Thank you for the opporunity to provide comments upon this draft document. As in most fields of study, external review and comment can only help to produce a clearer, more accurate and more relevant final result and I am pleased to be part of this.
It is disappointing, however, that neither this report and its supporting materials nor any other Caltrain studies are published on the World Wide Web. Limiting distribution to printed paper copies both limits the availability of the information and makes it more difficult for interested parties to locate and respond to relevant information.
Please consider making future reports available on the web. We would be pleased to assist if there are any technical difficulties in doing so.
It is very discouraging that Caltrain should presume to undertake an "assessment" of all available "options" of equipment and, upon that analysis, make procurement decisions which will affect the type of transportation service available on the San Francisco Peninsula for the next thirty or more years.
It would be far more appropriate for Caltrain to develop performance-based requirements for new rolling stock (and for other equipment and services), and leave it to the ingenuity and experience of the global rail industry to produce bids which meet those requirements. Caltrain electrification is a project of sufficient size and of sufficiently high profile to be of interest to and to attract innovative solutions from a number of engineering/construction consortia.
The authors of the report appear to have a strong belief that new locomotives hauling the existing Caltrain passenger car fleet is the correct technical and fiscal solution to the problem of electrifying Caltrain. However, this conclusion is not strongly supported -- or indeed supported in any fashion -- by the data presented in the document. In fact it is possible to reach exactly the opposite conclusion, as will be noted below.
It should be borne in mind is that the global rail market, responding to the actual needs of and actual costs incurred by rail operators around the world, has overwhelmingly reached a conclusion different from that of that in the Rolling Stock Assessment Draft.
The most telling lesson should be the present situation in the United Kingdom, where I am unaware of a single passenger locomotive on order, compared to hundreds or EMU and DMU vehicles. The privatized Train Operating Companies and Train Leasing Companies of the UK are intensely aware of the bottom line, and would not be ordering equipment which extremely high operating costs (as stated but not supported in the Caltrain Rolling Stock Assessment Draft) or high capital costs if there were not an excellent return on investment. The UK's passenger equipment procurement situation is hardly unique -- the same could be said of train operators in Japan or Italy or Australia or Norway or many others -- but it is most compelling in the case of the UK because of the forthrightness of the commercial pressures and the quantity of new equipment on order.
Locomotive-hauled trains for suburban and regional service is now largely perceived as a solution which responded to technical limitations of the past (which have changed as the size of primary power transformers in particular has shrunk) and at best a stop-gap solution in the present, one with little currency for a system such as Caltrain's. (For example, operators with large numbers of general purpose locomotives might choose to buy unpowered passenger cars because of an expectation that the locomotives might be used for freight at night, or operators with large immediate equipment deficits might choose to buy and refurbish existing passenger cars for use with existing locomotives.)
But to reiterate: it should not be the job of Caltrain or its consultants to determine one "solution" and then solicit bids (or, in the case of all of Caltrain's recently over-specified rolling stock and fare collection specifications, only one real bid) to meet that narrow technical goal. I am willing to believe that there may be information of which I am unaware which would tilt the decision in favour of locomotive-hauled trains, but I do not think that Caltrain has the commercial information or the technical expertise to dictate this decision.
This statement appears to be based solely on such factors as passengers per linear foot of coach and weight per seat; operational issues and passenger amenity count for nothing. The appraisal of Caltrain's fleet ignores the unacceptable and increasing dwell times incurred as passenger volumes increase: something which has become very apparent in the last few years.
Most observers of the operation of the line, including all train crews whom I have interviewed, believe that Caltrain's already unacceptable combination of slow travel speed and abysmal on-time performance will only continue to degenerate as passenger volumes increase. That this is an attribute of the equipment and not inherent to growth is readily apparent from comparison with other railways which have experienced double-digit percentage growth in recent years. BART, as a particular local example, has not had its entire timetable fall apart simply because record numbers of passengers are riding its trains.
Simply put, equipment which spends unacceptable amounts of time moving at zero speed (whether because of excessive station dwell times for all passengers, schedule-killing dwell times for wheelchair users, or idle due to a low numbers of runs per day) is not equipment which can be classified as "competitive" with alternate equipment which spends more time moving paying passengers.
It is hard not to read this assessment, together with the unsupported (and ungrammatical) assertion on page 2-9 that "The gallery cars offer comparable station dwell time [to what?], an important consideration in commuter service" as more a defense of regrettable procurement decisions by Caltrain staff than an objective appraisal of the role of Caltrain's passenger car fleet in the current and future operational situations.
