Has anyone else tried this? I thought I'd try it today and it does seem to work reasonably well.

The only phase control dimmer I have is rather old and isn't much good now, it dims a lamp to no output before the knob even makes half a turn which certainly isn't right but it's all I have to work with at the moment.

First of all I tried a 3ft 30W tube with a standard switchstart circuit. The results were disappointing, I barely managed to dim the tube before it started to flicker and the arc extinguished. Knowing that this dimmer isn't designed for inductive loads I decided to add a 100W incandescent bulb to the circuit which certainly helped, I'd say I probably got the tube down to 50% brightness before the arc extinguished.

My next test subject was a 6ft 85W tube with a semi-resonant start ballast. I first tried it without the 100W bulb and I got good results down to about 50% brightness before it started to strobe and flicker like mad, I continued to dim and the flickering seemed to stop at about 20% brightness and dimming was quite smooth down to a very low level. I then tried it with the 100W bulb in parallel and this gave some very good dimming results, I managed to get to about 40% brightness before the tube started to flicker at 50hz and continued down to about 20% brightness. Further dimming after this was quite effective down to about 5% before the tube started to flicker again and eventually go out. If it wasn't for the 50hz flicker problem with the 6ft tube on the SRS ballast I'd say that it dims perfectly fine.

I'm wondering if I buy one of those 'hard fired' dimmers that are capable of dimming inductive loads such as transformers for 12V halogen lighting It might work better?

Has anyone else tried dimming fluorescent lamps with cheap phase control dimmers?

   

There are special ballasts designed for such purpose.  I have one you can buy if you'd like. It's rated 120 to 277v so it should work there too

I once tried dimming some fluorescents with a standard cheap dimmer...
High-Power-Factor Rapid Start :: that one didn't work very well.
Low-Power-Factor Rapid Start :: seemed to work just fine.
In both cases i also had an incandescent lamp on the dimmer as well.

Fairly recently i found a old dimmable magnetic ballast and connected its "to dimmer" wire & a set of incandescent Christmas lights (in series) to a cheap dimmer...amazingly that worked (just not at full brightness)

When dimming, you have to provide separate heating for electrodes, as on reduced power setting they do not get the heat from the arc anymore, so loose emissivity and so make the arc unstable.

What work good (and i've seen it magnetic dimming ballast):
Use 2x4.5V transformer to supply filaments from the unregulated mains. This will provide filament heating.
Then wire the series choke ballast in series with the arc path (use one pin from each lamp end)  and with the dimmer, some dimmers would require incandescent preload
Use starter or starting switch parallel to the lamp to create high voltage spike to ignite the lamp.

To get best lifetime of all components, first turn ON the filament supply and after few seconds turn ON the arc circuit to full power setting and let the lamp ignite and warm up a bit (few seconds is enough). Then the dimming would work well, but the minimum setting would be on the position, where incandescent would have still quite high brightness, this is normal...

I can get down to about 5% output on purpose made electronic ballast meant to be used with two-wire phase controlled dimming. 

@Luminaire: With decent HF controller you might go down to 1..2% of light output (what is usually ~5% of lamp power), but i think when the lowest setting level is below 5% light output (~10% of power), it cover most of application needs.

With magnetic ballast it would be difficult to go below ~20% power level with phase-cut dimmer.
Lower levels are feasible only with separate series chokes designed for lower current level ranges, so if the dimming could be or only in discrete steps (e.g. 10, 20, 50 and 100%) or splited to multiple ranges (e.g. 10..30% and 25..100%).

@Luminaire: With decent HF controller you might go down to 1..2% of light output (what is usually ~5% of lamp power), but i think when the lowest setting level is below 5% light output (~10% of power), it cover most of application needs.

With magnetic ballast it would be difficult to go below ~20% power level with phase-cut dimmer.
Lower levels are feasible only with separate series chokes designed for lower current level ranges, so if the dimming could be or only in discrete steps (e.g. 10, 20, 50 and 100%) or splited to multiple ranges (e.g. 10..30% and 25..100%).

Two-wire phase control, wired like incandescent only dims to 5% as far as products I'm aware. Those using third wire (neutral, hot and dimmed hot for control) can get down to about 0.5%(not a typo, half a percent. (Lutron Hi-Lume).  

Since you have to continue supplying power to electrodes, dimming below 10% or so doesn't realize worthwhile energy savings and these deep dimming Hi-Lume ballasts are substantially costlier than 5 or 10% ECO model. 

I have used this kind of external "transformer" to provide heating for electrodes:

http://www.whitenightcape.com/image/helvart40st.JPG
http://www.whitenightcape.com/image/helvart40st1.JPG

I have few extra if you are interested.