A Letter I've Written for Work
(All of you Metro bus riders should appreciate what some of us have had to recently endure. I thought I'd share, and welcome any techie corrections or suggestions. -P.)
In response to a request from my union, ATL 587, I have decided to contribute my two cents about the recently converted Breda trolleys and the problems operators have bringing them to smooth stops. As background, I was at one time active in the Seattle Electric Vehicle Association, and absorbed as a direct result of that participation a smattering of understanding how electric propulsion systems operate. Based on that, I can say that the Breda dynamic braking system cannot be relied upon to smoothly bring coaches to a stop on Seattle city streets.
I had my initial experience with Bredas straight out of part-time training four years ago, running morning Sound Transit 550 routes for two and a half shakeups. I got used to the beasts eventually, but did notice some quirks, such as the fact that I only had problems in consistently bringing coaches to a smooth, feathered stop in electric mode when southbound servicing the University Street Tunnel zone.
Why? That zone was the only one with a slight downhill grade. To understand why that was a problem, one needs to understand the Breda's dynamic braking system, especially as it compares to the system equipped on the older 4000 series trolleys.
First, some electric basics. All electric drives consist of a power source (in our case the overhead wires), a motor and a controller. A controller acts to limit the power from the lines to the motor, similar to the way a carburetor or fuel injector limits the fuel to an internal combustion engine.
The 4000s use direct current (dc) traction motors, a technology that has been around for over a hundred years. To slow such a coach dynamically, the controller feeds a current to the motor in reverse, inducing drag on the motor. This system can deliver very smooth dynamic torque all the way to a stop.
By contrast, the Bredas use alternating current (ac) motors. The control systems for an ac drive are technically more complex and were thus not practical until powerful and inexpensive microprocessors became available. The dynamic braking system on Bredas, however, is a bit simpler. Instead of drawing current from the wire to induce the motor to slow, the Breda controller turns the spinning motor into a generator and directs (or "shunts") the power it generates from its forward momentum into a load bank. Look on the top of a Breda; just forward of the pole mount you will see three resistors, what look like coiled springs with air gaps between the coils. These resisters turn the electricity from the motor into heat exactly the same way an electric baseboard heater or electric range, for example, might turn your household electricity into heat.
The problem with this particular dynamic braking system lies in how it is executed in the Breda.
Rather than electronically regulate exact amounts of electricity shunted to the load bank resistors, the brakes of the Bredas actuate three simple switches. Lightly depress the service brake and one resistor connects to the motor's output; more pressure actuates two, and still more connects all three. This gives very rudimentary control over slowing.
As I mentioned above, the only problems I had with the system involved attempting to stop smoothly while servicing the only downhill grade in the tunnel. To explain why will take a bit of effort, so please bear with me.
Let's say you are traveling in the tube between Westlake and University, and need to slow your approach for the zone. Furthermore, let's say the motor spin at 30mph, the old speed limit in that tunnel section, has the potential to generate up to 80,000 watts if all of it were captured at once (all the numbers are speculative -- I made them up -- but the principle is the same), and that each of the load resistors can generate 1,000 watts of heat max. You gently depress the brake, actuating a resistor and diverting 1,000 watts of your spinning motor's energy to the load bank, and feel a gentle slowing, almost too little to notice. A bit more pressure on the pedal and 2,000 watts goes to heat the tunnel; more pressure increases that to 3,000 watts. Perhaps with help from the brake shoes and drums, the coach slows.
This is where things get tricky.
As the coach and the motor slow, the maximum potential of the motor is reduced; but the load resistors always try to draw the same amount. Therefore, as the coach gets to 20 mph, the potential motor power would only be, say 40,000 watts; at 10 mph, 10,000 watts. A 3,000 watt load on a 80,000 watt (30 mph) motor represents a bit over 4% drag; the same load at 40,000 watts (20 mph) equals about 7.5%; when we get down to 10 mph and 10,000 watts the drag is up to 30% of the motor's output! And as the percentage of motor output being drawn by the resistors increases, the greater slowing torque the dynamic braking exerts on the coach.
