I've been reading this discussion around mixed consists here and other forums for some time now and it drives me batty. I have asked commenters what the problem is, but I have yet to receive any legitimate argument against it. Most responses revolve around an uninformed one-off statement issued by a WMATA PR person.
I'd like to share my thoughts on it, drawing on my expertise as an (electrical) engineer and long career in transit rail car maintenance, including my current stint at WMATA.
First, let's talk about how the trains actually move and stop. WMATA's cars, like most modern systems, use what's known as a "master controller" as the exclusive input to select the train's tractive effort. Rail fans may already be familiar with this component, which is sometimes referred to as a T-handle, or (erroneously) a deadman. This input acts upon the rail car's propulsion and friction braking systems via binary electrical train lines to make the train start, maintain speed, and stop. Train lines are simple electrical circuits that are either energized or not based on some input and, as the name suggests, run the entire length of a train. The signals travel from car to car via automatic electric couplers at each end of the car. The propulsion system handles all acceleration and electro-dynamic braking. The friction braking system handles low-speed deceleration (at speeds under approx 7 MPH) and holding the train at a stop.
With the exception of the 7000 series, all of WMATA's cars utilize master controllers with five notches for braking rates, a coast notch, five notches for power rates, and an emergency friction brake-only notch that dumps the brake pipe. Each of these notches represents a predetermined rate that is matched across all cars and fleets. These rates hard coded and again are identical regardless of car series.
There are approximately 80 train lines on Metro's cars carrying various signals used to control train functions such as doors, lights, intercoms, PA, HVAC, and of course rate selection. About a dozen are dedicated to rate selection, and are energized by the master controller in various unique patterns depending upon the selected master controller notch. The friction brake and propulsion systems are connected to the train lines and constantly monitor the energization patterns.
When an operator selects the P1 notch, the master controller outputs a pattern that signals the friction brake system to fully release the brakes and the propulsion system to accelerate at fixed rate of 0.75 miles per hour per second. The propulsion system on each individual car (note this is by car and not married pair, and again regardless of car series) operates entirely independently and without regard for what any other car is doing. It maintains this rate by first sampling the air spring pressure to determine the load weight (pressure rises as more passengers board the the car) and then constantly sampling the axle tachometer speed sensors and motor current. By monitoring these values, the propulsion system ensures each car accelerates at a nearly identical rate as selected by the master controller. The friction braking system operates in a nearly identical fashion, also monitoring load weight pressures and its own redundant speed sensors to ensure rate consistency. THIS IS WHY IT DOES NOT MATTER WHEN CAR SEIRES ARE MIXED WITHIN A CONSIST.
There is a comment in this thread discussing the mixture of 2000 and 3000 series cars which is even more confounding. When those two fleets underwent midlife rehabilitation, they were combined in to a single fleet. All components are 100% interchangeable. All internal documentation, procedures, and policies refer to them as a single fleet. Also relevant is the fact that the rehabbed 2/3K and 6K fleets were designed to be nearly identical, sharing many of the same fully interchangeable components including door controls, ATC, VMS, and propulsion - the same Alstom propulsion system used on NYCT's R160 fleet by the way.
Now, that being said, there is one minor inconvenience associated with mixed consists. The interior LED "next station" signs will not operate when 1000 or 4000 series cars are included in a consist. This is because they don't carry the required equipment that carries the serial data link which transmits the station ID throughout the trains. There is also an issue where some of this equipment doesn't communicate correctly in the newer cars. This is because the hardware architecture changed somewhere along the line during production. Some cars delivered with legacy equipment and some with newer hardware (including all of the 6K cars). The good news is that engineering and the hardware vendor are approximately halfway through an on-going project that will update all 2/3/5/6K cars with unified equipment that will ensure reliable communication.
In the past, there was an issue with the passenger emergency intercoms on older cars failing to trigger the operator's communication panel on 6K cars. This was a very simple design issue that (in typical Metro fashion) wasn't correctly tested for or caught during the 6K prototyping phase. This problem was finally corrected after an incident about 4 years ago.
I look forward to any comments or feedback as I am far from perfect (particularly at my age) and may have overlooked something.