Page 2-12 section 184.108.40.206 notes that Caltrain's rolling stock meets the letter of ADA law, but fails to point out that deployment of wheelchair lifts incurs a severe and unnecessary time penalty: a single pair of deployments is will throw any train at least three of minutes off schedule (an interval which should be classified as "late"), enough to cause insuperable timetable conflicts once Caltrain starts to operate express trains timetabled to overtake same-direction locals along very regrettably restricted sections of passing track.
None of these significant factors appear to have entered into the report's appraisal.
It should always be borne in mind that we are seeking to solve the problem of "what is the most effective way to improve transportation in the Caltrain corridor in conjunction with system electrification?", not just the narrow one of "what is the cheapest way to buy electric locomotives and justify the existence of the existing fleet?"
My perception is that procurement of fast, powerful and expensive electric locomotives to haul a fleet of antique passenger cars whose design makes it inherently impossible to reliably adhere to any schedule will not result in much improvement to either the image or the reality of Caltrain operation.
There is a serious discrepancy between the $4.2m per locomotive reported on page 4-2 and the above reports. A clarification would be appreciated.
By that standard there is exactly one available electric locomotive, the $6.5 million per unit Amtrak/MARC HHP-8 which recently and belatedly entered service on the North East Corridor, and the Adtranz ALP-46 touted and finally recommended elsewhere in the report is not a viable option.
Contrarily, it may well be that it is but little more engineering effort for a rolling stock manufacturer to modify an existing EMU to meet FRA requirements than it was for Adtranz to derive the ALP-46 design, or to transform Bombardiers' existing unpowered bi-level FRA-compliant passenger car shell into an EMU unit. Without actual information -- in the form of engineering and costing studies performed by builders with a commercial interest in the outcome, i.e. bids -- it is both impossible and inappropriate to predetermine the choice of technology in this fashion.
Above all, consistent criteria should be applied when making comparative studies. I do not perceive consistency here and elsewhere where EMU operation is assessed.
A small note on Page 4-4 section 4.2.2: I was informed in personal correspondence with Bombardier that the capacity of the NYMTA LIRR M-7 cars will be 96 seated per car, not the 87 quoted here.
Yet nowhere else is this subjective criterion applied, and in particular not to the silk purse of a late 1990s design, $6m a piece electric locomotive hauling the sow's ear of a train of distinctly dated and "analog" early 1950s design gallery cars.
Again, this is not to acknowledge the validity of any of the criteria, subjective or putatively otherwise, used to assess rolling stock in the report, but to indicate a lack of consistency in application.
It is indicative of the tenor of the document that on page 4-12 the fact that the ALP-46 is "based on a successful German design and is manufactured by a builder with a history of North American experience" is presented as a major selling point, yet no place is found for mention of the experience in traction design and/or success in American sales of Siemens, Bombardier (now Adtranz' parent), CAF, Talgo or any other manufacturer.
Power system designs (locomotive- or EMU-based) are largely directly transferable from one existing successful design to another, and as such deployment within North America is largely irrelevant. Much of the difficulty experienced in commissioning new designs revolves around interactions between new power systems and old signalling systems; Caltrain is taking the opportunity to upgrade its signalling and should incorporate the lessons learned in Europe and Asia. As such, the direct if tenuously genealogical experience of ABB/ASEA/Adtranz/Bombardier in accommodating North East Corridor signalling systems is uncompelling and hardly unique.
In its other sense, "North American Experience" may be reduced to a familiarity with the unique regulatory burden which the Federal Railroad Administration imposes upon rail operators (to the huge fiscal disadvantage of passenger rail operators, and to no discernable safety advantage.) Experience in meeting regulated standards is a given for any of the rolling stock manufacturers, all of whom see "the world" as their market and are adept at adapting to local conditions. Adtranz has no advantage in this department over Talgo or any other manufacturer.
The analysis in this section again appears to display prejudice towards a particular outcome (an electric locomotive ordered by but undelivered to one North American operator) with a unique vendor (Bombardier-née-Adtranz.) The outcome of a bidding process with one pre-determined winner is not, as Caltrain or BART passengers know, generally advantageous to those who must fund and use the resulting services.
Such performance advantages not only provide better (faster) service to the travelling public, but may lead to substantial capital savings through increased fleet utilization. (BART, for example, is selling its extremely expensive and experimental AATC signalling system partly on the basis that the higher average speeds and consequent better fleet utilization it might permit might allow for a number of "virtual" new trains to appear without buying new rolling stock.) A few minutes' end-to-end running time saved might enable a train to run an extra trip a day, or more dramatically might make the difference in the total number of trains required to sustain a regular "clock-face" service pattern.
