HP question

I am sort-of new to this railfan thing, so please excuse my ignorance, but I have a question. When it is said that a locomotive has "4000 Horsepower", is that the HP of the diesel prime-mover, the HP that the alternator that is driven by the diesel engine puts out (converted from watts, or kilowatts), or is it the HP that the electric traction motors produce. Or is it possibly some sort of combination there-of? Thanks in advance for any answers.

Simple question, no simple answer!
The power ratings used for diesel-electric locomtives in North America are input to traction generator. For most other purposes, I gather internal combustion engines are rated by brake horse power, but some of the power goes to run assorted auxiliaries (radiator fans, the compressor for air brakes, ...): a locomotive with what other people would call a 2600 hp engine might deliver only 2400 hp to the main generator for traction purposes, and so be called a 2400 hp locomotive.
(Outside of North America, diesel electric locomotives are rated by bhp. Thus the "2400 hp" Alcos of ca. 1960 had the same engine as top-of-the-line Alco export locomotives advertised as 2600 hp, and the engine of a 3000 hp domestic EMD SD40 is the same as in a "3300 hp" export GT26C.)
Now for transmission losses: generators and traction motors are efficient, but (remember: ENTROPY INCREASES!) not perfect. Horsepower "at the rail" tended to be about 82% of the rated horsepower of the locomotive for the first several decades of American dieselaization. (I have been told that modern AC locomotives are somewhat better.)
Example: in 1979, "Trains" published a 4-part article on the basics of diesel locomotive technology, which Kalmbach subsequently reprinted as a pamphlet, "The Diesel from D to L," by Vernon L. Smith (sometime mechanical officer for the Belt Railway of Chicago). As an example, he discussed the GP38-2. The engine of this "2000 hp" locomotive "put[s] out about 2225 hp to cover auxiliary and traction requirements." The generator (note that this is an AC/DC locomotive with an AC generator and rectifiers to provide DC to the traction motors) efficiency is 94 to 95%, and the traction motor efficiency ("including cabling losses and gear losses at full speed") is 86%, so the horsepower at the rails is 1634.

In addition to Allen's excellent and comprehensive answer, you also have to consider what organization's rating methods are applied to the published data, as the calculation of auxiliary losses differs greatly between different standards organizations in their specifications for testing diesel engines. And the diesel manufacturers can sometimes play some games with ratings too, like specifying unrealistic air inlet temperatures and barometric pressures to make the performance of their engines look particularly good.

Mxdata--
Do you have any examples? The nearest thing I can recall is that the FM Erie-built* was rated at 2000 hp under certain ... optimistic ... assumptions as to ambient temperature and/or pressure, and was in practice often the equivalent of what would normally be rated as an 1800 hp locomotive. ...
More generally: is there any scuttlebut as to whether "X horsepower" from one locomotive builder tends to mean something different from "X horsepower" from another?
---
*Kevin B: if you are really a newcomer to railroady stuff, you can be forgiven for not knowing about this. After the Second World War, the Fairbanks-Morse company had a diesel engine design (used in many U.S. Navy submarines and destroyer escorts) that they wanted to market in locomotives, but no assembly plant suitable for building large locomotives. So they contracted with General Electric to have a large cab unit built at GE's plant in Erie, PA. Somewhat over a hundred were built between 1945 and 1949, and the last were scrapped decades ago. Kalmbach Books' "Diesel Locomotives- the first 50 years" (not sure that's the exact title) isn't the most complete or the most scholarly history in the world, but is a good start if you are interested in diesel dinosaurs.

Allen Hazen wrote:
Do you have any examples? The nearest thing I can recall is that the FM Erie-built* was rated at 2000 hp under certain ... optimistic ... assumptions as to ambient temperature and/or pressure, and was in practice often the equivalent of what would normally be rated as an 1800 hp locomotive. ...
More generally: is there any scuttlebut as to whether "X horsepower" from one locomotive builder tends to mean something different from "X horsepower" from another?

Allen,

There are standard conditions to compare the output of diesel engines.
for example AAR conditions (60 F ambient Temp. and 1000 feet altitude above sea level) and UIC conditions (25 C at sea level)
Engine suppliers render guidlines to calculate natural de-rate of their engines in form of graphic diagrams, ladder diagrams (EMD for instance), tables or formula

There is also an international standard (ISO 3046) to calculate natural de-rate of engines in general.

Natural derate mainly related to these parameters:
barometric pressure
air temperature
inter cooler water temperature
fuel viscosity
fuel temperature

Different engines are different in de-rate nature, for example those which has double cooling circuits and use LT (low temp.) water for intercoolers and oil coolers are far more better than single circuit ones. (apparently it helps to cool charge air more, and the cooler charge air means more density and more oxigen rich air to aspirate so lower derate in hot environments.)

One of the favorite salesman tricks is to apply the 60F ambient standard that Jamshid noted, to applications where it is totally irrelevant. Engine salesmen sometimes hand out fuel consumption figures for their products based on the 60F standard, to users who would operate the engines in 120F to 130F ambient conditions. It makes a BIG difference! Some of the engine manufacturers (particularly those building the smaller engines) have a substantial library of rating charts for their engines under various conditions (including non-standard conditions), which can supply a lot of information to a knowledgable customer, or cause a lot of confusion for one that does not understand the rating conditions.

One factor which is usually not taken into account in published ratings is overfueling transients during load fluctuations. Most engine tests are either steady state or specific duty cycle related. When you subject the engine to transients due to wheelslip, you can get much different results. This has tended to work to the benefit of EMD in many over the road tests, because the 645 and 710 recover from load fluctuations well.