Of course, we operators have been trained to "feather the brakes," to reduce service brake pedal pressure as the coach slows to a stop, and then to firmly engage the pressure once that stop has been realized. So I would lighten my touch, disengaging third and then the second resistor, and reducing the drag of the load bank.
Even with only one resistor engaged, though, the coach will eventually slow enough that 100% of the motor's output would be diverted to the load bank, resulting in very a sudden stop (and perhaps potential damage to either the motor or controller).
That's why the controller disengages the dynamic braking at between 3-4 mph.
Disengages.
Meaning that if you are rolling downhill, the dynamic drag that kept the coach from accelerating suddenly disappears, resulting in a sudden acceleration . . . that necessitates an equally sudden reaction from the operator, a panicked jamming of the service brake, feathering be damned.
For any driver accustomed to feathering brakes in equipment other than the Bredas, this is an operational contradiction. Any attempt to feather Breda brakes will result in a choppy stop.
Oh, and one more thing. As the slope becomes steeper, this unpredictable downhill acceleration at 3-4 mph dramatically increases, leading to stopping that gets even choppier and even sloppier. I remember considering southbound University Street the trickiest stop to make in my route. I cannot even imagine trying to make a smooth stop coming down the Counterbalance.
Make no mistake: This is a design problem. It can, like any design problem, be corrected. For starters, that three-stage actuation system should be scrapped. Instead, a dynamic braking controller could be inserted between the motor and the load bank that would reduce the current to the resistors as the coach slows. This would have to be custom designed and manufactured; but it would solve the problem.
However management chooses to deal with this design flaw, I would stress that more than just smooth stops of trolley routes is at stake. This is a serious safety issue, to be sure.
It is also a serious training situation for all operators who must drive these coaches. To understand how serious we should better understand the story of Pavlov's dogs.
Most everyone has heard of the famous experiments Ivan Pavlov conducted at the turn of the century, experiments and conclusions that earned him the Nobel Prize in 1904. To refresh your memories, Pavlov and his assistants attached dogs to salivary measuring equipment which measured how much each dog drools at any given time. A few minutes before dinner someone would sound a buzzer in the dog's ear. They were this way able to "condition," or train the dogs to get ready for dinner by drooling.
That's the part of the experiment about which most have heard. What of the rest of the testing?
After the dogs were conditioned the experimenters then "deconditioned" them. The buzzers would be sounded, but not at dinnertime. The dogs caught on quickly and stopped drooling after the buzzers. After deconditioning, they would then "recondition" them again, feeding consistently right after buzzing. Reconditioning, like retraining, didn't take nearly as long as the initial conditioning.
What they did later I find personally disturbing and enlightening.
Some dogs, those that had never been given drool training, were subjected to random buzzing. The person buzzing the dogs and the person feeding the dogs never spoke. Buzzing was just some weird distraction to these dogs, one best ignored.
When well conditioned dogs were subjected to random buzzing, however, the result was quite different.
Those dogs went completely insane.
Specifically, dogs trained for months to associate the feeding buzzer with the food bowl reacted to the random buzzings and feedings by developing a schizophrenic state known in psychology as "waxy catatonia." The dogs refused to move. If you wanted to, you could move a leg or a paw, hold it for a bit, and the dog would keep it there until it got too tired. These dogs stopped eating altogether and had to eventually be euthanized.
It's like they just gave up.
The dynamic braking system of the coaches now known to drivers as the FrankenBredas has the potential to do to drivers what random conditioning did to Pavlov's trained dogs. Every day, Atlantic operators sign in not knowing whether the dynamic brakes of their coach will operate until the coach stops or not, whether the torque of braking will increase linearly, or rise sharply and then end altogether in yet another brief uncontrolled downhill descent.
Yes, we are humans, not captive dogs, and can theoretically handle more stress than our canine cousins; but these trolley routes are already the most used in Metro's system, accounting for (if I remember correctly) 35% of the ridership systemwide. Every additional stop, every additional rider -- not to mention the very real antagonism some riders on these routes seem to exhibit to All Things Metro -- all adds up to stresses which threaten to undermine we Humans Behind the Wheel without considering if today's brakes will function properly or not.
