Hot-Data™ and Hot-Swappin’ at Agaté Technologies
Tom Woznicki

Don Mohundro, a top-notch PCB designer here in Silicon Valley, called me around Christmas last year. Agaté Technologies, one of his customers, needed a small flex circuit for a new product. The circuit turned out to be a great example of how flex circuits can be electrically simple but mechanically complex.

Agaté had developed software called Tioman™, an Enable_Device Driver. It enables the automatic configuration of a computer and allows the user to remove or replace most ISA or PCI devices without having to reboot the system. In other words, the Tioman™ device driver allows a notebook computer to have a device bay where you can insert and remove various devices like floppy drives, CD-ROMs, another hard drive, or another battery pack, and the computer will immediately recognize the new device.

To further take advantage of the Tioman™ software, Agaté developed the Hot-Data™ docking station for desktop computers. Designed around the new 3.0 inch hard drive form factor now being built by JTS Corporation and others, the Hot-Data™ dock allows the user to hot-swap a new hard drive without having to reboot. The picture above shows the Hot-Data™ dock.

When they showed the Hot-Data™ dock to their customers, they said that they liked the concept, but they wanted to be able to use 2.5-inch drives as well as 3.0 inch drives. So Agaté created a plastic cartridge that duplicated the shape of the 3.0-inch drive and used a flex circuit as an interposer between the 2.5-inch drive and the connector in the cartridge. The cartridge is called the Macau™. The picture below shows the flex and the drive in a clear plastic cartidge.

There were lots of mechanical design issues:

1. The flex had to make tight bends when the drive was inserted in the cartridge. Once in place, the circuit would not move.

2. A 44 pin, 2.0mm dual row receptacle would be at one end with a 50 pin surface mount PCMCIA receptacle at the other end to mate with the docking station. The alignment of the connectors was such that the solder surfaces for the SMT connector were on top side and the pads for the through hole connector were on the bottom side.

3. The routing was fairly straight, but there were several more ground pins on the PCMCIA connector that would have to somehow be connected to ground.

4. The length had to be just right - long enough to allow for assembly but short enough so that when in it’s final position it would not press against the plastic.

After trying several approaches to make this a single-sided flex, I decided the best approach was to make it a double-sided circuit. I made a small ground plane under the PCMCIA end to bus all the ground pins together. This ground plane also helps the polyimide stiffener support the surface mount solder joints.

The two copper layers simplified the design at the other end too. To pass a 0.006 trace between 0.028 diameter holes on 0.079 centers the coverfilm opening diameter could be no larger than 0.053 – not enough to guarantee a solderable annular ring of 0.001 minimum. By using a two-layer flex the traces go between the pads on the component side where they will be covered by both the coverfilm and the polyimide stiffener. On the solder side we won’t need coverfilm since the plated hole will hold the copper down during soldering.

The shape and length were determined first on the computer and verified by cutting paper dolls out of 5 mil Mylar. The bends are a little sharp – bend radius estimated to be 3-4 times the circuit thickness. Pads-only plating is used to protect the rolled-annealed copper in the bending area. The traces are laid out to be no more than 45° to the bend area – acceptable since these are one-time bends.

Finally, all the traces are carefully routed so that all traces are placed on the top side. The copper on the other side is etched away except for the ground plane under the PCMCIA connector and the solder pads for the through-hole connector. Laying the traces out in this fashion allows us to eliminate the coverfilm on the bottom side, because the polyimide stiffener covers the ground plane. This gives the circuit almost as much flexibility as a single-sided circuit.

So the final construction has a diclad base laminate with pads-only plating, coverfilm and a polyimide stiffener on one side and just a polyimide stiffener on the bottom side.

All this attention to detail paid off — the first prototypes fit just right and functioned perfectly. The picture above shows the assembled circuit. The PCMCIA connector is on top. The picture below shows how the cartridge interfaces with the docking station.

Production units of the Hot-Data™ dock with the Macau™ cartridge are just starting to ship. You can contact the smiling folks at Agaté Technologies at 510-492-5430 or through their website at

Don Mohundro, also a smiling fellow, can be reached at his service bureau, The Designers, at 408-984-8430.

Tom Woznicki is the owner of Flex Circuit Design Company, a consulting company in Silicon Valley that specializes in designing flex circuits. He can be reached at or 1-888-FLEXMAN.

This article was published in the September/October 1997 issue of Flexible Circuits Engineering magazine. It was edited slightly for this website.

Back to Flex Circuit Design Company homepage.

©1997 Tom Woznicki. All rights reserved.