Circuit Board Enhances Data Collection, Saves on Cabling

JLab employees pose with the flash ADC board they designed. From left are Fast Electronics Staff Engineers Fernando Barbosa, Benjamin Raydo and Hai Dong; Fast Electronics Group Leader Chris Cuevas, and Data Acquisition Staff Scientist Ed Jastrzembski and Fast Electronics Associate Coordinator Jeff Wilson.

Experiments in Jefferson Lab's experimental halls typically begin with a smash, as millions of electrons from the accelerator crash into an experimental target. Much of the resulting debris is collected and measured by large detectors.

Ideally, scientists would keep all of the information recorded by the detectors, but there's simply too much of it. So they sample portions of the data and use a computer, called the trigger, to process the samples to determine what data to keep and what to throw away. While the trigger is making that determination, the full data set is preserved by sending it through a loop of cable. This process takes just eight millionths of a second – but that's time enough for the data to travel through nearly 1.25 miles of cable!

Soon, though, all that cable may be replaced by a circuit board designed by Jefferson Lab's Fast Electronics Group in collaboration with the Data Acquisition Group.

"We spent a lot of time doing the simulations and talking to various groups to get the data that is important to them. This is a general-purpose board that anybody can use. And we are excited, because this is the first prototype," says Fernando Barbosa, a staff engineer in the Physics Division’s Fast Electronics group.

The circuit board, called a flash analog-to-digital converter, converts the data into a digital format and stores it while an onboard processor determines which parts to keep. The flash ADC can then send the data to other processors and/or to tape.

Physicists will use the flash ADC to convert and pare experimental data down their essentials.

"You practically don't lose anything for eight microseconds. But the trigger should take two microseconds max to make a decision. So you have plenty of time to make a decision and to start writing to tape what you really need," Barbosa explains. "So, you'll be writing less to tape, but you’ll get the same or more data out."

The group has been working on the project, when time permitted, for several years. According to Hai Dong, staff engineer, all the pieces the group needed to complete the board only came together recently, due to advances in processor technology and in time spent writing and verifying the computer code to make it all work.

"We couldn't have built this board two years ago. The processors were not available at the speed we needed. These are the analog-to-digital converters; these were not available," Dong says. "It's also about a year's worth of time, just for the code."

The flash ADC was designed to meet growing data rate needs in the Jefferson Lab experimental program.  It will have particular impact on the Hall D data acquisition systems following the 12 GeV Upgrade.  JLab will need several hundred of these custom-built boards once the design has been finalized.

In the meantime, group members are anxious to see how well the flash ADC performs in experiments. Once testing is complete, it’s likely that prototypes will be installed in one or more of the experimental halls. The group expects that to happen by the end of the year.

Chris Cuevas, Fast Electronics group leader, says researching and developing the flash ADC has been a great project.

"This is a big deal for JLab. There is nothing on the market that could have accomplished exactly what we need for digitizing the detector signals and performing first-level trigger processing. We don't normally design these things in-house, but we’ve shown we can do that."

By Kandice Carter
JLab science writer