Nothing is said about the yard operational advantage of EMU trains, which take up less storage space (which is and will be in cripplingly short supply at Caltrain), are easier to process and make/break in the yard, and will not require separate locomotive servicing facilities. Such ongoing total ownership costs should be realistically simulated and accounted for in a comprehensive Assessment.
Nothing is said about the extremely high levels of availability of modern EMU fleets maintained either by the their builder or any other organization contractually bound to provide a low-fault fleet. For example, London's Siemens-maintained Heathrow Express EMUs are reported (May 2000 Railway Gazette International) to be achieving a mean distance of 116000km (over 730 SF-SJ round-trips!) between mechanical faults resulting in any delay of five minutes. It is frankly impossible to conceive of a regime in which Caltrain's existing passenger car fleet alone, even excluding locomotives, could be made to provide even the same order of magnitude of in-service reliability.
High availability and new vehicles designed for maintainability translate directly into lower fleet spares ratios, better fleet utilization, and thus a substantially smaller fleet for a given level of service.
Yet none of these factors are mentioned in the Rolling Stock Assessment, let alone quantified.
It would be extraordinarily regrettable -- and in fact a substantial breach of fiducial responsibility to the tax-paying public -- if Caltrain were to mis-manage its fleet upgrade process in such a way as to engender an expensive and suboptimal "chronology of procurements" resulting in a more expensive, less flexible and less efficient fleet.
This is stating that locomotive-hauled trains are more expensive for SEPTA to operate, to the extent that the agency is willing to leave extremely expensive assets (electric locomotives) sitting idle and depreciating for somewhere around 18 hours a day. This is in direct contradiction to the extraordinary and never substantiated assertion elsewhere in the document (page 4-15) that EMUs cost "cost up to 60% more to operate", and is complete agreement with the experience of the entire global rail industry, which has been the in process of abandoning locomotive-hauled trains for much of the last two decades.
The global rail industry appears to have reached conclusions quite different from those of Caltrain's consultants.
The discussion of the price of a hypothetical AEM-7 is again irrelevant, since no such locomotive is or ever again will be available.
There is no reason to suppose that transit demand in Caltrain's Peninsula corridor should be any less than that in the considerably less transit-favorable Oakland-Fremont corridor, in which BART operates EMUs to meet what must be "extraordinarily high demand" along a line with far fewer and more widely-spaced stations than Caltrain's.
Without any quantitative data on what the "extraordinary" level of passenger demand which justifies EMU service might be, or on what factors might justify "optimizing trip schedules" and how these might relate to transit demand along Caltrain's right of way in the next 20 years, one can only read this as a statement of prejudiced preference of either the authors or commissioners of the Assessment, and not as any sort of basis for quantitative analysis or for sound economic guidance.
Yet conclusions derived from this statement form the core of the flimsy economic analysis of section 220.127.116.11 and drive the major (pre-determined?) conclusions of the Executive Summary of the document.
Sweeping statements of this nature, particularly statements which appear so very much in opposition to those reached by nearly every other rail operator on the planet, demand detailed and reliable supporting analysis. Unfortunately and damningly, the Assessment fails to supply any.
The only breakdown of any per-mile operating and maintenance costs for any class of equipment is that of the antique NICTD EMU operation in section 3.1. Yet the Assessment does not provide comparable data for New Jersey Transit, the Long Island Rail Road (which is in the process of procuring 1000 EMUs), STCUM, MARC, SEPTA, Metra, Metrolink or even for Caltrain! This is leaving aside the national rail operators of Europe, the private Train Operating Companies of Britain, the private operators of Japan, the state operators of Australia, the recently floated MTR of Hong Kong, Indian Railways or any of the other railway operators who could provide solid comparative financial data.
Worse still, the only operator of both EMUs and loco-hauled trains reported on in the document (SEPTA) was found on page 4-13 (section 4.3.2) to not find it economical to operate locomotives except at rush hour. This surely leads to a conclusion about operating costs radically different from that of the Assessment.
Without even token support of the central economic tenet of its conclusions the document can only appear rather amateurish, and can not and should not form the basis for any multi-hundred-million-dollar procurement decisions.
The same paragraph goes on to state that "[...] some commuter rail operators [estimate EMUs cost] up to 60% more than an equivalent locomotive hauled system." Which operators? Based on what data? How accurately accounted? How are the cost differentials explained? How are the discrepancies between the experience of these unnamed operators and hundreds of EMU operating systems elsewhere to be accounted for?
These are extraordinary claims, and demand solid supporting evidence.
