Unwinding the APA winding mystery

apa-lee
An APA frame at PSL with Lee Greenler, mechanical engineer and Darrell Hamilton, mechanical technician.

“Almost everyone who visits to see the APAs and wire-winding machine – physicists, engineers, everybody — asks about how the wires are wound,” said Bob Paulos, director of the Physical Sciences Laboratory (PSL) in Stoughton, WI, where the first three anode plane assemblies (APAs) for ProtoDUNE-SP are being fabricated. “We have to explain it better.”

Here I share my improved understanding of APA wiring after my visit to PSL on an unusually warm and beautiful November day.

The APAs, based on a design concept by Bo Yu of BNL, are 2.3 m wide, 6.3 m long, and, when fully outfitted with wire planes, 12 cm thick. Paulos’ team constructs each APA on a framework of rectangular cross section structural stainless steel tubing. The team stretches a fine wire mesh across each side of the completed frame, making good electrical connections between the mesh and the frame to ensure a uniform ground across the entire frame area.

On both sides of a completed APA, four layers of sense and shielding wires cover the mesh layer. The collection (X) wires form the innermost layer, which is followed by two induction wire layers (first V then U). A layer of shielding or grid (G) wires forms the outermost layer, i.e., the layer that an ionization electron would first encounter. All wire layers span the entire area of the APA frame and are connected and anchored on boards mounted along the short edges of the frame.

Wire-winding head with CuBe wire
Wire-winding head with CuBe wire

The wiring process uses an impressive-looking automated winding machine that first rotates an APA frame to the vertical landscape position, then circulates a head paying out copper beryllium (CuBe) wire at a constant 5 N of tension.

“Some fairly complex software controls the machine,” Paulos explained. “We’re lucky to have some mechanical engineers who are also knowledgeable about software.”

Winding was not in progress during my visit, but Paulos described it while pointing out the various features of the machine and the APA frame.

Detail of APA frame showing X layer.
Detail of APA test frame showing X layer (no mesh).

The machine winds the X layer first, a few millimeters above the mesh layer, with the windings oriented parallel to the long dimension of the frame. The X wire gets strung continuously up one side of the frame and down the other (the blue wires in the APA sketch, below), the head moving a bit to the side on each rotation to set the wire pitch. (In practice, certain conditions require stopping and restarting the winding of a layer, but “continuous” describes the general process.)

The X windings maintain their positions along the 6-meter length with help from spacers (called combs) located along the cross bars in the frame. Once the windings are in place, each one is soldered and anchored onto the end boards at both ends of the frame, where it is then clipped. The resulting single-frame-length wires on both sides of the frame are all disconnected electrically from each other.

Sketch of an APA showing portions of the collection layer X (blue) and the induction layers U (green), V (magenta). Each U and V layer has three segments that must be disambiguated in the reconstruction (see related article Getting LArSoft in shape for DUNE).
Sketch of an APA showing portions of the collection layer X (blue) and the induction layers U (green), V (magenta). Each U and V layer has three segments that must be disambiguated in the reconstruction (see related article Getting LArSoft in shape for DUNE).

The V layer (magenta) is wound next, followed by the U layer (green). These two layers are strung at oblique angles relative to the frame and to each other. The wire for each layer starts from a point on one short edge, advancing at an angle up one side of the frame, around the (long) edge to the other side, up this other side at the same angle, around the (other long) edge, back to the first side, then around the other short edge where it begins a complementary return journey.

Section of long edge of APA frame showing spacers and solder pads
Section of long edge of APA frame showing spacers and solder pads

Similar to the X layer, the machine winds a single V or U wire continuously until the entire surface area is covered on both sides of the frame, maintaining the appropriate pitch between the windings; the windings are then terminated and clipped at the ends to electrically isolate them from one another.

The figure shows only portions of each of these three wire layers to accentuate their angular relationships to the frame and to each other. The outermost G layer, not shown in the sketch, is strung last, in the same way as the X layer.

“Starting from simple structural steel tubing that has to be flat over 6 meters to a few mm is a daunting task. Add to that frame over 24 km of wire spaced to near perfect pitch, and the task goes beyond daunting,” Paulos said. “Input on many processes from the entire DUNE collaboration has helped make this possible. In just the next few days we will start to build the first ProtoDUNE APA. We are really looking forward to the day when this APA is functioning in a LArTPC.”