NOAA Active Region 1520 certainly was a very active sunspot group. During its transit across the solar disk, it produced 5 M-class solar flares and even one X-flare (the strongest class). When it crossed the solar limb on 18 July, it certainly didn't diminish in activity. Indeed, over the next few days, several coronal mass ejections could be seen billowing away from behind the solar limb. A particular interesting event occurred in the early morning hours of 23 July, when this active region produced another eruption. At that time, NOAA 1520 was already several days on the backside of the Sun, and no instruments from Earth could monitor the blast in x-ray or radio-wavelengths. However, the STEREO-A satellite was well positioned to observe this event. Together with its twin brother STEREO-B, these spacecraft are orbiting the Sun in such a way that they gradually get a better view of the Sun's backside. The schematic underneath shows which part of the solar surface was visible from Earth (green), from STEREO-A (red), and from STEREO-B (blue). As can be seen, the flare was directed almost straight towards STEREO-A, and occurred very close to the east limb for STEREO-B.
The flare itself occurred around 02:30UT in the morning, following some magnetic activity that had already been going on for an hour. From the brightness and the extent of the eruption, one can safely conclude that this was a high energetic event, probably belonging to the strongest X-class of solar flares. The ejection of material left its mark in the Sun's corona, being visible as an irregular dark hole to the bottom left of the flare. The ejected cloud was visible in images from SDO, PROBA2 (movie), SOHO, and both STEREO-satellites. STEREO-A saw this ejected material as a full halo emanating from the Sun, indicating that this particle cloud was on a collision course with the tiny spacecraft. Even though the source of the eruption was behind the limb as seen from Earth, bright, hot post-flare loops of magnetic field, tell-tale markers of the magnetic activity that drives coronal eruptions, were soon visible rising high above the solar surface. The powerful flare also produced what space weather experts call a proton event, generating a stream of high energy protons reaching STEREO-A in just a few hours. Starting at about 10:30UT, these high energy particles interfered with the spacecraft's onboard electronics, significantly degrading the quality of images from STEREO-A. Some of these protons managed to make it all the way to the Earth, causing a very weak proton event. This is not an uncommon occurrence, as in the past - there have been other events so far on the Sun's backside to produce proton effects here on Earth.
The most interesting part of this eruption, again testifying of its rare strength, occurred during the evening hours of the very same day. Indeed, the coronal mass ejection (CME) arrived at STEREO-A already around 20:00UT, with solar wind speed jumping from about 550 km/s to at least 1400 km/s. When a CME travels the Sun-Earth distance (in this case Sun - STEREO-A) in less than 22 hours, it gets a special name: Fast Transit Event. Only a handful of such events are known, and they are associated with severe geomagnetic storming. One of the last such events were the CME's from the Halloween events (29-30 October 2003), when speeds (near Earth) were recorded in excess of at least 1850 km/s, having travel times of about 19 hours.
Of course, one could ask with which speed the CME took off at the Sun. As a CME is usually decelerated by the much slower solar wind around it, this CME must have had a much higher speed than the 1400 km/s it had at STEREO-A. Usually, scientists take some key-points along the CME's outskirts to determine its speed. This is always a tricky case, because when the CME is traveling along the line-of-sight, the eruption's apparent speed might be significantly less than the true speed (see schematic underneath). With NOAA 1520 already well beyond the solar limb, this line-of-sight effect had a significant influence on the measurements by SOHO. STEREO-B on the other hand was best placed for this kind of calculations.
Combining images from the spacecraft (SOHO, STEREO-A and -B) in an advanced triangulation matrix, scientists from NASA's Space Weather Research Center were able to calculate the initial speed of the CME to be about 3400 km/s. Such a speed places it in the hall of CME's with the highest initial CME speed ever, comparable e.g. to the 20 January 2005 event. A performance worthy of a true Olympian!