feb 2020 note: after 20 years most of the external URLs died, surprise surprise. Rather than delete them outright I neutered them so you can see them as plain text. (fyi, in 2010 I converted the car back to gasoline, a Howell TBI EFI kit. As of Feb 2020 the car is still on the road, alas not in my name. I'd take it back in a minute.
Tom Jennings, 3 August 2003
In 1987 I embarked on a project to convert my car to run on LPG (aka propane) instead of gasoline. What little I knew at that time indicated it would be a great motor fuel. I built the car in 1989/1990, and wrote it up as a small booklet, "Do-It-Yourself Automotive LPG Conversion", and amazingly, sold about 1200 copies in two years mostly through the Real Goods Trading Company (Ukiah, CA).
It's now 2003, and I'm still driving the same car. Not only have I not had any problems with the fuel system (hardly any with the car at all), it's actually running substantially better -- on long highway trips I get nearly 21 MPG equivelant (see below for LPG/gasoline comparisons), nearly a 20% improvement over the claimed mileage of this car when new (18 MPG with standard transmission and overdrive). This isn't bad math or wishful thinking -- I have fuel consumption/mileage data back to 1998. Around town (Los Angeles city driving) I get approx. 15 MPG gasoline equivelent.
Below is the original booklet, updated in Mar 94, Apr 97, Nov 98, plus additional accumulated technical information. For a fuel that's been around since the 1920's there's shockingly little available information.
My conclusion to the question, is LPG any good as a motor fuel? An unqualified yes. It easily burns clean, being a small-molecule gas (products mainly water and carbon dioxide, and less of that per-mile) and cheap. It has no complex additives that I can determine (eg. no MTBE, lead, ethyl-anything). But not profitable for oil companies. Read on if you care. There might even be enough information for you to duplicate my efforts, for the very few people who care.
LPG is a wonderful internal-combustion fuel. For automotive iron produced before approximately 1980 the advantages are pretty much as I describe them I believe -- a hands-down winner. This is because these old cast iron engines are so terribly inefficient and broadly-tuned to begin with (eg. my '63 Rambler) that the enhancements made via LPG retrofit amount to "picking the low-hanging fruit".
However, modern designs like run-of-the-mill Honda Civics have already picked all of the low hanging fruit, easy to do when you have a crew of engineers, a budget and five-year plan: computerized closed-loop (OBD II, etc), excellent metallurgy, accurate hydrodynamic modelling, etc. A horsepower per cubic inch used to be the realm of $1M race cars; now every cheap import's got that. My Rambler's 232 cubic inch motor gets about .35 HP/cu in! I love it dearly, it's a pleasure to drive and work on, but it's as big and heavy as two Civic's complete drive trains.
Let's say LPG on a mid-60's engine gives a 50% improvement in tailpipe pollutants (just a hazardous guess). An equivelant 1995 motor with air mass, tailpipe O2, knock sensors and fuel injection has long ago grabbed up that "easy" 50% improvement, and then some; ULEV cars are barely measurable. Adding LPG to the mix would get it... 5% better? 10%? 1%? bang per buck, it's not worth the overall increased system cost.
But for retrofit on old stuff it's hands-down great. After 14 years I'm still happy with it.
a-href="http://www.autolpg.com" targ=Automotive LPG on the Electronic Highway, a superb resource for basic information, history, and detailed technical overview of automotive LPG fuel systems. Leave it to the Dutch to make such a clear, concise, zero-bullshit site! It's really a great site.