

deutsche Übersetzung
See also this comparison
page.
Thomas Avery has also performed measures on
5292 motors, see Lugnet
thread.
The roster

Electric Technic Motor 9V
Lugnet
Partsref 2838c01, Peeron
2838c01, Lego 74569
The older 9V Technic motor (1990).
Ungeared, it has a high rotation speed and low torque,
so for most applications it requires an external
gear reduction. 

Electric Technic Micromotor
Lugnet
Partsref 2986, Peeron
2986, Lego 70823
Appeared in 1993, this small, light
weight motor turns slowly and offer low torque 
but respectable torque for its size. Must be used
generally with pulley,
top
and base,
but other motion transmissions are possible (see
examples by Brian Sadowsky, with a 16t
gear/clutch or a pin
joiner) 

Electric Technic MiniMotor 9v
Lugnet
Partsref 71427c01, Peeron
71427c01, Lego 71427
Since 1997, this motor replaces
2838. Geared down and quite efficient, this is the
motor of choice for most applications. 

Electric Technic MiniMotor 9v
Lego 43362
In 2002, Lego replaced 71427 motor
with a new type, 43362. Externally almost identical,
its internal structure is
very different. Performances are almost as good,
and its weight is much lower. 

Electric RC Race Buggy Motor
Peeron
5292
(data
updated on Augus 15, 2012: some values were all
wrong!) Introduced in 2002, this
motor appeared with 8475
RC Race Buggy. Very powerful, it also consumes
a lot of energy. Not recommended for use with a
RCX which can't deliver the current needed by this
beast. The innermost shaft hole is geared up by
a 23/17 factor. Only the outermost output is tested
below. 

Electric Technic Motor 9V Geared
Peeron
47154
This motor was first included in
4094
Motor Movers set (2003). Provides an axle hole
with friction, allowing to choose axle length without
the need of an extender. Flat bottom allowing easy
mounting. 

NXT motor
This motor is specific to the NXT
set (2006). Includes a rotation encoder, returning
to the NXT the position of the shaft with 1°
resolution. Because of the special connector of
this motor (nonstandard phone plug type), a cable
adapter is required to drive this motor with regular
9V sources. Not recommended for use with a RCX which
can't deliver the high current that this motor can
consume. Slow rotation speed, minimizing the need
of external gear train. 

Power Functions Medium motor
This motor belongs to a new range
of motors and control elements introduced in 2007,
the Power Functions. It uses a new 4 pins 9V connector
that provides permanent 9V supply to control elements
as well as controlled power to the motors (compatibility
with old 9V system is built in extension cords).
The Medium motor has holes for studless constructions
as well as 6x2 bottom plate. 

Power Functions XL motor
Big brother of Power Functions Medium
motor, it provides a lot of mechanical power. Actuated
by the same core as NXT motors, it rotates slightly
faster (less internal gear reduction). Mounting
is done through several pin holes. 

9V Train motor
Stefan Vorst measured performances
of the 9V Train motor. The results are below. 

RC Train motor
This motor was introduced when LEGO
stopped the 9V train with metal tracks. The performances
of this motor are... not so good. 

Power Functions Train motor
As the train system unifies with
Power Functions, this motor , equipped with a PF
cable and connector, replaces the RC train motor.
Fortunately the performances are much improved,
with an efficiency and power even exceeding the
old 9V train motors. 

Power Functions Emotor
Introduced with LEGO Education 

Power Functions Large motor
This motor appeared first in Rock
Crawler set (9398). Significantly more powerful
than the medium motor, it packs a lot of power in
a light, small package with many convenient mounting
options. But it's efficiency at light load is pretty
low (large current at noload). 

EV3 Large motor
This motor is very similar to NXT
motor, but with slightly better fixing capability
(compatible with Technic frames, holes aligned with
hub). Its efficiency seems slightly lower. 

