links to Rowland's main page

The use of printed circuit boards for rapid prototyping of ion trap geometries

 

 

March 12, 2014, contact Alessandra ferzoco@rowland.harvard.edu

The image on the left depicts our third generation ion trap. Redesign has been to accommodate thermal contact with a cold head (not shown, the bottom copper piece is mounted to the second stage of a Sumitomo SRDK-408D2B-F70L ) and mounting and alignment of the individual boards relative to each other and the instrument axis.

We create the geometries in SolidWorks then import them into Altium. The process of converting 3D CAD files into Gerber files is not trivial and includes creating custom functions within Altium. The hope is that now that we've figure that out, the process should be fairly streamlined.

Based on previous work at Rowland, and this review of PCB manufacturers we chose to work with Advanced Circuits for our first three runs of boards. The most critical dimension is the inscribed radius of the center lobes. The goal is to hold 0.003" diameter tolerance over the whole stack of 16 boards. This is better than the advertised rout tolerances, but because of the relatively low board cost we have the option of selecting the best 16 out of a run of 50 (or more, but we chose 50). Their engineers also thought the tolerances would be feasible.

One advantage of Advanced Circuits is that they offer very fast turn around times, for a price of course. The price matrix to the right is for 1/8" thick FR4 with a lead free solder electrode finish and no solder mask or silkscreen.

Measurements of the dimensions of the boards are in the images below. The most critical dimension is the inscribed diameter of the quadrupole, which governs the trapping fields and board-to-board alignment mechanism. It is possible to get 16 boards within 0.003" of each other from this run, but the dimension is shifted smaller than specified. Conversations with Advanced Circuits about whether this is due to systematic error in the system that we can try to predict and account for in our drawings were unproductive. While this company can be quick and produce a high quality product, we found it quite difficult to get information from their engineers. It's also unclear how much a gold plated electrode layer, which we would use in a final product, will change the measured dimension. I think we can get a suitable product from Advanced Circuits by adjusting our drawings to account for inaccuracy, ordering a larger number of boards from which to choose the 16, and placing enough orders to test whether the errors are systematic. Before we went down this path, however, a second option came along.

Sierra Circuits was recommended to us by Edison Gieswein at DGF Tech. We just recently submitted a job to them and don't have the boards back yet (the fastest turn the offered was 5-day). Their engineers were responsive to emails, informative, and motivated to meet the requirements of our job. As a matter of fact they asked for the set of go/no-go gauges we use to measure the inscribed diameter so they can calibrate their tooling! Already this company is proving easier to work with, and I hope their effort is reflected in the quality of their boards.