After five years of development and more than a year of testing in the professional peloton, SRAM is launching its electric drivetrain this spring. A bold step. Shimano's Di2 has dominated the scene for seven years and still shifts by cable. According to initial images, the new Dura-Ace, which will probably be launched in 2017, will once again use cables. And Campagnolo's EPS, the prototype of which we rode more than ten years ago, also communicates via copper wire. Competitor Mavic had already ventured into wireless shifting in 1998 - it was the French company's second attempt after their first electrically operated gear changer (ZAP, 1993) failed due to leaking plug connections. The wireless groupset was waterproof, it also shifted fine, but unfortunately not absolutely reliable, and so this pioneering feat also came to nothing - not on, but under the wheels.
New logic
So now SRAM is daring to tackle the hot iron. The Americans are not only cutting the cables, but also reinventing the operating concept: Like the mechanical double-tap shifters, the eTap grips only have one shift paddle each. But the operating logic is completely different. The left handle moves the chain to the left (to a larger sprocket), while the right handle moves it to the right (to a smaller sprocket). If you press both levers at the same time, the derailleur shifts to the other chainring.
It sounds complicated, but it is impressively simple and intuitive. There were hardly any gear changes during the test drives. A lighter gear on the left and a heavier one on the right - you quickly get the hang of it. As the paddles are large, they are easy to operate even when wearing thick gloves - a clear advantage over Shimano's Di2, where you can't always hit the exact button you want. Operating the front derailleur is also easy; the shifters tolerate being pressed out of sync and shift the chain quickly and reliably to the other chainring.
The shift buttons have a distinctive pressure point and provide clear feedback. The chain changes on the sprocket not only feels a little slower - compared to the mechanical Red or even a Di2: In the laboratory, we measured a reaction time of between 11 and 16 hundredths of a second for the rear derailleur. The Di2, on the other hand, responds immediately. The shifting behaviour of SRAM and Shimano on the rear wheel is otherwise very similar. Both work with high precision, SRAM a little louder.
If you hold down an SRAM button, the gears change in one go until you let go again. We also tested this behaviour in the laboratory. The rear derailleur needs seven hundredths of a second to move from sprocket to sprocket. However, if you shift in series by applying continuous pressure, the rear derailleur remains in each gear position for three tenths of a second - presumably to give the chain time to climb fully onto the neighbouring sprocket. It takes around three seconds to shift through once completely. Shimano manages this comparably, but a tad faster, namely in 2.8 seconds. SRAM is faster if you hammer on the buttons with your index finger in a staccato pattern. Unlike Di2, the eTap rear derailleur does not over-shift, but positions itself exactly under the sprockets.
The eTap front derailleur has the so-called yaw technology of its mechanical counterpart - this means that the cage not only pushes but also swivels. The front derailleur therefore does not have to readjust when the chainline changes. The eTap only shifts two tenths of a millimetre when shifting to the larger chainring and then resets itself automatically. All 22 gear ratios can be shifted without the chain grinding.
The shifting performance is very good with the original SRAM derailleur, it just doesn't quite reach the level of Shimano under load. The front derailleur on the Venge Vias, one of our test bikes, shifts less convincingly under load with the Specialized blades; the chain is only lifted onto the larger blade reliably and without rattling on every second attempt. SRAM expressly recommends using only the original chainrings. In our experience, Shimano's Di2 is less sensitive in this respect. It is not due to the motor power, as the front derailleurs from SRAM and Shimano develop the same shifting force. In the test setup, both push against the side of the chain with up to 110 newtons.
Power supply
The rear derailleur and front derailleur draw their power from interchangeable lithium-ion batteries with 300 milliampere-hours (mAh). The 24 gram energy storage units are locked in place with a lever and can be changed in no time at all. The handles, on the other hand, each contain a CR-2032 button cell, which is much more fiddly to change. Three very small and short screws have to be loosened here, which can easily be lost.
SRAM says the batteries will last for 1,000 kilometres, the button cells for two years. Naturally, we wanted to find out whether this was true. To do this, we mounted the derailleur with full batteries and at an ambient temperature of 18 degrees on a test stand on which a servo motor operated the levers. On the front derailleur, the counter stopped after 8,000 gear changes, while the rear derailleur managed 16,000 gear changes every three seconds. Under these conditions, the button cells in the grips had enough juice for 70,000 rear gear changes.
What this means in kilometres depends on the riding style; the playfully light function certainly encourages more frequent shifting. However, SRAM's figures seem rather conservative in view of our measured values. Suddenly finding yourself on the track without power is pretty unlikely.
Separate LEDs in all components, which light up at each switching step, provide information about the charge status. In our tests, the circuit began to light up red when there was 10 to 25 per cent capacity remaining. Shortly before the end, the indicators start to flash. So unless you're cycling on the last groove in the Race Across America, there should always be enough time to change the battery. If the power fails, however, nothing works; the rear derailleur cannot be positioned by hand.
Interference radio
The question remains as to how secure the wireless communication is. SRAM claims to transmit in the 2.4 GHz band, using a proprietary protocol and, of course, encrypted so that only associated components respond. The Americans say that they have even brought hackers into the team to carefully encrypt the circuit and make it unassailable. It is therefore unlikely that someone from outside will take control of the circuit - for example to disrupt a competitor.