Management has an obligation to standardize its fleet if it ever expects to develop drivers with standard habits and skills that reflect well on all of us here at Metro.
In response to a request from my union, ATL 587, I have decided to contribute my two cents about the recently converted Breda trolleys and the problems operators have bringing them to smooth stops. As background, I was at one time active in the Seattle Electric Vehicle Association, and absorbed as a direct result of that participation a smattering of understanding how electric propulsion systems operate. Based on that, I can say that the Breda dynamic braking system cannot be relied upon to smoothly bring coaches to a stop on Seattle city streets.
I had my initial experience with Bredas straight out of part-time training four years ago, running morning Sound Transit 550 routes for two and a half shakeups. I got used to the beasts eventually, but did notice some quirks, such as the fact that I only had problems in consistently bringing coaches to a smooth, feathered stop in electric mode when southbound servicing the University Street Tunnel zone.
Why? That zone was the only one with a slight downhill grade. To understand why that was a problem, one needs to understand the Breda's dynamic braking system, especially as it compares to the system equipped on the older 4000 series trolleys.
First, some electric basics. All electric drives consist of a power source (in our case the overhead wires), a motor and a controller. A controller acts to limit the power from the lines to the motor, similar to the way a carburetor or fuel injector limits the fuel to an internal combustion engine.
The 4000s use direct current (dc) traction motors, a technology that has been around for over a hundred years. To slow such a coach dynamically, the controller feeds a current to the motor in reverse, inducing drag on the motor. This system can deliver very smooth dynamic torque all the way to a stop.
By contrast, the Bredas use alternating current (ac) motors. The control systems for an ac drive are technically more complex and were thus not practical until powerful and inexpensive microprocessors became available. The dynamic braking system on Bredas, however, is a bit simpler. Instead of drawing current from the wire to induce the motor to slow, the Breda controller turns the spinning motor into a generator and directs (or "shunts") the power it generates from its forward momentum into a load bank. Look on the top of a Breda; just forward of the pole mount you will see three resistors, what look like coiled springs with air gaps between the coils. These resisters turn the electricity from the motor into heat exactly the same way an electric baseboard heater or electric range, for example, might turn your household electricity into heat.
The problem with this particular dynamic braking system lies in how it is executed in the Breda.
Rather than electronically regulate exact amounts of electricity shunted to the load bank resistors, the brakes of the Bredas actuate three simple switches. Lightly depress the service brake and one resistor connects to the motor's output; more pressure actuates two, and still more connects all three. This gives very rudimentary control over slowing.
As I mentioned above, the only problems I had with the system involved attempting to stop smoothly while servicing the only downhill grade in the tunnel. To explain why will take a bit of effort, so please bear with me.
Let's say you are traveling in the tube between Westlake and University, and need to slow your approach for the zone. Furthermore, let's say the motor spin at 30mph, the old speed limit in that tunnel section, has the potential to generate up to 80,000 watts if all of it were captured at once (all the numbers are speculative -- I made them up -- but the principle is the same), and that each of the load resistors can generate 1,000 watts of heat max. You gently depress the brake, actuating a resistor and diverting 1,000 watts of your spinning motor's energy to the load bank, and feel a gentle slowing, almost too little to notice. A bit more pressure on the pedal and 2,000 watts goes to heat the tunnel; more pressure increases that to 3,000 watts. Perhaps with help from the brake shoes and drums, the coach slows.
This is where things get tricky.
As the coach and the motor slow, the maximum potential of the motor is reduced; but the load resistors always try to draw the same amount. Therefore, as the coach gets to 20 mph, the potential motor power would only be, say 40,000 watts; at 10 mph, 10,000 watts. A 3,000 watt load on a 80,000 watt (30 mph) motor represents a bit over 4% drag; the same load at 40,000 watts (20 mph) equals about 7.5%; when we get down to 10 mph and 10,000 watts the drag is up to 30% of the motor's output! And as the percentage of motor output being drawn by the resistors increases, the greater slowing torque the dynamic braking exerts on the coach.
Of course, we operators have been trained to "feather the brakes," to reduce service brake pedal pressure as the coach slows to a stop, and then to firmly engage the pressure once that stop has been realized. So I would lighten my touch, disengaging third and then the second resistor, and reducing the drag of the load bank.