No data are provided to support the assertion that the price differential between an unpowered single-level car and a single-level EMU is $1.25m ($3.0m - $1.75m). In contrast, the only recent North American EMU order, the admittedly competitive NYMTA LIRR M-7 car bid won by Bombardier, will deliver entire single-level EMUs for $1.85m a unit, or little more than the Assessment prices the car's traction equipment alone!! Even allowing a higher unit price closer to $2.5m (for a hypothetical 100 car order, compared to the 1000 orders and options of the NYMTA contract), it is difficult to believe that traction equipment alone could cost $1.25m, or over 50% of the total price.
The estimate of $2.8m per EMU, of which $1.25m is traction equipment, should be treated with the greatest caution up until such time as a vehicle builder signs a contract for that amount. In the meantime, more realistic comparisons and more realistic train composition scenarios may provide better guidance.
Data from European EMU orders provide similar cause for extreme skepticism. Over the last five years and predating the recent declines of the Euro, all of the major manufacturers have been part of consortia contracted to deliver entire EMU vehicles -- including the 180kmh trains in international service across the new Øresund road/rail link between Sweden and Denmark delivered for $4.8m per triplet or $1.6m per car -- for just slightly more than the "estimated" traction component cost of this Assessment. Note that these costs are largely independent of both the FRA regulatory burden and the "Buy American" trade protectionism as traction equipment does not determine train impact strength and since railway power control electronics, high power traction motors, pantographs and the like cannot be procured locally. No argument can be made that foreign-sourced traction equipment alone delivered to American could cost more than entire foreign-sourced vehicles delivered to Europe's open market. Note also that since traction systems and components are increasingly developed and sold as standard systems across many different product lines there will be little price premium for the power components of an 80-vehicle Caltrain order compared to those for in a 200-vehicle order in Germany or Britain or elsewhere.
Furthermore, this "estimate" appears to make the unrealistic assumption that all vehicles in an EMU fleet would be powered (to the tune of $1.25m each.) In suburban and regional service it is far more typical for EMUs to be formed into motor+trailer pairs or (for >160kmh operation) motor+trailer+motor triplets. In this case, the price of the trailer car is approximately the same as any other unpowered passenger car. Even pessimistically assuming $1.75m for each new trailer and taking the outrageous $1.25m per motor car traction premium suggested in the Assessment, we arrive at an average cost per of $2.4m per car in a motor+trailer pair, or $2.6m per car in a motor+trailer+motor triplet.
Assuming the same 6.7% "competitive bulk discount" for a fleet purchase as in section 4.4.2 of the Assessment, and assuming a more realistic but still extremely conservative $1.0m traction premium, one might estimate the cost of a motor+trailer pair at $4.2m (1.75+1.75+1.0 x 93.3% = average $2.1m/car) and a motor+trailer+motor at $6.8m (1.75x3 + 1.0x2 x 93.3% = average $2.3m/car)
These numbers are significantly below the "estimate" of the report, and though extremely conservative will still lead to economic conclusions at total variance with those of the Assessment.
The failure of the table in this section to display even roughly comparable numbers (consistent inclusion of annualized remanufacturing costs, comparison of equipment in similar service patterns) makes it actively misleading.
More informative would be to solicit operators of fleets of modern diesel and electric locomotives and EMUs and DMUs for their cost bases. While the absolute numbers may not translate directly to Caltrain's case, the comparative figures might be revealing, and would lead to quantitatively based conclusions quite different from those arrived at by other means in the Assessment.
This makes for very embarrassing reading.
To be even borderline realistic, it could be assumed that 20 or 21 electric locomotives could replace 23 diesels and 84 (more likely 80) EMUs replace 23 locos and 93 gallery cars and provide the same level of service.
Updating the electric loco cost to $6m/unit (in line with the
$6.2 to $6.5m which NJT and MARC are paying, not the $5m in the
table on page 4-16) and updating the EMU cost to $4.2m/pair (down
from the outrageous $5.6m of section 4.4.2) the table on
page 4-17 should be corrected as follows:
|Caltrain Draft Rolling Stock Assessment Scenarios|
23 locos and
with 93 EMUs
|Corrected scenarios assuming $4.2m per EMU pair, 20 $6m electric locos|
23 locos and
with 84 EMUs
($73.8m cheaper than locos!)
($84m cheaper than
|Corrected scenarios with numbers stacked
steep $7.2m per EMU triplet, only $5m per locomotive
23 locos and
with 84 EMUs
(Still $28.6m cheaper than locos,
even at a very pessimistic
$7.2m per EMU triplet)
Under these corrected assumptions, it appears to be fiscally insane -- to the tune of 40 to 70 million dollars -- to purchase electric locomotives rather than replace the fleet with EMUs.