EV3 Medium motor
This motor is one of the highlight
of EV3 set compared to NXT one: a motor of reduced
size and classical front axle hub. Reduced size
comes with reduced power, similar to PF medium motor
(but being more downgeared, it is slower and delivers
more torque). 
Weight
2838

2986

71427

43362

5292

47154

NXT

EMotor

48g 
10g 
42g 
28g 
55g 
40g 
80g 
50g 
PF Medium

PF XL

9V Train

RC Train

PF Train

PF Large

EV3 Large

EV3 medium

31g 
69g 
72g 
53g 
57g 
42g 
82g 
39g 
Supposed to be equivalent to 71427, 43362 motor is 30% lighter.
This is generally an advantage, except when the motor is used
as a counterweight, or to balance the structure, for example
in COGshifting walkers.
Noload characteristics
Test conditions: motor is powered by a variable, regulated
power supply. An ammeter measures current flowing through the
motor, a voltmeter monitors tension across. The rotation speed
is measured by a RCX equipped with a light sensor, looking at
an halfwhite/half
black cylinder.
9 V supply 
2838

2986

71427

43362

5292

47154

NXT

EMotor

Rotation
speed (rotations per minute) 
4100 rpm 
35 rpm 
360 rpm 
340 rpm 
1700 rpm
/ 1240 rpm 
460 rpm 
170 rpm 
780rpm 
Noload current 
35 mA 
6 mA 
3.5 mA 
9 mA 
160 mA 
31 mA 
60 mA 
17.5mA 
9 V supply 
PF Medium

PF XL

9V Train

RC Train

PF Train

PF Large

EV3 Large

EV3 Medium

Rotation
speed (rotations per minute) 
405 rpm 
220 rpm 
2000rpm 
2000rpm 
1900rpm 
390 rpm 
175rpm 
260 rpm 
Noload current 
65 mA 
80 mA 
90mA 
90mA 
90mA 
120 mA 
60mA 
80 mA 
43362 has a higher noload current than 71427, probably
caused by higher internal
friction. 47154 has a fairly high noload current, because
of its 5stages gear reduction. But it uses bigtooth gears
in the last stages, probably much more sturdy that the 2stages,
thintooth 71427/43362 internal gearing. The 5292 also exhibit
very high noload current, here again caused by internal friction.
This explains also the break in its speed/voltage curve. All
train motors show similar noload characteristics, especially
a relatively high current.
As is usual for DC motors, rotation speed is proportional
to voltage applied to them, this can be seen on graphs below.
Noload current depends little on voltage.
Stalled characteristics
Stalled current consumption is simply measured with motor
axle shaft locked by hand. Stalled torque is established from
the maximum weight that can be lifted by the machine described
below. Note that stalled torque measure is VERY imprecise.
9 V supply 
2838

2986

71427

43362

5292

47154

NXT

EMotor

Stalled
torque 
0.85 N.cm 
1.6 N.cm 
6 N.cm 
5.5 N.cm 
14 N.cm 
6 N.cm 
50 N.cm 
3.4N.cm 
Stalled current 
700 mA 
80 mA 
360 mA 
340 mA 
3.2 A 
580 mA 
2 A 
410 mA 
9 V supply 
PF Medium

PF XL

9V Train

RC Train

PF Train

PF Large

EV3 Large

EV3 Medium

Stalled
torque 
11 N.cm 
40 N.cm 
2.8 N.cm 
1.7 N.cm 
3.6 N.cm * 
18 N.cm 
43 N.cm 
15 N.cm 
Stalled current 
850 mA 
1.8 A 
950 mA 
750 mA 
1.3 A * 
1.3 A 
1.8 A 
780 mA 
Take care to avoid extended period stall condition, as power
dissipated in motor case is quite high (6 Watts for 2838, 3
W for 71427) will cause a rapid temperature rise. Note that
71427 and 43362 motors, equipped
with a thermistor, should be protected against frying (not
tested though !!!). 5292 motor is protected too, since stalled
current decreases quickly (It's the rectangular yellow component
on this photo.
47154 protection can be seen easily through
clear case.
The NXT motor is also protected by a thermistor
(Raychem RXE065 or Bourns MFR065).
That means that the high 2A current (and associated whooping
torque) can be sustained only for a few seconds. Same thing
for the Power Functions XL motor.
The train motors also contain thermistor limitations. For
the PF train motor, this protection trips too fast and prevents
direct measure of the stalled current. These values were obtained
by extrapolation.
Loaded characteristics
Here is the setup used to measure motors under load. Electrical
power is measured with voltmeter and ammeter. Mechanical power
delivered by the motor is evaluated from the time used to lift
the weight by some height (5 cylinder turns  the first two
turns are not counted to eliminate initial acceleration). Torque
applied is obtained from weight and cylinder radius.
Cylinder is directly placed on motor axle shaft, except for
2838 motor where a 1/5 gear reduction was used. Additionnal
friction introduced may have somewhat impacted 2838 efficiency,
but anyway this gearing is necessary for most applications.
Torque displayed for this motor corrects gear reduction. The
fast 5292 motor, the PF and RC train motors were also measured
with a 1/3 gear reduction.
In 2010 I updated my test
setup to the NXT platform: Mindsensors
launched the PowerMeter
sensor that allows the NXT to measure directly the voltage
applied to the motor and the current consumed. A light sensor
in front of a black and white cylinder reads the number of turns
done by the winch, and the time needed to lift the weight. Using
custom board with two electromechanical relays, the NXT can
control the motor under test: run, float or brake (this later
state is used to prevent the load to drop brutally on the floor
at the end of lifting). A laboratory power supply is used to
power the motor under test.
The photo below shows the NXT equipped with
PowerMeter
sensor and motor control board.
Here is a screen capture of the NXC
motor test program:
2838