But: The 2.4 GHz band is busy and not very regulated. WLAN, Bluetooth, ANT+ and much more can be found there. Against this background, how susceptible is the switching system to interference? What happens when others transmit on the same frequency? To investigate this, we went to the Laboratory for Electromagnetic Compatibility (EMC) at the University of Stuttgart and had the technology analysed there by EMC expert Malte Neumann (see "Radio interference" on page 71). In the judgement of the electrical engineer, SRAM did a good job within the bounds of what was technically possible. You have to hit the frequency of the circuit very precisely to jam it. If you then transmit at a high power level, the circuit no longer reacts, the frequency is "blocked" and switching steps are cancelled. However, as the data is only transmitted for a period of milliseconds, it is not very likely that other circuits will interfere with your own. Lost switching commands are also repeated as soon as - to put it bluntly - the line is free again. This can result in minimal delays, but does not mean that communication is cancelled.
We also tried disturbing the eTap's circuits with high power in the mobile phone frequency range - the most likely scenario according to expert Neumann. But the eTap was unimpressed by this, no matter how much we interfered. In racing mode, there is ultimately the residual risk of an interference transmitter that hits the frequency exactly - this could theoretically "freeze" the circuit. However, this would involve considerable effort and would be conspicuous because it would also interfere with a lot of other devices. As an amateur athlete, however, you don't have these worries and enjoy the wireless design and high switching comfort.
Installation
Assembly is easy without cables. Aligning the front derailleur takes the most time. All components must first be paired, which is done within a minute. Function buttons on the rear derailleur and front derailleur allow them to be operated without the handlebar grips, which is very practical for adjustment work (and allows you to change gear with empty grip batteries if necessary). Only for the fine adjustment of the rear derailleur do the function buttons on the grips have to be pressed, with which you can also adjust the gears on the move. If you press the shift paddle and function button at the same time, the rear derailleur changes its position in steps of 0.25 millimetres.
SRAM does not initially offer customisation of the derailleur via software, for example automatic shifting in series, as is possible with Shimano or Campagnolo. However, an ANT+ module is built in, which will allow information such as the gear engaged to be transmitted to Garmin computers in the future.
The mechanical customisation is probably more important: Up to two satellite switches, so-called blips, can be connected to each grip via cable. This allows you to shift gears from the upper handlebar or in the handlebar arch, as with Shimano's Sprint switches. The additional switches can be routed under the handlebar tape, in which case you have to press harder on them. The SRAM satellites do not quite match the ease of use of the Shimano sprint shifters.
It sparks!
SRAM's eTap is a powerful piece of technology with a convincing operating concept. If the shifting system is successful in racing, the other manufacturers will come under pressure to follow suit and replace cables with wireless systems. The official price for the complete eTap groupset is €2,692, retrofit kits for the mechanical Red are available from €1,545.
SHORT & SHORT
SRAM's eTap relies on wireless instead of cables. The batteries in the grips are sufficient for 7OO,OOO gear changes, the front derailleur can be shifted 8,OOO times on one battery charge, the rear derailleur 16,OOO times.
Switching endurance in comparison
With full batteries, the SRAM Red eTap drivetrain achieves just under half the shifting steps that Shimano's Dura-Ace Di2 with internal battery manages. In both groupsets, the front derailleurs draw twice as much energy as the rear derailleur. The button cells in the eTap grips last a good four times as long as the batteries. Ranges of 1,000-2,000 kilometres (SRAM) and 2,000-4,000 kilometres (Shimano) seem realistic - depending on how often you shift gears.
INTERFERENCE
How reliable is the radio circuit? The test in the EMC laboratory clarifies
To test the interference immunity of the radio technology, we moved the eTap circuit on a roller into an anechoic chamber, shielded from electromagnetic radiation from the environment. Firstly, the circuit is monitored during switching. The frequency scan shows that the switching commands are sent at 2.413 MHz. The transmission power of the group is comparable to a WLAN. Then the antenna in the anechoic chamber is used to transmit interference signals. For this purpose, a wide frequency range is traversed with increasing transmission power. Result: Between 2,409 and 2,420 MHz we cause the circuit to fail - from a field strength of 1.8 V/m. A powerful source of interference that transmits continuously can block the circuit. EMC expert Malte Neumann confirms that the eTap is well designed because it can only be interfered with in a narrow frequency band: "You can't do any better with this technology." Typical sources of interference such as overhead lines can therefore be ruled out. Nevertheless, it is not completely impossible for interference to occur. There is a lot going on in the 2.4 GHz band and there is little regulation. There are numerous potential sources of interference, from model aeroplanes to company radio.
The frequency scan shows the frequency and intensity of the signals emitted by the circuit in the anechoic chamber. The largest deflection is at 2,413 MHz - this is the data packet that transports the shift command from the handle to the gearshift mechanism. The frequency corresponds to WLAN channel 1 and the transmission power also corresponds to a typical WLAN. Outdoors, the range of the grips is considerable.
