https://youtu.be/cNv-3opP8pk

In this video you'll find a workflow for casting your own aluminium parts. This guy is making pretty high quality parts by 3d printing a copy and then making a sand mold from that. Also a bunch of outdoors lamps.

In the EM36ST i am thinking what is the actual role of those capacitors in the circuit ?

If the capacitor sees HF applied to it, then it is important to not exceed its ripple current ratings. Provided that PHE850 have about 10 times lower loss angle (tan d) than PME271, you have likely upgraded the ballast and won't have any issues

If the capacitor sees lamp ignition voltage peaks that reach its max ratings, then it may be more complicated because each capcitor type may react differently

I've noticed that the arctube of my Aliexpress 70W 10000K MH lamp is very dirty: https://www.lighting-gallery.net/gallery/displayimage.php?pos=-263254
What is all that dirt?

I've just been replacing some more of these Rifa caps this afternoon.
Firstly a pair of them in a 1999 Tridonic PC 30 C001 ballast, the originals appeared ok when I got the ballast but when I opened it to check today they had cracked so I replaced them with a pair of plastic PHE 850's.

Second was another Tridonic EM 36 ST emergency module from 2010 with a couple of different sizes of them, both replaced for PHE 850's of the same ratings which will hopefully prevent any future issues.

I've noticed something weird in several of the F34T12s I got:

Some of them that I've started using in everyday use are getting these little gray lines and a few dots on the inside along the whole length that look like scratches/dots and when the lamps are cold (on a magnetic ballast) and are just starting sometimes there are fairly bright white sparks on those scratches right as they start.

I would post a video but it's a bit hard to catch it since usually the other lamps are too bright to catch the effect, what is causing this?
It's not like the pink sparks that happen at the ends

Have a look at 2 series from Kemet :
https://content.kemet.com/datasheets/KEM_F3001_F861_X2_310.pdf
https://content.kemet.com/datasheets/KEM_F3092_F862_X2_310.pdf

One with the warning, the other rated specifically for the series applications. Same standards/approvals and ratings (except the automotive Q200, which is not directly an electrical rating, but only addresses the number of failures allowed/1 million units)

The big difference is in the "environmental test data" table

If you pick one capacitor value that exists in both series, you'll see the difference in its dimensions - They have different film thickness inside

The more common 275V X2 and X2 mini series from Kemet (for which i don't know whether a matching "series applications" series exists) also have this warning :
https://content.kemet.com/datasheets/KEM_F3093_R46_X2_275_110C.pdf
https://content.kemet.com/datasheets/KEM_F3094_R46_X2_275_110C_MIN.pdf

PME don't have the warning and dont mention specifically series applications either, they vaguely state "all X2 applications" :
https://content.kemet.com/datasheets/KEM_F3011_PME271M_X2_275.pdf

Got a link to one you recommend?

Kemet (the ones who took over production of the PMEs afaik) also mentions the across/in series with distinction in their product datasheets.

I kinda want to buy a batch of those SMD paper capacitors of them. Put a bunch in parallel with a fast fuse in series, and see how long they'll last. What stops me from doing that is that those things are still about €1,50 a pop in consumer amounts, and i want a decently large sample size...

I have seen one plastic dielectric cap that had gone up in smoke. It was a tubular one with 'straight' ends, no brand name on it. Oddly, in a Behringer PA subwoofer crossover filter. I don't know what went wrong in that PA rig, i wasn't there, but that capacitor had a big mushroom of carbonized resin bulging out of it. Many speakers were blown that night. But that still is the only catastrophic failure of a non-paper capacitor i've encountered.

I do agree with you, but that only goes so far when someone is eating glass or 80yo meat...

Lets have a look at some other capacitors

https://www.we-online.com/en/components/products/pbs/capacitors/film_capacitors/info

Go down to the FAQ at the bottom - "Can X2 caps be used in series with the mains ?"

Some applications called capacitive power supply or transformerless capacitive dropper use X2 capacitors in the input line (between mains power and the load). This use is called “in series with the mains”.
For this application, special THB X2 capacitors with explicit confirmation of this application must be used. We confirm this for certain types by specifying “Across the mains or in series with the mains” under “Application” in the data sheet. Würth Elektronik offers X2 THB film capacitors under the series WCAP-FTXH.

Seems very standards and ratings related, unless....

THB is not just some name of a series of capacitors. It stands for Temperature Humidity Bias - testing in which the capacitor life and reliability are tested under conditions of Temperature, Humidity, and Bias (applied voltage)

Yep, they are also concerned about humidity. From a manufacturer that does not make and never made paper capacitors

Wait a sec.. So to work in a capacitor/zener application, the capacitor must be one specifically made to last under THB conditions, but for plain X2 application for EMI filtering (which is definitely not any better working condition in terms of applied voltage/transients), all of that is not needed ?

EMI testing (to get certification for the product the capacitor is installed in) is performed on the 1st sample unit of the product and takes a few hours. (Assuming the unit may be powered for other testing and demonstrations before it arrives for EMI testing, make that a few months). Nobody will notice that it went missing afterwards. That is exactly as long as the EMI filter capacitors are required to last in the eyes of product manufacturers

This won't work for a capacitor which actually matters for the functioning of a product, which will result in mass warranty claims and recalls



For capacitors which design is older than some of the tests in question, yet actually last and stay in spec for the first 20 years, Rifa PME are anything but garbage

They are what they are - compact "metallised film" (assuming paper is a film) capacitors which are electrically fairly good, but have a unique other fault mechanism which is non electrical in its first stages, before it progresses into an electrical fault



The reason why old capacitors last long electrically is the low extent of process control in theirmanufacturing. Let me explain :

Life of film capacitors has a very strong relation to voltage. Life is proportional to (Vapplied/Vrated)^n, where values i seen for n = 7, 8, 10, 20 acc. to different sources (varies by film material, thickness, and individual manufacturers)

With older processes the film thickness uniformity, and risk of existence of pinholes, were not very precisely controlled. So they were made with overkill film thickness in the 1st place to guarantee that they work reliably

In PME capacitors it is also the fact that the dielectric is paper, naturally a fibrous and not very uniform material, which again results in some more margin taken in the design

All capacitors with metallised film technology (maybe except the earliest ones) have a self healing mechanism - Blowing up of metallisation around pinhole film breakdowns, allowing the capacitor to continue working after such event. In contrast, in foil capacitors every isolation breakdown is immediate full failure by shorting out

In modern capacitors the precision with which the film can be made (thickness, uniformity, cleanliness), and precision of models to estimate its life, allow to shrink it to exactly the thickness at which it will survive the rated voltage, with the self healing mechanism "ticking" just slow enough to be acceptable for an intended rated life

In most X2 capacitors nowadays the film is so thin that corona discharge around the edges of the metallization pattern (think of it, sharp edges of a metal layer of the thickness of microns-scale coating) becomes equally significant factor in the destruction of the metallization. They arent even "ticking" anymore, they just burn away continuously

The self healing mechanism really works exceptionally well if with all this, we don't see the modern capacitors explode and go up in smoke every day. But it too has its limits. Once in a while there will be a modern X2 capacitor completely melted