Torque 
Rotation
speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
6 V 
0.45 N.cm 
580
rpm 
0.32 A 
0.27 W 
1.9 W 
14 % 
7 V 
0.45 N.cm 
1000
rpm 
0.32 A 
0.46 W 
2.3 W 
20 % 
9 V 
0.45 N.cm 
2000
rpm 
0.32 A 
0.9 W 
3 W 
31 % 
12 V 
0.45 N.cm 
3300
rpm 
0.33 A 
1.5 W 
4 W 
39 % 
71427

Torque 
Rotation
speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
4.5 V 
2.25 N.cm 
57
rpm 
0.12 A 
0.13 W 
0.54 W 
24 % 
7 V 
2.25 N.cm 
160
rpm 
0.12 A 
0.38 W 
0.85 W 
45 % 
9 V 
2.25 N.cm 
250
rpm 
0.12 A 
0.58 W 
1.1 W 
54 % 
12 V 
2.25 N.cm 
375
rpm 
0.12 A 
0.88W 
1.5 W 
61 % 
43362

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
4.5 V 
2.25 N.cm 
50 rpm 
0.12 A 
0.12 W 
0.54 W 
22 % 
7 V 
2.25 N.cm 
140 rpm 
0.12 A 
0.33 W 
0.85 W 
39 % 
9 V 
2.25 N.cm 
219 rpm 
0.12 A 
0.51 W 
1.1 W 
47 % 
12 V 
2.25 N.cm 
333 rpm

0.12 A 
0.77W 
1.5 W 
54 % 
47154

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
4.5 V 
2.25 N.cm 
90 rpm 
0.19 A 
0.21 W 
0.85 W 
24 % 
7 V 
2.25 N.cm 
210 rpm 
0.19 A 
0.49 W 
1.33 W 
37 % 
9 V 
2.25 N.cm 
315 rpm 
0.19 A 
0.74 W 
1.7 W 
43 % 
12 V 
2.25 N.cm 
468 rpm

0.19 A 
1.1 W 
2.3 W 
48 % 
2986

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
9 V 
1.28 N.cm 
16 rpm 
0.04 A 
0.021 W 
0.36 W 
16 % 
12 V 
1.28 N.cm 
28
rpm 
0.04 A 
0.038W 
0.48 W 
28 % 
5292

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
4.5 V 
5.7 N.cm 
150 rpm 
1.36 A 
0.87 W 
6.12 W 
14 % 
6 V 
5.7 N.cm 
380 rpm 
1.38 A 
2.27 W 
8.28 W 
27 % 
7.5 V 
5.7 N.cm 
580 rpm 
1.37 A 
3.45 W 
10.3 W 
34 % 
9 V 
5.7 N.cm 
780 rpm

1.40 A 
4.61 W 
12.6 W 
37 % 
10.5 V 
5.7 N.cm 
1030 rpm 
1.46A 
6.16 W 
15.3 W 
40 % 
NXT

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
4.5 V 
16.7 N.cm 
33 rpm 
0.6 A 
0.58 W 
2.7 W 
21.4 % 
7 V 
16.7 N.cm 
82 rpm 
0.55 A 
1.44 W 
3.85 W 
37.3 % 
9 V 
16.7 N.cm 
117 rpm

