First, a request:
Does anyone have 3x F39T5HO bulbs that are no longer useful to you, but that still light, that you are willing to donate to this experiment? I will pay shipping/packaging costs.
Ok, now for the latest results.
The ballast did finally warm up. Using an IR thermometer, I waited for it to reach peak temperature, and measured:
129°F @ hottest point on case
111°F @ center of case
Is this too
hot? I don't have that spec for this ballast, so I looked up other guidelines:
* Underwriters' Laboratories sets an absolute maximum safe case temperature of 194°F.
* Several electronic ballast datasheets from various manufacturers state max operating temperatures ranging from 122°F-140°F. These do not specify where on the case to measure, so it probably means average case temperature (which corresponds best to my center of case measurement).
* The only ballast datasheet that specified a max temperature for a specific point on the case said 176°F.
While the ballast is producing extra heat as a result of me using it out-of-spec, it is not overheating
, at least not by any measure I can find.
I then remembered I have some more bulbs to test! I braved my attic (it's a scary place), and returned with 3x 25W 4' T12 shop lights, and hooked them up. The ballast also drives them nicely throughout the full dimming range:
20W @ 10%, 77W @ 120%
I cranked it up to 120%, waited for the temperature to peak, and measured:
126°F @ hottest point on case
109°F @ center of case
Even though we're now pulling 77W instead of 45W, the case temperature went down a few degrees. No surprise there; since these bulbs are closer to what it's intended to drive, it's operating more efficiently.
Thinking of nothing else to try, I then opened up the ballast to see what I could learn.
It was easy to open. The circuit is not potted (sealed in plastic). It could be serviced, perhaps even modified to better suit a particular bulb.
There is a separate output transformer for each bulb, and room for a fourth. Since it drives each bulb individually, it's not limited to three bulbs. I have verified it works fine with one or two bulbs. Nice!
The heat is being produced by 4x MOSFET transistors and the output transformers. The most efficient heat transfer is through the back of the ballast case. For my prior tests, I had this ballast on its back directly on a wooden table. I've since elevated it, and the center-point temperature dropped 4°F. It should operate even cooler if it were bolted to a metal surface, or fan-cooled.
I had conflicting info whether this ballast was a rapid-start or programmed-start design. After examining the circuitry, I'm fairly sure it's rapid-start. This is important.
Before I continue, here's a quote from Wikipedia regarding different starting methods which will put the significance of starting method into perspective:
An instant start ballast starts lamps without heating the cathodes at all by using high voltage (around 600 V). It is the most energy efficient type, but gives the least number of starts from a lamp as emissive oxides are blasted from the cold cathode surfaces each time the lamp is started. This is the best type for installations where lamps are not turned on and off very often.
A rapid start ballast applies voltage and heats the cathodes simultaneously. Provides superior lamp life and more cycle life, but uses slightly more energy as the cathodes in each end of the lamp continue to consume heating power as the lamp operates. A dimming circuit can be used with a dimming ballast, which maintains the heating current while allowing lamp current to be controlled.
A programmed-start ballast is a more advanced version of rapid start. This ballast applies power to the filaments first, then after a short delay to allow the cathodes to preheat, applies voltage to the lamps to strike an arc. This ballast gives the best life and most starts from lamps, and so is preferred for applications with very frequent power cycling such as vision examination rooms and restrooms with a motion detector switch.
All T5HO bulbs seem to be meant for use with programmed-start ballasts, not
rapid-start ballasts like this one. Using a T5HO on this ballast will shorten its life each time you start it up.
However, there are some mitigating factors.
First, an aquarium light is usually started only once a day. Problems from cathode degradation might not appear for some time. The phosphors might degrade, rendering the bulb useless due to reduction in brightness or spectrum shift, long before the cathodes die.
Second, because of the way the dimming circuit ramps the power up from 0% on startup, this ballast is actually gentler than a typical rapid-start. Practically, it can be considered somewhere between
a rapid-start and programmed-start.
In short, it may provide a useful life from a T5HO
, but the only way to know for sure is to test.
I am prepared to do such a test. I have X10 home automation under computer control, which will allow me to do an accelerated lifetime test by cycling the ballast on/off 1,000 times, and see what happens to the bulbs.
Thus my request at the top of the post for 3x used F39T5HO bulbs. It's a brutal test even on a perfectly matched ballast, and it would be an awful shame to do that to a set of new bulbs.
Very interesting. I have been interested in something dimmable but,, dimmable T5HO ballests are expensive. I was assuming I'd have to bite the bullet and invest money and time into DIY LEDs. I want to see what develops with this.
Seems our thoughts are the same on this.
My "dream light": The bulk of the light produced by dimmable fluorescents, with range covering any needed lighting intensity. Supplemented by a few red, green, and blue LEDs; allowing fine tuning of spectrum to please the eye, or moonlight operation.
I would be careful using the wrong type bulbs with the ballast. You could damage the ballast or bulbs and possibly both.
I'm glad you said that. It's a risk/reward situation. I understand that and am willing to risk this ballast and maybe a few bulbs to possibly achieve the reward of a very nice light. I posted this in DIY instead of lighting because the DIY crowd generally understands this too.
I'll also say that I'm not the first to use strange ballast/bulb combinations. They have been used with success in ODNO configurations.