My fizzy LPF40 has been on all day (approx 9 hours) in the kitchen and the ballast must be burning hot as the body of the fitting is very warm.  Is it normal for a ballast to run so hot after 9 hours of use?

The question is, what you mean by "hot"...
Many ballasts are designed with temperature rise about 50..70degC from the ambient, what mean 70..100degC operating temperature, you can never hold your hand on something such hot.
Other aspect: The ballast usually dissipate ~10..20W of losses. That 10..20W have to somehow escape to the surrounding environment and the only way is through the fitting case. So it have to warm it up, you can never avoid that. It depend on the fitting thermal conductivity, if it will be warmed up only locally on higher temperature, or larger part on a bit lower temperature.
Definitely the 10..20W losses will cause the fixture case to be noticeably warmer...

Thanks for the reply...   I guessed they get warm but not that hot, a comparison between a netaine and a magnetic ballast comes to mind - both are mini heaters in that aspect.  Made me think if there are 10 to 20 watt in losses before the 40 watt tube, how much energy is being used compared with a modern HF fitting..

So on the basis a good quality iron ballast has a loss of 10 watt I looked what the losses High-Frequency control gear are.  I found Philips HF Performer advertises a loss of 3 watts running a 36 watt tube

Based on a 4ft single tube fitting

40 watt T12 tube (2800 lumen) + ((10 watt iron ballast losses) = total 50 watt) = 56 lumen/watt
40 watt T12 tube (2800 lumen) + ((20 watt iron ballast losses) = total 60 watt) = 46 lumen/watt
36 watt T8 tube (2850 lumen) + ((3 watt HF gear losses) = total 39 watt) = 73 lumen/watt

36 watt T8 tube + HF gear gives an average of 19% more lumens per watt than 40 Watt T12 + Iron Ballast

Theoretical lumens per 100 watt

40 watt T12 tube c/w iron ballast = 5600 lumen / 100 watt
36 watt T8 tube c/w HF control = 7300 lumen / 100 watt

The electronic is indeed more efficient than "standard" magnetic ballasts, but it is extremely sensitive from many points of view, so it's real life reliability become a problem. And not as much for individual failures (usually the magnetic case), but rather for external events making complete installations to fail at once.
Typical evets for destroying the electronic are
- the failing lamps (the protection mechanism does not always shut down in time). And usually it mean the lamp just started the "downhill" track, so on other ballast still may appear working. But as the parameters are already deteriorating, it stresses the ballast.
- Electrical overstress coming from overvoltage events in the mains. The installations are usually quite well protected from disturbances coming from the external world (atmospheric discharges,...), but by far not from the overvoltages originating inside the installation itself, like combined magnetic and electronic ballast after one switch being just one example.
- Mechanical shocks from improper handling. This I see as quite severe problem: Electronic ballasts is quite fragile piece of electronic, as fragile as the e.g. glass tubes of the lamps. But they do not appear so (tough or flexible looking shell,...). So many people tend to trow them around, let them fall, etc. And as they are lightweight, they do not cause as apparent "problems with the boss" as the heavy magnetic do with such handling (noise, sore hands,...).
- Sensitivity towards vibrations (wind and surrounding traffic for outdoor installations,...). Again, these are fragile things, where most of the internal connections are made by just solder joints. And mainly the newly enforced lead-free materials, they are very prone to fatigue cracks...

Because of these problems with the electronic, some ballast makers revived the idea of a simple magnetic ballast, but redesigned them so, their losses become way lower, practically at the same level as common for the electronic. These are generally bulkier and pricier components (usually more expensive than an equivalent electronic), but deal way better in the lighting environment (e.g. 6kV pulse is of no problem at all). And compare to the "classic", they are usually equipped with an overtemperature protection (with the low losses there is plenty of temperature margin to fit the cut out device temperature tolerance), so even the events of mild overvoltages (like missing neutral in a 3-phase system,...; usually killing for magnetic, but quite well tolerated by well designed electronic), what should make them virtually indestructible. But only the time will tell, these are in use for too short time to draw any valid conclusion on their reliability.

ph yeah your spot on...   There does seem to be a trade off between reliability and efficiency.   I guessed there were losses in a magnetic ballast but I was surprised how much they were.  OK the total energy usage are more energy efficient than a GLS or Halogen bulb but I had 5% loss on a typical magnetic ballast in my mind and would have agreed the EU are just making a mountain out of a molehill, but 25-50% losses made me go Wowwww, with energy demands growing and set to outstrip supply, I can see their concern.

Reading on I understand there are three grades of magnetic ballast depending on their efficiency, the older ones losing more than more modern magnetic ballasts.  I haven’t looked into it but I understand using a capacitor across the live and return helps reducing the losses.

I agree electronic control gear is generally regarded as unreliable and it does seem a waste and a huge inconvenient if your plunged into darkness if it kicks the bucket not to mention a massive waste of components.  All I can think is there is a huge world wide knee-jerk reaction as the realisation of demand/supply of the worlds current and future energy demands sink in resulting in strict regulations being proposed with very tight time scales and targets, so lighting manufacturers are using unrefined/hurried solutions to comply.  Surely major players like Thorn, Philips, Fitzgerald don't expect everyone to settle for such unreliability in a "quality" fitting.

As you mention, modern magnetic ballasts have improved efficiency greatly compared with older designs maybe there is a scope to fine turn them further.  There is nothing I would love to see more than a magnetic ballast with a 10% loss

I haven’t looked into it but I understand using a capacitor across the live and return helps reducing the losses.

This does not influence the fixture operation at all, neither it's losses. On the contrary, it have losses on it's own.
But the purpose is a bit different: With a low power factor you are transferring quite limited real power, but loading the transmission lines and equipment like there was large power transferred. The main intention is to save money on the wiring.
It is true, the lower apparent power handled by the wiring mean lower losses in the wiring, but this usually just offsets the losses associated with the correction of the power factor.

So if you want to design an illumination of some larger hall without the capacitors, with 16A breaker per branch you are limited to about 24 F36T8 lamps on each such branch, as each of them draw 0.43A, but with real power only ~45W (assume you don't want to load one "16A" branch for more than 10A to avoid false tripping).
But if you add the power factor correcting capacitors, the real power stay the same 45W (assume no losses in the capacitor for the simplicity), but the current will be just 0.22A per lamp, so on the 16A branch you may hook up 47 lamps with the same design margin, so you suffice with half of the wiring and corresponding breakers. Quite some money savings...

Thanks.   Good example...  Its like the same number of cars only needing a single lane road instead of a duel carriageway without bottlenecks

Sounds perfectly normal if its a VS ballast in there!