0.55 A 
2.03 W 
4.95 W 
41 % 
12 V 
16.7 N.cm 
177 rpm

0.58 A 
3.10 W 
6.96 W 
44.5 % 
PF Medium

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
4.5 V 
3.63 N.cm 
73 rpm 
0.28 A 
0.27 W 
1.26 W 
22 % 
7 V 
3.63 N.cm 
185 rpm 
0.29 A 
0.70 W 
2.03 W 
34 % 
9 V 
3.63 N.cm 
275 rpm

0.31 A 
1.04 W 
2.79 W 
37 % 
12 V 
3.63 N.cm 
405 rpm

0.32 A 
1.53 W 
3.84 W 
40 % 
PF XL

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
4.5 V 
14.5 N.cm 
43 rpm 
0.52 A 
0.65 W 
2.34 W 
28 % 
7 V 
14.5 N.cm 
100 rpm 
0.54 A 
1.51 W 
3.78 W 
40 % 
9 V 
14.5 N.cm 
146 rpm

0.55 A 
2.21 W 
4.95 W 
45 % 
12 V 
14.5 N.cm 
214 rpm

0.56 A 
3.24 W 
6.72 W 
48 % 
9V Train

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
4.5 V 
0.90 N.cm 
375 rpm 
0.40 A 
0.36 W 
1.80 W 
20 % 
6 V 
0.90 N.cm 
667 rpm 
0.39 A 
0.62 W 
2.34 W 
27 % 
7.5 V 
0.90 N.cm 
1071 rpm

0.38 A 
0.99 W 
2.85 W 
35 % 
9 V 
0.90 N.cm 
1250 rpm

0.38 A 
1.11 W 
3.42 W 
33 % 
RC Train

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
3 V 
0.85 N.cm 
 
 
 
 
 
4.5 V 
0.85 N.cm 
 
 
 
 
 
6 V 
0.85 N.cm 
171 rpm 
0.43 A 
0.15 W 
2.59 W 
6 % 
7.5 V 
0.85 N.cm 
549 rpm

0.43 A 
0.49 W 
3.23 W 
15 % 
9 V 
0.85 N.cm 
990 rpm

0.43 A 
0.88 W 
3.91 W 
22 % 
10.5V 
0.85 N.cm 
1323 rpm

0.44 A 
1.18 W 
4.63 W 
25 % 
12 V 
0.85 N.cm 
1683 rpm

0.45 A 
1.50 W 
5.43 W 
27 % 
PF Train

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
3 V 
0.85 N.cm 
135 rpm 
0.35 A 
0.12 W 
1.05 W 
11% 
4.5V 
0.85 N.cm 
468 rpm 
0.36 A 
0.42 W 
1.62 W 
26 % 
6 V 
0.85 N.cm 
792 rpm 
0.37 A 
0.71 W 
2.22 W 
32 % 
7.5 V 
0.85 N.cm 
1107 rpm

0.38 A 
0.99 W 
2.85 W 
35 % 
9 V 
0.85 N.cm 
1458 rpm

0.38 A 
1.30 W 
3.42 W 
38 % 
10.5V 
0.85 N.cm 
1782 rpm

0.39 A 
1.59 W 
4.10 W 
39 % 
12 V 
0.85 N.cm 
2124 rpm

0.40 A 
1.90 W 
4.80 W 
40 % 
EMotor

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
4.5V 
1.32 N.cm 
63 rpm 
0.17 A 
0.087 W 
0.76 W 
11 % 
6 V 
1.32 N.cm 
186 rpm 
0.17 A 
0.26 W 
1.02 W 
25 % 
7.5 V 
1.32 N.cm 
300 rpm

0.17 A 
0.42 W 
1.27 W 
33 % 
9 V 
1.32 N.cm 
420 rpm

0.18 A 
0.58 W 
1.62 W 
36 % 
10.5V 
1.32 N.cm 
520 rpm

0.18 A 
0.72 W 
1.89 W 
38 % 
12 V 
1.32 N.cm 
640 rpm

0.18 A 
0.89 W 
2.16 W 
41 % 
PFlarge

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
4.5V 
6.48 N.cm 
67 rpm 
0.46 A 
0.46 W 
2.07 W 
22 % 
6 V 
6.48 N.cm 
138 rpm 
0.47 A 
0.94 W 
2.82 W 
33 % 
7.5 V 
6.48 N.cm 
203 rpm