What is not accurately stated are that the risks may be low -- particularly if sufficient contractual incentive is provided -- and that the alternatives may not be very undesirable. Moreover, the degree of risk should be contractually assumed by rolling stock manufacturers and objectively evaluated by Caltrain and its consultants as part of a bidding process, not predetermined in a partial and early Rolling Stock Assessment.
Certainly the preferred solution of the Assessment's authors -- that of putting a $6m electric locomotive at the head of a train of 1950s-era gallery cars with abysmal station dwell times -- strikes many observers of Caltrain as an alternative only in the academic sense: it does indeed clear the bar of providing an electrified Caltrain service, but does not clearly result in improved service.
The technical difficulty of introducing a new EMU design appears grossly overstated in view of contemporary rolling stock design. Several FRA-acceptable unpowered double-deck passenger car designs are available to provide the basis for a car shell, traction control electronics and pantographs are standard items, and traction bogies heavy enough to support even an overweight US passenger car are in service underneath thousands of locomotives already. Caltrain will not be making any extraordinary demands in terms of speed or internal or external noise or interior amenity, the only novelty is the large FRA weight penalty. An EMU for Caltrain might very well be a straight-forward adaptation of an existing unpowered design or of an existing foreign design; it should be up to rolling stock bidders to assess the risks and reflect them in the contracts they offer.
This is again misleading. The NYMTA LIRR M-7 Bombardier EMU cars, which are presumably the design referred to, do not "exist" in a legal and regulatory sense: they are also a new design, yet the report briefly considers them as proven and delivered (in the same way it "evaluates" the non-existent Siemens NJT locos) before dismissing them out of hand for not being double-deckers.
It is also far from uncontestable that multi-level cars are in fact required for Caltrain service. I'm prejudiced by personal history to believe so, but I'm also aware of the interesting high-capacity, rapid-loading, wide single-level EMUs ("Regina") being introduced in Scandinavia.
Again, the experience and the innovation of the global rail market should be allowed to propose innovative solutions which meet the performance criteria of Caltrain, rather than have Caltrain dictate design down to inappropriate levels of detail.
But then the Assessment goes on to complete mockery of "performance specifications", suggesting that exactly one of two locomotives from exactly one vendor (post Bombardier/Adtranz merger) is the limit of procurement possibility.
This is quite astounding conclusion, given that Adtranz has not delivered a "US proven locomotive" to New Jersey Transit and that Bombardier/Alstom was two years late in delivering the Acela fiasco.
There is no reason to suppose that any other qualified bidders could not deliver products as well as or better than the sole-source Bombardier-merged-with-Adtranz favored by the Assessment's authors. This prejudicial conclusion is unwarranted and counter-productive.
So much for competitive procurement. So much for "performance standards". Adtranz: name your price!
What incentive? There is to be one bidder! Just like for Caltrain's disastrous sole-source 1997 gallery car and locomotive procurement and the disastrous 1998 ticket vending machine procurement, in all of which both the public's fiscal interest and the public's opportunity to influence design were short-changed by unprofessional and irresponsible over-determination of design and pre-determination of bid winners.
The entire basis of the Assessment will have to be radically altered before any sort of competitive incentive for equipment construction and maintenance is possible.
The claim that performance requirements will take longer to develop in the absence of a "current" design is a complete non sequitur: specifications are completely independent of current designs -- unless, of course, the meaning of "the process should follow the same basic steps" is that a single qualified bidder be identified in advance!
Real performance based requirements could easily be developed today for any sort of equipment: it is the job of bidding manufacturers to evaluate design risk and ongoing maintenance costs and to offer contracts which reflect their commercial assessment. Caltrain's small in-house staff is both unable to and unqualified to pre-determine the rolling stock design which will best serve its customers.
Perhaps. But perhaps not. Certainly nothing written down as part of the Draft Assessment of Electrically Powered Rolling Stock Equipment would necessarily persuade a dubious observer.
It should also be constantly borne in mind that the purpose of electrifying the Caltrain line is to improve service to customers and improve the environment and economy of the region, not just to string wires and buy locomotives as an end in itself.
This statement is completely at variance with reality, reflecting a PCJPB staff view of the world rather than a Caltrain passenger's.
In reality, the gallery cars are loathed by all who have ever experienced any alternative, and are "accepted" solely by those who were responsible for their procurement.
Far from a positive outcome of locomotive procurement, the prospect of indefinite operation of the inefficient and passenger-hostile gallery car fleet is yet another excellent reason to consider a different solution. Let us hope that Caltrain has the wisdom to be open to a more open-ended manner of determining solutions.