Even with only one resistor engaged, though, the coach will eventually slow enough that 100% of the motor's output would be diverted to the load bank, resulting in very a sudden stop (and perhaps potential damage to either the motor or controller).
That's why the controller disengages the dynamic braking at between 3-4 mph.
Disengages.
Meaning that if you are rolling downhill, the dynamic drag that kept the coach from accelerating suddenly disappears, resulting in a sudden acceleration . . . that necessitates an equally sudden reaction from the operator, a panicked jamming of the service brake, feathering be damned.
For any driver accustomed to feathering brakes in equipment other than the Bredas, this is an operational contradiction. Any attempt to feather Breda brakes will result in a choppy stop.
Oh, and one more thing. As the slope becomes steeper, this unpredictable downhill acceleration at 3-4 mph dramatically increases, leading to stopping that gets even choppier and even sloppier. I remember considering southbound University Street the trickiest stop to make in my route. I cannot even imagine trying to make a smooth stop coming down the Counterbalance.
Make no mistake: This is a design problem. It can, like any design problem, be corrected. For starters, that three-stage actuation system should be scrapped. Instead, a dynamic braking controller could be inserted between the motor and the load bank that would reduce the current to the resistors as the coach slows. This would have to be custom designed and manufactured; but it would solve the problem.
However management chooses to deal with this design flaw, I would stress that more than just smooth stops of trolley routes is at stake. This is a serious safety issue, to be sure.
It is also a serious training situation for all operators who must drive these coaches. To understand how serious we should better understand the story of Pavlov's dogs.
Most everyone has heard of the famous experiments Ivan Pavlov conducted at the turn of the century, experiments and conclusions that earned him the Nobel Prize in 1904. To refresh your memories, Pavlov and his assistants attached dogs to salivary measuring equipment which measured how much each dog drools at any given time. A few minutes before dinner someone would sound a buzzer in the dog's ear. They were this way able to "condition," or train the dogs to get ready for dinner by drooling.
That's the part of the experiment about which most have heard. What of the rest of the testing?
After the dogs were conditioned the experimenters then "deconditioned" them. The buzzers would be sounded, but not at dinnertime. The dogs caught on quickly and stopped drooling after the buzzers. After deconditioning, they would then "recondition" them again, feeding consistently right after buzzing. Reconditioning, like retraining, didn't take nearly as long as the initial conditioning.
What they did later I find personally disturbing and enlightening.
Some dogs, those that had never been given drool training, were subjected to random buzzing. The person buzzing the dogs and the person feeding the dogs never spoke. Buzzing was just some weird distraction to these dogs, one best ignored.
When well conditioned dogs were subjected to random buzzing, however, the result was quite different.
Those dogs went completely insane.
Specifically, dogs trained for months to associate the feeding buzzer with the food bowl reacted to the random buzzings and feedings by developing a schizophrenic state known in psychology as "waxy catatonia." The dogs refused to move. If you wanted to, you could move a leg or a paw, hold it for a bit, and the dog would keep it there until it got too tired. These dogs stopped eating altogether and had to eventually be euthanized.
It's like they just gave up.
The dynamic braking system of the coaches now known to drivers as the FrankenBredas has the potential to do to drivers what random conditioning did to Pavlov's trained dogs. Every day, Atlantic operators sign in not knowing whether the dynamic brakes of their coach will operate until the coach stops or not, whether the torque of braking will increase linearly, or rise sharply and then end altogether in yet another brief uncontrolled downhill descent.
Yes, we are humans, not captive dogs, and can theoretically handle more stress than our canine cousins; but these trolley routes are already the most used in Metro's system, accounting for (if I remember correctly) 35% of the ridership systemwide. Every additional stop, every additional rider -- not to mention the very real antagonism some riders on these routes seem to exhibit to All Things Metro -- all adds up to stresses which threaten to undermine we Humans Behind the Wheel without considering if today's brakes will function properly or not.
Management has an obligation to standardize its fleet if it ever expects to develop drivers with standard habits and skills that reflect well on all of us here at Metro.