0.48 A 
1.38 W 
3.60 W 
38 % 
9 V 
6.48 N.cm 
272 rpm

0.49 A 
1.85 W 
4.41 W 
42 % 
10.5V 
6.48 N.cm 
338 rpm

0.49 A 
2.30 W 
5.15 W 
44 % 
12 V 
6.48 N.cm 
405 rpm

0.50 A 
2.75 W 
6.00 W 
46 % 
EV3 large

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
4.5 V 
17.3 N.cm 
24 rpm 
0.69 A 
0.43 W 
3.10 W 
14 % 
6 V 
17.3 N.cm 
51 rpm 
0.69 A 
0.92 W 
4.14 W 
22 % 
7.5 V 
17.3 N.cm 
78rpm 
0.69 A 
1.41 W 
5.17 W 
27 % 
9 V 
17.3 N.cm 
105 rpm

0.69 A 
1.90 W 
6.21 W 
31 % 
10.5 V 
17.3 N.cm 
132 rpm

0.69 A 
2.39 W 
7.24 W 
33 % 
12 V 
17.3 N.cm 
153 rpm

0.69 A 
2.77 W 
8.28 W 
33 % 
EV3medium

Torque 
Rotation speed 
Current 
Mechanical power 
Electrical power 
Efficiency 
4.5V 
6.64 N.cm 
24 rpm 
0.35 A 
0.17 W 
1.57 W 
10 % 
6 V 
6.64 N.cm 
72 rpm 
0.35 A 
0.50 W 
2.10 W 
24 % 
7.5 V 
6.64 N.cm 
120 rpm

0.35 A 
0.83 W 
2.62 W 
32 % 
9 V 
6.64 N.cm 
165 rpm

0.37 A 
1.15 W 
3.33 W 
34 % 
10.5V 
6.64 N.cm 
207 rpm

0.37 A 
1.44 W 
3.88 W 
37 % 
12 V 
6.64 N.cm 
249 rpm

0.37 A 
1.73 W 
4.44 W 
39 % 
The speed of 43362 motor is about 12 % lower than speed
of 71427. Though this is in the range of variations measured
by Steve Baker among a bunch of nine
71427 motors, my measures on three 71427 and two 43362 showed
the 12 % difference between the two groups.
The RC train motor had a poor efficiency and delivers little
torque at low voltage (it was not able to move under 6V loaded
with 0.85 N.cm). The PF train motor has a much improved efficiency,
even better than the old 9V train motor.
Speed and current vs.
torque
Synthesis
charts
(charts updated on July
4, 2012: 5292 motor values were wrong) These
charts summarize the above curves. The most meaningful shows
the various motor sorted by maximum power they are able to deliver
at 9V. Because rpm/torque curve is linear a motor provides maximum
power when load slows it down to half of noload speed.
The following chart sorts motors by torque and
by noload rotation speed (of course this depends a lot on internal
downgearing of the motors!!!). Torque chart lists torque at
half speed point.
Protections
71427 and 43362 motors are protected from abuses by two devices:
 a PTC thermistance (here an Epcos
B1056). This resistor, mounted in series with the motor,
has a low value when it is cold (about 1.7 ohms), rapidly
increasing as temperature rises. When large current flows
through the motor, self heating rises thermistance temperature
and resistance value, so the current is limited by voltage
drop across thermistance.
 a BZW0415B, bidirectional transient
voltage suppressor diode. This diode protects RCX from large
voltage spikes that could be generated by the motor. But
it also forbids applying more than 15V to the motor...
A similar protection is integrated in 47154 motors, as can
be seen on this photograph. NXT
motor is also protected with a PTC thermistance
and a transient voltage suppressor diode (D4 on this photograph).
Outputs of RCX are also protected from overload: the motor
driver chip used (Melexis MLX10402  datasheet)
has a current limitation set to 500 mA, and a thermal shutdown
which disable the output if die temperature rises too much.
Here is the curve limitation that I measured on a RCX. It
was powered by an external regulated power supply, and tested
at 9V (6 alkaline batteries) and 7.2V (6 NiCd or NiMH rechargeable
batteries).
There is a significant voltage drop before reaching
current limitation knee (at about 500 mA). So a stalled 71427
motor receives only about 7V at 300 mA, while two paralleled
71427 or a single 2838 almost reach current limitation (5.5V
/ 430mA).
Once current limitation is established (for
example with a dead short), power dissipation in the driver
is very high, and it quickly enters thermal shutdown mode. After
that, the output cycles on/off with a period of about
1 second: the driver circuit heats up, stops output, cools down,
reenables output, heats up again and so on.
You can also see on the graph that with a dead
short, the output can deliver slightly more than 500mA. So if
all three outputs are shorted, total consumed current is more
than 1.5A, exceeding rating of the fuse
that protects RCX. This condition should not happen in normal
circumstances, even with all three outputs loaded with 2 stalled
71427 motors...
I also had a look to current output capabilities
of 8475 RC
Race Buggy. Its main output drives two paralleled 5292 motors
that consumes more than 3A when stalled so it has to be beefy
! And indeed it seems to have a current limitation of about
4A, and a thermal shutdown providing on/off cycling like RCX
motor driver.
Getting maximum
mechanical power from RCX output
October 2012 update:
Similar curves for Power Functions motors driven by PF IRreceiver
are available here.
Using an illimited power supply (fresh batteries for example),
a DC motor provides maximum mechanical power when loaded at
half of its stall torque. This is also the load where rotating
speed is half of noload speed (this assumes ideal conditions
such as low internal friction, but according to load curves
showed above, this is exact enough to be useful).
But with RCX output, some voltage drop occur as current increases,
and current limitation can also trigger in if two motors under
heavy load are paralleled on the same output.
Here are the curves showing mechanical power versus load
torque for various motor combinations. The RCX was externally
powered from a regulated power supply, and I measured mechanical
power at 9V (equivalent to 6 alkaline batteries) and 7.2V (6
NiCd or NiMH rechargeable batteries).

