Three and even four and probably five (!) tubes will work, but you will not have a preheat for the extra tube ends. You can leave this as is (as Philips Certalume series ballasts shamelessly do) but it will drop the lifetime even further!

Just do not force tubes into EOL and stop the thing when they form some dense sputtered 'socks'.

Just curious, even if they are the wrong lamps, why wouldn’t it trigger eol protection on such a “smart” ballast? I mean, I feel like it would get upset when it notices the lamps current start violently changing right?

Jeez @RRK and @Medved you need to look a things a bit more pragmatically


I am quite sure the ICN-2S28 would run 2 F8T5's just fine, wired up in the 2xF14T5 configuration


we can tell from the wide range of tubes it runs from 1 F14T5 to 2 F28T5's that its a constant current ballast, and the official Max HF ballast current rating of a F8T5 is parity with what the ballast drive current likely is, so I dont forsee any issues there


and I dont see any problems with EOL either seeing as the ballast is rated for one singular F14T5, if it can safely detect and shutdown one of those at EOL, it should be able to handle 2 F8T5's

being rated for modern HE T5's plays in its favour here since they generally have pretty strict requirements for EOL detection and shutdown to prevent destructive  EOL


so @stillaintjeff24 while I cant gaurentee it wont rip a hole in the space time continuum, I am pretty sure you would be quite ok to use the ballast to drive 2 F8T5's, if you already have the ballast etc to hand no harm in trying it, if you have a kill-a-watt type plugin meter you can also messure total power draw to have some idea whats going on

I would wire it up as you would for 2 F14T5's so both lamps get proper cathode heating on start up

Well, no one will tell exactly. Usually, fluorescents are relatively tolerant to overcurrent, unlike incandescents, as cathode spots just grow a bit bigger, and lifetime drops linearly proportional to overcurrent. BUT you are dealing with recent orient imports, and can expect thin filaments, poor quality activation, so the tube may run on the edge even at nominal current of 145mA. 

as interesting aside with regards to that if you probe a lot of modern Chinese 8W T5's they often have quite low resistance cathodes, and I have wondered if they are being lazy and using filaments for HE T5's

if you browse @TL8W 's gallery and comments you can see a good a few examples of such

Yep, that is all more or less like I said, but there may be subtle nuances, like (though I am not 100% sure, probably James knows better) modern name brand T5HO/HE tubes likely include some kind of hydrogen release EOL harakiri, while F8T5 imports likely have none. Do not expect ballast circuitry to do miracles though, mostly they detect (1) continuity only at one of the filaments (2) excessive overall voltage of all of the tubes. Some, only some very advanced driver chips like Osram's or Tridonic may attempt to detect rectification at EOL. So if relatively low-voltage tube goes EOL this may be well ignored by a ballast, which will continue to pump an increasing power at constant current.



   

Just curious, even if they are the wrong lamps, why wouldn’t it trigger eol protection on such a “smart” ballast? I mean, I feel like it would get upset when it notices the lamps current start violently changing right?

Like I said before, do not expect miracles. Especially on such wide-range ballast, which should support lamp string voltage from circa 82V (a single 14W HE tube) to ~300V (two HE 28W) without any extra programming. So it is hard to tell a short tube going EOL from just two long ones running OK. Fluorescent ballasts do not usually go as far as monitoring dynamic tube voltage behavior. This is sometimes implemented in metal halide electronic ballasts, to decrease the chances of exploding or external arcing at EOL, and readily causes some side effects, like Osram Powertronics hate to run otherwise electrically compatible HPS lamps...

That makes me feel a bit better. I was just worried about having the tubes blow up or something.(not really) but I did want to be safe. Thanks! I will try it, hopefully it works for a while.

@LightBulbFun

So would it hurt the lamps to run in 2x14? Would it run them a lot over the intended current? I get the cathode heating, just is an interesting configuration. Is it better to have high resistance cathodes on the lamps? I got them today, and they measured 14.5 ohms. Does that really matter?

I have some ballasts that list anything from 2xF28T5's down to 1xF14T5 .. I've always meant to try it with a pair of F8T5's just to see what'd happen.
Can't see that it'd hurt anything. 2 F8's is slightly longer than 1 F14 (it'd also be easy enough to try with a F14, and then a pair of F8's...as long as total power draw was the same or more with the F8's I'd be comfortable running them.)

Oh and LOL the ballast is longer than a F8, I'm sure it'd look pretty ridiculous with then.

what is HE? What’s the difference?

There are 3 families of standardized T5 size tubes, all using the same sockets and looking very alike externally, but different wattage ratings and different lengths.

First, the oldest, is what is often called "miniature" T5. These came way more than half century ago (somewhere in 40's or 50's) and as the name suggest, were meant as miniature form of fluorescents, intended for small fixtures. Their current rating was about 0.145..0.17A at mains frequency, so they can share common ballast type (shorter lamps are designed at a bit higher currents, because that is how a simple choke use to behave). These are the 150mm F4T5, 212mm F6T5, 288mm F8T5 and 517mm F13T5.

