It's funny how different passenger and fright trains are.
If you are designing your tracks for passenger trains, you can use very steep grades that would be impossible for freight trains. If you are designing for electric passenger trains (with overhead wires for basically infinite energy) you can have trains climb incredibly steep tracks for a very long time, since passenger trains are almost always overpowered. Seriously, the locomotive to car ratio for passenger trains is rediculous compared to freight trains, and that doesn't even consider how light passenger cars are compared to freight cars (passenger cars, no matter how full of people they are, are still mostly full of air).
The one thing fast passenger trains can't handle is sharp curves, because then they have to slow down.
If you are designing your track for freight trains, you want to avoid grades at all costs, since you will be lugging millions of tons up and down that grade forever. Curves are great, since they flatten out the slope of your tracks.
So for freight trains, grades of 2% is really steep, and the tracks are usually very curvy. For passenger trains, grades can be
as steep as you like but curves are dealbreakers.
So the grades of I-15 in Cajon pass being 6-7% are not a problem. The real problem wth I-15 is the curves.
Road curves are different than railroad curves. Highway curves use a single radius, since road lanes are wide and drivers can steer within the lane to smooth out G forces.
Trains cannot steer within their track, so the curve itself needs to be laid out with a variable radius in order to minimize G forces (and avoid derailments). These curves are called
Spiral Curves, and have a very gradual entrance and exit that are extremely difficult to calculate. (The difficulty in calculations is a main reason why these curves are not common on roadways.)
So even though the speed limit of I-15 through Cajon pass is 60-70 mph, trains that follow that alignment will not be able to go that fast. Probably 45-50 mph at the most.
This applies more broadly than I-15 in Cajon; in most cases where rail transit had been shoehorned into road ROW's, the cars are allowed to go faster than the trains for the simple reason that the infrastructure was not designed for trains.
Let trains be trains and let cars be cars.
Sticking to the existing railroad ROW's will be just as fast, if not faster, than using I-15. (An all-new passenger rail ROW would be fastest, but let's be realistic here...) There are 3 rail ROW's through Cajon pass: One Union Pacific (formerly Southern Pacific) track furthest up the mountainside, a double-track shared UP/BNSF line halfway down the mountainside, and a single-track BNSF/UP line that runs mostly along the bottom of the canyon.
This third track at the bottom, called Mainline 3, is the best for passenger service. It has the fewest curves and what curves exist are really gentle compared to the other two options. Because of the few curves, it is also the steepest option at 3%, but that doesn't matter for passenger trains.
If I were the consulting engineer for this project (
) I would recommend that VTUSA builds a second track next to Mainline 3. Perhaps it would be possible to build a third track beside the existing double-track line (Mainlines 1 and 2) and then buy Mainline 3 from BNSF outright. Perhaps get the state of California to buy it as a dedicated passenger rail line to be used by VTUSA, Amtrak, and Metrolink (I can imagine Metrolink service between San Bernardino in the medium-far future.)
Anyway, Mainline 3 is the way to go. It is exactly 45 miles between San Bernardino and Victorville by rail -Amtrak already runs this route in 70 minutes. It would surprise me if a VTUSA train, with their much higher power-to-weight ratio, couldn't do it in 45 minutes.
PDF map of rail routs in Cajon Pass
As for cost, following Main 3, no new major bridges or tunnels would need to be built, but the I-15 bridge over Main 3 would need to be rebuilt in order to accommodate 2 (or more) tracks. The last 5 miles into San Bernardino are extremely tight and other adjacent infrastructure would need to be moved to accommodate another passenger-only track. But if we make a broad generalization and say $5 million per mile, (higher than the estimate of $2 million per mile on flat ground) we arrive at a price of
$225 million. This does *not* include the ROW costs, which may end up doubling my estimate, but even so I think this shows that a dedicated passenger track up Cajon is not all that far-fetched an idea.
Before people say that this is being unreasonably optimistic, in 2008 BNSF completed
16 miles of new track in Cajon Pass for $90 million. This project also included things like daylighting old tunnels and some very significant earthwork and slope stabilization projects because the new track is very high up the mountain (to keep grades low). $90 million/16 miles = $5.625 million/mile, so my estimate of $5 million/mile is not unrealistic.
For context, Brightline's Orlando extension is expected to cost
$2 BILLION. That extension is only 40 miles long, but there are a significant number of bridges and underpasses, and all the track is designed to run at 125 mph, which adds significantly to costs. The cost per mile of this project is not comparable to Cajon pass, but the significantly lower price of my estimate is very comparable; it shows that VTUSA would be well within its resources to extend their tracks into San Bernardino if they chose to do so.