since that time this lamp was continuously switched ON. It blew the 26th July 2011, after about 880 days or 21120 hours in use. This shows that CFL's work best when they are continuously ON; will they save energy under this scheme?

last edit: 28Jul11

go back