Then at the end of 20'th century, when phosphors capable to work at high temperatures and high intensities came, two more families of fluorescents using the same T5 tube size came, the T5HE and T5HO.
The first of the two were lamps optimized for maximum efficacy for the size and to share the common current rating (0.17A at high frequency) so like the "miniature", they could share the same ballast (ballast is then acting very close to an ideal current source; it is the ballast you most likely have here). These were 550mm F14T5HE, 850mm F21T5HE, 1150mm F28T5HE and 1450mm F35T5HE
The second group was designed with about double light output at theexpense of some efficacy,they operated at higher currents (0.46A at high frequency), so again one single ballast type can serve any of those lamps (or even multiple of them for some models). There belong the 550mm long F24T5HO, 850mm F39T5HO, 1150mm F54T5HO and 1450mm F80T5HO

Interesting. The way these ballasts and lamps work are very interesting. Again, I have another stupid question. When running the lamps in 1 lamp mode vs 2 lamp, wha is actually the difference? I know the wiring is different, but I have been thinking that the ballasts controlled the wattage, which doesn’t make a whole lot of sense, right? It’s the amps or current instead, so is the current separate for each lamp in 2 lamp mode or does it allow more current for the lamps run in semi parallel? I see they jump from red to green for programmed start. Basically it isn’t dangerous to run them in 2 lamp mode when the are 2 8 watt lamps, instead of 2 14 watt, because the actual current isn’t that different?

Looking more carefully, for some obscure reason, T5 HO tubes, unlike T5 HE, which all settled for the same 170mA (well, 165-175) rating, have a weird range of current ratings from 0.255A up to 0.53A! (T5 HO 80W)

Interesting. The way these ballasts and lamps work are very interesting. Again, I have another stupid question. When running the lamps in 1 lamp mode vs 2 lamp, wha is actually the difference? I know the wiring is different, but I have been thinking that the ballasts controlled the wattage, which doesn’t make a whole lot of sense, right? It’s the amps or current instead, so is the current separate for each lamp in 2 lamp mode or does it allow more current for the lamps run in semi parallel? I see they jump from red to green for programmed start. Basically it isn’t dangerous to run them in 2 lamp mode when the are 2 8 watt lamps, instead of 2 14 watt, because the actual current isn’t that different?

You have to know ballast topology to understand this. Basically, almost all modern mains powered electronic fluorescent ballasts begin with two-transistor halfbridge inverter running from a rectified DC bus, generating squarewave HF voltage in the range of approximately 25-50kHz. Next, in the case of multiple tubes, they each may run in parallel on their own combinations of a small HF choke and a capacitor, or be connected in series, running on a single common HF choke and a capacitor. Series circuit is cheaper, as magnetic elements are among the most expensive components in a circuit, but generally is considered less reliable as a failure of a single lamp will take out the whole circuit. Both variants are in use at the manufacturer's wish. Your variant is clearly marked as series in the datasheet. Sometimes (Helvar made for example) a hybrid circuit is used, running two tubes in parallel (or even two groups of series pairs!) from a single ballast choke combined with a smaller balancer choke which tries tho split the current in two halves.

Okay, now for the current regulation. Series LC circuit with a choke, and a capacitor connected in parallel with lamp, operated slightly above the main resonance has a natural current stabilizing behavior, running at about the same lamp current regardless of wild lamp voltage variations. This property is in wide usage with electronic ballasts. Some better circuits do more by sensing the current at a 'cold' (closer to the ground) tube (tube string) point and adjusting the frequency slightly to stabilize the current further. 

I see that’s interesting. The current stays relatively the same but the voltage is a lot more variable, and maybe not so important (arguably) that’s helpful, Thanks for that, quite interesting! When I set up the lamps I will have to test the current to see what it does, as I only recently acquired a proper T5 ballast as the two Advance Centiums. Wonder what a good current range is for 8 waters, as I have seen various results and reports.

I see that’s interesting. The current stays relatively the same but the voltage is a lot more variable, and maybe not so important (arguably) that’s helpful, Thanks for that, quite interesting! When I set up the lamps I will have to test the current to see what it does, as I only recently acquired a proper T5 ballast as the two Advance Centiums. Wonder what a good current range is for 8 waters, as I have seen various results and reports.

the official ANI C78.81 and IEC 60081 standard for double ended fluorescent tubes gives an an F8T5 a rating of 57V 145mA for 60Hz magnetic operation with a true lamp power of 7.2W, and IEC 60081 gives on HF control gear a specification of 50V 150mA for a true lamp power of 7.5W and allowable current range of 90mA-180mA

(fun-fact because of how manufactures somewhat f*cked up the short 4W/6W/8W T5 range to kludge them on all onto one ballast the 8W T5 is a rare example who's HF power rating is *higher* than that on magnetic gear, as on an IEC/ANSI magnetic reference ballast/specification, they only run at 7.1W/7.2W, but originally the F8T5 did run at 8W at 170mA off its own dedicated 8W T5 ballast)