What year did steam locos begin using oil for fuel ? - Model Railroader Magazine

The documentation I've seen indicated that the earliest practical application of oil fuel to locomotive firing dated to the early 1880s.  This involved heating broken rock or refractory pieces on a firebox grate and then spraying oil over it -- this both carbureted and cracked the oil and served as flameholding.  

Practical use of the grades of oil that were most "economical" (read: cheapest to source and use) were facilitated by the early development of steam-atomizing burners, often of the 'drooling' type like the famous von Boden-Ingles type used on ATSF.  These used steam both to heat and atomize the oil at the point of delivery and introduce primary air through long slots in the burner, the result being a long, flat plume of luminous flame filling the firebox volume with minimal impingement on "firepan" or internal firebox surfaces.  Usually this was installed at the throat, firing back into the firebox space, so that the heated combustion plume would pass around a masonry arch and back over the crown, increasing both the TOF available for combustion and the area exposed to luminous radiation (the uptake being proportional to the fourth power of temperature).

Some experiments were done with other burner types over the years, but comparatively little involved mechanical atomization (as in the Racer burners in marine practice).  The late version of Southern Pacific practice was a device called the Gyro-Jet burner, which used vanes to swirl the oil and primary air to get better atomization and entrainment with heavy oil.

SP did a careful comparison testing of different designs of burner in the early Fifties, right at the practical end of innovation in oil-fired steam in the United States, and wrote up the results in a study published in 1951.  This is available in a couple of formats (see for example the link in the recent thread on RyPN).  I have the PDFs but they are too large for conventional e-mailing, even if the e-mail feature here still worked.

Alternatively, light-oil firing could be a 'thing' -- some of this involving what we currently consider diesel #2 complete with its additives optimizing its use as a compression-ignition engine fuel.  PRR spent quite a bit of time and effort designing various engines to use this fuel, culminating in a detailed consideration of the Steamotive system in the mid-Thirties (this was the system used in the two Union Pacific 2500hp GE steam turbines of 1938).  Locomotives built to use heavy #5 (or bunker) oil could be effectively fired on this fuel with appropriate care and modification -- Frisco 4-8-2 1522 being a particularly well-documented example.  There is less heat content per gallon in the lighter oil, and its flame is not as luminous in spectra that give best radiant uptake, but it is far easier to handle and, in the modern world, much more easily obtainable.

As a note, both in England and in the United States there were 'scares' in the late Forties about a concerted coal strike eliminating coal as a locomotive fuel source.  Considerable engineering work, of interesting quality, was done by almost all American railroads into emergency oil conversion of coal-burning power -- amusingly, N&W prepared very detailed drawings of oil-firing equipment, including for the J class, and that remains at NWHS in case anyone wants a 'historically accurate' oil-firing system for 611.

Tender 'conversions' were often done with drop-in sheet-metal bunkers, but there were some conversions of 'water-bottom' tenders that used the existing side and slope sheets as part of the 'bunker' to save some weight and cost.  Note that a Vanderbilt tender, referenced earlier in the thread, only uses cylindrical construction for its water cistern space -- its coal bunkerage is wholly conventional, and I don't know of any Vandy that was oil-converted to use a cylindrical oil reservoir of any configuration (that doesn't mean there might have been a couple, probably 'drop-ins' where using the full capacity of the bunker space was less essential).