This contradicts what you said earlier about it happening when there is no regulator.
Perhaps I should clarify.
My first comment regarding the "real" regulator was unclear. Also, I may have misused the term "EOTD," adding to the confusion.
Let me define EOTD as when there remains no more liquid CO2 in the cylinder, and the gaseous CO2 begins to rush into the regulator, and it not being able to regulate the gas, will release it in increasingly large amounts into the aquarium (EOTD does not usually catastrophically release all the CO2 in at once).
Now, with most single stage regulators, the above phenomenon will occur.
With cheaper (I refer to this as "non-real") regulators, i.e. this one:
EOTD can occur, but it is more likely that they will not be able to regulate the pressure from the start (i.e. when there is still liquid CO2 in the cylinder), and will result in a catastrophic failure (as it did for that user).
Actually, in that thread, the paintball cylinder released all the gas (while the cylinder was new), killing all the shrimp.
Then (later in the thread), a single stage regulator end of tank dumped into another aquarium, killing off all German Blue Rams and their fry
Searching the web for a high pressure fine adjustment needle valve seems hopeless.
If you were open to running inline needle valves (rather than mounting them), you could use any standard needle/metering valve.