RCX powered by an external regulated power supply
through two fake batteries.
Caution: I shall not be held responsible
if you burn your RCX with incorrect voltage or bad
polarity ! 

The fake batteries where assembled from the sawed
ends of old alkaline batteries, maintained at the
right spacing with rods of hot melting glue.
Caution: batteries contain hazardous
chemicals that can be dangerous for your health.
Open them at your own risk and only if you know
what you are doing!!! 
You can see from the curves that although RCX can be operated
from NiMH batteries, the lower supply voltage translates in
a 40% cut down of available mechanical power .
Single motor curves
Paralleled motors curves
Two identical motor are powered from the same RCX output,
and their shaft are mechanically coupled.
Because of the higher current consumption of
47154 and 2838 motors, using two of them on the same RCX output
is not recommended, as they exceed RCX current limitation when
heavily loaded. At 0.8 W, tandem 71427 provide safely the greatest
mechanical power of all.
Conclusion
Each of these motors has unique characteristics which makes
it more or less suitable for different applications.
 Micromotor 2986 is at its best when space or weight
is at a premium. But its mechanical power is quite low.
 Technic motor 2838 is a real power hog, with poor efficiency,
but it can deliver 30% more power than Minimotor.
 Minimotor 71427 is probably the best performer of the
pack overall.
 The new 43362 is roughly equivalent to 71427, with slightly
degraded performances. But its light weight can be a boon
for many uses.
 Clear case 47154 provides a higher mechanical power
than 71427, but at the price of a somewhat lower efficiency.
 RC Race Buggy Motor 5292 is really powerful, but requires
a power supply up to the task. It's not a good idea to use
it with a RCX as the 500 mA current limitation won't let
it unleash its power...
 NXT motor delivers a high torque thanks to its internal
speed reduction gear train. Because of that, it also turns
slowly and efficiency is somewhat reduced. This motor could
be connected to RCX thanks to a compatibility cable, but
this is not recommended for use on a RCX because the high
current it can consume is too much for RCX 500 mA current
limitation.
 The Power Functions train motor has widely improved
characteristics compared to the older RC train motor.
Caution
! Though I tested motors with a 12V supply,
I can't guarantee that they bear the extra load
for extended time period. Use that
at your own risks ! 
deutsche Übersetzung 